From outer to inner space: traveling along a scientific career from astrochemistry to drug research.

This professional history describes my journey as a research scientist after my early training and experiences in the pre- and early post-World War II United States. My graduate training concentrated on a problem in astrochemistry: phenomena on comets. As my career developed, I felt confident enough in myself as an experimentalist to enter, and make contributions to, several different fields: structural biochemistry (via nuclear magnetic resonance spectroscopy), molecular immunology, pharmacology, neurochemistry, and cell biology. One emphasis is on the nature and quality of my scientific training that permitted me to do cross-disciplinary work. A second emphasis is on the technical and intellectual developments in late twentieth-century science and how, along with the changes in American society as it passed through three major wars, they influenced my life and thought.

The effects of morphine on cell proliferation.

There is increasing evidence that endogenous opioid peptides ("enkephalins") and other neurotransmitters have widespread, receptor-mediated roles as growth regulators in non-neuronal cells and tissues. For example, it is now believed that enkephalins produced in placental trophoblast giant cells have multiple roles in supporting embryo growth, and in maternal adaptation to pregnancy. Since plant and synthetic narcotics (e.g., morphine) bind to the same receptors, the questions immediately arise: Do narcotics also have actions as growth regulators? If so, do these actions have physiological significance in addicts? Recent work on the first of these questions is covered in this review. While the greatest volume of research has been focused on the proliferative effects of narcotics for cells of the immune system, the roles of opioid peptides and narcotics on the growth of a variety of other cells has come under study recently.

Differential effects of opioid and adrenergic agonists on proliferation in a cultured cell line.

A novel clonal cell line transfected with the delta-opioid receptor (delta-OR) encoding gene was used to study agonist-activated regulation of cell proliferation. In this cell line, endogenous beta2-adrenergic receptors (beta2-ARs) are coexpressed with the exogenous delta-ORs. Upon individual acute treatments with morphine and procaterol (a selective beta2-AR agonist), both the delta-OR and beta2-AR are coupled to differential modulation of cyclic AMP (cAMP) levels in accord with the classical second messenger response patterns to these agonists in the normal cellular settings of the receptors. But chronic morphine activation of the delta-OR inhibits cellular proliferation, while chronic procaterol activation of the beta2-AR stimulates it. Chronic treatment with the individual agonists is accompanied by differential activation of the mitogen-activated protein kinase (MAPK) isozymes, extracellular-regulated kinase (ERK) and c-Jun N-terminal kinase (JNK). The findings suggest that chronic beta2-AR activation stimulates proliferation by interacting with the ERK signalling cascade independent of a cAMP-mediated pathway. In contrast to treatment with individual agonists, chronic dual agonist treatment suppresses procaterol-induced stimulation of ERK activity and stimulation of proliferation indicating that a cross-regulatory interaction occurs between the delta-OR and beta2-AR signalling systems in the cells under these conditions.

Hormone-defined cell system for studying G-protein coupled receptor agonist-activated growth modulation: delta-opioid and serotonin-5HT2C receptor activation show opposite mitogenic effects.

G-protein-coupled receptor (GPCR) agonist-activated transformation of NIH/3T3 fibroblast cells has been documented by many workers. Our present interest is in the growth control exerted by these agonists. The mechanisms involved in GPCR agonist-activated growth regulation are not known and investigations using existing cell lines are complicated by the endogenous expression of numerous different GPCRs as well as by the fact that these cell lines are cultured in serum that contains naturally occurring agonists for these receptors. To study the agonist induced growth response of cells transfected with either delta-opioid or serotonin-5HT2C neurotransmitter receptor genes, we have developed new clonal cell lines derived from NIH/3T3 mouse fibroblast cells. These new cell lines, designated with the suffix 3T3DA, can be cultured stably in serum-free, hormone-defined medium: insulin is the only exogenous growth factor added to the culture medium of proliferating 3T3DA cell lines, and their proliferation can be stopped and started by the respective removal or addition of insulin. Micromolar concentrations of agonists were used to activate the corresponding opioid and serotonin receptors over periods extending to 6 days. We observed distinct patterns of GPCR-specific, agonist-activated growth regulation in serum-free cultures, but not in serum-supplemented cultures. At concentrations > 10 microM, morphine inhibits growth of delta-opioid receptor-expressing cells by 40% with respect to normal 3T3DA cells. Opioid agonist induced inhibition of cyclic AMP (cAMP) production as well as growth down-regulation are pertussis toxin sensitive indicating that the exogenously expressed delta-opioid receptors demonstrate classical opioid receptor signaling. The presence of 1 microM serotonin stimulates growth of serotonin-5HT2C receptor- expressing cells by approximately 100% with respect to normal 3T3DA cells. Neither the untreated nor the agonist-treated cells form colonies in soft agar, indicating that they retain anchorage-dependent growth control. These cell lines provide a simple system that could be used as a tool for probing the complex molecular mechanisms associated with GPCR agonist-activated growth control.

Theoretical analysis of a morphine withdrawal phenotype in a cultured cell line.

We have previously described a delta-opioid receptor-expressing cultured cell line that proliferates in a defined medium and responds to chronic morphine treatment with an inhibition of its rate of proliferation. To help provide an explanation for this behavior, we have used computer simulation of cell cycle kinetics to analyze the observed rates of proliferation of these cells in the presence and absence of morphine, and after withdrawal of morphine treatment. We questioned whether the difference in cell kinetics observed for the cell populations under the different treatments could be due to changes in the length of the cell cycle, withdrawal of cells from the cycle into a quiescent state, or differences in cell renewal. This was investigated by comparing observed cell numbers as a function of time with the results of different computer simulations using different values for these parameters. We found that we can provide a satisfactory explanation of the experimental observations on the basis of changes in a small set of parameters: Untreated cells experience a slowdown of cell proliferation at about the culture density where multiple cell-cell contacts are made and, beginning then, a large fraction are shunted from G1 into a quiescent state. Chronic morphine treatment inhibits proliferation by slowing passage through G1, but the cells remain as sensitive to cell-cell contacts as the untreated cells. After drug withdrawal following a 6 day treatment with morphine, the cells exhibit a large temporary increase in their rate of proliferation compared with control or chronically treated cells but about 48 hours after withdrawal, when cell-cell contacts just begin to be made, the cells return to almost their pre-treatment total cell cycle time and, as before, a large fraction are shunted into a quiescent state. Taken in conjunction with previously published results, the present ones indicate a possible interaction between morphine-induced and insulin-induced nuclear signaling pathways to the nucleus.

Co-localization of mu and delta opioid receptors on SK-N-SH cells detected by fluorescence microscopy using labeled anti-idiotypic antibodies.

Selective fluorescence labeling of opioid receptor subclasses on SK-N-SH cultured cells has been accomplished using labeled polyclonal anti-idiotypic antibodies along with subclass-selective opioid agonists (DPDPE, delta-selective; DAMGO, mu-selective) as blocking reagents. Labeling of the cells was examined using conventional fluorescence microscopy. Co-localization of mu- and delta-opioid receptors on SK-N-SH cells has been studied by double labeling fluorescence experiments. In agreement with our own, and other workers', previous observations on NG108-15 cells, a subpopulation of viable cells in asynchronous cultures are labeled. Among those SK-N-SH cells that are labeled, both subclasses of receptors are seen. On the basis of sequential blocking experiments we interpret our combined results to be consistent with a model where mu- and delta- binding sites reside on different subunits of a multimeric complex.

Two- and three-dimensional distributions of opioid receptors on NG108-15 cells visualized with the aid of fluorescence confocal microscopy and anti-idiotypic antibodies.

Polyclonal anti-idiotypic antibodies previously shown to be specific for mu- and delta-opioid receptor subclasses on rat brain membrane preparations using radioreceptor assays have been labeled with fluorescent conjugates of avidin. The resulting complexes have been used to study the distribution and properties of opioid receptors on living and fixed NG108-15 cells using normal fluorescence, and confocal fluorescence, microscopy. From the confocal data three-dimensional views of immunofluorescent distributions have been obtained using a computer transformation of the data. From experiments involving co-incubation with media containing naloxone, selective opioid peptides and stereoisomeric synthetic opioid analogs as blocking agents, the specificity and stereospecificity of immunofluorescent binding is reported. The effects on immunofluorescent binding by co-incubation with media containing a sulfhydryl reagent, 100 mM-NaCl and sodium azide are also reported. The results have implications concerning the use of immunoprecipitation in the biochemical isolation of the receptors. Finally, the confocal two-dimensional sections and two-dimensional reconstructions obtained from fixed cells show internal immunofluorescence in the form of cytoplasmic clusters. The reagents and methods should be applicable to visualizing native opioid receptor distributions on neurons.

Stereospecificity of opiate receptors in a synchronized culture of NG108-15 cells probed with fluorescent anti-idiotypic antibodies.

The stereospecificity of the binding of an anti-opiate receptor, anti-idiotypic antibody to receptors on a synchronized NG108-15 neuroblastoma x glioma cell culture has been examined by fluorescence labeling of the antibodies. We had previously found that unsynchronized NG108-15 cultures showed high specificity (reversibility by receptor-binding ligands) of antibody labeling. However, only a subpopulation additionally showed stereospecific labeling (nonreversibility by non-receptor-binding opiate enantiomers). In the present work we find that the percentage of cells that display stereospecific fluorescence is highly dependent upon position in the cell cycle. Furthermore, the properties of the bound fluorescence, with respect to sensitivity to photobleaching, also depend upon position in the cell cycle. The fluorescence behavior of labeled synchronized cell populations after treatment to inhibit de novo glycosylation is also reported. The results reported here may have implications concerning the structure of the opiate receptor complex and efforts to solubilize the binding protein in its native form.

Analysis of the binding site architecture of monoclonal antibodies to morphine by using competitive ligand binding and molecular modeling.

The structural features of mAb directed against the opiate morphine were analyzed by using competitive ligand analog-binding studies, examination of the V region amino acid sequence, and computer-aided molecular modeling of the fragment V region. The antibody response in BALB/c mice to morphine is relatively restricted, in that all of the mAb examined in this study contained the same lambda L chain and very similar H chain V regions. A three-dimensional model of the antimorphine-binding site was constructed by using computational and graphic display techniques. Each of the six complementary-determining regions was constructed by using fragment replacement methods employing canonical loop conformations of known "parent" structures. Experimental competitive ligand-binding data and theoretical modeling suggest that a charged glutamate residue at position H:50 and aromatic side chains of residues H:33W, H:47W, H:58F, H:95W, H:101iY, and L:91W are key features in ionic and hydrophobic interactions with the ligand. This study represents the first use of theoretical and experimental modeling techniques to describe the Ag-binding site of a mouse fragment V region containing a lambda L chain.

Direct production of Fv-fragments from a family of monoclonal IgGs papain digestion.

Fv fragments of four monoclonal antibodies specific for morphine binding have been produced from their divalent IgG forms by papain digestion using the classic procedure for Fab formation. The binding characteristics of one of the Fv fragments have been determined relative to the intact antibody by equilibrium dialysis. Its dissociation constant is a factor of five lower than the IgG. Previous work had resulted in the sequences of each the chains for the four Fv fragments. The light chains are all from the highly homologous lambda subclass while the gamma heavy chains are closely related except for their CDR regions. In this work optical molar extinction coefficients are predicted from amino acid sequences for each of the fragments. It is found that they differ significantly from each other and from the commonly used value for intact IgG. Detailed comparisons between our results and those reported previously on the molecular masses of Fv-derived light and heavy chains and hapten-Fv dissociation constants are given based on analytical gel electrophoresis and electroblotting experiments using dye and immunovisualization techniques. Isoelectric focusing experiments have been performed and the pIs obtained are compared to those predicted theoretically from the chain sequences. Gel filtration column chromatography, acrylamide gel electrophoresis and equilibrium dialysis experiments are consistent with significant aggregation of the Fv fragments in neutral solution with accompanying inactivation of the binding site. Comparison of sequences for the Fv light and heavy chains are made with those which have been proposed to be important for chain dimer association and for canonical hypervariable regions. This methods of Fv production is not regarded as a general one. However, it may be an approach which is general to lambda chain containing antibodies.

A glass-fiber filtration assay for solution phase hapten-antibody binding. Effect of surface treatment with a polycation.

Radioligands bound to murine monoclonal and polyclonal antibodies in solution are rapidly and effectively separated from free ligand by filtration through either untreated or polyethylenimine (PEI)-pretreated glass-fiber filters. Ligand binding to selected anti-morphine monoclonal antibodies, determined by filtration through untreated filters, was significantly greater (1.5-3.0-fold) than the binding activities obtained by gel filtration. However, radioactivity retained by filters that were pretreated with PEI (pH 4) was essentially the same as that obtained using the Sephadex G-25 short-column assay. The pH dependence of the retention on the coated fibers suggests that the mechanism of binding of antibodies to glass is different from that to the PEI-treated surfaces. The quantitative aspects of the assay are reported. The method should prove useful where quantitative, rapid and inexpensive binding radioassays need to be performed.

Comparative sequence and immunochemical analyses of murine monoclonal anti-morphine antibodies.

A more complete characterization is given for four previously reported anti-morphine monoclonal antibodies that bind the hapten with high affinity and to which anti-idiotypic antibodies have been raised that mimic opiates at receptor binding sites. The variable (V) region nucleotide sequences of the heavy (H) and light (L) chains of these murine antibodies have been determined by direct sequencing of the poly(A)+ mRNAs using specific oligonucleotide primers and dideoxynucleotide chain-termination, and the deduced amino acid sequences are compared. The primary sequences predicted for the VH segments of 10C3 and 11C7 antibodies are closely associated with the VHIIIB subgroup of mouse H-chain (80 to 82% homology), while those for the V-regions of 3B9 and 12D4 H-chains correlate well with the VHIIC subgroup (64 to 67% homology). The 11C7, 10C3, 3B9 and 12D4 hybridoma cell lines use JH1, JH2, JH3 and JH4 DNA segments, respectively. Since considerable variations in length and primary sequence in the CDR3 (complementarity determining region) peptides of all the H-chains are evident, conservation of the D-region structure does not appear to be necessary for effective hapten binding. However, sequence homologies of the CDR2 regions of all the antibodies indicate that residues Glu H-50, Ile H-51, Pro H-52a and Tyr H-59 are conserved, and that these segments may be more critically involved in binding than the other H and L-chain hypervariable regions. The marked VL sequence homology, greater than 93%, among the L-chains and consensus lambda sequence, suggests derivation from a similar or identical VL germ-line gene. The L-chain J-region peptides for all the antibodies are classified JL1 and no VL-JL junctional diversity was apparent. The antimorphine antibody L-chains are apparently generated by the joining of a specific J-gene segment to a single germ-line V-gene segment, and minor sequence variations are the result of somatic mutations within the coding region. The leader sequence for one of the H-chains was determined. The inhibition of morphine binding by phenoxybenzylation or iodination of the affinity-purified immunoglobulins indicates the involvement of a single tyrosyl residue within or close to the antibody-combining site for the opiate. This conclusion is supported by the sequence data and nuclear magnetic resonance measurements reported in the accompanying paper, in which the results are used to interpret nuclear magnetic resonance measurements on one of the ligand-antibody systems. The possible importance of additional contact amino acids, tryptophan, aspartic and/or glutamic acids, is discussed.

Nuclear magnetic resonance studies of flexible opiate conformations at monoclonal antibody binding sites. Quantitative interproton distances obtained from comparing theoretical and experimental transferred nuclear Overhauser enhancement: correlation with antibody sequence.

A previous publication described the use of qualitative intramolecular 1H-transferred nuclear Overhauser effect measurements to determine the conformations of flexible ligands at monoclonal anti-opiate antibody binding sites. This paper concentrates on the quantitative interpretation of experiments of this type using the ligand nalorphine (N-allyl morphine) and a single anti-opiate monoclonal antibody. I compare the experimental unidimensional driven nuclear Overhauser effect buildup curves to theoretical curves derived with a knowledge of the fixed interproton distances in the ligand. The discussion covers the potential accuracies of derived distances and concentrates on two problem areas associated with determining structures from this type of experiment. The most serious one is the case where, because of particular multiproton spatial distributions, spin diffusion is so rapid that it cannot be determined experimentally and where numerical fits of theoretical calculations are misleading. The results show that, while intraligand spin diffusion complicates the interpretation for some proton pairs, with many others accuracies within about 0.3 A for interproton distances from 2 to 4 A are attainable. The results confirm the earlier report that the conformation of nalorphine in this antibody binding site differs from the major one present in solution or in the crystal. An important aspect of the work is that theoretical prediction of nuclear Overhauser effect time-dependence is an important practical tool for recognizing cases where interpretation of experiments will be difficult. Initial data on protein-to-ligand transferred nuclear Overhauser effect are presented, which show that at least one aromatic amino acid residue is closely involved in the binding of the ligand. The companion paper presents the primary sequences of the variable regions of the antibodies being used in our studies. In this paper, these and associated immunochemical studies are correlated with the nuclear magnetic resonance results. The combination of data presented in the two papers provides a basis for future work on protein-ligand interproton distances in the range 1 to 5 A using both transferred nuclear Overhauser effect (for rapidly exchanging ligands) and isotope-edited, indirectly detected nuclear Overhauser effect (for tightly bound ligands).

Physical characterization of native opiate receptors. Additional information from detailed binding analysis of a radiation-inactivated receptor.

Target size analyses of the etorphine receptor were performed on frozen rat brain P2 homogenates using the radiation inactivation technique. Multi-point saturation curves at each radiation dose revealed that the apparent dissociation constant for the binding of this ligand to its receptor is a function of the dose. Analysis of the results shows clearly that the ligand-binding macromolecule is functionally coupled to at least one other macromolecule. When the coupling is destroyed the ligand dissociation constant becomes larger by over an order of magnitude. Thus, the variation of the dissociation constant with dose yields interesting new information on the nature of the native receptor which has implications with respect to the conformation of the binding site and to solubilization and cloning methods directed towards sequencing the ligand-binding component of opiate receptors.

NMR studies of flexible opiate conformations at monoclonal antibody binding sites. I. Transferred nuclear Overhauser effects show bound conformations.

500 MHz H, homonuclear, intra-molecular, transferred Nuclear Overhauser Effect measurements have been performed on the bound forms of a classical opiate antagonist, nalorphine and an agonist, levorphanol at their respective binding sites in two different specific anti-opiate monoclonal antibody fragments. Based upon previous studies of opiate conformations in solution the results clearly show without extensive interpretation that one of these flexible haptens has the opposite (from solution) isomeric conformation in its bound form. For nalorphine the axial isomer of the N-allyl substituent is the bound form whereas in solution the equatorial isomer dominates at a ratio of 5:1. For levorphanol the bound form is that of equatorial N-methyl in accord with the low energy conformation in solution. In this preliminary report we discuss the initial measurements and results and their implications with respect to the conformations of flexible ligands at macromolecular binding sites including opiate receptors.

Morphine-mimetic anti-paratypic antibodies: cross-reactive properties.

Polyclonal, anti-idiotypic, rabbit antibodies have been raised against four murine monoclonal anti-morphine Fab fragments. The antibody preparations, after affinity purification, have been shown to contain an anti-paratypic fraction which reversibly inhibits morphine binding to the anti-morphine antibodies and to cellular opiate receptors. Using some of the unique properties of this system, for the first time cross-reactivities of anti-paratypic antibodies with the monoclonal anti-morphine IgGs have been examined including competition for the same binding sites by a classical opiate agonist/antagonist pair.

Subclass specificity of anti-idiotypic anti-opiate receptor antibodies in rat brain, guinea pig cerebellum, & neuroblastoma x glioma (NG 108-15).

Receptor subclass recognition properties of affinity-purified rabbit polyclonal anti-idiotypic anti-opiate receptor antibodies in various membrane preparations have been determined. The anti-receptor immunoglobulins significantly decrease binding of 3H-[D-Ala2,-MePhe4,Gly-ol5]enkephalin, a highly selective mu agonist, in rat neural membrane. In the presence of a concentration of the unlabeled ligand sufficient to block existing mu sites, the antibodies compete, to a lesser degree with 3H-[D-Ala2,D-Leu5]enkephalin for delta site occupancy in both rat neural membrane, and neuroblastoma x glioma membrane preparations. The antibodies do not displace 3H-ethylketocyclazocine from rat brain or guinea pig cerebellum.