Juan S. Bonifacino received his Ph.D. from the University of Buenos Aires, Argentina, in 1981. He joined the Cell Biology and Metabolism Branch (CBMB), NICHD, in 1984 and currently heads its Unit on Intracellular Protein Trafficking.

Over the past few years, I have been interested in various aspects of the biogenesis and transport of integral membrane proteins within the secretory pathway. Many of my studies have examined the relationship between protein structure and intracellular trafficking. Research in my group focuses on two main topics that exemplify different mechanisms of protein localization within cells: 1) the assembly of multisubunit complexes in the endoplasmic reticulum (ER) and the retention of incompletely assembled complexes within the ER, and 2) the mechanisms of protein localization to a specific compartment of the secretory pathway, the trans-Golgi network (TGN).

My interest in the assembly of multisubunit complexes stems from work I did as a postdoctoral fellow in Richard Klausner's laboratory at NICHD. Our early studies of the assembly of the T-cell antigen receptor showed that only completely assembled complexes were efficiently transported to the cell surface, whereas free chains and incomplete complexes were largely retained intracellularly and in some cases, degraded. These phenomena have now been observed for many other protein complexes, but the molecular mechanisms involved are not well understood. Recent work from my laboratory on another family of hetero-oligomeric complexes -- class II antigens of the major histocompatibility complex (MHC) -- has provided insights into the early events in protein assembly in the ER and the mechanisms that control transport of newly made complexes from the ER. We found that incompletely assembled class II MHC molecules form large, heterogeneous aggregates in association with the ER chaperone, immunoglobulin-heavy-chain binding protein (BiP). Formation of such aggregates may be a determining factor in the process by which unassembled subunits are retained in the ER. Strikingly, we also found that BiP-associated aggregates exist transiently during assembly of normal class II MHC molecules in spleen cells. This suggests that aggregates are not necessarily aberrant products but are most likely true intermediates in the normal assembly process. These and other observations have allowed us to establish a sequence of events in the assembly of class II MHC molecules.

Whereas retention of unassembled subunits in the ER may depend on the general physicochemical properties of the proteins (i.e., aggregation and association with ER chaperones), localization to other compartments of the secretory pathway appears to be mediated by more specific signals. Work in my laboratory has led to the identification of a signal that mediates protein localization to the trans-Golgi network. The signal is borne within the cytoplasmic domain of two TGN-specific proteins, TGN38 and furin, and consists of a tyrosine-based motif related (but not identical) to internalization signal sequences. These observations, together with similar findings in yeast proteins, suggest the existence of a general mechanism for protein localization to the TGN that relies on specific recognition of cytoplasmic signals. We are now conducting studies to identify molecules that interact with such signal sequences and that control protein localization to the TGN.

Our laboratory studies the neurobehavioral consequences of chronic illness and its treatment in children and adults, particularly in patients with cancer and HIV infection. Our main efforts focus on characterizing disease-specific abnormalities in intellectual, mnemonic, attentional, socioemotional, and judgmental abilities and on documentating treatment-related changes in these functions.

Using existing test instruments and newly developed approaches, we can now comprehensively characterize central nervous system (CNS) manifestations and sensitively document treatment-related changes in neurobehavioral functioning. We then try to establish that these neurocognitive scores are associated with physiologic changes in brain images, cerebrospinal fluid (CSF), or laboratory markers that indicate abnormalities in the CNS resulting from disease or treatment. This is important in validating neurocognitive measures as markers of the effects of the disease on the CNS and ruling out other confounding factors (e.g., emotional and socioeconomic factors).

In this way, we have established that long-term survivors of childhood acute lymphoblastic leukemia (ALL) who received cranial irradiation as preventive therapy experienced adverse late sequelae both on intellectual and computed tomography (CT) brain-scan tests. We also showed that the effects on the intellect were associated with abnormalities in the CT scan, indicating that the neurobehavioral sequelae in ALL have an organic basis. We are currently evaluating long-term survivors who received less neurotoxic but equally protective CNS therapies.

We have also developed a technique for measuring the incidence and severity of CT brain-scan abnormalities in HIV infected children. We found that neurocognitive deficits and aberrant behavior are related to the degree of CNS abnormality, establishing the clinical significance of these lesions. Additional quantitative and longitudinal studies using magnetic resonance imaging (MRI) technology are under way. We also analyzed the CSF of children with symptomatic HIV infection for the possible presence of neurotoxins and found elevated concentrations of quinolinic acid (QUIN) that inversely correlate with the level of neuropsychological functioning. In ongoing studies, we are exploring CSF-to-serum ratios of QUIN and changes with therapy. A relation between disease stage, defined by CD4 measures and P24 levels, and CNS structure and function, indicated that advanced disease puts children at higher risk for significant HIV-associated CNS manifestations. Longitudinal studies that follow individuals over time and that use multiple regression models are in progress.

We further developed a methodology for evaluating neurobehavioral changes in chronically ill patients who are undergoing treatment in clinical trials. We were the first to show significant improvements in neurocognitive function with 3'-azida-3'-deoxythymidine (AZT) therapy in adult patients with AIDS dementia complex. We later extended these findings to children with HIV infection where significant improvements were observed in both encephalopathic and nonencephalopathic patients. Concurrent decreases in the size of enlarged ventricles and subarachnoid spaces on CT scans validated these findings. In addition, we observed a decrease in CSF QUIN with a concurrent increase in general cognitive function. In ongoing clinical trials, we have demonstrated and are further exploring pharmacokinetic correlations, relating changes in neurobehavioral function to dose effects, absorption area under the curve of the antiretroviral agent, and the agent's penetration into the CNS.

Researchers in my section are interested in the cellular and molecular mechanisms of development, neurotoxicity, and neurodegeneration. Our basic hypothesis is that oxygen radicals and other free radicals play an important role in these processes.

During cellular metabolism, aerobic organisms generate oxygen-derived free radicals, including superoxide and hydroxyl radicals. These substances can cause lipid peroxidation and oxidation of biomolecules, and are thought to be involved in myocardial infarction, strokes, and the neurodegenerative aging process. But aerobic organisms have evolved mechanisms that enable them to survive in the face of the ubiquitous presence of these free radicals. These mechanisms involve enzymes such as superoxide disumutase (SOD) and glutathione peroxidase. During the past few years, we have been studying the role of superoxide radicals in drug-induced neurotoxicity. Using transgenic mice that express elevated levels of human CuZnSOD, we have shown that these SOD Tg mice are protected against the neurotoxicity of N-methyl-4-phenyl-1,2,3,6-tetra-hydropyridine and methamphetamine -- compounds that are dopamine neurotoxins in mice. These results suggest that the neurotoxicity of these drugs is associated with superoxide radical formation.

We are also interested in the role of nitric oxide in the neurotoxicity of drugs of abuse. Using primary cultures from rat fetal mesencephalon, our laboratory has recently demonstrated for the first time that inhibition of nitric oxide synthase can attenuate the neurotoxicity of methamphetamine. We have also demonstrated for the first time that inhibitors of ADP-ribosylation can also protect against the toxicity of this drug in vitro. In addition, we have been able to establish a model for drug-induced gliosis in vitro. Using that model, we showed that inhibitors of ADP-ribosylation can prevent reactive gliosis. Using both in vitro and in vivo model, and several probes, we are continuing to dissect the specific pathways involved in neuronal cell death. We are also interested in finding out whether these same pathways are involved in apoptosis, or programed cell death.

My lab focuses on transmissible spongiform encephalopathies (TSEs), which are infectious and fatal neurodegenerative diseases occurring most prominently in sheep (scrapie), cattle (mad cow disease), and humans (kuru and Creutzfeldt-Jakob disease). The obscure infectious agent of these diseases resembles a virus in that it replicates in the host and has distinct strains, but so far, no agent-specific nucleic acid has been identified. The unusual biochemical properties of the agent led almost three decades ago to still-unproven hypotheses that it contains only protein. More recently, this putative "infectious protein," or "prion," was proposed to be PrP-res (or PrP^SC), a neuropathogenic, abnormally protease-resistant and amyloidogenic form of a host-encoded protein, PrP. We are interested primarily in how PrP-res is made, how its formation might be blocked, and what relationship its formation has to TSE-agent replication and pathogenesis.

Using scrapie-infected tissue culture cells developed by my LPVD colleagues Rick Race and Bruce Chesebro, my lab established that PrP-res is derived posttranslationally from normal, protease-sensitive PrP (PrP-sen) and that the subcellular site of conversion is the plasma membrane and/or along an endocytic pathway leading to the lysosomes. We also used these cells to identify some potent and selective inhibitors of PrP-res accumulation and scrapie-agent replication. These inhibitors appear to act by competitively inhibiting an interaction between PrP-res and an endogenous sulfated glycosaminoglycan (GAG). GAGs are components of pathogenic amyloids associated with many diseases, and our studies suggest that GAG-amyloid interactions may be attractive targets in designing drugs for these diseases. Some of these inhibitors have already been shown to have therapeutic value in preventing scrapie in animals. We hope that these compounds may be effective as drugs not only for TSEs but also for amyloidoses of greater clinical significance in humans, such as Alzheimer's disease.

One of the major difficulties in studying the underlying basis for TSE disease was the fact that no one had been able to convert the normal PrP to PrP-res in anything simpler than a scrapie-infected cell. In collaboration with Peter Lansbury's lab at the Massachusetts Institute of Technology, we recently overcame this problem by establishing a defined, cell-free reaction mixture that supports the formation of PrP-res. This discovery provided the first direct evidence that PrP-res formation results simply from an interaction of normal PrP with preexisting PrP-res. With this experimental system, we now have a unique opportunity to study the chemical details of this process, which is central to the TSE diseases, and to analyze the nature of the infectious agent that instigates this disease process in the host.