Optimal Convergence Rates Results for Linear Inverse Problems in Hilbert Spaces.

In this article, we prove optimal convergence rates results for regularization methods for solving linear ill-posed operator equations in Hilbert spaces. The results generalizes existing convergence rates results on optimality to general source conditions, such as logarithmic source conditions. Moreover, we also provide optimality results under variational source conditions and show the connection to approximative source conditions.

Seismic imaging of transition zone discontinuities suggests hot mantle west of Hawaii.

The Hawaiian hotspot is often attributed to hot material rising from depth in the mantle, but efforts to detect a thermal plume seismically have been inconclusive. To investigate pertinent thermal anomalies, we imaged with inverse scattering of SS waves the depths to seismic discontinuities below the Central Pacific, which we explain with olivine and garnet transitions in a pyrolitic mantle. The presence of an 800- to 2000-kilometer-wide thermal anomaly (ΔT(max) ~300 to 400 kelvin) deep in the transition zone west of Hawaii suggests that hot material does not rise from the lower mantle through a narrow vertical plume but accumulates near the base of the transition zone before being entrained in flow toward Hawaii and, perhaps, other islands. This implies that geochemical trends in Hawaiian lavas cannot constrain lower mantle domains directly.

Seismostratigraphy and thermal structure of Earth's core-mantle boundary region.

We used three-dimensional inverse scattering of core-reflected shear waves for large-scale, high-resolution exploration of Earth's deep interior (D'') and detected multiple, piecewise continuous interfaces in the lowermost layer (D'') beneath Central and North America. With thermodynamic properties of phase transitions in mantle silicates, we interpret the images and estimate in situ temperatures. A widespread wave-speed increase at 150 to 300 kilometers above the coremantle boundary is consistent with a transition from perovskite to postperovskite. Internal D'' stratification may be due to multiple phase-boundary crossings, and a deep wave-speed reduction may mark the base of a postperovskite lens about 2300 kilometers wide and 250 kilometers thick. The core-mantle boundary temperature is estimated at 3950 +/- 200 kelvin. Beneath Central America, a site of deep subduction, the D'' is relatively cold (DeltaT = 700 +/- 100 kelvin). Accounting for a factor-of-two uncertainty in thermal conductivity, core heat flux is 80 to 160 milliwatts per square meter (mW m(-2)) into the coldest D'' region and 35 to 70 mW m(-2) away from it. Combined with estimates from the central Pacific, this suggests a global average of 50 to 100 mW m(-2) and a total heat loss of 7.5 to 15 terawatts.

EEA1, a tethering protein of the early sorting endosome, shows a polarized distribution in hippocampal neurons, epithelial cells, and fibroblasts.

EEA1 is an early endosomal Rab5 effector protein that has been implicated in the docking of incoming endocytic vesicles before fusion with early endosomes. Because of the presence of complex endosomal pathways in polarized and nonpolarized cells, we have examined the distribution of EEA1 in diverse cell types. Ultrastructural analysis demonstrates that EEA1 is present on a subdomain of the early sorting endosome but not on clathrin-coated vesicles, consistent with a role in providing directionality to early endosomal fusion. Furthermore, EEA1 is associated with filamentous material that extends from the cytoplasmic surface of the endosomal domain, which is also consistent with a tethering/docking role for EEA1. In polarized cells (Madin-Darby canine kidney cells and hippocampal neurons), EEA1 is present on a subset of "basolateral-type" endosomal compartments, suggesting that EEA1 regulates specific endocytic pathways. In both epithelial cells and fibroblastic cells, EEA1 and a transfected apical endosomal marker, endotubin, label distinct endosomal populations. Hence, there are at least two distinct sets of early endosomes in polarized and nonpolarized mammalian cells. EEA1 could provide specificity and directionality to fusion events occurring in a subset of these endosomes in polarized and nonpolarized cells.

Intracellular routing of wild-type and mutated polymeric immunoglobulin receptor in hippocampal neurons in culture.

Certain epithelial cells synthesize the polymeric immunoglobulin receptor (pIgR) to transport immunoglobulins (Igs) A and M into external secretions. In polarized epithelia, newly synthesized receptor is first delivered to the basolateral plasma membrane and is then, after binding the Ig, transcytosed to the apical plasma membrane, where the receptor-ligand complex is released by proteolytic cleavage. In a previous work (Ikonen et al., 1993), we implied the existence of a dendro-axonal transcytotic pathway for the rabbit pIgR expressed in hippocampal neurons via the Semliki Forest Virus (SFV) expression system. By labeling surface-exposed pIgR in live neuronal cells, we now show (a) internalization of the receptor from the dendritic plasma membrane to the dendritic early endosomes, (b) redistribution of the receptor from the dendritic to the axonal domain, (c) inhibition of this movement by brefeldin A (BFA) and (d) stimulation by the activation of protein kinase C (PKC) via phorbol myristate acetate (PMA). In addition, we show that a mutant form of the receptor lacking the epithelial basolateral sorting signal is directly delivered to the axonal domain of hippocampal neurons. Although this mutant is internalized into early endosomes, no transcytosis to the dendrites could be observed. In epithelial Madin-Darby Canine Kidney (MDCK) cells, the mutant receptor could also be internalized into basolaterally derived early endosomes. These results suggest the existence of a dendro-axonal transcytotic pathway in neuronal cells which shares similarities with the basolateral to apical transcytosis in epithelial cells and constitute the basis for the future analysis of its physiological role.

Mechanisms of neuronal polarity.

Polarity is intrinsic to neuronal function. The somatodendritic domain receives and decodes incoming information and the axonal domain delivers information to target cells. Progressive loss of neuronal polarity is a major histopathological event in neural aging and neurodegenerative diseases, like Alzheimer's disease, preceding death and disappearance of nerve cells. Our laboratory is interested in the study of the pathways and mechanisms by which neuronal membrane polarity is established and maintained. Due to the lack of appropriate polarized neuronal cell lines for biochemical analysis, the molecular mechanisms underlying this phenomenon remain obscure. We use a neuronal culture system, hippocampal neurons from rat embryos, in which polarity is established in vitro, and the scientific rationale and experimental strategies proven useful in understanding the mechanisms of epithelial polarity. Here we review our own work on neuronal membrane polarity. The reader interested should consult any of the excellent reviews published recently (7,27,31,43).

Mapping of Ras-related GTP-binding proteins by GTP overlay following two-dimensional gel electrophoresis.

For identification of Rab, Rac, Rho, Ral, Rap, and Arf proteins on two-dimensional polyacrylamide gels, we have expressed full-length cDNAs of members of these protein families with the T7 RNA polymerase-recombinant vaccinia virus expression system. Membrane preparations from cells expressing the cDNAs were subjected to high-resolution two-dimensional polyacrylamide gel electrophoresis followed by [alpha-32P]GTP ligand blotting. We have mapped 28 small GTP-binding proteins relative to their isoelectric points and according to their molecular weights and by immunoblotting with specific antibodies. Rab and Rho proteins could be specifically identified by extraction of streptolysin O-permeabilized Madin-Darby canine kidney (MDCK) cells with Rab- and Rho-GDP dissociation inhibitor. We applied the reference mapping to analyze the GTP-binding patterns of synaptosome fractions from rat brain. The purified synaptosomes exhibited specific enrichment of Rab3a, Rab5a, Ral, and several other GTPases. This approach and the map we have produced should provide a useful aid for the analysis of the expression and localization of members of all families of small GTP-binding proteins in various cell types and subcellular fractions.

The involvement of the small GTP-binding protein Rab5a in neuronal endocytosis.

Rab5a is a small GTPase that regulates fusion of endocytic vesicles to early endosomes. We investigated whether Rab5a is involved in early endocytic traffic in both the axonal and the somatodendritic domains of polarized neurons. Using immunofluorescence, endogenous Rab5a was detected in axons and dendrites. Its localization in axons strongly overlapped that of the synaptic vesicle protein synaptophysin. Indeed, Rab5a co-immunoisolated with synaptophysin-containing vesicles, and antibodies against Rab5a labeled synaptic vesicle-like structures in nerve terminals. The functional association of Rab5a with dendritic and axonal early endosomes was assayed by electron microscopy after overexpression of wild-type and mutant Rab5a in cultured hippocampal neurons. This induced the formation of abnormal endosomes in both the somatodendritic and the axonal domains. These results show a role for Rab5a in axonal and dendritic endocytosis, and the presence of Rab5a on synaptic vesicles indicates that the axonal endosomes participate in the biogenesis of these vesicles.

Semliki Forest virus as a tool for protein expression in cultured rat hippocampal neurons.

We use the Semliki Forest Virus (SFV) as a tool for protein expression in primary cultures of rat hippocampal neurons. These cells develop in vitro into polarized neurons that can be infected with recombinant SFV with an efficiency of 60 - 80%. SFV-driven protein expression is detectable within 3-4 hours postinfection, at which time the newly-synthesized proteins are mainly present in the cell soma. By 6 to 8 hours postinfection foreign proteins are detectable in the neurites. Protein expression can continue for up to 48 hours. However, after 8 - 10 hours infected neurons start to suffer from cytopathic effects as evidenced by a change in morphology and detachment from the coverslip. The infection does not seem to affect the polarized distribution of proteins. Upon overexpression of rab8, a somatodendritic distribution is observed, similar to that of the endogenous protein. Therefore, the SFV expression system is suitable for short-term expression of proteins and can be used successfully to study the polarized distribution of heterologous proteins expressed in cultured hippocampal neurons.

Peripherin expression in hippocampal neurons induced by muscle soluble factor(s).

Previous studies have shown that neuronal cells in culture can switch neurotransmitters when grown in the presence of different target cells. To examine whether this plasticity extends to structural proteins, we cocultured hippocampal neurons and pituitary-derived neuroendocrine (AtT20) cells with astrocytes, kidney epithelial cells, or skeletal muscle cells. As a marker of phenotypic change we used the cytoskeletal protein peripherin, a type III intermediate filament (IF) subunit which is not expressed in hippocampal neurons and AtT20 cells. We show here that soluble factor(s) secreted specifically from skeletal muscle cells can induce the expression and de novo assembly of peripherin in a subset of post-mitotic neurons. We further demonstrate that one of these factors is the Leukemia Inhibitory Factor/Cholinergic Neuronal Differentiation Factor. The environmentally regulated expression of peripherin implies a remarkable degree of plasticity in the cytoskeletal organization of postmitotic CNS cells and provides a noninvasive model system to examine the de novo assembly of IF proteins under in vivo conditions.

Protein transport to the dendritic plasma membrane of cultured neurons is regulated by rab8p.

In the companion paper (Huber, L. A., S. W. Pimplikar, R. G. Parton, H. Virta, M. Zerial, and K. Simons. J. Cell Biol. 123:35-45) we reported that the small GTPase rab8p is involved in transport from the TGN to the basolateral plasma membrane in epithelia. In the present work we investigated the localization and function of rab8p in polarized hippocampal neurons. By immunofluorescence microscopy we found that rab8p localized preferentially in the somatodendritic domain, and was excluded from the axon. Double-labeling immunofluorescence showed that some of the rab8p co-localized in the dendrites with the Semliki Forest Virus glycoprotein E2 (SFV-E2). An antisense oligonucleotide approach was used to investigate the role of rab8p in dendritic transport of newly synthesized viral glycoproteins. Antisense oligonucleotides corresponding to the initiation region of the rab8 coding sequence were added to the cultured neurons for four days. This treatment resulted in a significant decrease in cellular levels of rab8p and transport of SFV-E2 from the cell body to the dendrites was significantly reduced. However, no effect was observed on axonal transport of influenza HA. From these results we conclude that rab8p is involved in transport of proteins to the dendritic surface in neurons.

Peroxisomal amine oxidase of Hansenula polymorpha does not require its SRL-containing C-terminal sequence for targeting.

Amine oxidase (AMO) is a peroxisomal matrix protein of Hansenula polymorpha, which is induced during growth of the yeast in media containing primary amines as a sole nitrogen source. The deduced amino acid sequence of the protein contains an SRL sequence at nine amino acids from the C-terminus. In this study, we have examined the possible role of the SRL motif in sorting of AMO to peroxisomes by mutating the corresponding gene sequence. For this purpose, we have developed a DNA construct that is specifically integrated into the AMO locus of the H. polymorpha genome, placing the mutant gene under the control of the endogenous AMO promoter and eliminating expression of the wild-type gene. Analysis of a stable transformant, containing the desired gene configuration, showed that mutation of the C-terminal sequence neither interfered with correct targeting of the protein into the peroxisome nor displayed significant effects on its activity. From this, it was concluded that the SRL-containing C-terminus is not essential for peroxisomal targeting of AMO in H. polymorpha.

Human catalase is imported and assembled in peroxisomes of Saccharomyces cerevisiae.

To study the conservation of peroxisomal targeting signals, we have determined the intracellular localization of human peroxisomal catalase when expressed in yeast. Using immunofluorescence, differential centrifugation and immunoelectron microscopy, we show that the protein is targeted to the peroxisomes of the heterologous cell and assembled in its active tetrameric form. These data show the conservation of the catalase targeting signal and import specificity between human and yeast peroxisomes.

Membrane traffic in polarized neurons in culture.

Fetal hippocampal neurons develop axons and dendrites in culture. To study how neurons form and maintain different plasma membrane domains, hippocampal neurons were infected with RNA viruses and the distribution of the viral glycoproteins was analyzed by light and electron microscopy. Infection of hippocampal cells with vesicular stomatitis virus (VSV) and fowl plague virus (FPV) resulted in the polarized distribution of the newly synthesized viral glycoproteins. The VSV glycoprotein appeared firstly in the Golgi apparatus and then in the dendrites. In contrast, the hemagglutinin of FPV, after accumulation in the Golgi apparatus, moved to the axons. These results suggest that the mechanism of sorting of viral glycoproteins might be similar in neurons and MDCK cells, a cell line of epithelial origin. In these cells the VSV glycoprotein and the hemagglutinin of FPV distribute to the basolateral and apical membranes, respectively. Transport of viral glycoproteins to both neuronal domains was microtubule dependent. Nocodazole treatment of infected neurons inhibited the delivery of axonal and dendritic viral glycoproteins equally. To investigate if the analogy between epithelial cells and neurons extended to include an endogenous plasma membrane protein, the distribution of Thy-1, a GPI-linked protein, was analyzed. By immunofluorescence and immunoelectron microscopy, Thy-1 was found exclusively along the axonal surface. In epithelial cells GPI-anchored proteins are located apically. The existence of a barrier on the neuronal plasma membrane that would prevent intermixing of axonal and dendritic proteins was analyzed by a liposomefusion assay.(ABSTRACT TRUNCATED AT 250 WORDS)

Overexpression of alcohol oxidase in Pichia pastoris.

The protein import capacity of peroxisomes in methylotrophic yeasts was studied using Pichia pastoris containing one or two extra copies of the gene encoding the peroxisomal protein alcohol oxidase. The alcohol oxidase overproduced in this strain was only partially imported and assembled into the active, octameric form of the protein. The excess remained in the cytosol as protein aggregates composed of monomers. These results are discussed in view of the possible application of peroxisomes as storage compartments for heterologous proteins.

Mutations in the FAD-binding fold of alcohol oxidase from Hansenula polymorpha.

Alcohol oxidase of methylotrophic yeast is an FAD-containing enzyme. When in its active form, the enzyme is an octamer and located in the peroxisomes. To study the importance of FAD-binding on the activity, octamerization and intracellular localization of the enzyme, alcohol oxidase of Hansenula polymorpha was mutated in its presumed nucleotide-binding domain, which is formed by the N-terminal sequence. Whereas mutations of a glutamic acid residue (E42) reduced the stability of the octamer, it hardly affected enzyme activity and expression. However, replacements of three conserved glycines (G13, G15 and G18) and a conserved glutamic acid (E39) within the fold had severe effects. The mutations not only resulted in loss of enzyme activity but in reduced protein levels as well, probably due to decreased stability of the mutant alcohol oxidase. However, octamerization of the protein still occurred. The existence of inactive octameric proteins provides information about the formation pathway of this octameric flavoprotein.

Induction kinetics and cell surface distribution of Escherichia coli lipoprotein under lac promoter control.

The induction kinetics and surface accessibility of the outer membrane lipoprotein were studied in an Escherichia coli strain with the lpp gene under control of the lac promoter. Free lipoprotein appeared rapidly after induction with isopropyl-beta-D-thiogalactopyranoside and reached a steady-state level after 30 min. The newly induced lipoprotein was slowly bound to the peptidoglycan layer. Immunological methods were developed to detect lipoprotein accessible at the cell surface after various pretreatments as well as peptidoglycan-bound lipoprotein at the surface of isolated peptidoglycan sacculi with specific antibodies in combination with 125I-protein A. With these methods an increase in lipoprotein molecules at the cell surface and bound to the peptidoglycan sacculus could be detected following induction. The topology of newly synthesized lipoprotein was examined in thin sections as well as at the cell surface and the surface of the peptidoglycan sacculus with immunoelectron microscopy. Ultrathin cell sections, whole cells, and isolated peptidoglycan sacculi showed lipoprotein distributed homogeneously over the entire surface.

Differential responses to osmotic stress of vasopressin-neurophysin mRNA in hypothalamic nuclei.

mRNA encoding the vasopressin-neurophysin precursor was quantitated in individual hypothalamic nuclei of rats by a liquid hybridization assay. Drinking of 2% saline for 14 days, a treatment that increased the plasma vasopressin concentration 9-fold, resulted in a 5- and 2-fold increase in mRNA levels in the supraoptic and paraventricular nucleus, respectively. This osmotic stimulus had no effect on vasopressin-neurophysin mRNA content of the suprachiasmatic nucleus. This dissociation in regulation of vasopressin-neurophysin mRNA in hypothalamic nuclei indicates the existence of two separate vasopressin systems that are independently activated.