Nanoimaging of ssRNA: Genome Architecture of the Hepatitis C Virus Revealed by Atomic Force Microscopy

The complex structures that RNA molecules fold into play important roles in their ability to perform various functions in the cell. The structure and composition of viral RNA influences the ability of the virus to implement the various stages of the viral lifecycle and can influence the severity of the virus effects on the host. Although many individual secondary structures and some tertiary interactions of the Hepatitis C virus genome have previously been identified, the global 3D architecture of the full 9.6 kb genome still remains uncertain. One promising technique for the determination of the overall 3D structure of large RNA molecules is nanoimaging with Atomic Force Microscopy. In order to get an idea of the structure of the HCV genome, we imaged the RNA prepared in the presence of Mg2+, which allowed us to observe the compact folded tertiary structure of the viral genome. In addition, to identify individual structural elements of the genome, we imaged the RNA prepared in the absence of Mg2+, which allowed us to visualize the unfolded secondary structure of the genome. We were able to identify a recurring single stranded region of the genome in many of the RNA molecules which was about 58 nm long. This method opens up a whole new avenue for the study of the secondary and tertiary structure of long RNA molecules. This ability to ascertain RNA structure can aid in drawing associations between the structure and the function of the RNA in cells which is vital to the development of potential antiviral therapies. Journal of Nanomedicine & Nanotechnology J o u r n a l o f N an om edicine & Natechn o l o g y ISSN: 2157-7439 Citation: Gilmore JL, Aizaki H, Yoshida A, Deguchi K, Kumeta M, et al. (2014) Nanoimaging of ssRNA: Genome Architecture of the Hepatitis C Virus Revealed by Atomic Force Microscopy. J Nanomed Nanotechol S5: 010. doi:10.4172/2157-7439.S5-010


Introduction
HCV is a worldwide epidemic, with about 150 million people worldwide infected and 350,000 deaths per year [1]. There is no vaccine for hepatitis C and treatments have met with limited success [2]. Given how pervasive HCV is, a better understanding of the virus may lead to important new discoveries which can help to highlight new avenues by which to treat individuals infected with the virus. Since HCV was originally cloned in 1989 [3], numerous studies have tried to understand the structure of the viral RNA, mainly in the conserved untranslated regions (UTRs) of the genome. Structural motifs that have been identified include the internal ribosome entry site (IRES) located in the 5'-UTR [4][5][6][7][8][9][10][11][12][13][14][15][16][17], the poly-U/UC region and 3'X RNA regions located in the 3'-UTR [18][19][20][21][22][23][24], and some stem loops in the coding region [25][26][27][28][29]. Additionally many long-range contacts between various regions of the genome have been reported [21,[30][31][32][33][34][35]. Despite the vast number of reports on various structural features of the HCV genome, the full 3D architecture of the full genome remains uncertain.
Most techniques generally study bits and pieces of a single stranded (ss)RNA genome and hope to eventually arrive at the final global genome structure over time, or they provide data about various interactions within the molecule without providing 3D information [36]. The ability of Atomic Force Microscopy (AFM) to directly visualize the nanostructure of the whole genome in a variety of configurations in a single experiment makes it a very useful technique to assess the folded structures formed by a variety of ssRNA molecules. The usefulness of AFM for evaluating the nanometer scale architecture of biological molecules [37] was realized soon after the inception of the technology [38]. AFM has been used extensively for imaging of DNA with applications ranging from the study of DNA dynamics [39], DNA-protein interactions [40][41][42][43][44][45], to DNA origami structures [46,47].
However, techniques for obtaining reproducible ssRNA images have been slower to develop, leading to far fewer AFM imaging studies on ssRNA, although the nanometer scale resolution of AFM makes it a valuable tool for revealing the organization of ssRNA structures. Most AFM studies which have visualized ssRNA have used Mg 2+ concentration of 4-10 mM [48][49][50][51][52]. However, a couple of early studies [53,54], as well as a couple others studies of viral RNA [55,56] have achieved more extended ssRNA configurations using low salt solutions. It is well documented that ions play a much larger role in the folding of RNA tertiary structures, but play a minimal role in the formation of secondary structures formed by Watson-Crick base pairing of the nucleotides in the RNA chain [57][58][59]. Thus, by imaging various ssRNA molecules prepared without the addition of Mg 2+ ions, we should be able to get valuable information about the individual secondary structural motifs in the genome and and connectivity of the molecule.
In this study, we developed a method to observe both the secondary and tertiary structure of the full 9678 nt HCV genome with nanometer resolution using AFM. By omitting Mg 2+ from the buffer used to dilute the RNA prior to AFM imaging, we were able to observe the secondary structure of the molecule. The molecules had a linear configuration with various appendages extending from the molecule. Some commonly observed structural features were present in these molecules, including one end with a 58 nm long single stranded region and another end with a sharp kink followed by a region with a beaded appearance lacking appendages. The single stranded region observed is likely to be the poly-U/UC motif which is conserved across HCV genomes [18][19][20][21][22][23][24]. The morphology of homopolymeric ssRNA was further verified by imaging an ssRNA transcript of the importin-β coding region prepared in vitro with and without a poly(A) tail.

RNA
The JFH-1 genomic RNA (accession no. AB047639) is a 9678 nt positive strand RNA [60]. The JFH-1 strain, belonging to genotype 2a, is the first strain of the Hepatitis C virus isolated from a patient which could be successfully replicated in cell culture [61]. The JFH-1 RNA used for these experiments was synthesized in vitro by T7 polymerase transcription and purified with the Zymoclean kit [62,63]. The polyadenylated importin-β mRNA was prepared in vitro. Mouse cDNA for importin-β was cloned into the pTD vector (Shimadzu Corp., Japan) and transcribed with the MEGAscript T7 kit (Ambion, USA). The RNA was polyadenylated with the Poly(A) Tailing kit (Ambion, USA) and purified by LiCl precipitation.

AFM sample preparation
All pieces of mica were cut with a hole punch and super-glued to a 15 mm metal specimen disc at least a day prior to sample deposition for AFM. The JFH-1 HCV RNA was prepared by diluting the RNA sample to 0.1-0.5 ng/µL in buffer EB (10 mM Tris, pH 8.5) from Qiagen or in a buffer containing 20 mM Tris, 1.5 mM MgCl 2 , and 1.5 mM dithiothreitol (DTT). The sample was then heated to 65°C for 2 mins, and then kept at room temperature until deposition. Just prior to sample deposition, a piece of mica affixed to a 15 mm metal specimen disc was freshly cleaved and 10 µL 10 mM spermidine was deposited on the mica and incubated for 3.5 mins. The spermidine was then washed away by washing three times with 1 mL of milli-Q water immediately followed by deposition of 10 µL of the RNA sample for 3.5 mins. The sample was then washed again three times with 1 mL of milli-Q water and dried with a stream of nitrogen. The sample was immediately used for AFM imaging.

AFM imaging and analysis
AFM imaging was performed with the Multimode AFM with a Nanoscope III or IV controller and with an E or a J scanner (Digital instruments, Inc., USA). The microscope was operated in the Tapping Mode™ at a scanning rate of 0.5-1.0 Hz. Rectangular silicon cantilevers with sharpened tetrahedral tips were used (OMCL AC160TS, Olympus Corp., Japan). These probes had a tip radius of about 7 nm, a resonant frequency of about 300 kHz, and a spring constant of about 26 N/m. Images displayed in this publication were flattened in the Nanoscope (v.5.31 r1) software and exported as jpegs. Images were analyzed with the Gwyddion software (Department of Metrology, Czech Metrology Institute, Czech Republic). The raw SPM images were opened in the Gwyddion software and processed by correcting the horizontal scars, correcting lines by matching the height median, leveling the data to make facets point upwards, leveling the data by mean plane subtraction, and removing the polynomial background (3 rd degree). The RNA molecules were marked by selecting all molecules above a 0.15-0.20 nm threshold and then removing all grains smaller then 60-70 px 2 . The edit mask tool in the subtract selection mode was used to deselect molecules from the periphery of the image. The distributions of various grain characteristics were then exported as a raw data file and then the measure individual grain tool was used to click on each molecule and record the zero basis volume and minimum circumcircle radius of each one. The values were inserted into the Origin Lab software (OriginLab Corp., USA), plotted into histograms, and fitted with Gaussian curves to obtain the center value(s) for each histogram. The errors reflect the standard deviation of the Gaussian distribution. The contour length of the molecules was measured and height profiles were generated with the Femtoscan software (Advanced Technologies Center, Moscow, Russia). The scatterplot was plotted and fitted with a linear regression curve with the y-intercept set to 0 in Microsoft Office Excel.

AFM imaging of HCV RNA tertiary structure
In our experiments, we used the full positive-stranded ssRNA HCV genome which is 9678 nt in length. When the RNA was heated for 2 mins, at 65°C in buffer containing 1.5 mM MgCl 2 just prior to deposition on the mica surface for imaging, the RNA molecules had a compact morphology with a variety of appendages extending from the structures ( Figure 1A and 1B). A histogram of the volume distribution of these molecules was fitted with a single peak at 3823 ± 1857 nm 3 ( Figure 1C). The diameter of the smallest enclosing circle for this Mg 2+containing buffer was 136 ± 40 nm ( Figure 1D). The diameter of HCV viral particles has been reported to be about 30-75 nm [64][65][66][67][68][69][70][71][72][73][74], so the RNA is likely to be compacted 2-5 times more when it is packaged into the capsid.

AFM imaging of HCV RNA secondary structure
Since Mg 2+ is required for many tertiary contacts to form, but not required for the formation of RNA secondary structures formed primarily by Watson-Crick base-pairing [57][58][59], we should be able to image the secondary structure without the formation of tertiary contacts by diluting the RNA in Mg 2+ -free buffer. After heating for 2 mins at 65°C in this buffer just prior to deposition, the morphology of the RNA was markedly different (Figure 2A and 2B). These RNA molecules had a linearized morphology with various small appendages along the molecule. When the volume of these molecules was analyzed, there was a main peak at 4136 ± 1768 nm 3 ( Figure 2C). This is close to the volume of the HCV genome molecules in the Mg 2+ buffer ( Figure  1C), suggesting that single molecules are observed in both set of conditions. There is also a larger peak in the volume histogram at 8691 ± 436 nm 3 . This suggests that some of the molecules may actually have dimerized in these images. The diameter of the smallest enclosing circle was 219 ± 63 nm ( Figure 2D), 1.6 times larger than the molecules in the Mg 2+ -containing buffer.

Structural Analysis of HCV RNA
To get a better idea of the secondary structure of the HCV RNA molecules, additional dimensional analysis was performed. In Figure 3, two molecules which appeared to have some structural similarities are shown. These two molecules show a common structural characteristic of one end with a blob with a height of about 2 nm followed by a 35-76 nm long region of the molecule with heights of about 0.5 nm ( Figure  3A Figure 3A, graphs). These values can vary considerably depending on the orientation of the molecule on the mica surface. The other end of the molecules has an appendage-free region with a 'beaded' appearance showing variable structures with heights of about 2-3 nm preceded by a sharp kinked region ( Figure 3A, left, arrows).
Additional molecules displaying this single-stranded morphology are displayed in Figure 3B. A histogram of the length of this single stranded region results in a single peak centered around the value of 58 ± 8 nm ( Figure 3C). It is likely that this region corresponds to the poly-U/UC region present in this RNA molecule. The poly-U/C region is conserved throughout all HCV genotypes, but varies in length [18]. In the JFH-1 genome used in this study, this region is 103 nt long. Varying the length or composition of this region has been reported to effect the efficiency of viral replication [19][20][21] as well as to influence recognition of the genome by the RIG-I protein which is necessary for the host cell to mount an Interferon (IFN)-mediated immune response to the virus [22][23][24]. Assuming that this single stranded regions is actually 103 nt long (the whole poly-U/UC region is unpaired and the regions immediately flanking it are paired), it would mean that, according to our measured length (58 nm), the length per nucleotide of the single stranded region is 0.56 nm. In a previous force measurement study with homopolymeric ribonucleotides, molecules 1500-4000 nt in long were stretched to a length of about 1600 nm. This means that the length per nucleotide should be about 0.4-1.0 nm, in good agreement with our calculated value [75,76]. The small blob at the end of the single stranded region is likely the 98 nt 3' X-RNA which is a well studied region immediately flanking the poly-U/UC region and has been reported to form either two or three stem loops [35].

Verification of RNA homopolymer morphology and ssRNA volume
To further verify the morphology of homopolymeric singlestranded RNA with AFM, an 3890 nt importin-β ssRNA transcript was also imaged (Figure 4). This transcript is not a true mRNA because it lacks the 5' and 3'-UTRs. The importin-β ssRNA was prepared by in vitro transcription in two different ways. Initially, the ssRNA was simply transcribed from cDNA with T7 polymerase and directly imaged ( Figure 4A). The same RNA was then treated with polyadenylate polymerase to add a random number of adenine residues to the molecules and imaged ( Figure 4B). The addition of the poly(A) tail shows a single stranded morphology similar to what was observed for the end of the viral RNA molecules. Along the mainly single stranded region, there are often kinks, or 'knobs" and some apparent doublestranded regions. This observation is corroborated by previous AFM imaging which also observed a 'knob' like structures along poly(A) RNA molecules [54]. Additionally, the crystal structure of a poly(A) duplex has been resolved to 1 Å resolution [77].
Measurement of the contour length of the importin-β poly(A) tails showed a very broad distribution of values ranging from 40-830 nm long ( Figure 4C). Attempting to fit the broad distribution with a Gaussian curve resulted in a value of 125 ± 88 nm, suggesting that the majority of the poly(A) tails were about twice as long as observed for the viral RNA and the standard deviation of the distribution is 11 times higher. Although the morphologies of the homopolymeric RNA molecules were similar, past studies have reported results suggesting that homopolymeric chains of different nucleotides do not behave exactly the same. For example, poly(G) forms tetrads [78]. Poly(C) and poly(A) have been reported to form single-stranded helical regions which are reflected by plateaus when the chains are stretched resulting from a helix-coil transistion. However, when poly(U) is stretched, there are no plateaus, suggesting that it does not form helical domains [76].
Verification that the viral ssRNA structures we observed represent the full viral genome comes from comparing the measured volumes in of various ssRNA molecules to their molecular weight. The graph in Figure 4D shows how the measured volumes of four different kinds of ssRNA molecules relates to their length in nucleotides. The four molecules analyzed include the HCV RNA (3823 ± 1857 nm 3 ), the importin-β ssRNA without the poly(A) tail (1930 ± 437 nm 3 ), the 28S ribosomal (r)RNA (2688 ± 655 nm 3 ), and the 18S rRNA (921 ± 527 nm 3 ). By fitting these points with a linear regression curve, the volume measured using our method increases by 0.4 nm 3 per nucleotide of RNA. Images of the 18S and 28S rRNA molecules are shown in Figure  4E for reference.

Perspective
The ability of AFM to visualize Mg 2+ -dependent structural changes of RNA opens up an exciting range of possibilities for what AFM can reveal about RNA in the future. For example, the ability of AFM to image molecules in aqueous environments opens up the exciting possibility of visualizing the folding/unfolding transitions of RNA using highspeed AFM (HS-AFM) [41][42][43][44][45]. Additionally, further refinement of the structural models could be gained by imaging individual portions of the RNA. Comparing these individual domains may aid in understanding how their structures relate to the structures observed in the full viral genome. Additionally, labeling strategies could be used to identify where a particular sequence may occur along an RNA backbone. In addition to labels, the interaction of various RNA-binding proteins with various RNA structures can also be studied. If a method to attach a RNA-binding protein or chemical to the AFM tip can be developed, recognition imaging [79][80][81][82] may also be used to assess which structural features of an RNA molecule the ligand interacts with. Also, proteins can be imaged along with the RNA to observe the effect they have on the viral structure. The ability of AFM to visualize global structural rearrangements of RNA may be a promising tool to investigate the role of RNA structures in viral processes for future studies, which could provide us with an understanding of how to control the spread of the virus in infected individuals.