A novel influenza A virus mitochondrial protein that induces cell death.

While searching for alternative reading-frame peptides encoded by influenza A virus that are recognized by CD8+ T cells, we found an abundant immunogenic peptide encoded by the +1 reading frame of PB1. This peptide derives from a novel conserved 87-residue protein, PB1-F2, which has several unusual features compared with other influenza gene products in addition to its mode of translation. These include its absence from some animal (particularly swine) influenza virus isolates, variable expression in individual infected cells, rapid proteasome-dependent degradation and mitochondrial localization. Exposure of cells to a synthetic version of PB1-F2 induces apoptosis, and influenza viruses with targeted mutations that interfere with PB1-F2 expression induce less extensive apoptosis in human monocytic cells than those with intact PB1-F2. We propose that PB1-F2 functions to kill host immune cells responding to influenza virus infection.

Physical and functional association of the major histocompatibility complex class I heavy chain alpha3 domain with the transporter associated with antigen processing.

CD8+ T lymphocytes recognize antigens as short, MHC class I-associated peptides derived by processing of cytoplasmic proteins. The transporter associated with antigen processing translocates peptides from the cytosol into the ER lumen, where they bind to the nascent class I molecules. To date, the precise location of the class I-TAP interaction site remains unclear. We provide evidence that this site is contained within the heavy chain alpha3 domain. Substitution of a 15 amino acid portion of the H-2Db alpha3 domain (aa 219-233) with the analogous MHC class II (H-2IAd) beta2 domain region (aa 133-147) results in loss of surface expression which can be partially restored upon incubation at 26 degrees C in the presence of excess peptide and beta2-microglobulin. Mutant H-2Db (Db219-233) associates poorly with the TAP complex, and cannot present endogenously-derived antigenic peptides requiring TAP-dependent translocation to the ER. However, this presentation defect can be overcome through use of an ER targeting sequence which bypasses TAP-dependent peptide translocation. Thus, the alpha3 domain serves as an important site of interaction (directly or indirectly) with the TAP complex and is necessary for TAP-dependent peptide loading and class I surface expression.

Class I molecules retained in the endoplasmic reticulum bind antigenic peptides.

We isolated major histocompatibility complex (MHC)-specific viral peptides from cells infected with influenza virus in the continuous presence of the drug brefeldin A, which blocks exocytosis of newly synthesized MHC class I molecules. MHC-specific peptides were also isolated from cells expressing mouse Kd class I MHC molecules whose cytoplasmic domain was substituted by that of the adenovirus E3/19K glycoprotein. This molecule was retained in an intracellular pre-Golgi complex compartment as demonstrated by immunocytochemical and biochemical means. Since we show that intracellular association of antigenic peptides with such retained class I molecules is necessary for their isolation from cellular extracts, this provides direct evidence that naturally processed peptides associate with class I MHC molecules in an early intracellular exocytic compartment.

Two novel routes of transporter associated with antigen processing (TAP)-independent major histocompatibility complex class I antigen processing.

Jaw1 is an endoplasmic reticulum (ER) resident protein representative of a class of proteins post translationally inserted into membranes via a type II membrane anchor (cytosolic NH2 domain, lumenal COOH domain) in a translocon-independent manner. We found that Jaw1 can efficiently deliver a COOH-terminal antigenic peptide to class I molecules in transporter associated with antigen processing (TAP)-deficient cells or cells in which TAP is inactivated by the ICP47 protein. Peptide delivery mediated by Jaw1 to class I molecules was equal or better than that mediated by the adenovirus E3/19K glycoprotein signal sequence, and was sufficient to enable cytofluorographic detection of newly recruited thermostabile class I molecules at the surface of TAP-deficient cells. Deletion of the transmembrane region retargeted Jaw1 from the ER to the cytosol, and severely, although incompletely, abrogated its TAP-independent peptide carrier activity. Use of different protease inhibitors revealed the involvement of a nonproteasomal protease in the TAP-independent activity of cytosolic Jaw1. These findings demonstrate two novel TAP-independent routes of antigen processing; one based on highly efficient peptide liberation from the COOH terminus of membrane proteins in the ER, the other on delivery of a cytosolic protein to the ER by an unknown route.

The human immunodeficiency virus type 1 (HIV-1) Vpu protein interferes with an early step in the biosynthesis of major histocompatibility complex (MHC) class I molecules.

The human immunodeficiency virus type 1 (HIV-1) vpu gene encodes a small integral membrane phosphoprotein with two established functions: degradation of the viral coreceptor CD4 in the endoplasmic reticulum (ER) and augmentation of virus particle release from the plasma membrane of HIV-1-infected cells. We show here that Vpu is also largely responsible for the previously observed decrease in the expression of major histocompatibility complex (MHC) class I molecules on the surface of HIV-1-infected cells. Cells infected with HIV-1 isolates that fail to express Vpu, or that express genetically modified forms of Vpu that no longer induce CD4 degradation, exhibit little downregulation of MHC class I molecules. The effect of Vpu on class I biogenesis was analyzed in more detail using a Vpu-expressing recombinant vaccinia virus (VV). VV-expressed Vpu induces the rapid loss of newly synthesized endogenous or VV-expressed class I heavy chains in the ER, detectable either biochemically or by reduced cell surface expression. This effect is of similar rapidity and magnitude as the VV-expressed Vpu-induced degradation of CD4. Vpu had no discernible effects on cell surface expression of VV-expressed mouse CD54, demonstrating the selectivity of its effects on CD4 and class I heavy chains. VV-expressed Vpu does not detectably affect class I molecules that have been exported from the ER. The detrimental effects of Vpu on class I molecules could be distinguished from those caused by VV-expressed herpes virus protein ICP47, which acts by decreasing the supply of cytosolic peptides to class I molecules, indicating that Vpu functions in a distinct manner from ICP47. Based on these findings, we propose that Vpu-induced downregulation of class I molecules may be an important factor in the evolutionary selection of the HIV-1-specific vpu gene by contributing to the inability of CD8+ T cells to eradicate HIV-1 from infected individuals.

Introduction of a glycosylation site into a secreted protein provides evidence for an alternative antigen processing pathway: transport of precursors of major histocompatibility complex class I-restricted peptides from the endoplasmic reticulum to the cytosol.

We found that the presentation of a H-2Kd-restricted determinant from influenza virus nucleoprotein (NP) to T cells is strictly dependent on expression of the transporter associated with antigen presentation (TAP), regardless of whether NP is expressed as a cytosolic or secreted NP (SNP). Introducing an N-linked glycosylation site into the determinant selectively reduced presentation of SNP. This indicates that glycosylation does not interfere with TAP-transported peptides, and therefore that cytosolic peptides derived from SNP must have been exposed to the glycosylation machinery of the endoplasmic reticulum (ER) before their existence in the cytosol. Based on these findings, we propose that TAP-dependent processing of at least some ER-targeted proteins entails the reimportation of protein from the secretory pathway to the cytosol, where the protein is processed via the classical pathway.

Dislocation of type I membrane proteins from the ER to the cytosol is sensitive to changes in redox potential.

The human cytomegalovirus (HCMV) gene products US2 and US11 dislocate major histocompatibility class I heavy chains from the ER and target them for proteasomal degradation in the cytosol. The dislocation reaction is inhibited by agents that affect intracellular redox potential and/or free thiol status, such as diamide and N-ethylmaleimide. Subcellular fractionation experiments indicate that this inhibition occurs at the stage of discharge from the ER into the cytosol. The T cell receptor alpha (TCR alpha) chain is also degraded by a similar set of reactions, yet in a manner independent of virally encoded gene products. Diamide and N-ethylmaleimide likewise inhibit the dislocation of the full-length TCR alpha chain from the ER, as well as a truncated, mutant version of TCR alpha chain that lacks cysteine residues. Cytosolic destruction of glycosylated, ER-resident type I membrane proteins, therefore, requires maintenance of a proper redox potential for the initial step of removal of the substrate from the ER environment.

Intracellular localization of proteasomal degradation of a viral antigen.

To better understand proteasomal degradation of nuclear proteins and viral antigens we studied mutated forms of influenza virus nucleoprotein (NP) that misfold and are rapidly degraded by proteasomes. In the presence of proteasome inhibitors, mutated NP (dNP) accumulates in highly insoluble ubiquitinated and nonubiquitinated species in nuclear substructures known as promyelocytic leukemia oncogenic domains (PODs) and the microtubule organizing center (MTOC). Immunofluorescence revealed that dNP recruits proteasomes and a selective assortment of molecular chaperones to both locales, and that a similar (though less dramatic) effect is induced by proteasome inhibitors in the absence of dNP expression. Biochemical evidence is consistent with the idea that dNP is delivered to PODs/MTOC in the absence of proteasome inhibitors. Restoring proteasome activity while blocking protein synthesis results in disappearance of dNP from PODs and the MTOC and the generation of a major histocompatibility complex class I-bound peptide derived from dNP but not NP. These findings demonstrate that PODs and the MTOC serve as sites of proteasomal degradation of misfolded dNP and probably cellular proteins as well, and imply that antigenic peptides are generated at one or both of these sites.

Strategies for tumor elimination by cytotoxic T lymphocytes.

Despite differences in their tissue of origin, many tumors share high level expression of certain tumor-associated proteins. Our laboratory has focused on the possibility of utilizing antigenic components of these proteins as a focus for T-cell immunotherapy of cancer. The advantage of targeting such commonly expressed proteins is the fact that such therapy could be of value in eliminating many different types of tumors. A potential barrier in the identification of T-cell epitopes derived from these proteins and presented by tumor cells is the fact that these proteins are also expressed at low levels in some normal tissues, and therefore, self-tolerance may eliminate T cells that are capable of recognizing these epitopes with high avidity. We have utilized two different murine model systems to explore the extent to which self-tolerance may limit the immune response to a tumor-specific antigen. The first compared the ability of mice deficient in expression of murine p53 (p53 knock-out mice) and normal mice, to respond against several epitopes of the p53 protein. The second model compares the ability of conventional mice with transgenic mice that express the influenza hemagglutinin in the periphery to respond to a dominant antigenic peptide of this transgene product. In both models we have investigated the effect self-tolerance has on elimination of tumors expressing the toleragen.

Electrophoresis of the complementary strands of the double-stranded Kemerovo virus RNAs in agarose-urea gel.

Single-stranded (ss)RNAs derived from 10 double-stranded (ds)RNA segments of Kemerovo virus (KV) were separated into 13 RNA bands by agarose-urea gel electrophoresis. The complementary strands of the dsRNA segments 1, 9 and 10 displayed different electrophoretic mobility. An attempt was made to determine the origin of the ssRNA bands. The ssRNA bands originating from the dsRNA segments 1, 2, 3, 9 and 10 were identified unequivocally, while those originating from the dsRNA segments 4, 5, 6, 7 and 8 were characterized partially. The minus RNA strands of the dsRNA segments 9 and 10 exhibited higher electrophoretic mobilities as their complementary plus RNA strands.

Evidence of 5'-terminal modification in the kemerovo virus double-stranded RNA segments and its removal by treatment with alkaline phosphatase.

Only one strand of each double-stranded (ds) RNA segment of the Kemerovo virus genome was 5'end-labelled using gamma-32P-ATP and T 4 polynucleotide kinase after preceding dephosphorylation of 5'ends by calf intestinal alkaline phosphatase. This suggests a 5'-terminal modification of the one of complementary strands in the ds RNA segments.

Polypeptides induced in chick embryo cells by Kemerovo virus.

Fifteen polypeptides induced by Kemerovo virus were detected in chick embryo cells (Mr 140, 98, 89, 72, 65, 62, 57, 54, 50, 47, 43, 41, 39, 31 kD, and 30 kD). Nine of them, namely the 140, 98, 65, 62, 57, 54, 50, 47 kD, and 41 kD polypeptides were also found in the partially purified virus. However, the latter contained also considerable amount of host cell proteins, predominantly the 205 kD, 45 kD, and 37 kD polypeptides. In the electron microscope the spherical viral particles exhibited a poorly defined surface structure of a diameter of 70-75 nm.

Polyacrylamide gel electrophoresis of the Kemerovo virus RNA.

Ribonuclease (RNase)-resistant RNA was isolated from partially purified Kemerovo virus by gel chromatography and or sucrose density gradient centrifugation. Double-stranded (ds) RNA only was found in the purified viral cores. The RNAs from both sources exhibited the same pattern of distribution in polyacrylamide gels. Ten dsRNA segments were identified. According to the results of coelectrophoresis of the Kemerovo virus and reovirus dsRNAs, the size of Kemerovo virus genome was estimated to be of about 11.7 X 10(6). The grouping of Kemerovo virus double-stranded segments according to their size in polyacrylamide gels corresponded to the 2:4:3:1 pattern.

Northern blot hybridization analysis of polyoma virus-specific RNA synthesized under the block of virus replication by 5-bromo-2'-deoxyuridine.

Polyoma (Py) virus-specific RNA, synthesized at reduced level in infected cells in the presence of antiviral compound 5-bromo-2'-deoxyuridine (BrdUrd) was characterized in more detail by Northern blot hybridization analysis. The results obtained with total, cytoplasmic and poly(A)RNA, isolated from mouse embryo cell cultures 42 hrs p.i. indicate that BrdUrd (6.34 micrograms/ml) lowers the level of typical classes of major virus DNA transcripts to a similar extent and that no new, atypical transcription products are formed.

Replacement of glycoprotein B gene in the herpes simplex virus type 1 strain ANGpath DNA by that originating from nonpathogenic strain KOS reduces the pathogenicity of recombinant virus.

Herpes simplex virus type-1 (HSV-1) strain ANGpath and its recombinants, in which the 8.1 kbp BamHI G restriction fragment (0.345-0.399) containing the glycoprotein B (gBpath) gene (UL27) or its subfragments-coding either for cytoplasmic or surface domains of gB-had been replaced with the corresponding fragments from nonpathogenic KOS virus DNA (gBKOS), were tested for their pathogenicity for DBA/2 mice and rabbits. The recombinant ANGpath/B6KOS prepared by transferring the 2.7 kbp SstI-SstI subfragment (0.351-0.368) of the BamHI GKOS fragment still had the original sequence of ANGpath DNA coding for the syn3 marker in the cytoplasmic domain of gB and was pathogenic for mice as well as for rabbits. Virological and immunohistological studies in DBA/2 mice infected with the latter pathogenic recombinant and with ANGpath showed the presence of infectious virus and viral antigen at inoculation site (epidermis, subcutaneous connective tissue and striated muscle in the area of right lip), in homolateral trigeminal nerve and ganglion, brain stem, midbrain, thalamic and hypothalamic nuclei. In contrast, nonpathogenic recombinants ANGpath/syn+B6KOS (prepared by transferring the whole BamHI GKOS fragment) and ANGpath/syn+KOS (prepared by transferring the 0.8 kbp BamHI-SstI subfragment of the BamHI GKOS fragment) showed limited haematogenous and neural spread, but no evidence of replication in CNS; thus, their behaviour resembled that of the wild type strain KOS. The recombinant ANGpath/syn+KOS, which was not pathogenic for mice, still remained pathogenic for rabbits, a phenomenon indicating the presence of an additional locus in the gB molecule participating on virulence. Sequencing the 1478 bp SstI-SstI subfragment of the BamHI G(path) fragment (nt 53,348-54,826 of UL segment) showed the presence of at least 3 mutations as compared to the KOS sequence, from which the change of cytosine to thymine at nt 54,251 altered the codon for arginine to that for histidine (amino acid 515) in the gB polypeptide chain.

Expression of a membrane protease enhances presentation of endogenous antigens to MHC class I-restricted T lymphocytes.

We find that expression of the membrane dipeptidyl carboxypeptidase angiotensin-converting enzyme (ACE) enhances presentation of certain endogenously synthesized peptides to major histocompatibility complex (MHC) class I-restricted cytotoxic T lymphocytes. ACE appears to function only in an intracellular secretory compartment of antigen-presenting cells. ACE-enhanced antigen presentation requires the expression of the putative antigenic peptide transporters, TAP1 and TAP2. These findings demonstrate that a protease can influence the processing of endogenously synthesized antigens and strongly suggest that longer peptides can be transported from the cytosol to a secretory compartment where trimming of antigenic peptides to the lengths preferred by MHC class I molecules can occur if the appropriate protease is present.

MHC-encoded proteasome subunits LMP2 and LMP7 are not required for efficient antigen presentation.

LMP2 and LMP7 are proteins encoded by MHC genes that are tightly linked to the genes encoding TAP, the transporter that conveys peptides from the cytosol to the endoplasmic reticulum for assembly with MHC class I molecules. LMP2 and LMP7 are subunits of a subset of proteasomes, large molecular assemblies with multi-proteolytic activities believed to degrade damaged and unwanted cellular proteins. Like TAP and class I molecules themselves, expression of LMP genes is enhanced after exposure of cells to IFN-gamma. These findings implicate LMP2 and LMP7 in the cytosolic production of antigenic peptides. Doubts have been cast, however, on the role of LMP2 and LMP7 in Ag processing, because cells lacking these proteins possess class I molecules that contain peptides quantitatively and qualitatively indistinguishable from the peptides bound to class I molecules derived from normal cells. In this paper we show that cells lacking LMP2 and LMP7 present seven TAP-dependent determinants derived from viral proteins. For two determinants, the kinetics of presentation are shown to be similar for LMP-expressing and -nonexpressing cells. We also demonstrate biochemically that peptide is not limiting in the assembly of class I molecules in LMP-nonexpressing cells. These findings provide additional evidence that LMP2 and LMP7 are not required for efficient Ag presentation, and suggest that these proteins have either a more specialized role in the production of class I-associated peptides, or are not involved in the processing of proteins for association with class I molecules.

TAP (transporter associated with antigen processing)-independent presentation of endogenously synthesized peptides is enhanced by endoplasmic reticulum insertion sequences located at the amino- but not carboxyl-terminus of the peptide.

Under most circumstances, cell surface MHC class I molecules display peptides derived from a cytosolic pool of proteins. The efficient presentation of such peptides requires the functioning of two MHC gene products [TAP1 and TAP2 (transporter-associated with Ag processing 1 and 2)] that form a complex that facilitates transmembrane movement of peptides from the cytosol to the endoplasmic reticulum, the site of peptide association with class I molecules. It has been previously shown that peptides can be presented in a TAP-independent manner in association with HLA A2.1 or H-2 Kd if they are expressed COOH-terminal to an endoplasmic reticulum insertion/signal sequence derived from the adenovirus E3/19K glycoprotein (Anderson et al., 1991. J. Exp. Med. 174: 489; Eisenlohr et al., 1992. Cell 71: 963). We show that: 1) the E3/19K signal sequence greatly enhances the presentation of each of four additional peptides tested in association with H-2 Kb or Kk, 2) the E3/19K signal sequence can be substituted by a signal sequence derived from beta-IFN, and 3) the E3/19K signal sequence does not function when located at the COOH terminus of antigenic peptides. These findings indicate that first, many peptides require TAP for efficient presentation to T cells, second, expression of peptides COOH-terminal to signal sequences is a generally applicable method of bypassing the TAP-dependence of peptide presentation and third, the leader sequence does not act to bypass TAP simply by increasing the hydrophobic nature of peptides.

Promiscuous liberation of MHC-class I-binding peptides from the C termini of membrane and soluble proteins in the secretory pathway.

TAP can efficiently transport peptides up to twice as long as those bound to MHC class I molecules, suggesting a role for endoplasmic reticulum (ER) proteases in the trimming of TAP-transported peptides. To better define ER processing of antigenic peptides, we examined the capacity of TAP-deficient cells to present determinants derived from ER-targeted proteins encoded by recombinant vaccinia viruses. TAP-deficient cells failed to present antigenic peptides from internal locations in secreted proteins to MHC class I-restricted T lymphocytes. The same peptides were liberated from the C termini of a secreted protein and the lumenal domains of two membrane proteins delivered to the ER via different routes. These findings suggest that proteases in the secretory compartment can liberate C-terminal antigenic peptides from virtually any context. We propose that this activity often participates in the removal of N-terminal extensions from TAP-transported peptides, thereby creating optimally sized products for MHC class I binding. We further demonstrate that ER trimming of C termini can occur if we express an appropriate carboxypeptidase in the secretory pathway. The absence of such trimming under normal circumstances suggests that carboxypeptidase activity is generally deficient in the ER, consistent with the concordance between the specificity of TAP and MHC class I molecules for the same types of C-terminal residues.

Generating MHC class I ligands from viral gene products.

MHC class I molecules function to present peptides comprised of eight to 11 residues to CD8+ T lymphocytes. Here we review the efforts of our laboratory to understand how cells generate such peptides from viral gene products. We particularly focus on the nature of substrates acted on by cytosolic proteases, the contribution of proteasomes and non-proteasomal proteases to peptide generation, the involvement of ubiquitination in peptide generation, the intracellular localization of proteasome generation of antigenic peptides, and the trimming of peptides in the endoplasmic reticulum.