SEARCH

SEARCH BY CITATION

Keywords:

  • correlative microscopy;
  • fluorescence;
  • human immunodeficiency virus;
  • imaging;
  • subdiffraction microscopy;
  • super-resolution microscopy

Abstract

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

The replication of HIV-1, like that of all viruses, is intimately connected with cellular structures and pathways. For many years, bulk biochemical and cell biological methods were the main approaches employed to investigate interactions between HIV-1 and its host cell. However, during the past decade advancements in fluorescence imaging technologies opened new possibilities for the direct visualization of individual steps occurring throughout the viral replication cycle. Electron microscopy (EM) methods, which have traditionally been employed for the study of viruses, are complemented by fluorescence microscopy (FM) techniques that allow us to follow the dynamics of virus–cell interaction. Subdiffraction fluorescence microscopy, as well as correlative EM/FM approaches, are narrowing the fundamental gap between the high structural resolution provided by EM and the high temporal resolution and throughput accomplished by FM. The application of modern microscopy to the study of HIV-1–host cell interactions has provided insights into the biology of the virus which could not easily, or not at all, have been gained by other methods. Here, we review how modern fluorescence imaging techniques enhanced our knowledge of the dynamic and structural changes involved in HIV-1 particle formation.

image

Today's virologists can draw on a constantly increasing abundance of sophisticated molecular, physical and bioinformatic methods to analyze their tiny objects of interest. Nevertheless, the concept of ‘seeing is believing’, i.e. the direct visual observation of biological structures and processes, does play a prominent role in cell biological and virological research. Viruses present some specific challenges for the application of imaging techniques. Viral particles are smaller than the diffraction limit of light microscopy, preventing the resolution of individual virions or subviral structures by conventional FM. Viral replication occurs within the context of a complex and dynamic intracellular environment and individual replication steps can occur on a timescale of minutes. Finally, viral replication cycles proceed through an eclipse, during which only the viral genome is present in the host cell. Structural transitions occurring upon particle disassembly or during assembly of virus progeny may render the virus morphologically unrecognizable throughout most of its replication cycle, when compared with the architecture of cell-free virions.

Consequently, these obstacles impede the analysis of the retrovirus HIV-1 by imaging techniques. With a diameter of ∼140 nm and a complex and heterogeneous architecture [1, 2], the enveloped virion itself (Figure 1A) cannot be resolved by light microscopy and is not amenable to crystallographic structure analyses. HIV-1 is a comparatively simple virus encoding only 15 mature proteins on its ∼9.8 kb ssRNA genome [3], but its replication involves numerous complex and dynamic interactions with host cell factors. Upon cell entry, the virion disintegrates in a stepwise, ill-characterized manner, leaving only its genomic information integrated into the host cell DNA to direct the synthesis of virus progeny. New HIV-1 virions assemble from individual components at the plasma membrane and gain their mature structure only during release from the virus producing cell. Thin-section EM has classically been employed to visualize HIV-1 [4] and provided numerous insights into HIV-1 biology during the past 30 years. However, the information gained from these images could often only be used in a supportive manner. More recently, cryo-EM and cryo-electron tomography (cryo-ET) yielded detailed information on HIV-1 particle architecture [1, 2]. However, EM methods provide static images that do not convey information on dynamic alterations. Driven by technology development in the field of microscopic techniques as well as by new fluorescent labeling methods, the possibilities for applying FM to the investigation of viruses have substantially increased over the last decade. Novel techniques have overcome some of the limitations inherent to the classical approaches and enhanced the usefulness and relevance of microscopic analyses for the study of HIV-1 cell interactions.

image

Figure 1. Gag is the main structural component of HIV-1. A) Schematic structure of the mature HIV-1 virion. Major virion components are indicated. B) Scheme of the immature Gag protein. Arrowheads indicate cleavage sites for the viral protease between individually folded subdomains. The right hand scheme shows a proposed structural model of Gag polyprotein generated by superimposing structures of the individual folded domains (from reference [1], with modifications). Positions where FPs, a TC-tag or a SNAP-tag (SNAP), respectively, have been inserted for live-cell imaging studies are indicated. See main text for references.

Download figure to PowerPoint

The HIV-1 replication cycle – is characterized by the dynamic formation, rearrangement, intracellular transport and dissociation of macromolecular structures. This is particularly the case for early replication steps from cell attachment to genome delivery into the nucleus and for the late phase of virion assembly, release and maturation (Figure 2). A full understanding of these events thus requires the analysis of both the structural rearrangements involved and the dynamics of these processes. No single imaging technique available today comprises the spatial and the temporal resolution needed to capture both of these aspects; different techniques display specific strengths and limitations (Figure 3). Results obtained using various approaches need to be gathered in a complementary manner to yield a complete picture of processes involved in HIV-1 replication. Here, we provide an overview over modern fluorescence imaging techniques that yield visual representations of virus–cell interactions and outline how they have contributed to our current understanding of HIV-1 particle assembly and release. For reviews on the application of other fluorescence-based methods, e.g. fluorescence resonance energy transfer (FRET), fluorescence recovery after photobleaching, fluorescence correlation spectroscopy or image correlation microscopy methods to the study of HIV-1, we refer the reader to other recent reviews [5, 6].

image

Figure 2. Model of HIV-1 morphogenesis. The HIV-1 Gag polyprotein is expressed in the cytoplasm of an HIV-1 infected cell (1) and trafficks to the plasma membrane by a not yet fully understood pathway. The N-terminal myristoyl moiety of Gag, basic residues in the MA domain and Gag oligomerization, mediated by CA-CA interactions as well as by NC-gRNA binding, play important roles in this process. Particle assembly is initiated by Gag dependent tethering of two molecules of gRNA to the plasma membrane (2). This complex nucleates the assembly of the immature Gag lattice, accompanied by recruitment of other virion components (3). Membrane fission, mediated by a part of the cellular ESCRT machinery, results in release of an immature, non-infectious virion (4). Concomitant with virus budding, the viral PR cleaves the Gag and GagPol polyproteins into their mature subunits. This proteolytic maturation leads to a dramatic rearrangement of structural proteins within the particle, resulting in formation of the conical capsid encasing the viral nucleoprotein complex (5). Infection of a new target cell is initiated by binding to the receptor CD4 and a co-receptor molecule (CCR5 or CXCR4) (6), which triggers conformational rearrangements in the viral Env proteins mediating fusion of the viral envelope with the host cell membrane (7) and release the viral core into the cytoplasm (8). The fusion process may occur at the plasma membrane or from within intracellular vesicles following endosomal uptake; the relative functional importance of both pathways in different cell types is still under investigation. The viral core structure disassembles (‘uncoating’), giving rise to a reverse transcription complex in which the ssRNA genome is converted into a dsDNA form by the viral RT and a pre-integration complex which targets the cDNA to the nucleus and mediates its integration into the host cell genome. The processes of uncoating, reverse transcription and formation of the pre-integration complex depend crucially on particle maturation, are functionally linked and need to be tightly controlled, but the sequence of events is currently not well understood. The integrated viral cDNA then directs the expression of virion components which serve as building blocks for new virus progeny.

Download figure to PowerPoint

image

Figure 3. Modern fluorescence imaging techniques used for the study of HIV-1 replication. The figure summarizes ranges of temporal and spatial resolution achieved in studies on HIV-1, as well as some current key advantages and limitations of the methods with respect to virological applications. Examples of images generated using the respective techniques are shown (images are modified from the indicated references).

Download figure to PowerPoint

Fluorescent HIV-1 Derivatives

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

Fluorescence imaging in HIV research comprises the application of different modern FM based approaches, as well as combinatorial approaches applying both FM and EM methods to the same sample. As a prerequisite, the structure of interest needs to be tagged with a fluorescent label. Selecting the appropriate strategy from the ever increasing plethora of labels for cell biological applications not only depends on the virological question to be addressed, but also on the microscopic technique chosen. For many applications the well characterized derivatives of natural auto fluorescent proteins (FPs), i.e. the enhanced green fluorescent protein (eGFP) and many related proteins [7], represent the best choice. The use of FPs for imaging of HIV-1 was initiated by the expression of GFP tagged versions of the HIV-1 structural polyprotein Gag (Figure 1B) by itself [8, 9]. In a pioneering study, McDonald et al. [10] used HIV-1 particles carrying an eGFP-tagged version of the Viral protein R (Vpr) to follow the intracellular transport of entering subviral complexes along the microtubule network. Since then, FP-tagged derivatives of various HIV proteins have been employed by many labs to capture the virus in various stages of replication [11-27]. FPs are perfectly suited for live-cell imaging. They do not require cumbersome staining steps and the analysis of a genetically altered virus with respect to its functionality is straightforward. However, their large size of ∼27 kDa, the intrinsic oligomerization propensity of some FPs, the relatively slow maturation of the fluorophore or its sensitivity towards acidic pH may present problems. Furthermore, some microscopic approaches demand the use of brighter or more photostable labels, or require controllable fluorescence emission (see below), which may necessitate using synthetic dyes. While this can be accomplished by immunofluorescence or direct chemical labeling of the virus, these strategies can affect virus functionality in a difficult to control manner and limit live-cell analyses. An alternative is presented by non-fluorescent genetically encoded tags which can be specifically labeled with a synthetic fluorophore. Representatives from this diverse group of genetically engineered peptides or proteins already successfully employed in the context of HIV-1 are the tetracysteine tag [TC-tag; [28]] and the SNAP-tag [29]. The TC-tag is only 6–12 amino acids long, minimizing the risk for functional impairment of the target protein. Insertion of TC-tag coding sequences is indeed tolerated at several positions within the HIV-1 genome without abolishing virus functionality, in certain cases even without impairing virus infectivity in tissue culture [30-33]. However, availability of a limited set of cognate dyes, the photochemical properties of these dyes and a propensity towards unspecific staining represent limitations. The SNAP-tag [29] is a self-labeling enzyme derived from alkyl-guanyl-transferase with a mass of ∼20 kDa. Its insertion into HIV Gag affects viral replication capacity to a lesser degree than FPs in the corresponding position [34] and allows very specific labeling of the fusion protein with a variety of synthetic dyes.

Viral genomic RNA (gRNA) cannot be directly fused to protein labels. However, tagging of gRNA surrogates with small RNA motifs representing specific recognition sites for bacteriophage or Escherichia coli RNA binding proteins (e.g. the coat protein of phage MS2) allows intracellular labeling of the RNA in living cells by association with FP-tagged versions of the cognate RNA binding proteins [35]. Using this and other RNA labeling systems, retroviral RNA trafficking can be analyzed by live-cell imaging [36].

Microscopic Methods

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

The study of virus–cell interactions has greatly profited from the introduction of single virus tracing (SVT) approaches [37]. The observation of individual virions or subviral complexes in live cells with second or sub-second time resolution allows for the detailed analysis of rapid events occurring in a non-synchronized manner. This was crucial, for example, to investigate the dynamics of virion assembly at the plasma membrane [16, 17]. SVT can show the order of events occurring at a particular region of interest, e.g. the timing of recruitment of cellular factors to a viral assembly site [12, 19]. It can also dissect different events occurring in parallel, e.g. the cell entry of viruses into the same host cell via different pathways [38]. Furthermore, it yields information on transient interactions between viral and other viral or cellular components, on the intracellular localization of interactions or on trafficking pathways. Depending on the question to be addressed, wide-field (WF) or confocal FM may be used. The investigation of HIV-1 assembly has benefited from the use of total internal reflection fluorescence microscopy (TIR-FM), which effectively restricts excitation of fluorophores to a distance of ∼150 nm from the cover slip, thereby allowing sensitive observation of events at the ventral plasma membrane.

While the high temporal resolution of SVT is ideal to capture rapid events in HIV replication, its spatial resolution is restricted to the diffraction limit of light microscopy. This does not allow visualization of subviral features and completely excludes obtaining information on molecular architecture. Detailed information on HIV architecture has been provided by EM techniques, in particular by electron tomography (ET) and cryo-ET [1, 2]. While traditional thin section EM is restricted to two-dimensional sections, ET captures the information on the 3D structure. For cryo-ET, unfixed and unstained samples are imaged embedded in a layer of vitreous ice, yielding a view of the specimen in a near-natural state. A limitation of cryo-ET is that it is restricted to thin samples (<0.5 µm), so that only isolated virions or intracellular structures near the edges of flat cellular protrusions can be analyzed. Ion-abrasion scanning EM (IA-SEM) can be employed to obtain 3D images of the cell interior [39]. For this, the surface of the sample is abraded in a stepwise manner using a focused ion beam and the exposed surfaces are visualized by SEM. These serial sections then allow reconstruction of a 3D image.

With their potential to capture the dynamics of rapid individual events on one hand and to provide molecular resolution of 3D structures on the other, SVT and EM techniques excellently complement each other. However, the fundamental gap separating these technologies with respect to temporal and spatial resolution makes it difficult to connect dynamic and structural aspects of a particular event. Recent advances in both FM and EM technologies have the potential to bridge this gap, offering exciting new glimpses into HIV-1 biology.

One step towards this is the increase in the spatial resolution of FM. Super-resolution FM techniques developed in recent years can overcome the diffraction limit of light microscopy. Their promise lies in combining increased spatial resolution with the advantages of FM: rapid analysis of large numbers of individual events, identification of proteins of interest through fluorescent labeling and – in principle – the option to perform live-cell analyses. Currently available techniques can be divided into software-based and hardware-based methods, depending on the principle applied for image generation. Software-based methods include variants of photoactivated localization microscopy (PALM) [40, 41] and of stochastic optical reconstruction microscopy (STORM) [42, 43]. They rely on serial acquisition of diffraction limited images by WF or TIRF microscopy, while ensuring activation of only a small subset of well separated fluorophores in each acquisition cycle. The xy-position of individual fluorophores is determined with high precision by Gaussian fitting of the diffraction limited signals and calculated localizations from all individual frames are computationally assembled into a final image. This requires controlled conversion of fluorophores between an ‘on’ and an ‘off’ state, which cannot be achieved using eGFP. Thus, depending on the method applied, labeling has to be performed either using specific pairs of chemical dyes [43], suitable individual dyes [42] or photoactivatable or photoswitchable FPs like PA-GFP or mEOS [40, 44]. The most important hardware-based technique is stimulated emission depletion (STED) microscopy [45]. On the basis of confocal FM, it employs two lasers that concurrently scan the sample: one pulse that excites the fluorophores and a second, red-shifted beam with a doughnut-shaped focus. This second beam de-excites fluorophores, except for those in a zero-intensity point in the center of the doughnut, thereby decreasing the size of the excited area. Currently, this technique imposes restrictions on the labeling strategy; most applications require synthetic fluorophores with high quantum yield and high photostability. The requirement for high laser power (STED) or comparatively long image acquisition times (PALM and STORM) currently also limit the use of subdiffraction microscopy for time-resolved live-cell approaches.

While subdiffraction microscopy can resolve individual virions and subviral entities, its current lateral resolution in the range of 20–40 nm does not permit visualization of detailed molecular architecture. A direct link between dynamic information obtained by FM with structural information down to nanometer resolution is offered by correlative microscopy, in which the same sample is consecutively analyzed by FM and (cryo)-ET or IA-SEM, respectively. The use of this approach for the study of HIV–cell interactions still presents technical challenges, but if successfully applied, it could address two fundamental problems of EM analysis of virus–cell interactions. First, localization of a particular event by FM guides ET analyses to a specific region of interest. This makes it possible focus on comparatively small and rare structures, e.g. an individual virus particle against the complex background of the cell [20, 46]. Second, it has the potential to combine information on the dynamics of an event with the molecular structures involved and should thus allow capturing distinct transient stages of a given process.

While these technologies undergo constant development, they have already contributed to our understanding of HIV–cell interaction. The following sections focus exemplary on late stages in the HIV replication cycle to illustrate how current imaging techniques can fill gaps in our knowledge on HIV replication.

HIV-1 Particle Assembly

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

HIV-1 particle formation occurs at the plasma membrane of a virus producing cell. The viral structural proteins, enzymes and genome are targeted to a budding site, where they assemble into a spherical virus bud. This complex process is directed by the 55 kDa Gag polyprotein (Figure 1B). Its matrix (MA), capsid (CA) and nucleocapsid (NC) domains contribute the driving forces of particle formation, namely membrane targeting, protein–protein-interaction and RNA binding functions, respectively [47]. Gag can assemble into virus like particles in the absence of other viral components. Specific interactions with Gag are essential for targeting the viral envelope glycoproteins (Env), the viral RNA genome (gRNA) and several other viral and cellular proteins to the budding site. The HIV-1 enzymes protease (PR), reverse transcriptase (RT) and integrase (IN) are incorporated in balanced amounts due to expression as parts of a Gag–Pol fusion polyprotein. With ∼2400 molecules per particle [48], Gag also accounts for about 50% of total virion mass and is the main structural determinant of the immature virus shell. Nascent HIV-1 buds are morphologically distinct structures found in large numbers at the membrane of virus producing cells, making them amenable for EM based analyses. Thus, thin section EM has been employed from the beginnings of HIV-1 research to characterize the overall architecture of HIV buds and particles [4]. More recently, Cryo-ET yielded detailed models of the 3D architecture of Gag at HIV-1 budding sites and in the immature virion. These studies showed that Gag molecules are arranged in parallel underneath the membrane in a hexameric lattice, whose curvature is attained by irregular defects [1, 2]. A surprisingly large gap in this lattice, which had not been appreciated based on 2D EM sections, covers ∼1/3 of the viral envelope [48, 49]. This flexible lattice arrangement, as the flexible nature of the polyprotein itself, entails tolerance towards the addition of genetically encoded tags. Successful fluorescent labeling strategies initially focused on the C-terminus of Gag expressed alone [8, 9]. In the complete viral context, a flexible region separating the MA and CA domains of the polyprotein can accomodate even bulky insertions [26] (Figure 1B). Using the latter approach, various tagged HIV-1 derivatives retaining partial or even full replication capacity in specific cell types were obtained [14, 26, 32, 34]. The small TC-tag was also fused to the NC [33] or CA [31, 33] domains of Gag in the viral context. Most of these strategies affect the replication capacity of the virus to some degree, but complementation with wild-type Gag can be used to produce labeled particles with wild-type morphology and entry competence. Fluorescently labeled HIV-1 Gag forms punctuate assemblies at the plasma membrane displaying a size of 100–140 nm [27, 34], in agreement with the virion diameter determined by cryo-EM [1, 2] and fluorescence intensities similar to that of individual labeled virions [16]. Correlative FM/scanning EM analysis demonstrated that Gag.eGFP punctae correspond to spherical virion structures bulging from the membrane [23]. Equimolar complementation with unlabeled molecules still yields incorporation of ∼1200 fluorophores per particle. This labeling density allows for the visualization of individual budding sites and virions with sub-second time resolution and investigation of HIV-1 assembly kinetics [16, 17]. SVT analyses employing a combination of FP-tagged surrogates of the viral gRNA with differently labeled Gag [18] indicated that bud formation is initiated through Gag dependent docking of two molecules of gRNA bound to a small number of Gag molecules to the plasma membrane. The complex then serves as a nucleation site for the gradual accumulation of Gag [18], which appears to be recruited mainly from the cytoplasm rather than by lateral movement from the surrounding membrane [16]. Assembly is completed within ∼7–10 min from the initial detection of Gag [16, 17]. The subsequent membrane abscission cannot be unambiguously assigned using TIR-FM. The resistance of Gag tagged with a pH-sensitive FP towards experimental acidification of the cytoplasm indicated that most budding particles had separated from the cell within a period of 30 min [17]. However, live observation of release by TIR-FM is hindered by the fact that virions budding at the ventral membrane can become trapped in the narrow space between the cell and the coverslip. For a subset of particles, budding into pockets underneath the membrane allowed detection of a sudden motility change indicative of release. These events were observed with significant delay after the exponential assembly phase, suggesting a lag phase of ∼15 min between completion of the bud shell and virus release [16]. In the absence of direct live-microscopy readouts for membrane fission, more detailed investigation of release kinetics remains challenging.

With dynamics of Gag assembly and the 3D structure of Gag assemblies established, imaging can address many additional aspects of HIV-1 particle formation. This entails the temporal recruitment of cellular factors or virion components to the assembly site as well as the spatial arrangement of these factors in relation to the Gag shell. With its potential to resolve subviral details at large numbers of individual budding sites and to identify selected components based on fluorescence, subdiffraction microscopy complements the analyses summarized above. Early reports describing PALM used HIV-1 Gag as a model to demonstrate the potential of the method to analyze clustering and motility of membrane-bound proteins [40, 50]. More recent analyses using direct STORM (dSTORM, [42]) showed distinct Gag clusters with diameters in the range of ∼100–150 nm, as expected for individual budding sites [27, 34]. While these findings mainly confirmed what had already been established with higher resolution by EM, they validate the approach and provide a basis for a more detailed description of Gag distribution at the cell membrane. They are also a starting point to investigate co-localization of Gag with other molecules using dual-color super-resolution microscopy. This is of particular interest with respect to plasma membrane lipids, which are known to play a role in HIV-1 morphogenesis. Assembly and budding occurs from specific raft-like lipid microdomains, but the exact nature of these domains and the pathway of their generation is incompletely understood [51, 52]. While the size of lipid microdomains is below the resolution of conventional FM, super-resolution microscopy would allow probing for the recruitment and presence of specific lipids to the budding site without using experimental antibody clustering of lipid domain markers for visualization. An initial attempt towards this has been recently made by characterization of cholera toxin stained GM1 patches using dSTORM, reporting lack of significant GM1 co-localization with Gag assembly sites [27]. A current limitation lies in the difficulty to tag lipids with labels suitable for subdiffraction microscopy while preserving their natural localization. Another interesting aspect concerns recruitment of the viral Env protein. The myristoylated Gag protein and the transmembrane containing Env traffic via separate routes to the plasma membrane, where interactions between the Env gp41 cytoplasmic tail and the MA domain of Gag mediate Env particle incorporation. The exact mechanism of Env recruitment to the Gag assembly site and the relevance of local membrane lipid composition for the process are currently not understood. A surprisingly low number of ∼7–14 Env trimers per virion has been detected in biochemical and cryo-ET analyses [53, 54] and cryo-ET revealed an uneven distribution of Env spikes on the viral envelope [54]. Conceivably, Env clustering on the virus surface might enhance efficiency of receptor contact and membrane fusion and thus may play a role in HIV-1cell entry. Cryo-ET images indicated a proteinaceous structure (‘entry claw’) at HIV-1-cell contact sites [55], suggestive of clustered Env-receptor interactions, but the molecular nature of these complexes was not revealed and dynamic information is lacking. Subdiffraction microscopy may provide complementary information, since it allows for the identification of proteins contributing to subviral structures and for analysis of a statistically relevant number of particles. Exemplary dSTORM visualization of clusters of Env surrounding Gag assembly sites and on individual virions [27], demonstrated the feasibility of the approach for further analyses of Env recruitment and distribution.

For the investigation of the still partially enigmatic steps following HIV-1 cell entry by imaging, it is crucial to discriminate between complete virions – bound to the cell surface or internalized through the endosomal pathway – and subviral structures resulting from membrane fusion. This can be accomplished by incorporating two different labels into the virus particle: one in the viral core, the second in the lumen outside the core, attached to the matrix or at the lipid envelope, respectively [13, 21, 22, 24]. Selective loss of the outer label may then be taken as indication of lipid mixing or cytoplasmic entry of the core. This strategy has been exploited to visualize individual retroviral membrane hemifusion or fusion events by SVT [e.g. [21, 24, 25]]. Using this approach individual, dynamin-dependent fusion events were observed to occur from endosomal structures within the cell [25], contributing to the proposition that endocytosis rather than fusion at the plasma membrane represents the pathway of productive HIV-1 entry [25]. While the double-labeling strategy appears straightforward, it relies on accurate determination of intensity changes in rapidly moving particles and careful controls for differential bleaching and quenching; interpretation may also be confounded by uncertainties about the distribution of the labeled components before and after fusion. Super-resolution microscopy might contribute by resolving morphologies of viral core structures characteristic for particular replication intermediates. PALM and dSTORM visualization of IN within HIV-1 particles revealed compact foci with a diameter smaller than that of the virion [27, 56], in agreement with the confinement of IN within the mature capsid [57, 58]. Computational modeling and analysis of IN distribution suggested distinctive morphologies of IN clusters in mature and immature virions [56], indicating that super-resolution microscopy may be used for morphological discrimination of viral replication stages.

HIV-1 Release

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

HIV-1 particle formation is completed by abscission of the viral lipid envelope from the plasma membrane. This step is mediated by a part of the cellular endosomal complex required for transport (ESCRT) machinery [59]. ESCRT is involved in various cellular membrane fission events, most importantly cytokinesis and the formation of multivesicular bodies, and is deployed by HIV-1 and a number of other enveloped viruses to propagate virus particle release. It consists of four protein complexes (ESCRT-0, -I, -II and -III) and associated proteins. HIV-1 Gag can recruit ESCRT through interaction with the ESCRT-I component TSG101 or the ESCRT associated protein ALIX [59]. This in turn leads to recruitment of ESCRT-III, whose assembly at the membrane is assumed to be the main driving force for bud neck constriction and membrane fission. The potential contributions of the assembling Gag lattice and of specific lipids to membrane deformation and fission have not yet been clearly defined. Different models for the mechanism of membrane constriction though ordered assembly/disassembly of ESCRT-III multimers have been proposed [47]. Biochemical and virological studies established the involvement of the ESCRT machinery in HIV-1 release and defined ESCRT-I, ALIX, the ESCRT-III proteins Chmp2 and Chmp4 and the ATPase Vps4 as essential components [59]. Complementing these data SVT yielded insights into the order of events at the retroviral budding site and allowed conclusions on the role of individual ESCRT components. The ESCRT associated protein ALIX, consistent with its role in connecting Gag to the ESCRT-III complex and its incorporation into the virion [59], gradually accumulated at nascent assemblies of the retrovirus equine infectious anemia virus (EIAV) together with the viral Gag protein [19]. In contrast, ESCRT-III components were only transiently recruited after completion of the assembly phase. Taken together, these data support a model in which gradual Gag dependent accumulation of early ESCRT factors to a threshold level triggers subsequent recruitment and assembly of ESCRT-III [19]. Transient recruitment for only ∼35 s was also observed for the ATPase Vps4A [12, 19], which catalyzes disassembly of ESCRT-III multimers at a late stage of the process [59]. Quantitation of Vps4A at the plasma membrane by image correlation microscopy suggested recruitment of 2–5 dodecamers to individual budding sites [12]. SVT challenged the previous view that the main function of Vps4 in HIV-1 release was to recycle ESCRT-III components for further use. Both the observation that Vps4A was recruited before a detectable increase in particle movement indicative of release [12], and the finding that ESCRT-III complexes were still recruited to the budding site in the presence of dominant negative Vps4, while unable to mediate particle release under these conditions [19], suggested an active involvement of Vps4 in the fission process. The results from SVT are consistent with a model, in which a dome-shaped assembly of Chmp4 [59], possibly transiently stabilized by binding of Vps4 oligomers to its inner surface, mediates gradual membrane constriction and fission. While SVT lacks the potential to characterize the structure of ESCRT assemblies, super-resolution microscopy could provide the resolution necessary to discriminate between models proposing a dome-shaped assembly within the bud neck and those that propose larger ESCRT spirals surrounding the budding site. A more detailed molecular view awaits visualization of ESCRT complexes at the HIV-1 bud neck by ET, as already accomplished in the case of ESCRT complexes involved in cytokinesis [60]. Since ESCRT-III and Vps4 are only briefly recruited to HIV-1 assembly sites, this might require the experimental arrest of budding structures. Thin-section EM already indicated the presence of electron-dense structures, possibly corresponding to Chmp4 assemblies, in the neck of HIV-1 buds arrested upon simultaneous depletion of ESCRT-III components Chmp2A and B [61]. Alternatively, correlative microscopy might be employed to identify sites of transient ESCRT recruitment, followed by rapid fixation of the sample for ET or IA-SEM analyses. A recent study demonstrated that entering HIV-1 particles can be captured during endocytic uptake by correlating SVT with time resolution in the seconds range and cryo-ET structure analysis [20]. However, this challenging approach has not yet been applied to virus assembly.

Following membrane fission, particles are released from the cell surface. This step is blocked by the cellular restriction factor tetherin, which inhibits the dissociation of HIV-1 and other enveloped viruses from the plasma membrane unless counteracted by a specific viral defense mechanism [62, 63]. Tetherin is an external membrane protein, bound by an unusual combination of a transmembrane domain and a C-terminal glycosyl–phosphatidylinositol anchor. It is also incorporated into the HIV-1 lipid envelope and interactions between tetherin molecules and the viral and cellular membrane connect the budded virion with the cell [62, 63]. Dual-color STORM has recently been employed to characterize the distribution of tetherin molecules and HIV-1 tetherin interactions at the membrane, revealing clusters of a low number of tetherin molecules at HIV-1 budding sites [27]. While the formation of compact membrane-localized tetherin clusters depended on both membrane anchoring moieties of the protein, the transmembrane domain appeared to be particularly important for the localization of tetherin to the vicinity of Gag assemblies [27].

HIV-1 Cell-to-cell Transmission

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

Release and maturation are typically displayed as final steps in textbook schemes of HIV-1 replication. Changing perspective, they may also be considered as initial steps of the next cycle preparing the virus for entering a new target cell (Figure 2). Infection in tissue culture is usually performed by adding cell-free HIV-1 particles to target cells. However, direct transmission from infected to non-infected T-cell through a so-called virological synapse is more efficient in vitro and probably represents the relevant pathway of viral spread in vivo [64]. Direct transmission between T-cells does not fundamentally differ from cell-free infection in the sense that it involves budding of infectious particles from the producer cell, followed by uptake into the target cell [65]. This results in an intimate and dynamic connection between late and early steps of HIV-1 replication. Consequently, cell-to-cell transmission is difficult to study by bulk methods, rendering visualization of the process an important aim. SVT was in fact crucial for discovering extracellular ‘surfing’ of retroviruses along filopodial extensions towards the cell body [66]. Since then, it has been employed by a number of groups to further investigate intercellular transport of HIV-1 via filopodial bridges or nanotubes [67]. A replication competent TC-tagged HIV-1 derivative [32] and an eGFP labeled derivative displaying replication competence in MT-4 T-cells [14] have been used to follow Gag transfer to the contact site between infected macrophages and T-cells [32] as well as across the virological synapse between two T-cells [15] by SVT. Time-resolved 3D confocal microscopy of T-cell synapses revealed movement of Gag patches at the membrane towards the cell adhesion site, the formation of button-shaped Gag accumulations at the synapse and finally the transfer of Gag from these buttons into the target cell [15]. The temporal and spatial resolution did, however, not allow dynamic analysis of individual particles. This system was recently extended to study maturation using Gag.eGFP/Gag.mRFP labeled reporter particles whose FRET signal intensity changed upon proteolytic Gag maturation [68]. Based on the average FRET signals of intracellular Gag punctae after synaptic transmission, the authors proposed that HIV-1 passes the synapse as immature, non-infectious particle, which is taken up into endosomes where it is activated for fusion by slow proteolytic maturation [68]. However, this proposition is not easily reconciled with the general observation that mature virions are detected without difficulty in the vicinity of producer cell membranes by EM whereas morphological maturation intermediates are not observed, suggesting that HIV-1 maturation is relatively rapid. Furthermore, ET analyses of T-cell virological synapses revealed numerous mature HIV-1 particles in the intercellular cleft [69, 70]. The time course of proteolytic maturation thus requires further investigation.

While events at HIV-1 virological synapses have been characterized in recent years from the structural as well as from the dynamic side by EM and FM, respectively, [64] direct experimental connection between both aspects has not been achieved. The capture of HIV-1 particles on or within cells by correlative microscopy following cell free infection [20, 46] represents a step towards this aim. An even more challenging goal is to follow HIV-1 spread through organotypic cell cultures or live tissue. Besides further advances in FM technology, this will require the development of a fluorescently labeled HIV-1 derivative which replicates with high efficiency in different cell types.

Conclusion

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

During the past decade, technological developments have greatly enhanced the potential of fluorescence microscopy methods to study virus cell interactions. They provided virologists with novel tools to tackle questions addressed only with difficulty, or not at all, by other methods. While SVT is firmly established as a tool to investigate the dynamics of HIV-1 cell interactions, super-resolution FM and correlative microscopy are currently emerging as complementing options whose potential for the investigation of HIV-1 replication still waits to be fully exploited. For correlative microscopy, current challenges lie in improving methods that allow precise spatial correlation between FM and EM images and the rapid transfer of samples from live FM to EM observation. Developments that would further enhance the attractiveness of super-resolution FM comprise expanded options for multi-color detection, enhanced resolution in three dimensions and an improved temporal resolution, which would allow combining high spatial resolution with the advantages of SVT.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References

Work in the lab of B. M. has been supported by the Deutsche Forschungsgemeinschaft (grant MU885/4) and European Union FP7 grant HEALTH-F3-2008-201095 (HIV-ACE). We thank Hans-Georg Kräusslich for continuous support.

References

  1. Top of page
  2. Abstract
  3. Fluorescent HIV-1 Derivatives
  4. Microscopic Methods
  5. HIV-1 Particle Assembly
  6. HIV-1 Release
  7. HIV-1 Cell-to-cell Transmission
  8. Conclusion
  9. Acknowledgments
  10. References