Positive-strand RNA, (+)RNA, viruses include many human pathogens, such as the families picornaviridae (e.g., poliovirus [1
]), togaviridae (e.g., rubella virus [2
]), and flaviviridae (e.g., Dengue virus [3
], Zika virus [4
], and hepatitis C virus [5
]). While these viruses pursue a wide range of replication strategies in diverse hosts, they also share key features and are defined by their use of a (+)RNA genome. This genomic RNA fulfils three distinct functions: (1) it is the mRNA from which proteins are produced; (2) it is the template from which the negative-strand RNA ((−)RNA) is transcribed, to serve as the replicative intermediate; and (3) it codes the genetic information that must be packaged into assembling particles and transferred to target cells. At any given time, an infected cell will contain many (+)RNA molecules performing these roles, although any single (+)RNA molecule may only function in one capacity at one time. As the (+)RNA molecules are identical to one another, they cannot be differentiated by sequence and their separate roles can only be determined through simultaneous analysis of the interacting cofactors.
Following infection of a target cell, the viral RNA first serves as mRNA, exploiting the host-cell translational machinery to direct synthesis of the viral proteins; these proteins include enzymes that are responsible for synthesis of first the negative, then the positive strand of the genome, and proteins that modify the environment of the cell to support viral replication. Positive-strand RNA viruses replicate their genomes in the cytoplasm of infected cells, in association with virus-induced membrane structures, often termed the “membranous web” [6
]. These membranes provide a foundation on which to anchor the viral replication complex (RC), and in combination with viral proteins, may provide protection against surveillance by the innate immune system. In the RC, the virus synthesizes new (+)RNA. At early times post-infection, the new (+)RNA will be used to generate more viral proteins and (−)RNA; at later time points, (+)RNA is packaged into particles made up of the viral structural proteins, and released from the cell. To better understand the replication of these viruses, their interactions with the host cell, and ultimately, how to combat them, it is necessary to consider both (+) and (−)RNA, and their interaction with proteins, viral and cellular.
While there are thousands of antibodies available to specifically identify viral and cellular proteins, and sufficient fluorescent tags to allow co-visualization of multiple proteins in a single sample, these imaging approaches are often incompatible with conventional methods for visualization of nucleic acids (e.g., hybridization of fluorescently labeled oligonucleotide probes). Branched DNA (bDNA) in situ hybridization is a technique that exploits sequence specific probes, and branching preamplifier and amplifier DNAs, to produce an intense localized signal [7
]. Unlike conventional FISH methods, bDNA FISH is readily compatible with immunofluorescence, allowing simultaneous analysis of nucleic acids and proteins (Figure 1
). Various bDNA approaches have been developed for commercial use, including RNAscope [8
], PrimeFlow [9
], and ViewRNA [10
]. These techniques have been applied to quantify and localize specific nucleic acids and the cells that harbor them. The compatibility of PrimeFlow with flow cytometry has proven particularly useful for analysis of human immunodeficiency virus (HIV)-1 latency and reservoirs [11
], while RNAscope and ViewRNA have been employed diagnostically for histological staining [14
], and in cell biology to visualize cellular and viral RNAs [20
]. These techniques have frequently been employed in low resolution imaging approaches, such as histology and flow cytometry, that exploit the robust signal to clearly identify rare infected cells. A variation on the robust detection of an abundant RNA in rare cells is the detection of less abundant targets; these FISH methods have sufficient sensitivity to identify individual nucleic acid molecules, sometimes referred to as single molecule FISH. One such application of this sensitivity has been to modify the experimental conditions to achieve labelling of the viral nucleic acids in HIV-infected cells [22
]. Most infected cells in the clinical context only contain a small number of integrated proviruses, typically just one [26
]; thus, the signal amplification of bDNA FISH renders it ideally suited to the visualization of such a low abundance target.
Similar to the integrated provirus of retroviruses, the (−)RNA of (+)RNA viruses is present in low amounts relative to the (+)RNA [27
]. It is, nevertheless, essential for the synthesis of (+)RNA and, as it lacks the multiple functions of (+)RNA, it is a more reliable marker of the RC. Here we use hepatitis C virus (HCV)-infected human hepatoma cells and Zika virus (ZIKV)-infected Vero cells to demonstrate the specific labelling of (+)RNA and (−)RNA, allowing analysis of viral activity at an individual cell level. To validate the approach, we applied bDNA FISH to a phenomenon we had previously examined [31
], that of the rapid response of HCV-infected cells to a variety of anti-HCV direct-acting antiviral agents (DAAs). Using this approach we visualized the rapid decline in HCV RNA associated with the use of NS5A inhibitors [31
]. This new assay represents a novel approach to evaluate other RNA viruses in future studies.
Studying the (−)RNA of (+)RNA viruses is particularly challenging, owing to its relatively low abundance in infected cells [27
], and the low sensitivity of traditional fluorescent in situ hybridization imaging techniques when applied to low abundance transcripts [47
]. Nevertheless, understanding the behavior of (−)RNA is of critical importance to elucidating the HCV replication mechanism, as it is a most critical component of the RC due to its function as the template for the production of new (+)RNA. Consequently, disruption of the RC may best be understood by study of the (−)RNA. The low amount of (−)RNA in cells might reflect the limited role it plays in the viral replication cycle and/or may be a viral strategy to avoid forming dsRNA, a potential inducer of innate immunity [48
In the current study, we specifically labeled the (−)RNA of HCV, both in singly and dual labeled infected cells. In addition, we performed cell-level quantitation of the two RNA species, and compare this imaging-based quantitation to RT-qPCR; we found an approximately five-fold excess of the HCV (+)RNA by both techniques, consistent with a previous report that used RT-qPCR [27
In contrast with HCV, ZIKV (+)RNA was so abundant in infected cells that it was not possible to count foci, making direct comparisons to the abundance of (−)RNA unreliable. We also found that the ZIKV (−)RNA required more rigorous pre-treatment with protease than HCV (−)RNA to permit FISH staining. We assume this reflects differences in the nature of the RC, the amount of protein bound, and accessibility to protease treatment; these properties very likely vary extensively between viruses. Nevertheless, both bDNA FISH and strand-specific RT-qPCR approaches revealed a large excess of (+)RNA over (−)RNA in Zika virus-infected cells, comparable to a previous estimate using another flavivirus, the West Nile virus [51
]. Collectively, the data obtained from HCV and ZIKV studies suggest that RT-qPCR and bDNA FISH are both suitable to follow transcript abundance, however, care must be taken when comparing transcript numbers, as efficiency of amplification and labelling can vary significantly between viruses and methods.
It was striking that our ability to detect colocalized RNA strands was extremely limited. In the case of ZIKV we saw very few (−)RNA strands by bDNA FISH, while in HCV harboring cells, (−)RNA was reasonably abundant, but rarely colocalized with the (+)RNA signal. As the replication of the viruses requires transcription of one strand from the other, we would anticipate a certain proportion of the (+)RNA and (−)RNA signals should colocalize. While we have no single definitive explanation for the lack of colocalized RNA molecules, there are potential explanations. In the Zika imaging, we found that it was necessary to increase the concentration of protease. This may indicate that the (−)RNA, in particular, is sequestered and unavailable for labeling, potentially due to extensive interactions with the RC. A related potential contributing factor is the extremely high melting temperature reported for long double-stranded RNA [52
]. This could lead to a situation where the dsRNA is generally refractory to FISH labelling (which requires a single-stranded RNA molecule), reducing the incidence of apparent colocalization, and reducing the apparent abundance of (−)RNA, a greater proportion of which is likely to be double-stranded in replication complexes. This interpretation would suggest that the (−)RNA molecules seen are predominantly “free” single-stranded RNAs; the abundance of this species may vary greatly between viruses and cell types. Efficient labeling of dsRNA species, such as may be found in the RC, may require the use of aggressive denaturation to release single-stranded RNAs for FISH.
An area of particular importance in antiviral research is the mechanism of action of DAAs, understanding their impact on virus replication and their role in therapeutic regimens. A strength of this imaging-based technique is to permit analysis of viral RNA and protein in the same cell, and to select those cells desired for analysis. All DAAs will prevent the spread of infection over time, irrespective of their specific mechanism, so long as they inhibit some part of the replication cycle. In contrast to other methods, such as RT-qPCR or Western/Northern blot analyses, which can phenotypically characterize only mixtures of cell populations, the use of an imaging-based approach allows us to specifically examine cells that are already infected, and determine how different DAAs influence on-going viral replication. Recently, we used another imaging-based strategy to assess the effects of DAAs against HCV by following NS5A staining [31
]. Similar to others [44
], we were able to show that the NS5A-targeting drugs have the fastest effect on suppression of viral proteins and total viral RNA and on redistribution of NS5A. Here, we wished to determine whether staining for viral RNA could provide additional insights. To this end, we followed the effects of antivirals targeting the protease, polymerase and NS5A on viral (+) and (−)RNA. We confirmed that the NS5A inhibitors are the most potent and fast acting inhibitors, both by counting RNA foci per cell and RT-qPCR. Whereas the extent of (−)RNA suppression was found to be similar with both methods (Figure 5
C,E, gray bars), reductions in the (+)RNA appeared less pronounced when RNA foci per cell were counted. This was likely due to underestimation of the foci number when a large number of (+)RNA was present (such as in the DMSO-treated control), leading to underestimating the effect of the inhibitors. As such, the focus counting approach may be more suited to viruses or conditions where less RNA is present.
As a solution to the difficulties in counting discrete RNA foci, we used an alternative form of quantification, measuring the total fluorescence in the field of view, expressed per cell. This greatly improved the consistency of the data, and confirmed that the NS5A inhibitors were particularly effective at rapidly inhibiting HCV, as we have previously reported [31
]. The lack of suppression by the NS3 inhibitor Danoprevir is not a surprise in this context, given that inhibiting the viral protease will have a limited impact on the function of extant RCs in the short time frame (8 h) of these experiments [55
]. It is noteworthy that the NS5A inhibitors also suppress viral RNA to a greater extent than Sofosbuvir, a potent inhibitor of the HCV polymerase, NS5B [57
]. As the concentration of Sofosbuvir used should completely inhibit polymerase function, the apparently greater potency of the NS5A inhibitors as rapid suppressors of the viral RNA is consistent with them functioning through a distinct or additional mechanism that affects the quantity of viral RNA [32
]; it is possible that these inhibitors destabilize the RC and hasten the degradation of the viral RNA, particularly the (−)RNA that is normally protected from nucleases by the RC [59
]. This hypothesis is consistent with the clear disruption of NS5A localization following treatment with the NS5A inhibitors seen in this study and previously by us and others [31
It has been previously reported that NS5A inhibitors suppress levels of (−)RNA to a greater degree than the levels of (+)RNA [61
]. In our assays we did not observe significant differences in the impact of NS5A inhibitors on (+) versus (-)RNA. The discordance in the reported results may come from the disparate timing of the two experiments. The study of Ramanan and colleagues added the inhibitors 20 h post infection [61
]; we allowed infection to proceed for 48 h before the addition of inhibitors, as we were interested in the effects of these inhibitors in the context of an established infection. It is possible that the sensitivity of (−)RNA to NS5A inhibitors is more pronounced before the infection is fully established.
Collectively our data demonstrate the utility of branched DNA in situ hybridization as a tool to help address questions relating to the biology of RNA viruses. The high sensitivity of bDNA FISH allows sensitive visualization of incoming viral RNA, before any amplification has taken place, and early stages of viral genome replication. The ability to select specific cell populations for analysis enables in-depth investigation of viral RNA strand-specific effects, localization, and interactions with proteins, owing to the compatibility with immunofluorescence. Thus, this approach may be particularly suitable, not only to the study of drug mechanisms of action, but also to studies involving over-expression or silencing of host cell-factors. The high specificity and sensitivity of staining offered by these methods should help illuminate the critical role of (−)RNA, until now a concealed player in the biology of (+)RNA viruses.