Biomarking Infection with BSI
By: Carol A. Rouzer, VICB Communications
Published: April 1, 2013
Back-scattering interferometry offers a new approach to detecting and quantifying pathogen-derived RNA in the infected host.
Infection with most human bacterial and viral pathogens is accompanied by the production of single-stranded RNA that is specific to the invading organism. In the case of some viruses, the amount of foreign RNA synthesized during the course of infection can be quite high, suggesting that these molecules could serve as biomarkers of the disease. Exploiting this opportunity, however, requires a rapid, reliable, sensitive, and easy method to detect and quantify viral RNA in complex biological samples. Currently available methods for RNA detection, including reverse transcription-polymerase chain reaction (RT-PCR), and assays based on fluorescent or other chemical probes, require extensive sample processing, lack adequate sensitivity, and/or require chemical synthesis of a complex probe molecule. To solve this problem, Vanderbilt Institute of Chemical Biology members Darryl Bornhop and David Wright along with collaborator Rick Haselton explored the application of back-scattering interferometry (BSI) to the detection and quantification of viral RNA in biological samples [N. M. Adams, et al. (2013) Nuc. Acids Res., published online March 21, DOI:10.1093/nar/gkt165].
BSI uses a relatively simple instrument comprising a helium-neon laser, a microfluidic chip bearing a semi-circular channel, and a CCD (charge-coupled device) camera (Figure 1). The laser irradiates the sample located in the channel of the microfluidic chip. As the light passes through the sample, it is reflected and refracted multiple times by the walls of the channel, so that the light reflected back from the channel is in the form of a set of high contrast interference fringes, which is recorded by the camera. The fringe pattern is determined by the refractive index (RI) of the sample, so changes in RI resulting from interactions between molecular species in the sample result in changes in the fringe pattern. The magnitude of the change is dependent both on the structural alterations occurring during the molecular interaction and on the concentrations of the interacting species. Thus, BSI can be used both to monitor molecular interactions and to quantify the molecules involved in those interactions.
Figure 1. Diagrammatic representation of a back-scattering interferometer. Light from a helium-neon laser is focused on the channel of a microfluidic chip. Passage of the light through the sample in the channel results in its refraction to form an interference fringe pattern that is recorded by a CCD camera.
RNAs are defined by their unique sequences, and the relatively simple structure, uniform charge, and high water solubility make nucleic acid-based probes promising tools for BSI applications. To test the use of BSI for the detection of viral RNAs as potential biomarkers of infection, the Wright, Bornhop, and Haselton labs focused on the 1300 nt RNA transcript of the respiratory syncytial virus (RSV) nucleocapsid gene (RSVN) (Figure 2). They generated complementary DNA probes of 15, 22, 44, and 88 nt in length, all targeted to the same region of the RSVN transcript, and used BSI to monitor the association of varying quantities of the transcript with an excess quantity of each DNA probe. In all cases, BSI detected the probe-transcript interaction; however, the strength of the observed signals varied.
Figure 2. The respiratory syncytial virus (RSV) is an irregularly shaped single stranded RNA virus. It is a common cause of respiratory illness, especially in young children. Image reproduced from Wikimedia Commons under the Creative Commons Attribution-ShareAlike 3.0 Unported License.
Based on both the slope of the plot of change in fringe pattern (measured in radians) versus RNA concentration, and on the calculated lower limit of detection (LOD) of the viral RNA, the investigators concluded that the 22 nt probe provided the best overall response, followed by the 44 nt, 88 nt, and 15 nt probes, respectively (Figure 3). Of particular interest was the finding that four consecutive probes provided a better signal than a corresponding single 88 nt probe, though dispersing the four probes across the entire transcript increased the response even more. The best signal was derived from the use of nine probes dispersed across the entire transcript (Figure 3). Separate experiments using a fluorescent dye that intercalates into double-stranded oligonucleotides confirmed that BSI was, indeed, detecting hybridization of the DNA probes with the RSVN transcript.
Figure 3. Diagrammatic representation of the probes used to detect the presence of RSVN RNA (red). DNA probes are indicated in black, and LNA probes are in blue.
The initial results indicated that the BSI signal was dependent on the length, number, and distribution of probes used to detect the target RNA. The investigators next tested the response to 22 nt DNA probes bearing 0, 1, 3, 5, 7, or 10 mismatched bases. They found that 1 and 3 mismatches were tolerated, producing no statistically significant change in BSI response. In contrast, 5 or more mismatches produced significantly reduced signals, although some binding was still detected with up to 7 mismatches. When RSVN RNA was added to samples containing HEP-2 cell lysate, BSI easily detected the presence of the viral RNA if total RNA in the sample was extracted prior to analysis. However, in unextracted samples, signal strength was suppressed by increasing amounts of cell lysate.
To better understand the basis for the response of BSI to the interaction of DNA probes with RSVN RNA, the investigators designed seven different 22 nt probes targeting different regions of the RSVN sequence. They then used mfold software to predict the secondary structure of the RSVN RNA in the region where each probe should bind. They discovered that, in general, BSI signal intensity correlated with the extent of single-stranded secondary structure in the probe target region. However, there were two probes that gave a much stronger signal than predicted based on this structural characteristic, leading the investigators to propose that binding of these probes resulted in a change in tertiary structure of the RSVN RNA.
Figure 4. Structure of an LNA nucleotide showing the methylene bridge between the 4’-carbon and 2’-oxygen of the deoxyribose.
Locked nucleic acids (LNAs) are RNA analogs in which a methylene bridge connects the 4’-carbon to the 2’-oxygen of the ribose ring (Figure 4). The bridge locks the ribose in the 3’-endo conformation, thereby enhancing base stacking and backbone organization. As a result, LNA probes tend to exhibit enhanced hybridization with complementary DNA or RNA targets. Thus, the researchers tested the ability of LNA-based probes to generate a BSI signal upon binding to the RSVN RNA. The results revealed that LNA-based probes produced a stronger signal than DNA-based probes targeting the same sequence in the viral RNA transcript (FIgure 2). However, the investigators were surprised to find that this increase in signal did not correlate with increased hybridization in the case of the LNA probes. This led them to hypothesize that the observed increase in BSI signal was due to a higher content of A-form structure in the LNA-RNA hybrids as compared to DNA-RNA hybrids. The A-form helix is favored both in RNA-RNA hybrids and by the 3-endo conformation found in LNAs, whereas DNA-DNA hybrids favor the B-form helix (Figure 5). To test their hypothesis, the investigators explored the effect of addition of trifluoroethanol to a sample of double-stranded DNA, which gradually converts the B-form helix to the A-form. They observed a marked shift in BSI signal that accompanied the conformational change, confirming that the strong response to LNA-based probes could at least in part, be explained by formation of hybrids that adopt the A-form conformation.
Figure 5. Comparison of the A-form double helix (left) favored by RNA-RNA hybrids and dehydrated DNA-DNA hybrids and the B-form double helix favored by DNA-DNA hybrids. Image reproduced from Wikimedia Commons under the Creative Commons Attribution-ShareAlike 3.0 Unported License.
Together the results confirm the potential for using BSI to detect viral and other foreign RNAs in complex biological samples. Further work will be required to maximize specificity, define the parameters for optimal probe design, and develop the best methods to avoid interference from sample matrix components. Considering the sensitivity, economy, and ease of use of BSI, exploration of ways to fully exploit this technology will be well worth the effort.