A collaborative framework for 3D alignment and classification of heterogeneous subvolumes in cryo-electron tomography.

The limitation of using low electron doses in non-destructive cryo-electron tomography of biological specimens can be partially offset via averaging of aligned and structurally homogeneous subsets present in tomograms. This type of sub-volume averaging is especially challenging when multiple species are present. Here, we tackle the problem of conformational separation and alignment with a "collaborative" approach designed to reduce the effect of the "curse of dimensionality" encountered in standard pair-wise comparisons. Our new approach is based on using the nuclear norm as a collaborative similarity measure for alignment of sub-volumes, and by exploiting the presence of symmetry early in the processing. We provide a strict validation of this method by analyzing mixtures of intact simian immunodeficiency viruses SIV mac239 and SIV CP-MAC. Electron microscopic images of these two virus preparations are indistinguishable except for subtle differences in conformation of the envelope glycoproteins displayed on the surface of each virus particle. By using the nuclear norm-based, collaborative alignment method presented here, we demonstrate that the genetic identity of each virus particle present in the mixture can be assigned based solely on the structural information derived from single envelope glycoproteins displayed on the virus surface.

Automatic feature extraction and statistical shape model of the AIDS virus spike.

We introduce a method to automatically extract spike features of the AIDS virus imaged through an electron microscope. The AIDS virus spike is the primary target of drug design as it is directly involved in infecting host cells. Our method detects the location of these spikes and extracts a subvolume enclosing the spike. We have achieved a sensitivity of 80% for our best operating range. The extracted spikes are further aligned and combined to build a 4-D statistical shape model, where each voxel in the shape model is assigned a probability density function. Our method is the first fully automated technique that can extract subvolumes of the AIDS virus spike and be used to build a statistical model without the need for any user supervision. We envision that this new tool will significantly enhance the overall process of shape analysis of the AIDS virus spike imaged through the electron microscope. Accurate models of the virus spike will help in the development of better drug design strategies.

Recovery of fitness of a live attenuated simian immunodeficiency virus through compensation in both the coding and non-coding regions of the viral genome.

We have analyzed a SIV deletion mutant that was compromised both in viral replication and RNA packaging. Serial passage of this variant in two different T-cell lines resulted in compensatory reversion and the generation of independent groups of point mutations within each cell line. Within each group, single point mutations were shown to contribute to increased viral infectivity and the rescue of wild-type replication kinetics. The complete recovery of viral fitness ultimately correlated with the restoration of viral RNA packaging. Consistent with the latter finding was the rescue of Pr55 Gag processing, also restoring proper virus core morphology in mature virions. These seemingly independently arising groups of compensatory mutations were functionally interchangeable in regard to the recovery of wild type replication in rhesus PBMCs. These findings indicate that viral reversion that overcomes a genetic bottleneck is not limited to a single pathway, and illustrates the remarkable adaptability of lentiviruses.

Cryo-electron tomographic structure of an immunodeficiency virus envelope complex in situ.

The envelope glycoprotein (Env) complexes of the human and simian immunodeficiency viruses (HIV and SIV, respectively) mediate viral entry and are a target for neutralizing antibodies. The receptor binding surfaces of Env are in large part sterically occluded or conformationally masked prior to receptor binding. Knowledge of the unliganded, trimeric Env structure is key for an understanding of viral entry and immune escape, and for the design of vaccines to elicit neutralizing antibodies. We have used cryo-electron tomography and averaging to obtain the structure of the SIV Env complex prior to fusion. Our result reveals novel details of Env organisation, including tight interaction between monomers in the gp41 trimer, associated with a three-lobed, membrane-distal gp120 trimer. A cavity exists at the gp41-gp120 trimer interface. Our model for the spike structure agrees with previously predicted interactions between gp41 monomers, and furthers our understanding of gp120 interactions within an intact spike.

Compensatory substitutions restore normal core assembly in simian immunodeficiency virus isolates with Gag epitope cytotoxic T-lymphocyte escape mutations.

The evolution of human immunodeficiency virus type 1 (HIV-1) and simian immunodeficiency virus (SIV) as they replicate in infected individuals reflects a balance between the pressure on the virus to mutate away from recognition by dominant epitope-specific cytotoxic T lymphocytes (CTL) and the structural constraints on the virus' ability to mutate. To gain a further understanding of the strategies employed by these viruses to maintain replication competency in the face of the intense selection pressure exerted by CTL, we have examined the replication fitness and morphological ramifications of a dominant epitope mutation and associated flanking amino acid substitutions on the capsid protein (CA) of SIV/simian-human immunodeficiency virus (SHIV). We show that a residue 2 mutation in the immunodominant p11C, C-M epitope (T47I) of SIV/SHIV not only decreased CA protein expression and viral replication, but it also blocked CA assembly in vitro and virion core condensation in vivo. However, these defects were restored by the introduction of upstream I26V and/or downstream I71V substitutions in CA. These findings demonstrate how flanking compensatory amino acid substitutions can facilitate viral escape from a dominant epitope-specific CTL response through the effects of these associated mutations on the structural integrity of SIV/SHIV.

Distribution and three-dimensional structure of AIDS virus envelope spikes.

Envelope glycoprotein (Env) spikes on AIDS retroviruses initiate infection of host cells and are therefore targets for vaccine development. Though crystal structures for partial Env subunits are known, the structure and distribution of native Env spikes on virions is obscure. We applied cryoelectron microscopy tomography to define ultrastructural details of spikes. Virions of wild-type human immunodeficiency virus 1 (HIV-1) and a mutant simian immunodeficiency virus (SIV) had approximately 14 and approximately 73 spikes per particle, respectively, with some clustering of HIV-1 spikes. Three-dimensional averaging showed that the surface glycoprotein (gp120) 'head' of each subunit of the trimeric SIV spike contains a primary mass, with two secondary lobes. The transmembrane glycoprotein 'stalk' of each trimer is composed of three independent legs that project obliquely from the trimer head, tripod-like. Reconciling available atomic structures with the three-dimensional whole spike density map yields insights into the orientation of Env spike structural elements and possible structural bases of their functions.

Molecular and biological characterization of simian-human immunodeficiency virus-like particles produced by recombinant fowlpox viruses.

Virus-like particles (VLPs) mimicking the simian-human immunodeficiency virus SHIV89.6P (VLPSHIV) were produced by co-infection of Vero cells with fowlpox SIVgag/pol (FPgag/polSIV) and fowlpox HIV-1env89.6P (FPenv89.6P) recombinant viruses. As a necessary prerequisite for a more efficient vaccine approach, ultrastructural, functional and molecular characterizations of VLP(SHIV) were performed in the SHIV-macaque model to verify the similarity of these particles to SHIV89.6P. Here we show that VLPSHIV can infect T cells by fusion without replication, as demonstrated by the absence of new viral progeny in VLPSHIV-infected C8166 cells. Biochemical characterization showed identical protein profiles of VLPSHIV and SHIV89.6P, and ultrastructural analysis of Vero cells releasing VLPSHIV also confirmed the morphological similarity of these pseudovirions to SHIV89.6P particles. Viral mRNAs were also found packaged inside the core of VLPSHIV by RT-PCR and reverse transcriptase assays.

Conserved serines in simian immunodeficiency virus capsid are required for virus budding.

The simian immunodeficiency virus (SIV) capsid protein (CA), a constituent of the Pr55Gag polyprotein, is phosphorylated in virions but not in virus-producing cells (Rue, S.M., Roos, J.W., Tarwater, P.M., Clements, J.E., Barber, S.A., 2005. Phosphorylation and proteolytic cleavage of gag proteins in budded simian immunodeficiency virus. J. Virol. 79 (4), 2484-2492.). Using phosphoamino acid analysis of CA, we show that serine is the primary phosphate acceptor. A series of substitution mutants of serines in the CA domain of Pr55Gag were constructed in the infectious viral clone SIVmac239. These virus mutants were examined for defects in virus replication and virion infectivity, release, and morphology, as well as alterations in phosphorylation of CA-containing proteins. Although the virus mutants exhibited a number of replication defects, none of these defects could be directly attributed to aberrant CA phosphorylation. A novel defect was a block in early budding, which was common among several virus mutants with substitutions in the CA N terminus. Together, these results indicate that certain residues in the CA N terminus are crucial for early budding events.

Electron tomography analysis of envelope glycoprotein trimers on HIV and simian immunodeficiency virus virions.

We used electron tomography to directly visualize trilobed presumptive envelope (env) glycoprotein structures on the surface of negatively stained HIV type 1 (HIV-1) and simian immunodeficiency virus (SIV) virions. Wild-type HIV-1 and SIV virions had an average of 8-10 trimers per virion, consistent with predictions based on biochemical evidence. Mutant SIVs, biochemically demonstrated to contain high levels of the viral env proteins, averaged 70-79 trimers per virion in tomograms. These correlations strongly indicate that the visualized trimers represent env spikes. The env trimers were without obvious geometric distribution pattern or preferred rotational orientation. Combined with biochemical analysis of gag/env ratios in virions, these trimer counts allow calculation of the number of gag molecules per virion, yielding an average value of approximately 1400. Virion and env dimensions were also determined. Image-averaging analysis of SIV env trimers revealed a distinct chirality and strong concordance with recent molecular models. The results directly demonstrate the presence of env trimers on the surface of AIDS virus virions, albeit at numbers much lower than generally appreciated, and have important implications for understanding virion formation, virus interactions with host cells, and virus neutralization.

[Electron microscopic investigation of CD4+ lymphocyte cell line C8166 after infection with simian immunodeficiency virus (SIV)]. / Elektronenmikroskopische Untersuchung an der CD4+ Lymphozytenzellkultur C8166 nach Infektion mit dem Simian-Immunodefiency-Virus (SIV).

The SIV infection of rhesus macaques (Macaca mulatta) is the most appropriate animal model in HIV research. The permanent human T-cell line C8166 is used for in vitro SIV propagation. This paper describes ultrastructural features of the cells after infection with SIVmac. The C8166 cells are ultrastructurally characterized by a heterogenous morphology which is independent of the infection. SIV induced cell syncytia are observed 18 hours after infection. Viral particles and budding occur 48 hours p.i with a peak at the day 8. Viral particles present the typical lentiviral morphology. Using the monoclonal antibody anti SIVp28 and ultra small (0.8 nm) immunogold-silver enhancement technique, we are able to demonstrate SIV antigen immunoelectron microscopically. Therefore, this ultrastructural method is suitable to detect SIV antigen in in vivo experiments with C8166 cells from day 8 p.i. serving as positive control.

Comparative efficacy of recombinant modified vaccinia virus Ankara expressing simian immunodeficiency virus (SIV) Gag-Pol and/or Env in macaques challenged with pathogenic SIV.

Prior studies demonstrated that immunization of macaques with simian immunodeficiency virus (SIV) Gag-Pol and Env recombinants of the attenuated poxvirus modified vaccinia virus Ankara (MVA) provided protection from high levels of viremia and AIDS following challenge with a pathogenic strain of SIV (V. M. Hirsch et al., J. Virol. 70:3741-3752, 1996). This MVA-SIV recombinant expressed relatively low levels of the Gag-Pol portion of the vaccine. To optimize protection, second-generation recombinant MVAs that expressed high levels of either Gag-Pol (MVA-gag-pol) or Env (MVA-env), alone or in combination (MVA-gag-pol-env), were generated. A cohort of 24 macaques was immunized with recombinant or nonrecombinant MVA (four groups of six animals) and was challenged with 50 times the dose at which 50% of macaques are infected with uncloned pathogenic SIVsmE660. Although all animals became infected postchallenge, plasma viremia was significantly reduced in animals that received the MVA-SIV recombinant vaccines as compared with animals that received nonrecombinant MVA (P = 0.0011 by repeated-measures analysis of variance). The differences in the degree of virus suppression achieved by the three MVA-SIV vaccines were not significant. Most importantly, the reduction in levels of viremia resulted in a significant increase in median (P < 0.05 by Student's t test) and cumulative (P = 0.010 by log rank test) survival. These results suggest that recombinant MVA has considerable potential as a vaccine vector for human AIDS.

The simian immunodeficiency virus envelope glycoprotein contains multiple signals that regulate its cell surface expression and endocytosis.

The cell surface expression of the envelope glycoproteins (Envs) of primate immunodeficiency viruses is, at least in part, regulated by endocytosis signal(s) located in the Env cytoplasmic domain. Here, we show that a membrane proximal signal that directs the simian immunodeficiency virus (SIV) Env to clathrin-coated pits, and is conserved in all SIV and human immunodeficiency virus Envs, conforms to a YxxØ motif (where x can be any amino acid and Ø represents a large hydrophobic residue). This motif is similar to that described for a number of cellular membrane proteins. By surface plasmon resonance we detected a high affinity interaction between peptides containing this membrane proximal signal and both AP1 and AP2 clathrin adaptor complexes. Mutation of the tyrosine in this membrane proximal motif in a SIV Env with a prematurely truncated cytoplasmic domain leads to a > or = 25-fold increase in Env expression on infected cells. By contrast, the same mutation in an Env with a full-length cytoplasmic domain increases cell surface expression only 4-fold. We show that this effect results from the presence of additional endocytosis signals in the full-length cytoplasmic domain. Chimeras containing CD4 ecto- and membrane spanning domains and a full-length SIV Env cytoplasmic domain showed rapid endocytosis even when the membrane proximal tyrosine-based signal was disrupted. Mapping experiments indicated that at least some of the additional endocytosis information is located between residues 743 and 812 of Env from the SIVmac239 molecular clone. Together, our findings indicate that the cytoplasmic domain of SIV Env contains multiple endocytosis and/or trafficking signals that modulate its surface expression on infected cells, and suggest an important role for this function in pathogenesis.

Nucleocapsid protein zinc-finger mutants of simian immunodeficiency virus strain mne produce virions that are replication defective in vitro and in vivo.

All retroviruses (except the spumaretroviruses) contain a nucleocapsid (NC) protein that encodes one or two copies of the Zn2+-finger sequence -Cys-X2-Cys-X4-His-X4-Cys-. This region has been shown to be essential for recognition and packaging of the genomic RNA during virion particle assembly. Additionally, this region has been shown to be involved in early infection events in a wide spectrum of retroviruses, including mammalian type C [e.g., murine leukemia virus (MuLV)], human immunodeficiency virus type 1 (HIV-1), Rous sarcoma virus, and other retroviruses. Mutations in the two Zn2+-fingers of the NC protein of simian immunodeficiency virus strain Mne [SIV(Mne)] have been generated. The resulting virions contained the normal complement of processed viral proteins with densities indistinguishable from wild-type SIV(Mne). All of the mutants had electron micrograph morphologies similar to those of immature particles observed in wild-type preparations. RNA packaging was less affected by mutations in the NC protein of SIV(Mne) than has been observed for similar mutants in the MuLV and HIV-1 systems. Nevertheless, in vitro replication of SIV(Mne) NC mutants was impaired to levels comparable to those observed for MuLV and HIV-1 NC mutants; replication defective NC mutants are typically 10(5)- to 10(6)-fold less infectious than similar levels of wild-type virus. One mutant, DeltaCys33-Cys36, was also found to be noninfectious in vivo when mutant virus was administered intravenously to a pig-tailed macaque. NC mutations can therefore be used to generate replication defective virions for candidate vaccines in the SIV macaque model for primate lentiviral diseases.

Substitution of leucine 8 in the simian immunodeficiency virus matrix protein impairs particle formation without affecting N-myristylation of the Gag precursor.

In the late stages of replication of the simian immunodeficiency viruses (SIV), the matrix protein (MA) plays a central role in the transport of Pr55gag to the plasma membrane, assembly of virus particles, and incorporation of the envelope glycoprotein into particles. Targeting of Pr55gag to the plasma membrane is mediated by two motifs within the MA protein: the N-terminal myristate and a cluster of positively charged amino acids. In this report, we characterized the assembly phenotype of an SIV Gag mutant (L8Q) carrying the single amino acid substitution of glutamine for leucine at position 8 in the MA domain. The hydropathic profile of the mutated MA protein indicated that the L8Q amino acid change disrupts the hydrophobic character of the region comprising the first 10 residues of the protein. Expression of mutant L8Q Gag protein in CV-1 cells, by means of the vaccinia virus vector system, resulted in efficient synthesis and N-myristylation of Pr55gag. However, this mutation severely impaired particle production, as inferred from both biochemical and electron microscopy analyses. Cellular fractionation assays revealed that in cells expressing mutant L8Q, the proportion of cytosol-associated Pr55gag was significantly increased compared to that observed upon expression of wild-type Gag. Furthermore, mutant L8Q Gag partitioned onto cytosol and membrane fractions in a manner similar to nonmyristylated Gag polyprotein. Taken together, these results indicate that the L8Q mutation reduces the membrane-binding capacity of the Gag precursor. It is therefore likely that in the SIV MA, in addition to the N-myristate group, the hydrophobicity of the neighboring region is important for efficient association of Pr55gag with the plasma membrane.

Assembly and release of SIV env proteins with full-length or truncated cytoplasmic domains.

We have used recombinant vaccinia viruses expressing full-length or truncated gag or env genes of SIVmac239 to investigate the requirements for assembly of SIV proteins. We observed that assembly of virus-like particles (VLPs) was found to be 3- to 5-fold higher with full-length Env than with the truncated forms, or than VLPs containing only Gag proteins, in primary monkey cells or various human cell lines. When cells expressing Env proteins in the absence of Gag were examined by immunoelectron microscopy, clusters of Env protein and membrane vesicles containing Env proteins were observed at cell surfaces. A low level of vesicles was released from cells expressing full-length Env, but about a 10-fold higher level was released in cells expressing a truncated form of Env [Env733(t)] in which the cytoplasmic domain is only 17 amino acids in length. Another truncated protein, Env718(t), with a short cytoplasmic tail of 3 aa, was also incorporated into VLPs at a 10-fold higher level than the full-length Env protein and was more efficiently released in vesicles. The mature SU and TM proteins were predominantly incorporated into VLPs with full-length Env, but both cleaved and uncleaved precursor proteins were present in VLPs with truncated Env as well as in Env and Env(t) vesicles. A more prominent layer of spikes was seen by electron microscopy in VLPs with truncated Env than in VLPs containing full-length Env. These results indicate that truncated Env proteins have the ability to self-associate on the cell surface and are assembled into a more closely packed array than full-length Env, which could explain the preferential incorporation of Env proteins with short cytoplasmic tails into virions.

SIV infection of monkey spleen cells including follicular dendritic cells in different stages of disease.

Immunoaffinity enriched spleen follicular dendritic cells (FDCs), lymphocytes, and macrophages from SIVsm-inoculated cynomolgus monkeys (Macaca fascicularis) at different stages of disease were compared for latent and productive SIV infection. Analysis of FDCs by in situ hybridization, electron microscopy, and coculture assays indicated that comparatively high levels of virus were associated with the FDC fraction. Polymerase chain reaction (PCR) and RT-PCR results revealed that the levels for SIVpol DNA did not correlate with the level of env mRNA in the various cell subsets, suggesting differences in latency. Limiting dilution assays for spliced env mRNA showed a 10-100-fold higher amount of env mRNA in FDCs than in other spleen cell subsets early during SIV infection. At late stages of disease, the number of productively infected FDCs significantly decreased in parallel with a marked reduction of the FDC network and follicular involution. Our findings indicate that destruction of FDCs probably reflects a cytopathic effect of SIV and/or the activity of specific antiviral cytotoxic T lymphocytes.

Assembly of SIV virus-like particles containing envelope proteins using a baculovirus expression system.

The requirements for SIV particle assembly and envelope incorporation were investigated using a baculovirus expression system. The Pr56gag precursor protein expressed under control of the polyhedrin promoter (pPolh) produced high levels of immature retrovirus-like particles (VLP) upon expression in Sf9 insect cells. To determine the optimal conditions for envelope protein (Env) incorporation into VLP, two recombinant baculoviruses expressing the SIV envelope protein under control of a very late pPolh or a hybrid late/very late capsid/polyhedrin (Pcap/polh) promoter and a recombinant expressing a truncated form of the SIV envelope protein (Envt) under the hybrid Pcap/polh promoter were compared. We have observed that utilization of the earlier hybrid promoter resulted in higher levels of Env expression on the cell surface and its incorporation into budding virus particles. We have also found that the Envt protein is transported to the cell surface of insect cells and incorporated into VLP more efficiently than full-length Env. In addition, we examined the effect of coexpression of the protease furin, which has been implicated in the proteolytic cleavage of the Env precursor gp160 in mammalian cells. Coexpression of furin in insect cells resulted in more efficient proteolytic cleavage into gp120 and gp41, and the cleaved proteins were incorporated into VLP.

Morphological changes in lymph nodes and expression of VCAM1 and cytokines at the late stages of SIV-induced disease in rhesus monkeys.

Four patterns of structural alterations were found in lymph nodes (LNs) from rhesus monkeys 17 to 34 months after infection with simian immunodeficiency virus (SIV-mac251). SIV p27gag antigen and viral particles were localized either between the processes of follicular dendritic cells (FDCs) or in the cytoplasm of macrophages. In hyperplastic follicles, enlarged germinal centres contained numerous Ki67+ proliferating centroblasts which were rather rare in light zones occupied by the CD23+ FDC network. Involuted follicles contained a small number of Ki67+ centroblasts and the CD23 labelling was limited to a very small apical zone. A correlation was found between the morphological characteristics of the follicles (hyperplasia-involution) and the level of expression of the vascular cell adhesion molecule 1 (VCAM1) on FDCs. A gradient in VCAM1 intensity with no expression in the subcapsular-intermediary sinuses, low membrane labelling in the mantle and strong expression in the FDC network was observed. IL1 alpha+ and IL6+ (interleukin) cells (lymphocytes and macrophages) were detected in the mantle, the interfollicular area and the medulla of LNs. Expression of the tumour necrosis factor alpha and ultrastructural markers of interferon alpha production were found in a few FDC and macrophages. Our findings indicate a close relationship between the morphofunctional properties of FDC and the LN structure in SIV infection.