Search

Aberrant accumulation of neurofilaments is a feature of human motor neuron diseases. Experimentally motor neuron disease can be induced in transgenic mice by overexpressing the mouse neurofilament light subunit (NF-L), the human heavy subunit (NF-H), or mouse peripherin. Here we describe that mice harboring a bacterial artificial chromosome (BAC) transgene containing the human midsized neurofilament subunit (NF-M) gene develop a progressive hind limb paralysis associated with neurofilamentous accumulations in ventral horn motor neurons and axonal loss in ventral motor roots. Biochemical studies revealed that all three mouse neurofilament subunits along with the human NF-M contributed to filament formation, although filaments contained less peripherin. In addition the endogenous mouse NF-M became less phosphorylated in the presence of the human protein and accumulated in the cell bodies of affected neurons even though phosphorylated human NF-M did not. Remaining motor axons contained an increased density of neurofilaments and morphometric studies showed that principally small myelinated axons were lost in the transgenic animals. Removing half of the mouse NF-M by breeding the transgene onto the mouse NF-M heterozygous null background offered no protection against the development of disease, arguing that the effect is not simply due to elevation of total NF-M. Collectively these studies argue that the human and mouse NF-M proteins exhibit distinct biochemical properties and within mouse neurons are not interchangeable and that indeed the human protein may be toxic to some mouse neurons. These studies have implications for the use of human neurofilament transgenic mice as models of amyotrophic lateral sclerosis.

The formation of the labyrinth layer is a critical step of placental development. The transcription factor glial cells missing 1 (Gcm1) plays a pivotal role in labyrinth development, but the sequence of events controlling its expression has not been identified yet. Our studies presented herein show that Gcm1 expression occurs in three distinct phases during placental development, each specific to a particular stage of chorio-allantois interaction. In the first, the pre-fusion phase, Gcm1 mRNA is expressed in isolated clusters of chorionic cells, but not efficiently translated. Upon allantois–chorion fusion, the second phase, Gcm1 expression is greatly induced in clusters of chorionic cells separated by non-expressing cells and the Gcm1 mRNA is translated to protein. In the third phase, the labyrinth formation, cells expressing Gcm1 proliferate, involute in the chorionic plate and branched villi formation begins.

Mammalian neurofilaments are assembled from the light (NF-L), midsized (NF-M), and heavy (NF-H) neurofilament proteins. While NF-M and NF-H cannot self-assemble into homopolymers, the data concerning NF-L has been more contradictory. In vitro bovine, porcine, and murine NF-L can homopolymerize in the absence of other subunits. However, in vivo studies suggest that neither rat nor mouse NF-L can form filaments when transfected alone into cells lacking endogenous intermediate filaments. By contrast, human NF-L forms homopolymers in similar cell lines. Recently we generated mice with null mutations in the NF-M and NF-H genes. To determine if mouse NF-L can homopolymerize in mouse axons, NF-M and NF-H null mutants were bred to create a line of double mutant animals. Here we show that axons in NF-M/H double mutant animals are largely devoid of 10-nm filaments. Instead, the axoplasm is transformed to a microtubule-based cytoskeleton-although the lack of any increase in tubulin levels per unit length of nerve or of increases in microtubule numbers relative to myelin sheath thickness argues that microtubules are not increased in response to the loss of neurofilaments. Thus in vivo rodent neurofilaments are obligate heteropolymers requiring NF-L plus either NF-M or NF-H to form a filamentous network.

Neurofilaments are central determinants of the diameter of myelinated axons. It is less clear whether neurofilaments serve other functional roles such as maintaining the structural integrity of axons over time. Here we show that an age-dependent axonal atrophy develops in the lumbar ventral roots of mice with a null mutation in the mid-sized neurofilament subunit (NF-M) but not in animals with a null mutation in the heavy neurofilament subunit (NF-H). Mice with null mutations in both genes develop atrophy in ventral and dorsal roots as well as a hind limb paralysis with aging. The atrophic process is not accompanied by significant axonal loss or anterior horn cell pathology. In the NF-M–null mutant atrophic ventral root, axons show an age-related depletion of neurofilaments and an increased ratio of microtubules/neurofilaments. By contrast, the preserved dorsal root axons of NF-M–null mutant animals do not show a similar depletion of neurofilaments. Thus, the lack of an NF-M subunit renders some axons selectively vulnerable to an age-dependent atrophic process. These studies argue that neurofilaments are necessary for the structural maintenance of some populations of axons during aging and that the NF-M subunit is especially critical.

Neurofilaments (NFs) are prominent components of large myelinated axons. Previous studies have suggested that NF number as well as the phosphorylation state of the COOH-terminal tail of the heavy neurofilament (NF-H) subunit are major determinants of axonal caliber. We created NF-H knockout mice to assess the contribution of NF-H to the development of axon size as well as its effect on the amounts of low and mid-sized NF subunits (NF-L and NF-M respectively). Surprisingly, we found that NF-L levels were reduced only slightly whereas NF-M and tubulin proteins were unchanged in NF-H–null mice. However, the calibers of both large and small diameter myelinated axons were diminished in NF-H–null mice despite the fact that these mice showed only a slight decrease in NF density and that filaments in the mutant were most frequently spaced at the same interfilament distance found in control. Significantly, large diameter axons failed to develop in both the central and peripheral nervous systems. These results demonstrate directly that unlike losing the NF-L or NF-M subunits, loss of NF-H has only a slight effect on NF number in axons. Yet NF-H plays a major role in the development of large diameter axons.

Paralytic tremor (pt) in rabbits and shaking pup (shp) in dogs are allelic dysmyelinated mutants of the proteolipid protein (Plp) gene. Both mutations affect the same amino acid, histidine36, which is replaced by glutamine in pt and by proline in shp. Phenotypic expression of these two mutations is very different. Paralytic tremor presents a much milder form of dysmyelination than shaking pup. The number of oligodendrocytes in the mutant rabbit is normal, while in the dog, the oligodendrocyte number is reduced due to early death or incomplete maturation. We have previously reported an abnormal intracellular transport of the PLPpt, whereas DM-20pt was normally transported to the cell membrane. In the present study, we show that the transport of the two isoforms containing the shp mutation is impaired in transfected Cos-7 cells. Cotransfecting cells with different ratios and combinations of mutated PLP and DM-20 cDNAs, we demonstrated that DM-20pt, but not DM-20shp, facilitates intracellular trafficking and integration into the plasma membrane of either of the two mutated PLPs. The phenotypic difference between these two allelic mutations can result from differences in DM-20 protein trafficking and sorting. These results show that the loss of function of PLP is not position-dependent but depends on the nature of the mutation.

Myelin basic proteins (MBPs) are major proteins of central nervous system (CNS) myelin, where they facilitate the apposition of cytoplasmic faces of myelin lamellae. Myelin-bearing oligodendrocytes transport MBP mRNA to myelin, where newly translated protein is directly inserted into the myelin sheath. An apparent absence of MBPs in oligodendrocyte perikarya has suggested that MBP localized to the soma is translationally inert. We now demonstrate by confocal immunofluorescence microscopy that not only are MBPs present in the majority of oligodendrocyte perikarya but oligodendrocytes are heterogeneous with respect to their localization of MBPs; MBPs are concentrated in some cells at the plasmalemma and distributed in others throughout the cytoplasm and, surprisingly, the nucleus. MBPs are present in the nuclei of over half of oligodendrocytes in the adult, but in almost all MBP+oligodendrocytes during myelinogenesis. Transport of MBPs into nuclei appears to be a regulated process since some cells exhibit robust MBP accumulation in their cytoplasm but exclude MBPs from their nuclei. We show that oligodendrocyte nuclei contain all four major MBP isoforms, but that in transgenic mice, the epitope-tagged 14 kD MBP isoform preferentially segregates to the plasmalemma. Our data demonstrate that oligodendrocytes are not required to exclude MBPs from their perikarya and suggest that MBPs have a specific function in the oligodendrocyte perikarya and nucleus.

Pelizaeus-Merzbacher disease (PMD) is a leukodystrophy linked to the proteolipid protein gene (PLP). We report a cellular basis for the distinction between two disease subtypes, classical and connatal, based on protein trafficking of the two PLP gene products (PLP and DM20). Classical PMD mutations correlate with accumulation of PLP in the ER of transfected COS-7 cells while the cognate DM20 traverses the secretory pathway to the cell surface. On the other hand, connatal PMD mutations lead to the accumulation of both mutant PLP and DM20 proteins in the ER of COS-7 cells with little of either isoform transported to the cell surface. Moreover, we show that transport-competent mutant DM20s facilitate trafficking of cognate PLPs and hence may influence disease severity.

Neurofilamentous changes in select groups of neurons are associated with the degenerative changes of many human age-related neurodegenerative diseases. To examine the possible effects of aging on the neuronal cytoskeleton containing human proteins, the retinas of transgenic mice expressing the gene for the human middle-sized neurofilament triplet were investigated at 3 or 12 months of age. Transgenic mice developed tangle-like neurofilamentous accumulations in a subset of retinal ganglion cells at 12 months of age. These neurofilamentous accumulations, which also involved endogenous neurofilament proteins, were present in the perikarya and proximal processes of large ganglion cells and were predominantly located in peripheral retina. The presence of the human protein may thus confer vulnerability of the cytoskeleton to age-related alterations in this specific retinal cell type and may serve as a model for similar cellular changes associated with Alzheimer's disease and glaucoma.

In ten transgenic lines, expression of a human mid-sized (M) neurofilament (NF) transgene was restricted to neurons in the central and peripheral nervous systems. However, no two lines gave identical expression patterns and none exactly matched the expression of mouse NF(M). These varied expression patterns within the neural compartment likely result from interactions of the transgene with enhancer elements located in the regions flanking the insertion site. Unexpected patterns of enhancer activity included an enhancer active in subsets of cerebellar basket cells as well as others preferentially active in subsets of motor or sensory neurons.

The primary structure of the proteolipid protein (PLP) from the central nervous system (CNS) myelin of mammals is highly conserved with only three amino acid differences between the mouse, rat, dog, bovine, and human proteins. Furthermore, within a particular species no polymorphisms in the protein have been identified. Recent interest has focused on the targeting of PLP in oligodendrocytes and the role that mutant forms of this protein play in generating dysmyelinating or hypomyelinating diseases. We previously expressed the human cDNA encoding PLP in transiently transfected Cos-7 cells and characterized the subcellular distribution of the protein in this simple heterologous system. In the current study we have used the same paradigm to examine the effect of five missense mutations in the PLP gene on processing of the encoded protein. The mutations chosen span the carboxy-terminal half of PLP and encompass that part of the protein in which most mutations have been identified. Our results show that transport of all mutations examined was arrested in the secretory pathway at an early stage, causing the mutant proteins to accumulate in the endoplasmic reticulum. Thus, a common mechanism of protein misfolding and failure of PLP to reach the cell surface of oligodendrocytes rather than the inability of the mutant protein to perform some crucial function at the cell surface may be responsible for the diseases caused by many PLP mutations. Our results, together with those of others, prompt us to speculate that the pathobiology observed in PLP mutants may result from oligodendrocyte cell death caused by the accumulation of misfolded protein in the endoplasmic reticulum. This speculation is consistent with the observations that oligodendrocytes bearing misfolded PLP, as in the jimpy mutant, proliferate but die rapidly while oligodendrocytes from PLP deletion survive and produce a myelin-like membrane which lacks PLP. © 1994 Wiley-Liss, Inc.

Transgenic mice of line MβP3 express bacterial β-galactosidase in oligodendrocytes but not other cells of the CNS. The marker enzyme, demonstrated histochemically or by immunostaining in oligodendrocyte cell bodies and along myelin internodes, appears at the time of myelination and persists thereafter; in transplantation experiments, the marker may serve to indicate both the source of particular cells and their state of differentiation. The subventricular zone of the lateral ventricle, grafted from transgenic to wild-type perinatal recipient mice, yields histochemically labeled oligodendrocytes in surrounding host tissue. When grafts are placed in cerebral cortex near callosal radiations, graft-derived oligodendrocytes are found in cerebral cortex and subcortical white matter as far as 1.5 mm from the site of implant but not in nearby caudoputamen. This study is the first to document differentiation of transplant-derived oligodendrocytes in normal developing CNS. Our results are consistent with the well- established notion that oligodendrocyte progenitors migrate during normal development and suggest that such migration might be guided or restricted by mechanisms yet to be identified.

By a combination of DNase I footprinting, methylation interference, and gel shift analyses we have identified multiple binding sites for nuclear proteins within the promoter region of the human neurofilament H gene. Two sites likely bind the transcription factor Sp1 while two others may be targets for previously unrecognized DNA binding proteins. One site, PAL, occurs within the 10 bp sequence GGGGAGGAGG. Two copies of the PAL sequence form an interrupted palindrome around one of the Sp1 sites. A second site, PROX, is found within the sequence GGTTGGACC. Nuclear extracts prepared from both neural and non-neural cell lines, mouse brain, and mouse liver contain proteins that recognize and bind to the PROX and PAL sequences indicating that proteins which bind to these target sequences are widespread. The appearance of these target sequences in the 5′ upstream region of several neuron specific genes suggests that they play key roles in the transcription of neuron specific genes. The functional activity of these target DNA sequences was demonstrated by transfection assays using a reporter gene fused to nested deletions of the NF(H) promoter region. Interestingly, these assays revealed that maximal transient expression was obtained with DNA fusion genes containing the PAL, PROX and TATA sequences. Inclusion of the Sp1 sites into the fusion genes failed to enhance the expression of the reporter gene. To determine if the NF(H) promoter can be activated in a tissue specific manner during development transgenic mice containing the promoter region linked to a β-galactosidase reporter gene were generated. In one line sporadic expression of the transgene occurred in the CNS and testis while in four other lines no expression occurred. Collectively these results suggest that the NF(H) gene promoter is active in a tissue specific manner only by interactions with regulatory elements that lie further upstream or downstream of the start site of initiation.

The human neurofilament (H) promoter contains multiple binding sites for nuclear proteins including a Proximal (Prox) site centered around the sequence GGTTGGACC and an adjacent pyrimidine (Pyr) tract site centered around the sequence CCCTCCTCCCC. Surprisingly binding to a probe containing the Prox/Pyr region of the NF(H) promoter was competed in gel shifts by an oligonucleotide containing only an Sp1 binding site (GGGGCGGGG). Supershift assays with a polyclonal anti-Sp1 antisera confirmed that Sp1 was part of the complex formed with the Prox/Pyr probe. However neither bacterially expressed Sp1 516C or vaccinia virus expressed full-length Sp1 778C bound to the Prox or Pyr sequences in DNase I footprints or gel shift assays. Gel shift competitions and supershift assays with probes containing either Prox or Pyr tract sites alone demonstrated targeting of Sp1 to the Prox binding site and identified a non-Sp1 containing complex which contains a Prox binding protein. Adding exogenous Sp1 to a HeLa nuclear extract enhanced the Sp1-containing complex but had no effect on the Prox complex. These studies show that Sp1 can be targeted to a non-Sp1 site in the human NF(H) promoter through protein/protein interactions with a distinct sequence specific DNA-binding protein.

Brief ectopic expression of Gcm1 in mouse embryonic tail bud profoundly affects the development of the nervous system. All mice from 5 independently derived transgenic lines exhibited either one or both of two types of congenital spinal cord pathologies: failure of the neural tube to close (spina bifida) and multiple neural tubes (diastematomyelia). Because the transgene is expressed only in a restricted caudal region and only for a brief interval (E8.5 to E13.5), there was no evidence of embryonic lethality. The dysraphisms develop during the period and within the zone of transgene expression. We present evidence that these dysraphisms result from an inhibition of neuropore closure and a stimulation of secondary neurulation. After transgene expression ceases, the spinal dysraphisms are progressively resolved and the neonatal animals, while showing signs of scarring and tissue resorption, have a closed vertebral column. The multiple spinal cords remain but are enclosed in a single spinal column as in the human diastematomyelia. The animals live a normal life time, are fertile and do not exhibit any obvious weakness or motor disabilities.

We used the yeast two-hybrid assay to identify novel proteins that interact with the D1dopamine receptor. The third cytoplasmic loop (residues 217–273) of the rat D1receptor was used as bait to identify clones encoding interacting proteins from a rat brain cDNA library. This identified two clones encoding the C terminus of rat neurofilament-M (NF-M) (residues 782–846). The NF-M clone did not interact with the third cytoplasmic loops of the rat D2, D3, or D4receptors, but showed weak interaction with that of the D5receptor. Coexpression of full-length NF-M with the D1receptor in HEK-293 cells resulted in >50% reduction of receptor binding accompanied by a reduction in D1receptor-mediated cAMP accumulation. NF-M had no effect on the expression of other dopamine receptor subtypes. Using a D1receptor-green fluorescent protein chimera and confocal fluorescence microscopy, we found that NF-M reduced D1receptor expression at the cell surface and promoted accumulation of the receptor in the cytosol. Interestingly, the D1receptors that were expressed at the cell surface in the presence of NF-M were resistant to agonist-induced desensitization. Cellular colocalization of NF-M and the D1receptor in the rat brain was examined by epifluorescence microscopy. These experiments showed that ∼50% of medium-sized striatal neurons expressed both proteins. Colocalization was also observed in pyramidal cells and interneurons within the frontal cortex. Similar immunohistochemical analyses using NF-M-deficient mice showed decrements in D1receptor expression compared with control mice. These results suggest that NF-M interacts with the D1receptorin vivoand may modify its expression and regulation.

Schwann cells and oligodendrocytes produce myelin sheaths of widely varying sizes. How these cells determine the size of myelin sheath for a particular axon is incompletely understood. Axonal diameter has long been suspected to be a signal in this process. We have analyzed myelin sheath thickness in L5 lumbar root and spinal cord white matter of a series of mouse mutants with diminished axonal calibers resulting from a deficiency of neurofilaments (NFs). In the PNS, average axonal diameters were reduced by 20 ‐37% in the NF mutants. Remarkably, the average myelin sheath thickness remained unchanged from control values, and regression analysis showed sheaths abnormally thick for a given size of axon. These data show that a genetically induced reduction in axonal caliber does not cause a reduction in myelin sheath thickness in PNS and indicate that Schwann cells read some intrinsic signal on axons that can be uncoupled from axonal diameter. Interestingly, myelin sheaths in the spinal cord of these animals were not abnormally thick, arguing that axonal diameter may contribute directly to the regulation of myelination in the CNS and that oligodendrocytes and Schwann cells use different cues to set myelin sheath thickness. J. Neurosci. Res. 65:493‐ 499, 2001. © 2001 Wiley-Liss, Inc.

Trophoblast cells of the placenta are established at the blastocyst stage and differentiate into specialized subtypes after implantation1,2. In mice, the outer layer of the placenta consists of trophoblast giant cells that invade the uterus and promote maternal blood flow to the implantation site by producing cytokines with angiogenic3 and vasodilatory4 actions. The innermost layer, called the labyrinth, consists of branched villi that provide a large surface area for nutrient transport and are composed of trophoblast cells and underlying mesodermal cells derived from the allantois. The chorioallantoic villi develop after embryonic day (E) 8.5 through extensive folding and branching of an initially flat sheet of trophoblast cells, the chorionic plate, in response to contact with the allantois. We show here that Gcm1, encoding the transcription factor glial cells missing-1 (Gcm1), is expressed in small clusters of chorionic trophoblast cells at the flat chorionic plate stage and at sites of chorioallantoic folding and extension when morphogenesis begins. Mutation of Gcm1 in mice causes a complete block to branching of the chorioallantoic interface, resulting in embryonic mortality by E10 due to the absence of the placental labyrinth. In addition, chorionic trophoblast cells in Gcm1-deficient placentas do not fuse to form syncytiotrophoblast. Abnormal development of placental villi is frequently associated with fetal death and intrauterine growth restriction in humans, and our studies provide the earliest molecular insight into this aspect of placental development.

Although gcm was first recognized for its role in specifying glial cell fate in Drosophila melanogaster, its mammalian counterparts are expressed predominantly in non-neural tissues. Here we demonstrate expression of the mouse and human GCM 1 proteins in placenta. We have prepared a highly specific antibody that recognizes the GCM 1 protein and have used it to assess the temporal and spatial expression profile of the protein. In both mouse and human placenta, the protein is associated with cells that are involved with exchange between maternal and fetal blood supplies: the labyrinthine cells of the mouse placenta and the syncytio- and cytotrophoblasts of the human placenta. Using the full-length h Gcm 1 cDNA as a probe, we have mapped the gene on human chromosome 6p12 by fluorescent in situ hybridization.

The 129 mouse strain has become of increasing interest to neurobiologists due to its importance in gene targeting studies. However it has been pointed out that 129 mice suffer from a number of neuroanatomical idiosyncrasies that may make them less attractive as animal models in neurobiology. Here we show that 129 mice also differ from other commonly used strains in possessing large numbers of unmyelinated axons in their lumbar motor roots. By contrast in all other strains of mice (C57BL/6, C3H, Swiss-Webster) that we studied the axons in the L5 roots are all myelinated. Additionally we show that 129 mice have smaller myelinated axons than other mouse strains and perform poorly in the rotorod test. These characteristics must be kept in mind in studies of mutant mice that are frequently performed on a mixed genetic background containing a129 contribution.

Oligodendrocyte-specific protein (OSP)/claudin-11 is a recently identified transmembrane protein found in CNS myelin and testis with unknown function. Herein we demonstrate that Osp null mice exhibit both neurological and reproductive deficits: CNS nerve conduction is slowed, hindlimb weakness is conspicuous, and males are sterile. Freeze fracture reveals that tight junction intramembranous strands are absent in CNS myelin and between Sertoli cells of mutant mice. Our results demonstrate that OSP is the mediator of parallel-array tight junction strands and distinguishes this protein from other intrinsic membrane proteins in tight junctions. These novel results provide direct evidence of the pivotal role of the claudin family in generating the paracellular physical barrier of tight junctions necessary for spermatogenesis and normal CNS function.

Neurofilaments (NFs) are prominent components of large myelinated axons and probably the most abundant of neuronal intermediate filament proteins. Here we show that mice with a null mutation in the mid-sized NF (NF-M) subunit have dramatically decreased levels of light NF (NF-L) and increased levels of heavy NF (NF-H). The calibers of both large and small diameter axons in the central and peripheral nervous systems are diminished. Axons of mutant animals contain fewer neurofilaments and increased numbers of microtubules. Yet the mice lack any overt behavioral phenotype or gross structural defects in the nervous system. These studies suggest that the NF-M subunit is a major regulator of the level of NF-L and that its presence is required to achieve maximal axonal diameter in all size classes of myelinated axons.

Previous studies have shown that rodent neurofilaments (NF) are obligate heteropolymers requiring NF-L plus either NF-M or NF-H for filament formation. We have assessed the competence of human NF-L and NF-M to assemble and find that unlike rat NF-L, human NF-L is capable of self-assembly. However, human NF-M cannot form homopolymers and requires the presence of NF-L for incorporation into filaments. To investigate the stage at which filament formation is blocked, the rod domains or the full-length subunits of human NF-L, human NF-M, and rodent NF-L were analyzed in the yeast "interaction trap" system. These studies demonstrated that the fundamental block to filament formation in those neurofilaments that do not form homopolymers is at the level of dimer formation. Based on theoretical biophysical considerations of the requirements for the formation of coiled-coil structures, we predicted which amino acid differences were likely to be responsible for the differing dimerization potentials of the rat and human NF-L rod domains. We tested these predictions using site-specific mutagenesis. Interestingly, single amino acid changes in the rod domains designed to restore or eliminate the coiled-coil propensity were found respectively to convert rat NF-L into a subunit capable of homopolymerization and human NF-L into a protein that is no longer able to self-assemble. Our results additionally suggest that the functional properties of the L12 linker region of human NF-L, generally thought to assume an extended beta-sheet conformation, are consonant with an alpha-helix that positions the heptad repeats before and after it in an orientation that allows coiled-coil dimerization. These studies reveal an important difference between the assembly properties of the human and rodent NF-L subunits possibly suggesting that the initiating events in neurofilament assembly may differ in the two species.

Pelizaeus-Merzbacher disease (PMD) is a dysmyelinating disease resulting from mutations, deletions, or duplications of the proteolipid protein (PLP) gene. Distinguishing features of PMD include pleiotropy and a range of disease severities among patients. Previously, we demonstrated that, when expressed in transfected fibroblasts, many naturally occurring mutant PLP alleles encode proteins that accumulate in the endoplasmic reticulum and are not transported to the cell surface. In the present communication, we show that oligodendrocytes in an animal model of PMD, the msd mouse, accumulate Plp gene products in the perinuclear region and are unable to transport them to the cell surface. Another important aspect of disease in msd mice is oligodendrocyte cell death, which is increased by two- to threefold. We demonstrate in msd mice that this death occurs by apoptosis and show that at the time oligodendrocytes die, they have differentiated, extended processes that frequently contact axons and are expressing myelin structural proteins. Finally, we define a hypothesis that accounts for pathogenesis in most PMD patients and animal models of this disease and, moreover, can be used to develop potential therapeutic strategies for ameliorating the disease phenotype.

To understand the function of the myelin oligodendrocyte glycoprotein (MOG), a myelin specific protein of the central nervous system, transgenic mice were produced. The transgene is a fusion gene containing 1.9 kb of murine myelin basic protein promoter, 430 bp of rat MOG cDNA in the reverse orientation and 4.5 kb of human proteolipid protein gene. In spite of high expression of antisense MOG mRNA in the oligodendrocytes, MOG synthesis was not inhibited in transgenic mice. This lack of inhibition of MOG underlines the difficulties encountered with antisense transgenic strategies.

qkI, a newly cloned gene lying immediately proximal to the deletion in thequakingviablemutation, is transcribed into three messages of 5, 6, and 7 kb. Antibodies raised to the unique carboxy peptides of the resulting QKI proteins reveal that, in the nervous system, all three QKI proteins are expressed strongly in myelin-forming cells and also in astrocytes. Interestingly, individual isoforms show distinct intracellular distributions: QKI-6 and QKI-7 are localized to perikaryal cytoplasm, whereas QKI-5 invariably is restricted to the nucleus, consistent with the predicted role of QKI as an RNA-binding protein. Inquakingviablemutants, which display severe dysmyelination, QKI-6 and QKI-7 are absent exclusively from myelin-forming cells. By contrast, QKI-5 is absent only in oligodendrocytes of severely affected tracts. These observations implicate QKI proteins as regulators of myelination and reveal key insights into the mechanisms of dysmyelination in thequakingviablemutant.

Previous studies have revealed that the transgenic mouse line expressing the human neurofilament-mid-sized (NF-M) gene evidences age-dependent and cell-specific pathological neurofibrillary accumulation in the central nerve system. In the current study, we investigated the learning and memory processes of NF-M transgenic mice at 3 and 8 months of age in a modified Morris water maze using a series of tasks including those primarily related to reference memory (i.e., spatial learning, reversal learning and probe trials) and to working memory (i.e., matching to sample tasks with or without delays). At 3 months of age, NF-M transgenic mice were indistinguishable from age- and litter-matched non-transgenic wild-type controls on any of the tests of reference and working memory. At 8 months of age, however, the NF-M transgenic mice exhibited significantly poorer performance than the age- and litter-matched wild-type control mice on both reference and working memory tasks. Immunohistohistological study of the brains of the 8-month-old NF-M transgenic mice revealed spherical and tangle-like neurofilamentous accumulation in their cerebral cortices. These results suggest that NF-M transgenic mice express both age-related histopathological changes and age-dependent learning and memory deficits. Whether NF-M transgenic mice exhibit even more severe behavioral impairments when they become aged is currently under study.

Alterations in neurofilaments are a common occurrence in neurons of the human nervous system during aging and diseases associated with aging. Such pathologic changes may be attributed to species-specific properties of human neurofilaments as well as cell-type-specific regulation of this element of the cytoskeleton. The development of transgenic animals containing human neurofilament subunits offers an opportunity to study the effects of aging and other experimental conditions on the human-specific form of these proteins in a rodent model. The present study shows that mice from the transgenic line NF(M)27, which express the human midsized neurofilament subunit at low levels (2–25% of the endogenous NF-M), develop neurofilamentous accumulations in specific subgroups of neurons that are age dependent, affecting 78% of transgenic mice over 12 months of age. Similar accumulations do not occur in age-matched, wild-type littermates or in 3-month-old transgenic mice. In 12-month-old transgenic mice, somatic neurofilament accumulations resembling neurofibrillary tangles were present predominantly in layers III and V of the neocortex, as well as in select subpopulations of subcortical neurons. Intraperikaryal, spherical neurofilamentous accumulations were particularly abundant in cell bodies in layer II of the neocortex, and neurofilament-containing distentions of Purkinje cell proximal axons occurred in the cerebellum. These pathological accumulations contained mouse as well as human NF subunits, but could be distinguished by their content of phosphorylation-dependent NF epitopes. These cytoskeletal alterations closely resemble the cell-type-specific alterations in neurofilaments that occur during normal human aging and in diseases associated with aging, indicating that these transgenic animals may serve as models of some aspects of the pathologic features of human neurodegenerative diseases.

Delineating the properties and functions of the major central nervous system myelin proteins has been the focus of intensive research for decades. For PLP, this task has been confounded by its unusual properties, the complexity of the cellular membrane in which it resides, and the absence of a functional assay for the protein. The development of new experimental paradigms in which to study PLP may shed fresh light on the properties and functions of this intrinsic membrane protein. In the present communication we have used indirect, double label, immunofluorescence, and confocal microscopy to examine the distribution of PLP in Cos-7 cells transfected with an expression vector bearing the human PLP cDNA. Our results show that PLP is synthesized in the rough endoplasmic reticulum of transfected cells and passes through the Golgi apparatus to the cell surface. These results are consistent with previous studies showing PLP reaches the cell surface by transport through the secretory pathway. Levels of PLP at the cell surface are modest, most likely because protein deposited in this compartment can be endocytosed and subsequently transported to perinuclear lysosomes. Similar results are reported in the companion communication by Sinoway et al. (J Neurosci Res, 37:551-562, 1994). Using transfected HeLa cells they show that DM20 alone and PLP coexpressed with DM20 assume appropriate conformations for transport to the cell surface. The presence of PLP in subcellular compartments beyond the endoplasmic reticulum in Cos-7 cells indicates that the protein achieves a conformation appropriate for transport in the absence of other oligodendrocyte-specific factors; however, accumulation of large amounts of PLP in the cytoplasmic membrane compartment may require interactions with such glial-specific factors. Thus, the transfection paradigm described herein should prove a useful tool for investigating the folding and sorting of wild type and mutant forms of PLP as well as its membrane topology and posttranslational processing.

Bacterial beta-galactosidase is widely used as a marker for gene expression and in cell tracing experiments. In a survey of three transgenic mouse lines expressing beta-galactosidase in the central nervous system (CNS) under the control of different promoters, we find substantial variation in the intracellular distribution of the lacZ protein. In line M beta P5, transgene beta-galactosidase expression is driven by a promoter/enhancer fragment from the oligodendrocyte-specific myelin basic protein gene; however, electron microscopy of histochemically stained preparations reveals transgene expression not only in oligodendrocytes but also in some neurons. Immunofluorescence and immunoperoxidase staining show the beta-galactosidase protein distributed throughout the perikaryal cytoplasm of oligodendrocytes and in processes reaching to myelin sheaths. By contrast, immunoreactive protein appears restricted in neurons to one or a few small perikaryal immunoreactive granules. The granules are visible in the electron microscope as amorphous inclusion bodies of moderate electron density and lack a limiting membrane. Histochemical staining patterns with X-gal and Bluo-gal echoed the protein distribution: diffuse distribution of enzyme protein yielded cells filled with substrate, while punctate enzyme distribution yielded restricted or punctate histochemical staining. Examination of two other lines using different promoter/enhancers to drive expression in the CNS showed both diffuse and punctate beta-galactosidase immunolocalization and histochemical staining. The amount of protein synthesized or other properties, yet unidentified, intrinsic to the target cells may determine the intracellular distribution of beta-galactosidase. In retroviral marking studies, clone members have been identified as those cells filled with X-gal reaction product. This approach may underestimate both clone size and the minimum number of divisions separating the members of each clone.

In order to identify potential regulatory elements of the human mid-sized (M) neurofilament (NF) gene we preformed DNase I footprinting, gel mobility shift assays and methylation interference studies with probes from the NF(M) immediate 5′ flanking region. These studies identified multiple sites for DNA-binding proteins including four Sp1 sites, and single sites each for members of the NF-1 and AP-1 families of DNA binding proteins. In addition a binding site within a pyrimidine tract likely binds a novel DNA-binding protein which also interacts with the human NF(H) gene promoter. Factors that bind to these sites are found in both neural and non-neural cells suggesting that the NF(M) promoter may not contain tissue specific regulatory signals. In transient assays, addition of these binding sites to an NF(M) minimal promoter containing only a TATA box lead to a greater than 40-fold activation of transcription over background. Progressive 5′ deletions reduced expression in a step wise manner suggesting that all the factors likely act synergistically as positive regulators of transcription.

The induction of autoantibodies and their possible role in the pathogenesis of autoimmune disease are poorly understood. Involvement of infectious agents has been suspected, but direct evidence is sparse. Whether immunological unresponsiveness to self by antibody-forming B cells is maintained by clonal abortion, clonal anergy or suppression, or how the scenario of interactions between helper T cells, B cells and antigen-presenting cells is distorted in autoantibody responses, is being analysed and widely debated. To evaluate tolerance of neutralizing B-cell responses we used transgenic mice expressing the cell membrane associated glycoprotein (G) of vesicular stomatitis virus (VSV) as self-antigen. We show that autoantibodies to VSV-G cannot be induced by VSV-G in adjuvant or by recombinant vaccinia virus expressing VSV-G, but are triggered by infection with wild-type VSV. The data show that helper T-cell tolerance is crucial in maintenance of B-cell non-reactivity and that cognate T-B recognition is necessary to break tolerance of self-reactive B cells. These results may help to understand mechanisms of virus-induced autoimmunity.

Despite differences in size, genetic content, and transcribing abilities, four types of defective interfering particles (DIs) interfered with the replication of the homologous (Indiana) helper virus to the same extent. Similarly, no significant difference was observed when the abilities of the DIs to be replicated with the Indiana helper were compared. However, when the New Jersey (Hazlehurst) strain was employed as a helper virus, significant differences in DI replication were observed. Three of the four particles were replicated as well as or better than in the completely homologous system; the fourth was replicated only to a very limited extent. In none of the infections with the Hazlehurst strain of VSV(NJ) was there detectable interference, which was measured as a reduction in the number of progeny infectious particles. Thus, depending on the helper virus used, these particles can be classified as defective interfering (DI) or defective noninterfering (DNI) particles. This curious result was obtained not only with the Hazlehurst strain but also with the Prevec strain of VSV(NJ).

Serial subculture of Vero cells infected with the chick embryo adapted Enders strain of mumps virus gives rise to a low level productive though persistent infection. Persistently infected cultures exhibit minimal cytopathology; however, widely dispersed foci of multinucleate cells are almost always present. Neither infected cell nor virus growth is temperature sensitive. Biological and biochemical evidence indicate that defective interfering particles are replicated along with infectious (nondefective) virus during the course of the persistent infection, although the plaque-purified inoculum virus stock contained only genome size RNA. With continued cell passage a heterogeneous, changing population of subgenomic sized viral RNA accumulates, suggesting that defective interfering (DI) RNA species are evolving from virion RNA, with no single DI RNA predominating. Since such factors as interferon, antibody, and temperature-sensitive mutants are absent from this system, DI particles are the likely factor modulating this persistent infection.

The 3′ termini of RNA from vesicular stomatitis virus and from three widely dissimilar defective interfering particles of the virus are PypGpU-OH. The possible relevance of these findings to autointerference and to replication of vesicular stomatitis virus is discussed.

The base sequence at the 3' end of vesicular stomatitis virus RNA was determined by using terminal labels and chemical RNA sequencing. The leader RNA was complementary to 47 bases at the 3' terminus, whereas the nucleocapsid gene (N) began 51 nucleotides from the 3' end of the genomic RNA. The intervening bases were 3'...GAAA...5' for the Indiana serotype and 3'...GAAAA...5' for the New Jersey serotype. The complements of these bases did not appear in either the leader RNA or the N mRNA. This sequence may function as a stop signal or cleavage site during transcription. Furthermore, processing or termination at this sequence must be inhibited during the production of full-length RNA plus-sense strands (replication). We recently found similar sequences approximately 46 to 48 nucleotides from the 3' ends of several defective interfering particle RNAs where the short defective interfering particle transciption products terminate. This sequence is present also at the end of the polymerase (L) gene.

The RNA isolated from a new defective interfering particle of vesicular stomatitis virus is 80 to 90% double-stranded. Annealing and electron microscopic analyses indicate that the complementary sequences of this duplex are covalently linked in a hairpin structure. Approximately half of the sequences found in this RNA are normal genomic sequences found in the RNA of the infectious virus. The other half, the complement of the former, are message sequences that are not found in the infectious virus. Within the particle, the RNA exists as a single-stranded ribonucleoprotein complex, 0·85 μm in length. Upon deproteinization the RNA collapses into a self-complementary double stranded duplex, 0·33 μm in length. The possible relevance of covalently linked message and genomic sequences to the mechanism of vesicular stomatitis virus replication is discussed.

We have isolated cDNA clones encoding the four different forms of mouse myelin basic protein (MBP) and have analyzed the structure of the MBP gene. The three larger forms of MBP differ from the smallest by the inclusion of either or both of two short amino acid sequences at positions 57 and 124 of the smallest protein. The mouse genome contains a single MBP gene comprised of seven exons. The two amino acid sequences present only in the larger MBPs are encoded by separate exons. Furthermore, all exons in the coding region begin or end in complete codons so that alternative splicing does not alter the reading frame. We conclude that the four forms of this myelin protein are encoded in separate mRNAs, each derived by a simple alternative splicing of the primary MBP gene transcript. Comparison of the amino acid sequence encoded by each exon with a recent model of the secondary structure of MBP suggests that each of the seven exons encodes one or two of the predicted structural motifs of the protein.

The 5'-terminal nucleotide sequence from positions 50 to 130 of vesicular stomatitis virus RNA was determined indirectly by using a defective interfering particle RNA which contains covalently linked genomic minus and antigenomic plus sense RNAs. The last 18 nucleotides of the L gene coding for in the viral polymerase were identified and isolated by specific duplex formation between 5' terminally labeled oligonucleotides from a small single-stranded defective interfering particle RNA and L gene mRNA. The L gene ends at position 60 from the 5' terminus of the vesicular stomatitis genome. The data demonstrated that the first seven adenine residues in the polyadenylic acid tail of L gene mRNA may be coded for in the genome and suggested that the viral transcriptase itself may carry out polyadenylation, possibly by chattering at the uridine-rich sequence at the end of the L gene. Analysis of the 5'-terminal sequence of vesicular stomatitis virus genomic RNA revealed that it might fold into a complex secondary structure with possibly 62% of the bases paired.

The 28 and 13–18 S messenger RNA classes of vesicular stomatitis virus contain the same number of methyl groups, all of which were present at the blocked 5′-ends of the molecules. This suggests that the low efficiency of translation of the 28 S mRNA relative to that of the 13 S mRNA is not due to undermethylation.

The levels of intracellular viral RNA have been measured in BHK-21 cells undergoing standard and autointerfered infections with vesicular stomatitis virus (VSV). The total amount of intracellular nucleocapsid RNA is almost unaffected by autointerference. However, the composition of this class of RNA changes with increasing autointerference: Control cells undergoing standard infections contain only the 140 S, full genomic length nucleocapsid, while autointerfered cells contain, in addition, a smaller nucleocapsid which corresponds in size to that found in the defective interfering (DI) particle. There is a systematic decrease in viral mRNA with increasing autointerference. This decrease in mRNA closely parallels the reduced yield of total progeny particles eventually released from these cells. A causal relationship between the two is proposed. These results are compatible with explanations of autointerference that depend upon a competition between the DI and VSV genomes for the viral replicase.

Two temperature-sensitive (ts) mutants of the M protein of vesicular stomatitis virus (tsG31 and tsG33) are defective in viral assembly, but the exact nature of this defect is not known. When infected cells are switched from nonpermissive (40 degrees C) to permissive (32 degrees C) temperatures in the presence of cycloheximide, tsG33 virus release increased by 100-fold, whereas tsG31 release increased only by 10-fold. Thus, the tsG33 defect is more reversible than that of tsG31. Therefore, we investigated how the altered synthesis and cellular distribution of tsG33 M protein correlates with the viral assembly defect. At 32 degrees C tsG33 M protein is stained diffusely in the cell cytoplasm and later at the budding sites. In contrast, at 40 degrees C the mutant M protein formed unusual aggregates mostly located in the perinuclear regions of virus-infected cells and partially colocalized with G protein in this region. In temperature shift-down experiments, M can be disaggregated and used to some extent for nucleocapsid coiling and budding, which correlates with the virus titer increase. M aggregates also formed after shift-up from 32 to 40 degrees C, indicating a complete dependence of M aggregation on the temperature. Biochemical analysis with sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotting revealed that at 40 degrees C M protein is detected exclusively in pellet fractions (nuclear and cytoskeleton components), whereas at 32 degrees C M protein is mainly in the cytoplasmic soluble fractions. Furthermore, when the temperature is raised from 32 to 40 degrees C, the distribution of M protein tends to shift from the soluble to the pellet and cytoskeletal fractions. Electron micrographs of immunoperoxidase-labeled M protein showed that at 40 degrees C M aggregates are often associated with the outer nuclear membranes as well as with vesicular structures. No nucleocapsid coiling was observed in these cells, whereas coiling and budding were seen at 32 degrees C in cells where M protein was partly associated with the plasma membrane. We suggest that the tsG33 M protein mutation may produce a reversible conformational alteration which causes M protein to aggregate at 40 degrees C, therefore inhibiting the proper association of M protein with nucleocapsids and budding membranes.

The nucleotide sequences encoding the matrix (M) proteins of measles virus (MV) and canine distemper virus (CDV) were determined from cDNA clones containing these genes in their entirety. In both cases, single open reading frames specifying basic proteins of 335 amino acid residues were predicted from the nucleotide sequences. Both viral messages were composed of approximately 1,450 nucleotides and contained 400 nucleotides of presumptive noncoding sequences at their respective 3' ends. MV and CDV M-protein-coding regions were 67% homologous at the nucleotide level and 76% homologous at the amino acid level. Only chance homology was observed in the 400-nucleotide trailer sequences. Comparisons of the M protein sequences of MV and CDV with the sequence reported for Sendai virus (B. M. Blumberg, K. Rose, M. G. Simona, L. Roux, C. Giorgi, and D. Kolakofsky, J. Virol. 52:656-663; Y. Hidaka, T. Kanda, K. Iwasaki, A. Nomoto, T. Shioda, and H. Shibuta, Nucleic Acids Res. 12:7965-7973) indicated the greatest homology among these M proteins in the carboxyterminal third of the molecule. Secondary-structure analyses of this shared region indicated a structurally conserved, hydrophobic sequence which possibly interacted with the lipid bilayer.