Isolation of paramyxovirus glycoproteins. Association of both hemagglutinating and neuraminidase activities with the larger SV5 glycoprotein

Crystallization of biologically active hemagglutinin-neuraminidase glycoprotein dimers proteolytically cleaved from human parainfluenza virus type 1

We isolated, purified, and characterized the hemagglutinin-neuraminidase (HN) of human parainfluenza virus type 1, with the ultimate goal of producing crystals suitable for three-dimensional X-ray structure analysis. Pronase was used to cleave the globular head of the HN molecule directly from virus particles, forming HN monomers and dimers. The purified dimers retained neuraminidase and hemadsorption activity and were recognized by 14 anti-HN monoclonal antibodies, demonstrating intact HN antigenic structure and function. N-terminal sequence analysis of the dimers showed that cleavage had occurred at amino acid 136 or 137, freeing the C-terminal 438 or 439 amino acids. On electron micrography, the dimer appeared as two box-shaped structures, each approximately 5 by 5 nm. When the purified HN dimers were crystallized in hanging drops by vapor diffusion against 20% polyethylene glycol 3350, they formed both rectangular plates and needlelike crystals. The rectangular crystals diffracted X-rays, indicating an ordered atomic structure. However, the resolution was approximately 10 A (1 nm), insufficient for three-dimensional structural analysis. Experiments to improve the resolution by increasing the size and quality of the crystals are in progress.

The haemagglutinin-neuraminidase glycoprotein of the porcine paramyxovirus LPMV: comparison with other paramyxoviruses revealed the closest relationship to simian virus 5 and mumps virus

The complete nucleotide sequence of the haemagglutinin-neuraminidase (HN) gene of the porcine paramyxovirus LPMV, was determined from cDNA derived from viral genomic RNA. The gene was 1906 nucleotides long including a putative gene end and poly A signal. One long open reading frame was found encoding a protein of 576 amino acids with a calculated molecular weight of 63,324. The protein contains four potential N-glycosylation sites and a major hydrophobic region near the N-terminal, suggesting a membrane anchor domain. Comparison of the deduced amino acid sequence of the LPMV HN protein with that of other paramyxovirus HN proteins, revealed the highest amino acid identity to simian virus 5 of 43% and mumps virus of 41%.

Distinct hemagglutinin and neuraminidase epitopes involved in antigenic variation of recent human parainfluenza virus type 2 isolates

A panel of fourteen neutralizing anti-HN monoclonal antibodies (mAbs) to the prototype Greer strain of human parainfluenza virus type 2 (PI2) was used to determine the extent of antigenic variation in recent virus isolates. Competitive binding analysis with the mAbs indicated the presence of at least five distinct antigenic sites (I to V) on the HN glycoprotein molecule. MAbs recognizing different antigenic sites were found to be associated with the hemagglutinin (sites I, IV and V), hemagglutinin and neuraminidase (site II), or neuraminidase (site III) activities. The location of two distinct epitopes identifying the neuraminidase sites (II and III) was further verified from the generation of escape mutants. Antibodies directed to sites I and III failed to show any detectable binding or neutralizing activity against a number of natural PI2 virus isolates collected in Texas between 1986 and 1987. Interestingly, these natural variants, unlike the prototype virus, did not show any detectable neuraminidase activity with fetuin as a substrate and the enzyme activity was only detected with N-acetylneuramin-lactose as an alternative substrate. Despite the observed variation in the antigenic sites, primary infection with the prototype virus or the natural variants generated a protective immune response against challenge infection with the other virus strains.

Function and immunogenicity of human parainfluenza virus 3 glycoproteins expressed by recombinant adenoviruses

Human parainfluenza virus type 3 fusion (F) and hemagglutinin-neuraminidase (HN) cDNA sequences were inserted into the E3 region of the adenovirus type 5 genome. Cells infected with recombinant adenoviruses containing HPIV3 F (AdF) and HN (AdHN) sequences were shown to express HPIV3 F and HN proteins that were functional and immunogenic. The HN protein produced following AdHN infection was glycosylated, expressed on the surface of infected cells and exhibited both hemagglutinin and neuraminidase activities. AdF infection led to the synthesis of both the HPIV3 F0 precursor and its proteolytic cleavage product, F1. F proteins produced by AdF were glycosylated and expressed on the infected cell surface. Syncytium formation was observed in HeLa T4 cell monolayers upon coinfection with AdF and AdHN. The F and HN proteins expressed by recombinant adenoviruses were recognized by HPIV3 F- and HN-specific monoclonal antibodies. Mice injected intraperitoneally with AdF or AdHN produced antibodies that immunoprecipitated the appropriate HPIV3 glycoproteins and sera from immunized mice effectively neutralized HPIV3 virions. These results support future work using recombinant adenoviruses to study the immune response to individual HPIV3 glycoproteins as well as in protection studies using animal models.

Biophysical characterization of alpha-crystallin aggregates: validation of the micelle hypothesis

The size of alpha-crystallin aggregates, as well as the structural organization of each particle's subunits, is currently unknown, although a number of different laboratories have suggested both structures and average molecular weights (Thomson, J.A. and Augusteyn, R.C. (1984) Proc. Int. Soc. Eye Res. 3, 152). One hypothesis, compatible with literature reports and consistent with what is known of subunit primary structure and physiological function, is that alpha-crystallin exists in vivo as a naturally occurring protein micelle (Sen, A.C. and Chakrabarti, B. (1991) Biophysical J. 59, 108a.) To test this hypothesis, experiments were performed on this protein to determine its behavior under increased hydrostatic pressure and the effect of its concentration on aqueous surface tension. With increasing hydrostatic pressure, the turbidity of an alpha-crystallin solution increases exponentially to a plateau at about 6000-8000 psi; upon release of pressure, the samples slowly return to their original turbidity level. Other naturally aggregating proteins, such as skeletal muscle myosin, demonstrate a decrease in turbidity under the same conditions. The surface tension of alpha-crystallin in aqueous solution decreases to a plateau with increasing protein subunit concentration, with an inflection point over the range 0.18-0.25 mM; cholate and other amphiphiles exhibit similar behavior. In contrast, plots of surface tension over the equivalent concentration range for other protein aggregates in the same buffer more closely approximate the types of curve obtained with short chain aliphatic acids. These results indicate that alpha-crystallin behaves like the protein version of a micelle.

Simian virus 5 membrane protein maturation: expression in virus-infected cells and from a eukaryotic vector

Properties of the membrane protein (M) of the paramyxovirus simian virus 5 (SV5) isolated from purified SV5 virions, in SV5-infected cells or when expressed from cDNA using a eukaryotic vector (SV40-M) were examined. Kinetic (pulse-chase radiolabeling) studies showed that M protein expressed in SV5-infected and SV40-M recombinant virus-infected cells underwent maturation, detectable as time-dependent acquisition of reactivity with anti-M protein monoclonal antibodies. Kinetic studies using radiolabeled phosphate and studies with the alkylating agent N-ethylmaleimide indicated that the antigenic maturation of the M protein was not due to phosphorylation or disulfide bond formation, respectively. Immunofluorescent antibody staining studies showed a significant difference in staining patterns between SV40-M recombinant virus-infected cells and SV5-infected cells. SV40-M recombinant virus-infected cells exhibited an intensely staining cytoplasmic fibrillar network, whereas in SV5-infected cells, villar and some small granular structures were the only strongly staining structures.

The fusion and hemagglutinin-neuraminidase glycoproteins of human parainfluenza virus 3 are both required for fusion

Recombinant vaccinia viruses, VF and VHN, expressing the fusion (F) and hemagglutinin-neuraminidase (HN) glycoproteins of human parainfluenza virus 3 (HPIV3) were constructed. Infection of HeLa T4 cells with VF and VHN led to the synthesis of glycoproteins, with the correct apparent molecular weights, that were recognized by monoclonal antibodies specific for HPIV3F and HN. The HN glycoprotein was present on the surface of cells infected with VHN and these cells demonstrated both hemadsorbing and neuraminidase activities. The F glycoprotein was present in cleaved and uncleaved forms and was also expressed on the surface of VF-infected cells. Fusion activity, however, as evidenced by syncytium formation and lysis of human erythrocytes, could only be demonstrated when HeLa T4 cells were coinfected with VF and VHN. Fusion events that are mediated by HPIV3, therefore, require both the F and HN glycoproteins.

Protection of chickens with a recombinant fowlpox virus expressing the Newcastle disease virus hemagglutinin-neuraminidase gene

A recombinant fowlpox virus expressing the hemagglutinin-neuraminidase (HN) protein of Newcastle disease virus (NDV) strain Texas was generated. Immunoprecipitation with chicken anti-NDV serum confirmed authentic expression of the HN protein. Protection of chickens from infection with NDV was observed when birds were immunized with the recombinant HN fowlpox virus by the intramuscular route after one or two inoculations. Vaccination by the ocular route with a mixture of fowlpox recombinants expressing the fusion and HN proteins did not show added protection over that seen with the individual viruses.

Folding and oligomerization properties of a soluble and secreted form of the paramyxovirus hemagglutinin-neuraminidase glycoprotein

The paramyxovirus SV5 hemagglutinin-neuraminidase (HN) glycoprotein (a type II integral membrane protein) was converted into a soluble and secreted form (HN-F) by replacing the HN signal/anchor domain with a hydrophobic domain that can act as a cleavable signal sequence. Approximately 40% of the HN-F synthesized was secreted from cells (t1/2 approximately 2.5-3 hr). The extracellular HN-F molecules were identified as disulfide-linked dimers and the majority of the population of molecules were resistant to endoglycosidase H digestion. Examination of the oligomeric form of the secreted HN-F, by sucrose density gradient sedimentation, indicated that under conditions where HN was a tetramer, HN-F was found to be a dimer, and no extracellular HN-F monomeric species could be detected. Secreted HN-F was fully reactive with conformation-specific monoclonal antibodies and was enzymatically active as shown by HN-F having neuraminidase activity. Examination of the intracellular HN-F species indicated that HN-F monomers were slowly converted to the disulfide-linked form and that under the sucrose density gradient sedimentation conditions used the HN-F monomers aggregated. Some of the HN-F monomers were degraded intracellularly. These data are discussed in relationship to the seemingly different folding and oligomerization requirements for the intracellular transport of soluble and membrane bound forms of a glycoprotein. The soluble and biologically active form of HN may be suitable for further structural and enzymatic studies.

Parainfluenza virus type 1 reduces the affinity of agonists for muscarinic receptors in guinea-pig lung and heart

Membrane preparations of guinea-pig lung (containing multiple muscarinic receptor subtypes) and heart (containing M2 receptors only) were incubated with either neuraminidase, parainfluenza virus (which contains neuraminidase), or virus plus 2,3-dehydro-2-deoxy-N-acetylneuraminic acid, a neuraminidase inhibitor. None of these treatments affected [3H]quinuclidinyl benzilate [( 3H]QNB) binding. In the lung and heart, carbachol displaced 0.2 nM [3H]QNB from two sites. After treatment with either neuraminidase or virus the high affinity site was shifted to the right, and carbachol displaced QNB from one site with low affinity in the lung. In contrast, neuraminidase or virus decreased the affinity of carbachol for both sites in the heart. The neuraminidase inhibitor completely blocked virus-induced changes in carbachol affinity in both tissues. These results suggest that parainfluenza virus decreases the affinity of agonists for some of the muscarinic receptors in the lung, and for all of the muscarinic receptors in the heart due to its neuraminidase activity, which results in removal of sialic acid. The decreased agonist affinity in the lung may be responsible for the increased vagally induced bronchoconstriction seen in viral respiratory infections.

Sequence characterization of the membrane protein gene of paramyxovirus simian virus 5

The complete nucleotide sequence of the membrane (M) protein gene of the paramyxovirus simian virus 5 (SV5) was determined from cDNA clones of viral mRNAs. The M gene boundaries were determined by (i) primer extension sequencing on M mRNA; (ii) nuclease S1 analysis; and (iii) primer extension sequencing on viral genomic RNA. The M gene mRNA consisted of 1371 templated nucleotides. It contains a single large open reading frame that can encode a protein of 377 amino acids with a predicted Mr = 42,253. The authenticity of the predicted M protein coding sequence was confirmed by synthesis of the M protein from mRNA synthesized from cDNA. The predicted M amino acid sequence indicated it is an overall hydrophobic protein carrying a net positive charge. Alignment of the SV5 protein amino acid sequence with the M protein sequences of other paramyxoviruses indicated that these viruses fall into the following two groups: (1) SV5, mumps virus, and Newcastle disease virus; or (2) Sendai, parainfluenza virus type 3, measles virus, and canine distemper virus, with mumps virus M sequence being the most closely related to SV5.

Sequence of the hemagglutinin-neuraminidase gene of human parainfluenza virus type 1

The nucleotide sequence of the gene encoding the hemagglutinin-neuraminidase (HN) glycoprotein of human parainfluenza virus type 1 (PI1) was determined from cDNA clones derived from poly(A)+ RNA extracted from infected cells. A single open reading frame in the sequence was found to encode a putative protein of 575 amino acids with a calculated molecular weight of 63,960. The predicted amino acid sequence contains ten potential sites for N-linked glycosylation and one major hydrophobic region located 35 amino acids from the amino terminus, which appears to be the signal-anchor domain of HN. Comparison of the amino acid sequence with the HN glycoproteins of other paramyxoviruses indicated that the PI1 HN protein is most closely related to the Sendai virus (SV) HN protein.

Newcastle disease virus fusion protein expressed in a fowlpox virus recombinant confers protection in chickens

A cDNA copy of the RNA encoding the fusion (F) protein of Newcastle disease virus (NDV) strain Texas, a velogenic strain of NDV, was obtained and the sequence was determined. The 1,792-base-pair sequence encodes a protein of 553 amino acids which has essential features previously established for the F protein of virulent NDV strains. These include the presence of three strongly hydrophobic regions and pairs of dibasic amino acids in the pentapeptide Arg-Arg-Gln-Arg-Arg preceding the putative cleavage site. When inserted into a fowlpox virus vector, a glycosylated protein was expressed and presented on the surface of infected chicken embryo fibroblast cells. The F protein expressed by the recombinant fowlpox virus was cleaved into two polypeptides. When inoculated into susceptible birds by a variety of routes, an immunological response was induced. Ocular or oral administration of the recombinant fowlpox virus gave partial protection, whereas both intramuscular and wing-web routes of inoculation gave complete protection after a single inoculation.

Construction of expression plasmids for the fusion protein of Sendai virus, and their expression in E. coli cells and eucaryotic cells

To examine the properties and the role of the fusion protein (F) of Sendai virus at the molecular level, a plasmid, pUC-F, was constructed by inserting cDNA for the F protein into a pUC vector. Upon induction of E. coli cells transformed with pUC-F, a new protein was obtained, which was identified as Fo on Western blot analysis. The cDNA fragment for the F gene was excised from pUC-F and inserted into an eucaryotic expression vector, pSVL, to yield pSVL-F. COS-1 cells transfected with pSVL-F gave a band on SDS-gel electrophoresis which corresponded to the size of the Fo proteins.

Intracellular maturation and transport of the SV5 type II glycoprotein hemagglutinin-neuraminidase: specific and transient association with GRP78-BiP in the endoplasmic reticulum and extensive internalization from the cell surface

The hemagglutinin-neuraminidase (HN) glycoprotein of the paramyxovirus SV5 is a type II integral membrane protein that is expressed at the infected cell surface. The intracellular assembly and transport of HN in CV1 cells was examined using conformation-specific HN mAbs and sucrose density sedimentation analysis. HN was found to oligomerize with a t1/2 of 25-30 min and these data suggest the oligomer is a tetramer consisting primarily of two noncovalently associated disulfide-linked dimers. As HN oligomers could be found that were sensitive to endoglycosidase H digestion and oligomers formed in the presence of the ER to the Golgi complex transport inhibitor, carbonylcyanide m-chlorophenylhydrazone (CCCP), these data are consistent with HN oligomerization occurring in the ER. Unfolded or immature HN molecules that could not be recognized by conformation-specific antibodies were found to specifically associate with the resident ER protein GRP78-BiP. Immunoprecipitation of BiP-HN complexes with an immunoglobulin heavy-chain binding protein (BiP) antibody indicated that newly synthesized HN associated and dissociated from GRP78-BiP (t1/2 20-25 min) in an inverse correlation with the gain in reactivity with a HN conformation-specific antibody, suggesting that the transient association of GRP78-BiP with immature HN is part of the normal HN maturation pathway. After pulse-labeling of HN in infected cells, it was found that HN is rapidly turned over in cells (t1/2 2-2.5 h). This led to the finding that the vast majority of HN expressed at the cell surface, rather than being incorporated into budding virions, is internalized and degraded after localization to endocytic vesicles and lysosomes.

Proteins of Newcastle disease virus envelope: interaction between the outer hemagglutinin-neuraminidase glycoprotein and the inner non-glycosylated matrix protein

Using linear sucrose-density ultracentrifugation analysis of Triton-solubilized Newcastle Disease Virus envelopes, we have evidenced, for the first time, the existence of interactions between the outer hemagglutinin-neuraminidase transmembrane glycoprotein and the inner non-glycosylated peripheral matrix protein. Such interactions seem to be electrostatic. These conclusions are based on the behavior of both proteins at different ionic strengths. When in low ionic strength buffer, hemagglutinin-neuraminidase and matrix proteins band together in the sucrose gradient, whereas at high ionic strength both proteins band at different rates in the gradient. The behavior of the inner matrix protein in our conditions was the expected one for a peripheral protein. The results of these 'in vitro' studies are also discussed in terms of the possible 'in vivo' role of such interactions.

Cytocidal effect caused by the envelope glycoprotein of a newly isolated avian hemangioma-inducing retrovirus

We isolated a field strain of avian hemangioma retrovirus (AHV) which induces a cytopathic effect (CPE) on cultured avian and mammalian cells shortly after infection. The kinetics of cell killing were dependent on the multiplicity of infection. The CPE on avian and mammalian cells was independent of virus replication, because UV-irradiated virus led to cell death as well. Biochemical and genetic experiments indicated that AHV env gene products were responsible for the CPE. Partially purified AHV envelope glycoproteins (gp85), but not those of the Rous sarcoma virus Prague C strain, induced a CPE. Rous-associated virus type 1, in which the env region was replaced by the AHV gp85 region, induced a CPE on avian and mammalian cultured cells. Therefore, we suggest that CPE is induced by AHV via interaction between viral gp85 and the cell membrane. This mode of CPE is unique among avian sarcoma-leukemia viruses.

Virus-specific polypeptides of human parainfluenza virus type 4 and their synthesis in infected cells

We have studied the structural components of human parainfluenza virus type 4A (PIV-4A) and identified some virus-specific polypeptides by immunoprecipitation with polyclonal and monoclonal antibodies followed by one- or two-dimensional SDS-PAGE. HN polypeptides existed as monomer, disulfide-linked dimer, and disulfide-linked larger oligomer in cells infected with PIV-4A. Interestingly, the nonreduced NP, the nonreduced fusion, and the reduced F1 proteins migrated as doublets. Two F1 polypeptides were derived from different F1 + 2 proteins which migrated separately under nonreducing condition. In Vero cells infected with two strains of PIV-4A, two lower-molecular-weight proteins related to NP were detected. Oligopeptide patterns of the lower-molecular-weight protein were similar to those of NP protein synthesized in primary monkey kidney cells. The NP-related low-molecular-weight protein(s) was immunoprecipitated by 1 of 11 monoclonal antibodies against mumps virus NP protein. The MAb also reacted with NP proteins of PIV-2 and SV5. Thus, the epitope recognized by the MAb was common among PIV-2, PIV-4, mumps virus, and SV5, suggesting that the epitope might have an important biological function. However, the MAb did not react with the intact NP protein from cells infected with PIV-4, indicating that the epitope of PIV-4A was presented only when NP was cleaved. Phosphorylation was demonstrated for NP and P proteins.

Crystallization of Sendai virus HN protein complexed with monoclonal antibody Fab fragments

The hemagglutinin-neuraminidase (HN) protein of Sendai virus has been isolated from virus particles in a biologically active soluble form after removal by proteolytic digestion of the hydrophobic amino-terminal anchor sequence (S. D. Thompson, W. G. Laver, K. G. Murti, A. Portner, J. Virol., 62, 4653-4660, 1988). The soluble HN exists as both dimers and tetramers, and crystallization trials with each of these forms have so far yielded amorphous material. Dimers complexed with Fab fragments of a monoclonal antibody formed long needle crystals. So far, these are not suitable for X-ray diffraction analysis but the results suggest that HN molecules from paramyxoviruses, even if not crystallizable, may, when complexed with Fab fragments, in some cases yield crystals suitable for X-ray diffraction analysis.

Conformation and stability of Sendai virus fusion protein

The conformation and stability of Sendai virus fusion (F) protein were studied by circular dichroism spectroscopy, and the protein predictive models of Chou and Fasman and Robson and Suzuki were used to elucidate the secondary structure of Sendai virus F protein. The F protein conformation is predicted to contain 33% alpha-helix, 53% beta-sheet and 15% beta-turn by the Chou and Fasman model, and 30% alpha-helix, 55% beta-sheet, 9% beta-turn and 7% random coil by the Robson and Suzuki model. C.d. studies of F protein purified in the presence of the non-ionic detergent, n-octylglucoside, indicated the presence of 49% alpha-helix and 31% beta-sheet at pH 7.0, 54% alpha-helix and 28% beta-sheet at pH 9.0 and 50% alpha-helix and 23% beta-sheet at pH 5.4. A small change in conformation of the protein occurred when the pH was titrated from 7.0 to 5.4 and from 7.0 to 9.0 and a more pronounced conformational change occurred when the pH was changed from 9.0 to 5.4. The F protein in 0.2% n-octylglucoside was resistant to denaturation by 4 M guanidine hydrochloride, the reducing agent 20 mM mercaptoethanol, and to increases in temperature from 5 to 80 degrees C. Monoclonal anti-F protein antibody showed an increased binding to whole virus when the pH was changed from 7.0 to 9.0. The antibody binding was decreased when the pH was shifted from 9.0 to 5.4 Maximum haemolytic activity was observed with virus that was preincubated at pH 8.0.(ABSTRACT TRUNCATED AT 250 WORDS)

Analysis of the relationship between cleavability of a paramyxovirus fusion protein and length of the connecting peptide

The relationship between the length of the connecting peptide in a paramyxovirus F0 protein and cleavage of F0 into the F1 and F2 subunits has been examined by constructing a series of mutant F proteins via site-directed mutagenesis of a cDNA clone encoding the simian virus 5 F protein. The mutant F proteins had one to five arginine residues deleted from the connecting peptide. The minimum number of arginine residues required for cleavage-activation of the simian virus 5 F0 protein by host cell proteases was found to be four. F proteins with two or three arginine residues in the connecting peptide were not cleaved by host cell proteases but could be cleaved by exogenously added trypsin. The mutant F protein possessing a connecting peptide consisting of one arginine residue was not cleaved by trypsin. The altered F proteins were all transported to the infected-cell plasma membrane as shown by cell surface immunofluorescence or cell surface trypsinization. However, the only mutant F protein found to be biologically active as detected by syncytium formation was the F protein which has four arginine residues at the cleavage site. The results presented here suggest that in the paramyxovirus F protein the number of basic amino acid residues in the connecting peptide is important for cleavage of the precursor protein by host cell proteases but is not the only structural feature involved. In addition, the data indicate that cleavage of F0 into F1 and F2 does not necessarily result in biological activity and that the connecting peptide may affect the local conformation of the F polypeptide.

A recent field isolate of Sendai virus has a temperature-sensitive HN glycoprotein

A recent field isolate of Sendai virus was found to have a temperature-sensitive (ts) hemagglutinin-neuraminidase (HN) glycoprotein. The ts phenotype was manifested as a loss of cell binding, reduced replication, and virions that were lacking surface HN after growth at the nonpermissive temperature (38 degrees). Low neuraminidase activity and failure of the field isolate to remove sialic acid from receptors on the surface of erythrocytes indicated that rapid elution of the field isolate virions from erythrocytes at the nonpermissive temperature was not due to neuraminidase activity but to a proposed conformational change in the HN molecule. The specific amino acids responsible for the ts phenotype could not be determined due to the number of amino acid differences between the field isolate and Enders strain. Heat inactivation and monoclonal antibody inhibition of HN functions indicated that the HN protein of this isolate was, in addition to ts, an unstable molecule.

Antigenic and functional organization of human parainfluenza virus type 3 fusion glycoprotein

Twenty-six monoclonal antibodies (MAbs) (14 neutralizing and 12 nonneutralizing) were used to examine the antigenic structure, biological properties, and natural variation of the fusion (F) glycoprotein of human type 3 parainfluenza virus (PIV3). Analysis of laboratory-selected antigenic variants and of PIV3 clinical isolates indicated that the panel of MAbs recognizes at least 20 epitopes, 14 of which participate in neutralization. Competitive binding assays indicated that the 14 neutralization epitopes are organized into three nonoverlapping antigenic sites (A, B, and C) and one bridge site (AB) and that the 6 nonneutralization epitopes form four sites (D, E, F, and G). Most of the neutralizing MAbs were involved in nonreciprocal competitive binding reactions, suggesting that they induce conformational changes in other neutralization epitopes. Fusion-inhibition and complemented-enhanced neutralization assays indicated that antigenic sites AB, B, and C may correspond to functional domains of the F molecule. Our results indicated that antibody binding alone is not sufficient for virus neutralization and that many anti-F MAbs neutralize by mechanisms not involving fusion-inhibition. The degree of antigenic variation in the F epitopes of clinical strains was examined by binding and neutralization tests. It appears that PIV3 frequently develops mutations that produce F epitopes which efficiently bind antibodies, but are completely resistant to neutralization by these antibodies.

Adsorption of LLCMK2 cell-grown Sendai virus onto human red blood cells and its release from the virus adsorbed cells

An early stage of virus adsorption was studied in a system of Sendai virus metabolically labeled with [3H]leucine in LLCMK2 cells and of human red blood cells (RBCs). The efficiency of viral release from the virus-bound RBCs by incubation at 37 C depended on the number of virus particles which had been used for adsorption onto the RBC at 4 C. When 7.8 x 10(2) virus particles were previously adsorbed onto the RBC at 4 C, most of the viruses were dissociated from the RBC at 37 C. In the case of adsorption of 3 to 12 virus particles per RBC, however, most of the viruses were not dissociated from the RBC by incubation at 37 C. Such RBC-bound viruses were released by incubation with various bacterial neuraminidases (Clostridium perfringens, etc.) or with a large number of LLCMK2 cell-grown Sendai virus (LLCMK2-Sendai) particles, but not released by treatment with hemagglutinin-neuraminidase protein (Sendai-gp) isolated from egg-grown Sendai virus.

Isolation of a biologically active soluble form of the hemagglutinin-neuraminidase protein of Sendai virus

As a first step in establishing the three-dimensional structure of the Sendai virus hemagglutinin-neuraminidase (HN), we have isolated and characterized a potentially crystallizable form of the molecule. The sequence of HN, a surface glycoprotein, predicts a protein with an uncharged hydrophobic region near the amino terminus which is responsible for anchorage in the viral envelope. To avoid rosette formation (aggregation), which would preclude crystallization, this hydrophobic tail was removed from a membrane-free form of HN by proteolytic digestion. This digestion resulted in a single product with a molecular weight of about 10,000 less than native HN. N-terminal amino acid sequence analysis of cleaved HN (C-HN) indicated a single cleavage site at amino acid residue 131, resulting in a product consisting of the carboxyl-terminal 444 amino acids of HN. Functional analyses revealed that C-HN retained full neuraminidase activity and was able to bind erythrocytes, indicating that the N-terminal 131 residues were not necessary for these biological activities. Furthermore, this cleavage product retained the antigenic structure of intact HN, since monoclonal antibodies still bound to C-HN in enzyme-linked immunosorbent assay and Western (immuno-) blot analysis. Viewed by electron microscopy, the dimeric and tetrameric forms of intact HN form rosettes while C-HN maintains the oligomeric structure but no longer aggregates. Furthermore, the electron micrographs revealed a C-HN tetramer strikingly similar to the influenza virus neuraminidase in both size and gross structural features.

The E3 protein of bovine coronavirus is a receptor-destroying enzyme with acetylesterase activity

In addition to members of the Orthomyxoviridae and Paramyxoviridae, several coronaviruses have been shown to possess receptor-destroying activities. Purified bovine coronavirus (BCV) preparations have an esterase activity which inactivates O-acetylsialic acid-containing receptors on erythrocytes. Diisopropyl fluorophosphate (DFP) completely inhibits this receptor-destroying activity of BCV, suggesting that the viral enzyme is a serine esterase. Treatment of purified BCV with [3H]DFP and subsequent sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the proteins revealed that the E3 protein was specifically phosphorylated. This finding suggests that the esterase/receptor-destroying activity of BCV is associated with the E3 protein. Furthermore, treatment of BCV with DFP dramatically reduced its infectivity in a plaque assay. It is assumed that the esterase activity of BCV is required in an early step of virus replication, possibly during virus entry or uncoating.

Molecular cloning and sequence determination of the fusion protein gene of human parainfluenza virus type 1

Undegraded mRNA transcripts were isolated from human parainfluenza virus type 1 (hPIV-1)-infected LLC-MK2 cells and their size was determined through denaturing agarose electrophoresis. The two predominantly represented mRNA species (1.65 and 1.87 kb) are similar in size to other paramyxoviral mRNAs that encode their respective glycoproteins. The cDNA transcripts corresponding to these two mRNAs were used to construct two size-restricted cDNA libraries. A cDNA clone, containing a 1.87-kb insert, was identified as encoding the hPIV-1 fusion protein by positively hybridizing with a synthetic oligonucleotide mix whose sequence was derived from the conserved sequences of other paramyxoviral F0 genes. The nucleotide sequence of the cDNA insert was determined and found to contain a single, large open reading frame encoding a putative protein of 60,795 Da consisting of 556 amino acids. Comparison of the amino acid sequence with the fusion proteins of other paramyxoviruses enabled the identification of the highly conserved amino acids of the F1 N-terminus. In addition, the positions of the hydrophobic signal and transmembrane regions, cysteine, and proline residues are all conserved. These analyses confirm that the cDNA sequence is that of the F0 protein. The 5' end of the fusion protein mRNA was determined by primer extension to lie 155 bases beyond the 5' end of the cDNA insert.

Introduction of macromolecules into mammalian cells by cell fusion

Proteins with molecular weights of up to 500K can be enclosed in erythrocyte ghosts by exposing the ghosts to hypotonic solution containing these proteins. The proteins can then be introduced into recipient cells by fusing the ghosts with the cells using HVJ, PEG, or influenza virus. Some applications of this method are described. By an improved method, 15 kbp DNA and IgM (900 kDa) can be entrapped in erythrocyte membranes and these are then treated with liposomes containing gangliosides and HVJ. These treated membranes containing large macromolecules fuse with almost 100% of the recipient cells used. Naked liposomes infrequently fuse with cultured cells, so introduction of their contents into cells is very inefficient. However, liposomes constituted from lipid and glycoproteins (HN and F) of HVJ (Sendai virus), by removing a nonionic detergent, fuse with cells about 200 times more efficiently than naked liposomes. Naked liposomes can fuse with specific cells, such as cells infected with subacute sclerosing panencephalitis virus or with human immunodeficiency virus. Plasmid DNA and mRNA of up to about 40 kbp can be entrapped efficiently in liposomes associated with gangliosides formed by reverse-phase evaporation, and then reacted with HVJ. The contents of the resulting liposomes with HVJ can be introduced efficiently into cultured cells in a suspended or plated state, and nearly all the cells then express the gene transiently. This procedure is also effective for obtaining stable transformants of many kinds of cultured cells.

Structure, function, and intracellular processing of paramyxovirus membrane proteins

Paramyxoviruses are a fascinating group of viruses with diverse hosts and disease manifestations. They are valuable systems for studying viral pathogenesis, molecular mechanisms of negative strand viral replication, and glycoprotein structure and function. In the past few years this group of viruses has received increased attention and as a result there is a wealth of new information. For example, most of the genes of many paramyxoviruses have been cloned and sequenced. The recent availability of sequence information from a number of paramyxoviruses now allows the direct comparison of the amino acid sequence and determinants of secondary structure of analogous genes across the family of viruses. Such comparisons are revealing for two reasons. First, results provide clues to the evolution of these viruses. Second, and more importantly, comparisons of analogous genes may point to sequences and structural determinants that are central to the function of the individual proteins. Below is a comparison of five of the paramyxovirus genes with a discussion of the implications of common structural determinants for function, intracellular processing, and evolutionary origin. The focus is on the paramyxovirus membrane proteins, although other proteins are discussed briefly.

Immunization with vaccinia virus recombinants that express the surface glycoproteins of human parainfluenza virus type 3 (PIV3) protects patas monkeys against PIV3 infection

Patas monkeys (Eryphrocebus patas) were immunized intradermally with two vaccinia virus recombinants that individually express the hemagglutinin-neuraminidase glycoprotein or the fusion glycoprotein of human parainfluenza virus type 3 (PIV3). These immunizations induced a high titer of PIV3 serum-neutralizing antibodies. At 1 month after immunization, monkeys were challenged intratracheally with PIV3. Subsequent virus replication was reduced in these monkeys by 3.2 log10 and 1.9 log10 (mean peak virus titers) in the upper and lower respiratory tracts, respectively, compared with control animals. The average duration of virus shedding was also reduced from 9.0 to 3.4 days in the upper respiratory tract and from 5.3 to 1.2 days in the lower respiratory tract. These findings demonstrate that a single intradermal dose of live recombinant vaccinia viruses can significantly restrict the replication of a virus which primarily infects the epithelial cells of the respiratory tract.

Role of individual glycoproteins of human parainfluenza virus type 3 in the induction of a protective immune response

Affinity-purified hemagglutinin-neuraminidase (HN) and fusion (F) glycoproteins of human parainfluenza virus type 3 (P13 virus) were used to investigate their role in the induction of a protective immune response following immunization of hamsters. The efficacy of immunization with the glycoprotein antigens was tested by challenge infection. Results of virus recovery from lungs and trachea demonstrated that although immunization with HN or F alone induced an antibody response to the respective glycoproteins, it did not provide a significant level of protection. However, immunization with a mixture of both purified glycoproteins induced higher virus-neutralizing activity in bronchial lavages and afforded complete protection from challenge infection. Similarly, incomplete protection was observed after passive transfer of monospecific rabbit antibody to the purified HN or F in baby hamsters. On the other hand, passive transfer of a mixture of antibodies to HN and F conferred a higher level of protection. Thus, the presence of antibody to both glycoproteins of P13 virus may be essential for protective immunity.

Antigenic relationships between avian paramyxoviruses. III. A mathematical model of antigenic drift and a computer-assisted approach for construction of a phylogenetic tree

The suggested model of antigenic kinship between related paramyxoviruses is based on another concept of antigenic determinant, as compared to the previously suggested combinatorial mathematical model by the authors. According to it, antigenic changes of any determinant do not proceed by "leaps" but can be changed gradually. Such changed determinant can induce a correspondingly changed type of antibodies which still preserve a certain kinship to the original type of the determinant (before its changing) revealed by cross reaction serological tests. Accordingly, there can be "families" of the determinants differing by degree of relatedness to (or, reversely, by antigenic distance from) the "original" ("ancestor") determinant. In addition to another interpretation of the antigenic kinship, the new mathematical model was used as an approach for revealing phylogenetic relationships between antigenically related viruses.

Dynamics of functional maturation and inactivation of HN glycoprotein in NDV-infected chick embryo fibroblasts

In avirulent NDV strain-infected chick embryo cells treated with cycloheximide at different intervals post infection a decrease of the level of hemagglutinating (HA) and neuraminidase (Nase) activities was observed. Studies on this system led to conclusion that the HA-Nase (HN) glycoprotein molecules are unstable and the actual amount of the functionally active (mature) HN entities is determined by a dynamic equilibrium between the antidromic processes of the HN functional maturation and inactivation. Kinetic studies on the actual intracellular levels of the HA and Nase activities using 5 min intervals of their detection after the cycloheximide treatment permitted to uncouple the processes of the HN maturation and inactivation. Analytical part of the studies made it possible to compute quantitative parameters of the involved processes: (a) pool size of the functionally nonactive HN precursors, (b) time needed for their functional maturation, and (c) rate of their inactivation.

Expression of the F and HN glycoproteins of human parainfluenza virus type 3 by recombinant vaccinia viruses: contributions of the individual proteins to host immunity

cDNA clones containing the complete coding sequences for the human parainfluenza virus type 3 (PIV3) fusion (F) and hemagglutinin-neuraminidase (HN) glycoprotein genes were inserted into the thymidine kinase gene of vaccinia virus (WR strain) under the control of the P7.5 early-late vaccinia virus promotor. The recombinant vaccinia viruses, designated vaccinia-F and vaccinia-HN, expressed glycoproteins in cell culture that appeared to be authentic with respect to glycosylation, disulfide linkage, electrophoretic mobility, cell surface expression, and, in the case of the HN protein, biological activity. Cotton rats inoculated intradermally with vaccinia-HN developed serum neutralizing antibody titers equal to that induced by respiratory tract infection with PIV3, whereas animals receiving vaccinia-F had threefold lower neutralizing antibody titers. A single immunization with either recombinant vaccinia virus induced nearly complete resistance in the lower respiratory tract of these animals. With regard to protection in the upper respiratory tract, animals immunized with vaccinia-HN or vaccinia-F exhibited reductions in PIV3 replication of greater than 3,000-fold and 6-fold, respectively. This large difference (greater than 500-fold) in reduction of PIV3 replication in the upper respiratory tract was in contrast to the relatively modest difference (3-fold) in serum neutralizing antibody titers induced by vaccinia-HN versus vaccinia-F. This dissociation between the level of neutralizing antibodies and protection suggested that immunity to PIV3 is complex, and that immune mechanisms other than serum neutralizing antibodies make important contributions to resistance to infection. Overall, under these experimental conditions, vaccinia-HN induced a substantially more protective immune response than did vaccinia-F.

The influenza C virus glycoprotein (HE) exhibits receptor-binding (hemagglutinin) and receptor-destroying (esterase) activities

A cDNA copy of RNA segment 4 of influenza C/Cal/78 virus was cloned into an SV40 vector and expressed in CV-1 cells. The gene product expressed from the SV40 recombinant virus was immunoprecipitated by monoclonal antibodies directed against the influenza C virus glycoprotein. Cells infected with the recombinant virus also exhibited C virus-specific hemagglutinin and O-acetylesterase activity. This suggests that the same C virus protein is associated with receptor-binding as well as receptor-destroying activity. The latter viral activity was measured using as substrates bovine submaxillary mucin or a low molecular weight compound p-nitrophenylacetate. In analogy to the parainfluenza virus HN protein, the influenza C virus glycoprotein was termed HE, because it possesses hemagglutinin and esterase (receptor-destroying) activity.

Expression of biologically active and antigenically authentic parainfluenza type 3 virus hemagglutinin-neuraminidase glycoprotein by a recombinant baculovirus

The hemagglutinin-neuraminidase (HN) gene of human type 3 parainfluenza virus has been inserted into a baculovirus expression vector under the control of the polyhedrin promoter. HN protein produced in insect cells by the recombinant baculovirus appeared to be glycosylated, was transported to the cell surface, and was biologically active. All of the HN epitopes previously mapped functionally to a region(s) involved in neuraminidase and/or hemagglutination activities were conformationally unaltered on the recombinant protein. The HN produced in this system also induced a protective immune response in immunized cotton rats. From these studies we conclude that the HN expressed in insect cells represents a source of authentic HN glycoprotein suitable for structural analysis and immunization.

Localization of functional sites on the hemagglutinin-neuraminidase glycoprotein of Sendai virus by sequence analysis of antigenic and temperature-sensitive mutants

To locate the various functions associated with the hemagglutinin-neuraminidase (HN) glycoprotein of Sendai virus in the primary structure of the protein, a temperature-sensitive (ts) mutant and seven antigenic mutants were sequenced. The ts mutant was defective in its ability to agglutinate erythrocytes and infect host cells, while its neuraminidase activity was normal. Its sequence revealed two closely spaced amino acid substitutions (residues 262 and 264) and one distant substitution (residue 461). Revertants could not be isolated, suggesting that more than one of the substitutions is responsible for the defective hemagglutinating activity. The antigenic mutants were selected with monoclonal antibodies that delineate four nonoverlapping antigenic sites (I-IV) and separately inhibit hemagglutinating, neuraminidase, and hemolysis activities. Mutants selected with antibodies to antigenic sites I-III were used to map these functions on the primary sequence of HN. Each antigenic mutant had a single point mutation in the HN gene that resulted in an amino acid substitution in the protein. A site II mutant selected with an antibody which inhibits hemolysin activity had a substitution at amino acid 420, while a mutant selected with antibody that inhibits only erythrocyte binding (site III) had a substitution at amino acid 541. Two antigenic mutants selected with an antibody that inhibits hemagglutination and neuraminidase activities (site I) had amino acid substitutions in close proximity (residues 277 and 279) to the two closely spaced substitutions of the ts mutant. These findings suggest that the region defined by the ts mutant and these two antigenic mutants is involved in host cell binding. Antigenic mutants selected with another site I antibody had amino acid changes at residue 184, indicating that antigenic site I is discontinuous in the primary sequence. This antibody blocks only hemagglutination, but mutants selected with it had a decreased neuraminidase activity. This finding supports the idea that the neuraminidase site is close to, but distinct from, the hemagglutination site.

Newcastle disease vaccine (La Sota) strain specific monoclonal antibody. Brief report

Newcastle disease virus vaccine strain (La Sota) specific monoclonal antibody (La-1) was produced by immunizing mice with isolated glycoproteins of strain La Sota. This antibody was recognized only in the ELISA test in which it bound exclusively to La Sota strain out of a range of over 300 lentogenic, mesogenic and velogenic strains examined.

Resistance of a measles virus mutant to fusion inhibitory oligopeptides is not associated with mutations in the fusion peptide

The nucleotide sequence and predicted amino acid sequence has been obtained for the fusion (F) protein gene of the R93 strain of measles virus and compared to that of the parental strain, Edmonston B. The R93 strain is a mutant measles virus which is able to grow and induce cell fusion in the presence of the fusion inhibiting oligopeptide, Z-D-Phe-L-Phe-L-(NO2)Arg (SV4814). Primer extension sequencing on isolated R93 mRNA demonstrated the presence of three nucleotide changes leading to three amino acid changes, none of which are in the hydrophobic NH2-terminal region of the F1 polypeptide.

Comparison of the relative roles of the F and HN surface glycoproteins of the paramyxovirus simian virus 5 in inducing protective immunity

To compare the relative roles of the paramyxovirus simian virus 5 (SV5) major surface glycoproteins, fusion (F) and hemagglutinin-neuraminidase (HN), in inducing protective immunity, two recombinant vaccinia viruses were constructed. The F and HN polypeptides expressed by the recombinant viruses were indistinguishable from their authentic SV5 counterparts in electrophoretic mobility, glycosylation, and, for the F protein, cleavage of the precursor, F0, to the disulfide-linked subunits F1 and F2. Injection of rabbits and hamsters with live recombinant virus elicited an antibody response to either F or HN and provided a source of monospecific polyclonal antisera to the SV5 proteins. The vaccinia virus-SV5 F (vaccinia-F) recombinant induced higher levels of neutralizing antibody than did the vaccinia-HN recombinant, but animals inoculated with vaccinia-HN were better protected from challenge with SV5. Animals infected with both the vaccinia-HN and vaccinia-F viruses were nearly as well protected from challenge as were animals infected with SV5.

Detergent-treated Newcastle disease virus as an agar gel precipitin test antigen

A soluble Newcastle disease virus (NDV) agar gel precipitin (AGP) antigen was prepared by treating 100-fold concentrated NDV with a nonionic detergent. Virus concentration prior to detergent treatment was best accomplished by ultracentrifugation or by a simple, less expensive, and more practical method involving acid (HCl) precipitation of NDV. Virus concentrated by polyethylene glycol precipitation was found to have a low antigen titer and was not considered suitable as an AGP antigen. Antigens derived from the LaSota, Roakin, and Texas GB strains formed at least two lines of identity in the AGP test as early as 24 hr after inoculation of the agar gels. Virus used for AGP antigen production could be grown in chicken embryos from an NDV-immune as well as susceptible breeder flock. The NDV AGP antigen was found to be stable after 20 consecutive freezing and thawing cycles and storage at -20 C or 4 C for at least 6 months. Detergent-treated NDV was used as an AGP test antigen to determine serum antibody responses of chickens following infection and vaccination. Hemagglutination-inhibition, virus neutralization, and enzyme-linked immunosorbent assay antibody production was also evaluated for comparative purposes. The AGP test was found to be useful as an aid in diagnosing field infections and assessing inactivated virus vaccination responses. These purposes were achieved by demonstrating an increase in the number of AGP positive chickens between preinfection and postinfection or vaccination bleedings. The ease of performance and low cost of the AGP test favors its use for screening large numbers of serum samples, perhaps in conjunction with a quantitative serological test.(ABSTRACT TRUNCATED AT 250 WORDS)

Nucleotide sequence of the gene encoding the Newcastle disease virus hemagglutinin-neuraminidase protein and comparisons of paramyxovirus hemagglutinin-neuraminidase protein sequences

The nucleotide sequence of cloned cDNA copies of the mRNA encoding the Newcastle disease virus (NDV), strain A-V, hemagglutinin-neuraminidase (HN) protein was determined. A single open reading frame in the sequence encodes a protein of 570 amino acids with a calculated molecular weight of 62,280. The predicted protein sequence contains only one obvious potential membrane spanning region, located 27 amino acids from the amino terminus of the sequence. The predicted sequence contains 6 glycosylation sites and 14 cysteine residues. Comparison of the NDV HN protein sequence with three other paramyxovirus HN protein sequences reveals two regions that have homologies in all four sequences. The conserved cysteine residues are clustered in these two regions. One conserved region is located near the middle of the predicted sequence while the second region is in the carboxy terminal third of the molecule. The presence of conserved regions suggests the importance of these areas of the molecule in the structure or function of the protein.

Distinct functions of antigenic sites of the HN glycoprotein of Sendai virus

Monoclonal antibodies specific for the hemagglutinin-neuraminidase (HN) glycoprotein of Sendai virus were used to examine the antigenic structure of HN and its role in the initiation of infection and immunity. Using 10 anti-HN antibodies, four distinct antigenic sites designated I-IV were topographically mapped on the HN molecule by competitive-binding assays. To relate the biological functions of HN to its antigenic structure, anti-HN antibodies were analyzed for their inhibitory activity in neuraminidase, hemagglutination, and hemolysis inhibition tests. Antibodies to antigenic site I inhibited hemagglutination and one of these antibodies also inhibited neuraminidase activity. Antibodies to site II inhibited neither activity. However, hemolysis an F protein activity was inhibited, suggesting that these antibodies which bind to HN interfere with F-mediated fusion. Antigenic sites III and IV had different effects on the hemagglutinating and neuraminidase functions of HN: Site III antibodies inhibited hemagglutination while antibodies to site IV only inhibited neuraminidase activity. Antibodies to each antigenic site inhibited virus production. Since antibodies to sites I and III inhibited hemagglutination, it is likely that they block virus adsorption. Antibodies to HN site II only inhibited hemolysis, and therefore, may prevent virus penetration. Antibodies reacting with site IV inhibited virus production after virus penetration. Since neuraminidase activity was the only function inhibited, the viral enzyme may be involved in virus release. The fact that site IV antibodies inhibited neuraminidase but not hemagglutination suggests that these sites are distinct.

The fusion glycoprotein of Sendai virus: sequence analysis of an epitope involved in fusion and virus neutralization

To localize the amino acid residues on the F glycoprotein that are involved in Sendai virus fusion and virus neutralization, an anti-F monoclonal antibody which inhibits these functions was used to select three antigenic variants. Sequence analysis of the entire F gene of the three variants identified a single mutation that was responsible for the loss of antibody binding. The mutation, a proline to glutamine substitution at residue 399, was at a position in the primary sequence far removed from the hydrophobic F1-NH2 terminus thought to be directly involved in fusion. A synthetic peptide, comprising amino acid sequences in the region of the mutation, bound to the antibody used to select the variants, suggesting that the site of mutation is also the site of antibody binding. This information suggests that in the three-dimensional structure of the F molecule the amino acid residues around proline 399 are located close to the F1-NH2 terminus, and that fusion is directly inhibited by antibody binding. Other less likely alternatives are discussed.

Determination of the orientation of an integral membrane protein and sites of glycosylation by oligonucleotide-directed mutagenesis: influenza B virus NB glycoprotein lacks a cleavable signal sequence and has an extracellular NH2-terminal region

The membrane orientation of the NB protein of influenza B virus, a small (Mr, approximately 18,000) glycoprotein with a single internal hydrophobic domain, was investigated by biochemical and genetic means. Cell fractionation and protein solubility studies indicate NB is an integral membrane protein, and NB has been shown to be a dimer under nonreducing conditions. Treatment of infected-cell surfaces with proteinase K and endoglycosidase F and immunoprecipitation with a site-specific antibody suggests that the 18-amino-acid NH2-terminal region of NB is exposed at the cell surface. Oligonucleotide-directed mutagenesis to eliminate each of the four potential sites of N-linked glycosylation and expression of the mutant NB proteins in eucaryotic cells suggest that the two sites adjacent to the NH2 terminus are glycosylated. This provides further evidence that NB, which lacks a cleavable NH2-terminal signal sequence, has an exposed NH2 terminus at the cell surface.