SMN mRNA and protein levels in peripheral blood: biomarkers for SMA clinical trials.

BACKGROUND: Clinical trials of drugs that increase SMN protein levels in vitro are currently under way in patients with spinal muscular atrophy. OBJECTIVE: To develop and validate measures of SMN mRNA and protein in peripheral blood and to establish baseline SMN levels in a cohort of controls, carriers, and patients of known genotype, which could be used to follow response to treatment. METHODS: SMN1 and SMN2 gene copy numbers were determined in blood samples collected from 86 subjects. Quantitative reverse transcription PCR was used to measure blood levels of SMN mRNA with and without exon 7. A cell immunoassay was used to measure blood levels of SMN protein. RESULTS: Blood levels of SMN mRNA and protein were measured with high reliability. There was little variation in SMN levels in individual subjects over a 5-week period. Levels of exon 7-containing SMN mRNA and SMN protein correlated with SMN1 and SMN2 gene copy number. With the exception of type I SMA, there was no correlation between SMN levels and disease severity. CONCLUSION: SMN mRNA and protein levels can be reliably measured in the peripheral blood and used during clinical trials in spinal muscular atrophy, but these levels do not necessarily predict disease severity.

Protein kinase C gamma mutations in spinocerebellar ataxia 14 increase kinase activity and alter membrane targeting.

The protein kinase C gamma (PKCgamma) gene is mutated in spinocerebellar ataxia type 14 (SCA14). In this study, we investigated the effects of two SCA14 missense mutations, G118D and C150F, on PKCgamma function. We found that these mutations increase the intrinsic activity of PKCgamma. Direct visualization of labelled PKCgamma in living cells demonstrates that the mutant protein translocates more rapidly to selected regions of the plasma membrane in response to Ca2+ influx. These results point to specific alterations in mutant PKCgamma function that could lead to the selective neuronal degeneration of SCA14.

Androgens regulate the mammalian homologues of invertebrate sex determination genes tra-2 and fox-1.

Androgens, like other steroid hormones, exert profound effects on cell growth and survival by modulating the expression of target genes. In vertebrates, androgens play a critical role downstream of the testis determination pathway, influencing the expression of sexually dimorphic traits. Among cells of the nervous system, motor neurons respond to trophic effects of androgen stimulation, with a subpopulation of spinal motor neurons exhibiting sexually dimorphic survival. To study the mechanisms of androgen action in these cells, we performed a subtractive screen for genes upregulated by androgen in a motor neuron cell line. We show androgen-inducible expression of two RNA-binding proteins that are the mammalian homologues of invertebrate sex determination genes. Androgens upregulate the expression of tra-2alpha, an enhancer of RNA splicing homologous to Drosophila tra-2, and promote redistribution of the protein from a diffuse to a speckled pattern within the nucleus. Similarly, androgens upregulate the expression of a novel gene homologous to Caenorhabditis elegans fox-1. These data indicate that androgens exert their effects, in part, by modulating the expression and function of genes involved in RNA processing, and identify homologues of invertebrate sex determination genes as androgen-responsive genes in mammals.

Defective HIV-1 provirus encoding a multitarget-ribozyme inhibits accumulation of spliced and unspliced HIV-1 mRNAs, reduces infectivity of viral progeny, and protects the cells from pathogenesis.

A HeLa T4 cell line containing a defective human immunodeficiency virus type 1 (HIV-1) DNA (HD4) was isolated. After transactivation with Tat, the HD4 DNA was transcribed into a single 3.7-kb mRNA that encodes a chimeric CD4/Env protein and a multitarget-ribozyme directed against multiple sites within the gp120 coding region of HIV-1 RNA (Chen et al., 1992). Early steps in HIV infection such as entry, reverse transcription, and proviral DNA formation were not affected in HD4 cells, and HD4 was efficiently transactivated after either HIV-1 or HIV-2 infections. HIV-2, which lacks all of the HIV-1-specific ribozyme target sites, replicated to high levels in HD4 cells whereas HIV-1 replication was selectively inhibited. Despite a reduced accumulation of all HIV-1 transcripts, transactivation of HD4 was efficient. Surprisingly, the most abundant, multiply spliced mRNAs were reduced even though they lack all of the ribozyme target sites. These results strongly suggest that the ribozyme co-localizes with unspliced HIV-1 pre-mRNA and/or genomic HIV-1 RNA in the nucleus. Cleavage of these precursor RNAs explains the reduction of all spliced and unspliced HIV-1 RNAs. Cleavage of genomic RNA probably contributed to the three-fold reduction in the infectivity of viral progeny. Thus, the HD4 ribozyme RNA functioned as a ribozyme in the nucleus and as a mRNA for a chimeric CD4/Env protein in the cytoplasm. Its unusual large size for a ribozyme (3.7 kb) indicates that, in the future, other antiviral proteins, like negative transdominant mutant HIV-1 proteins, may also be encoded to increase its antiviral potential in a gene therapy approach.

Transduction of nondividing cells using pseudotyped defective high-titer HIV type 1 particles.

The use of Moloney murine leukemia virus (Mo-MLV)-based vectors to deliver therapeutic genes into target cells is limited by their inability to transduce nondividing cells. To test the capacity of HIV-based vectors to deliver genes into nondividing cells, we have generated replication-defective HIV type 1 (HIV-1) reporter vectors carrying neomycin phosphotransferase or mouse heat stable antigen, replacing the HIV-1 sequences encoding gp160. These vectors also harbor inactive vpr, vpu, and nef coding regions. Pseudotyped HIV-1 particles carrying either the ecotropic or the amphotropic Mo-MLV envelope proteins or the vesicular stomatitis virus G protein were released after single or double transfections of either human 293T or monkey COS-7 cells with titers of up to 10(7) colony-forming units per milliliter. A simple ultrafiltration procedure resulted in an additional 10- to 20-fold concentration of the pseudotyped particles. These vectors along with Mo-MLV-based vectors were used to transduce primary human skin fibroblasts and human peripheral blood CD34+ cells. The HIV-1 vector system was significantly more efficient than its Mo-MLV-based counterpart in transducing human skin fibroblasts arrested at the G0/G1 stage of the cell cycle by density-dependent inhibition of growth. Human CD34+ cells were transduced efficiently using HIV-1 pseudotype particles without prior stimulation with cytokines.

Inducible and conditional inhibition of human immunodeficiency virus proviral expression by vesicular stomatitis virus matrix protein.

Besides its role in viral assembly, the vesicular stomatitis virus (VSV) matrix (M) protein causes cytopathic effects such as cell rounding (D. Blondel, G. G. Harmison, and M. Schubert, J. Virol. 64:1716-1725, 1990). DNA cotransfection assays demonstrated that VSV M protein was able to inhibit the transcription of a reporter gene (B. L. Black and D. S. Lyles, J. Virol. 66:4058-4064, 1992). We have confirmed these observations by using cotransfections with an infectious clone of human immunodeficiency virus type 1 (HIV-1) and found that the amino-terminal 32 amino acids of M protein which are essential for viral assembly were not required for this inhibition. For the study of the potential role of M protein in the shutoff of transcription from chromosomal DNA, we have isolated stable HeLa T4 cell lines which encode either a wild-type or a temperature-sensitive (ts) VSV M gene under control of the HIV-1 long terminal repeat promoter. Transcription of the M mRNA was transactivated after HIV-1 infections. A cell line which encodes the wild-type M protein was nonpermissive for either HIV-1 or HIV-2. A cell line that encodes the ts M gene was transfected with the infectious HIV-1 DNA or was infected with HIV-1 or HIV-2. In all cases, at 32 degrees C, the permissive temperature for M protein, the cells were nonpermissive for HIV replication. At 40 degrees C, the ts M protein was nonfunctional and both HIV-1 and HIV-2 were able to replicate at high levels. A comparison of the amounts of proviral HIV-1 DNAs and HIV-1 mRNAs at 10 and 36 h after HIV-1 infection demonstrated that proviral insertion had not been prevented by M protein and that the block in HIV-1 replication was at the level of proviral expression. The severe reduction of HIV-1 proviral transcripts demonstrates that the VSV M protein alone can inhibit expression from chromosomal DNA. These results strongly support the hypothesis that the VSV M protein is involved in the shutoff of host cell transcription. M protein was able to attenuate HIV-1 infections and protect the cell population from HIV-1 pathogenesis. The temperature-dependent switch from a persistent to a lytic HIV-1 infection in the presence of ts M protein could be useful for studies of HIV-1 replication and pathogenesis.

Multitarget-ribozyme directed to cleave at up to nine highly conserved HIV-1 env RNA regions inhibits HIV-1 replication--potential effectiveness against most presently sequenced HIV-1 isolates.

Several mono-, di-, tetra-, penta- and nonaribozymes were developed. These multitarget-ribozymes were targeted to cleave HIV-1 env RNA at up to nine different conserved sites. Each multitarget-ribozyme consisted of a chain of up to nine hammerhead motifs, each flanked by a different targeting sequence. The multitarget-ribozymes were functional in vitro and gave rise to multiple, specific partial and/or complete RNA digestion products. Per RNA copy, multitarget-ribozymes were more efficient than monoribozymes or ribozymes targeting a subset of the same sites. In contrast to monoribozymes, a 400nt nonaribozyme, targeted to cleave at nine different sites within a 1.3kb HIV-1 env RNA substrate, was active and showed the same specificity of cleavage when it was part of a large 3.3kb transcript. We conclude that multitarget-ribozymes retain the specificity of monoribozymes, but they are more efficient per ribozyme RNA copy and they remain active when they are part of a large transcript. A tetra-, penta- or nonaribozyme under control of the SV40 late promoter, the beta-actin gene promoter or the HIV-1 LTR, respectively, were cotransfected with the infectious HIV-1 DNA clone pNL4-3 into permissive HeLa T4 cells. Each cotransfection resulted in a specific inhibition of HIV-1 replication as determined by syncytia formation and p24 antigen release. In addition, coexpression of the nonaribozyme with an HIV-1 env RNA transcript resulted in the specific dramatic reduction of the env transcript. We conclude that the multitarget-ribozymes are also functional intracellularly. A nucleotide sequence comparison of the target sites indicates that the multitarget-ribozymes could potentially be effective against all thirty HIV-1 isolates presently sequenced. Their use may help to slow the selection of viral escape mutants and thereby prolong their effectiveness. We anticipate that multitarget-ribozymes will also be more effective in the successful targeting of less variable cellular RNAs.

Insertion of the human immunodeficiency virus CD4 receptor into the envelope of vesicular stomatitis virus particles.

Enveloped virus particles carrying the human immunodeficiency virus (HIV) CD4 receptor may potentially be employed in a targeted antiviral approach. The mechanisms for efficient insertion and the requirements for the functionality of foreign glycoproteins within viral envelopes, however, have not been elucidated. Conditions for efficient insertion of foreign glycoproteins into the vesicular stomatitis virus (VSV) envelope were first established by inserting the wild-type envelope glycoprotein (G) of VSV expressed by a vaccinia virus recombinant. To determine whether the transmembrane and cytoplasmic portions of the VSV G protein were required for insertion of the HIV receptor, a chimeric CD4/G glycoprotein gene was constructed and a vaccinia virus recombinant which expresses the fused CD4/G gene was isolated. The chimeric CD4/G protein was functional as shown in a syncytium-forming assay in HeLa cells as demonstrated by coexpression with a vaccinia virus recombinant expressing the HIV envelope protein. The CD4/G protein was efficiently inserted into the envelope of VSV, and the virus particles retained their infectivity even after specific immunoprecipitation experiments with monoclonal anti-CD4 antibodies. Expression of the normal CD4 protein also led to insertion of the receptor into the envelope of VSV particles. The efficiency of CD4 insertion was similar to that of CD4/G, with approximately 60 molecules of CD4/G or CD4 per virus particle compared with 1,200 molecules of VSV G protein. Considering that (i) the amount of VSV G protein in the cell extract was fivefold higher than for either CD4 or CD4/G and (ii) VSV G protein is inserted as a trimer (CD4 is a monomer), the insertion of VSV G protein was not significantly preferred over CD4 or CD4/G, if at all. We conclude that the efficiency of CD4 or CD4/G insertion appears dependent on the concentration of the glycoprotein rather than on specific selection of these glycoproteins during viral assembly.

Role of matrix protein in cytopathogenesis of vesicular stomatitis virus.

The matrix (M) protein of vesicular stomatitis virus (VSV) plays an important structural role in viral assembly, and it also has a regulatory role in viral transcription. We demonstrate here that the M protein has an additional function. It causes visible cytopathic effects (CPE), as evidenced by the typical rounding of polygonal cells after VSV infection. We have analyzed a temperature-sensitive mutant of the M protein of VSV (tsG33) which is defective in viral assembly and which fails to cause morphological changes of the cells after infection at the nonpermissive temperature (40 degrees C). Interestingly, this defect in viral assembly as well as the CPE were reversible. Microinjection of antisense oligonucleotides which specifically inhibit M protein translation also inhibited the occurrence of CPE. Most importantly, when cells were transfected with a cDNA encoding the temperature-sensitive M protein of tsG33, no CPE was observed at the nonpermissive temperature. However, when these cells were shifted to the permissive temperature (32 degrees C), they rounded up and detached from the dish. These results demonstrate that M protein in the absence of the other viral proteins causes rounding of the cells, probably through a disorganization of the cytoskeleton. The absence of CPE at the nonpermissive temperature is correlated with an abnormal dotted staining pattern of M in these cells, suggesting that the mutant M protein may self-aggregate or associate with membranes rather than interact with cytoskeletal elements.

Homotypic and heterotypic exclusion of vesicular stomatitis virus replication by high levels of recombinant polymerase protein L.

The recombinant polymerase protein L of vesicular stomatitis virus (VSV) expressed in COS cells is able to transcribe and replicate the viral genome, resulting in complementation of temperature-sensitive polymerase mutants of VSV at the restrictive temperature (M. Schubert, G. G. Harmison, C. D. Richardson, and E. Meier, Proc. Natl. Acad. Sci. USA 82:7984-7988, 1985). Here we report that the efficiency of complementation is dependent on the level of L protein expression. Unexpectedly, only cells expressing low levels of recombinant L protein efficiently complemented tsL gene mutants, whereas cells with high levels of L protein did not. In fact, in all cells with high levels of L protein expression, which at 40 h posttransfection represented almost the total number of transfected cells, viral replication not only of the temperature-sensitive mutant but also of wild-type VSV was excluded. The inhibition of VSV appeared to occur at an early stage of the infectious cycle, and wild-type virus of the same serotype (Indiana) as the recombinant L protein as well as wild-type virus of a different serotype (New Jersey) was affected. Measles virus, on the other hand, was not arrested in cells with high levels of recombinant L protein, demonstrating that these cells were still capable of supporting a viral infection. The expression of high levels of only the amino-terminal half of the L protein from a recombinant mutant L gene that contains a small out-of-frame deletion in the middle of the L gene did not inhibit a VSV infection. Since the level of amplification for both L- and truncated L-encoding vectors is similar, we conclude that the arrest of VSV was caused by high levels of functional full-length L protein itself and not by high levels of vector-encoded L mRNA or other vector products or by side effects of vector amplification. These data strongly support the idea that the highly conserved gene order of nonsegmented negative-strand viruses and the sequential and attenuated mode of transcription are important regulatory elements which balance the intracellular concentration of viral proteins. They both assure that the L gene is the last and the least frequently transcribed gene, giving rise to low levels of L protein necessary for efficient replication.

Expression of a cDNA encoding a functional 241-kilodalton vesicular stomatitis virus RNA polymerase.

The large gene, L, of vesicular stomatitis virus (VSV), which codes for the multifunctional RNA-dependent RNA polymerase, was assembled from five overlapping cDNA clones. The sequence of the 6.4-kilobase gene of the final construct was identical to the consensus sequence reported earlier. The gene was inserted into the simian virus 40 transient expression vector pJC119. Antibodies directed against synthetic peptides corresponding to the amino and carboxyl termini of the L protein were raised in rabbits. Both antibodies specifically immunostained the cytoplasm of COS cells that had been transfected with the vector DNA. The expressed L protein was immunoprecipitated from cell extracts and it was identical in size to the L protein of the virion (241 kilodaltons). Most importantly, COS cells that expressed the recombinant L protein transcribed, replicated, and consequently complemented and rescued temperature-sensitive RNA polymerase mutants of VSV at the nonpermissive temperature. The kinetics of virus release were similar to those of a wild-type VSV infection. We conclude that the recombinant RNA polymerase protein L is indistinguishable in its size and its functions from the VSV polymerase.

Sites of copy choice replication involved in generation of vesicular stomatitis virus defective-interfering particle RNAs.

The copy choice model for the generation of defective interfering (DI) particles of vesicular stomatitis virus suggests that during replication the polymerase prematurely terminates, moves with the nascent daughter strand to another site on the same or a different template molecule, and resumes elongation of the nascent chain. We have analyzed the sites where premature termination or resumption of replication has occurred during the generation of the deletion DI particle LT, the snapback DI particle 011, and the panhandle DI particles T, T(L), and 611. The recombination sites were identified by comparing the nucleotide sequences of the relevant regions of these DI particle RNAs to those of the vesicular stomatitis virus L gene (Schubert et al., J. Virol. 51:505-514, 1984). Sequence homology was not detected between these sites, which rules out the existence of a general terminator or promoter sequence involved in copy choice replication. In several cases, however, premature termination or resumption of RNA replication may be favored by specific signal sequences. The sequences immediately before the start and at the end of the deletion in DI LT contain two hexanucleotides, ATCTGA and GATTGG, in a similar spacing. In these case of DI T and 611, but not of DI T(L), the end of the 5'-terminal region bears the hexanucleotide CCUCUU. This sequence is also repeated in the stem region in all three DI particle genomes. In addition, we present data that the added 3'-terminal regions of the panhandle DI particle RNAs may differ by only one base and are 46 [DI T(L) and 611] or 45 (DI T) bases long. We suggest that each site of the vesicular stomatitis virus genome has the potential to give rise to DI particle RNAs. Specific sequences, however, may modulate this process in a quantitative way, and they favor the generation of certain types of DI particle genomes like those of the panhandle type.

Primary structure of the vesicular stomatitis virus polymerase (L) gene: evidence for a high frequency of mutations.

A consensus sequence of the polymerase (L) gene of vesicular stomatitis virus, derived from three genomic cDNA copies, is presented. This analysis completes the primary structure of the vesicular stomatitis virus genome, totaling 11,162 bases. The L gene alone spans 6,380 nucleotides and codes for a basic 2,109-amino-acid protein with a molecular weight of 241,012. Sixteen point mutations were detected among cDNA clones prepared from viral RNA of the same strain, representing direct evidence for either the high mutability of vesicular stomatitis virus, the infidelity of reverse transcription during cDNA synthesis, or a combination of both. Some mutation, if present in the viral genome, would result in the translation of incomplete L proteins. For example, two out of four cDNA copies which covered the same region of the L gene had a single-base deletion in the exact same position, whereas the other two clones did not, strongly suggesting that a subpopulation of the genomic RNA may contain this lethal mutation. These lethal mutants define a new class of defective and most likely interfering particles which are indistinguishable in size from the parental virus and can be distinguished only by direct sequencing. We suggest that because of its infidelity, the viral polymerase itself introduces mutations and because of its size, most of these mutations are localized within the polymerase gene. In persistently infected cells in which the selective pressures on the polymerase are different, some of these L gene mutations may further erode the accuracy of the polymerase and thereby lead to the increased mutation rate that is characteristic of this type of infection.

In vitro transcription of vesicular stomatitis virus: initiation with GTP at a specific site within the N cistron.

In vitro transcripts of vesicular stomatitis virus (VSV) were either 5'-terminally labeled by incorporation of [beta-(32)P]GTP or were selected on Hg-agarose after incorporation of gamma-thio-GTP. Capped RNAs ranged in size from 23 nucleotides, the shortest capped RNA detected, to full-length message size. The 5'-terminal sequences corresponded to those of N message and to a small amount of NS message. Approximately 14% of the capped N gene transcripts were terminated at positions 86 to 90 of the VSV genome, giving rise to specific, 36 to 40-nucleotide-long, capped RNA species. The GTP-initiated RNAs were short with a predominant 28-nucleotide-long RNA species. A minor portion was as large as mRNAs. Nucleotide sequence analyses of the short RNA revealed that it was specifically initiated at positon 91 of the VSV genome, 41 nucleotides within the N cistron. This corresponds exactly to the site where transcription of the 40-nucleotide-long, capped RNA terminated. Initiation with GTP at position 91 occurred at approximately the same frequency as termination of the capped RNA at position 90, suggesting that intracistronic initiation at position 91 may depend upon termination of transcription of the 5'-proximal region and therefore may be sequential. This unique RNA represents the first transcript of VSV which was initiated at an intracistronic site with GTP, and may also represent the first example of a transcript derived from a stop/start mechanism of VSV transcription in vitro. Although initiation occurred frequently at the beginning of the N cistron yielding 11 to 14-nucleotide-long, [beta-(32)P]ATP-labeled transcripts (D. F. Pinney and S. U. Emerson, J. Virol. 42:889-896, 1982), capping of these short RNAs was not detected. This suggests that transcripts may have to be 15 to 23 nucleotides long to be accepted as substrates by the guanyltransferase.

Immunocytochemical localization of aspartate aminotransferase immunoreactivity in cochlear nucleus of the guinea pig.

There is substantial evidence supporting the role of aspartate or glutamate as the neurotransmitter of the auditory nerve. The concentration of aspartate aminotransferase (L-aspartate:2-oxoglutarate aminotransferase, EC 2.6.1.1), an enzyme associated with the metabolism of these amino acids, is high in axons and terminals of the auditory nerve. Antibodies were raised against aspartate aminotransferase and used in immunocytochemical studies to determine its localization in the cochlear nucleus of the guinea pig. Indirect immunofluorescence techniques were used for light microscopic localization of aspartate aminotransferase-like immunoreactivity in normal guinea pigs and guinea pigs with auditory nerve lesions. Fluorescent rings of aspartate aminotransferase-like immunoreactivity were seen around spherical cells in the anteroventral cochlear nucleus. In animals with auditory nerve lesions, rings were no longer seen in the ipsilateral cochlear nucleus. Immunoreactivity was also seen on cells in the posteroventral cochlear nucleus and in auditory nerve fibers. Ultrastructural studies were done in the rostral anteroventral cochlear nucleus, using the peroxidase-antiperoxidase technique. Aspartate aminotransferase-like immunoreactivity was seen at axosomatic synapses on large spherical cells in terminals with the morphological characteristics of auditory nerve terminals. Other classes of terminals on the soma of large spherical cells showed no immunoreactivity. It was concluded that aspartate aminotransferase-like immunoreactivity is present in axons and terminals of the auditory nerve. These findings indicate that aspartate aminotransferase-like immunoreactivity may serve as a marker at terminals where aspartate or glutamate is a neurotransmitter.

Disulfide-disulfide interchange catalyzed by a liver supernatant enzyme.

An enzyme widely distributed in rabbit tissues which catalyzes an interchange between N,N-di-dinitrophenyl-L-cystine and oxidized glutathione to form the mixed disulfide is described. D-Penicillamine disulfide can be substituted for oxidized glutathione and the mixed disulfide of cysteine and glutathione can serve as the sole substrate giving as one product of interchange, oxidized glutathione. The enzyme is very labile and only limited purification of it has been achieved. The activity increases with increasing pH above 6.6, the Km for N,N-di-dinitrophenyl-L-cystine is 0.2 mM and for oxidized glutathione 0.8 mM. The enzyme is inhibited by SH reagents with protection against iodoacetamide inactivation provided by N,N-di-dinitrophenyl-L-cystine. Evidence is presented that disulfide-disulfide interchange enzyme is a different activity from the previously described protein disulfide isomerase and thiol transferase.