Development of concentration dependent sensors

In this section the construction and evaluation of Spinach-SAM sensors with modified concentration dependence is presented. Combining several constructs with significant differences in K_d is expected to broaden the dynamic range of the full construct.⁵ This is illustrated in Figure 1.

Figure 1: Concept; broadening the dynamic range of the sensor by combining two aptamers with different detection ranges. If one construct is able to detect the target molecule in the blue range, while another construct is able to detect the target molecule in the green range, combining the two constructs broadens the dynamic range of the biosensor, and enables detection of the target molecule in the full red range.

Design

The basis of our study is the original construct by Paige et. al in which the Spinach aptamer is linked to a SAM aptamer through a linker region.³ We hypothesized that changing the linker region between the aptamers or introducing mutations in the SAM aptamer would influence the concentration dependency of the biosensor. Four constructs were created with mutations in the linker region (T constructs) and six constructs were created with mutations in the aptamer recognizing SAM region (S constructs). Two different linker regions were used for the two different classes. The most effective linker was just one GC base pair and was used in the S constructs.³ The second most effective linker region was 2 base pairs; a GC and a CG bp. This region was mutated for the T constructs.
The purpose of the mutations in the linker region was to lower the stability of the stem thereby affecting the concentration dependency. Hence one of the GC pairs was exchanged with an AU base pair (construct T2 and T3) or an AC mismatch (construct T4).
The wild type SAM detecting aptamer is called S1. In the construct S2 and S3 one base in the metabolite recognition loop was mutated. This was expected to affect the affinity for the binding of the SAM metabolite. In the S4 and S5 constructs, one base in the second loop was mutated. This was expected to affect the stability of the tertiary structure of the aptamer-SAM metabolite complex. In the S6 construct two GC-basepairs were switched, which was expected to affect the folding of the aptamer. The different mutations are shown in Figure 2.

Figure 2: Illustrations of the secondary structure of the Spinach-SAM construct including mutation sites. Mutations made in the detection aptamer is shown (blue boxes) and named. Mutations made in the linker region is also shown (grey boxes) and named. The DFHBI binding-loop (green box) and the SAM binding-loop (red box) is shown.

Assembly PCR and transcription

Due to the length of the RNA aptamers, smaller DNA fragments were purchased and assembled into the full-lenght DNA counterpart. This was done by assembly PCR and is illustrated in the Figure 3, left. The DNA was then amplified by PCR. In the DNA sequence a T7 promotor was included. This was used for transcription of the DNA into the RNA aptamers.

Figure 3: left) Illustration of assembly PCR reaction; right) 1% agarose gel of PCR reaction for the different constructs (T1-T4 and S1-6). The expected lenght of DNA was 139-141 base pairs

Figure 4: Transcription of the DNA template into RNA. The RNA constructs were analyzed and purified by 8% denaturing PAGE.

Measurement of K_d values

The mutated RNA constructs were tested for concentration dependency by a series of measurements of fluorescence in response to the concentration of SAM.

The RNA and DFHBI concentration were kept constant at 0.26 µM and 2.6 µM respectively, while the SAM concentrations were increased from 100x to 10000x the concentration of RNA. The relative fluorescent enhancement for the 10 constructs are shown in Figure 5.

Figure 5: The relative fluorescent enhancement of the constructs. The fluorescence of DFHBI was measured (light blue), then RNA was added (orange), followed by SAM in 3 different concentrations: 100x[RNA] (grey), 1000x[RNA] (yellow) and 10000x[RNA] (dark blue). The last measurement is a control with no added RNA. All measurements were done after 10 minutes of incubation.

Some mutations induced a fluorescence response while others did not. The constructs with mutations in the detection aptamer showed no enhancements in fluorescence. This may be the result of the high binding specificity of the aptamer, which means none of the bases can be interchanged without disrupting binding of SAM.
The constructs with mutations in the linker region were more promising. Two constructs responded to variations in SAM concentrations; T1 and T2. T1 showed a fluorescence enhancement of 3.8 while T2 was enhanced by a factor of 9.6. T1 has an unmutated linker region while T2 has a mutated GC to AU-basepair. These two constructs were selected for further investigation. The K_d value was determined based on fluorescent measurements made with small increments in the addition of SAM. The results are shown in Figure 6.

Figure 6: left) Fluorescence intensity as a function of SAM concentration for T1 and T2. right) Fluorescence intensity as a function of SAM concentration for T1 and T2 on a logarithmic scale. The series were normalized to the highest total fluorescence of either T1 or T2 before the average was taken. All measurements were triplicates and DFHBI fluorescence was subtracted before plotting.

To determine the K_d value the measurements were fitted with a degree three polynomium. The K_d value was found as the concentration of SAM that lead to fluorescence at half of the maximum fluorescent intensity - that is, the concentration of SAM corresponding to 50% saturation of the construct. The K_d values were 286 µM and 202 µM for T1 and T2, respectively. The mutation in T2 made its linker region less stable than that of T1, and surprisingly more dependent on the target molecule. Counterintuitively, the least stable construct was most responsive towards the target molecule.
The two structures were mixed in a 1:1 ratio with a total RNA concentration of 0.26µM and measurements of the affinity were carried out. The result is shown in Figure 7. The total fluorescence was less than for the individual measurements of the constructs. It is possible that this problem could be helped by fusing the two together, as a defined construct could remove uncertainties in stoichiometry of the constructs.
The apparent K_d value was calculated to be 222 µM, which shows that it was possible to modulate the apparent K_d value by combining multiple constructs.

Figure 7: left) Fluorescence intensity as a function of SAM concentration for a 1:1 mixture of T1 and T2. right) Fluorescence intensity as a function of SAM concentration for a 1:1 mixture of T1 and T2 on a logarithmic scale. The series were normalized to the highest total fluorescence before the average was taken. All measurements were duplicates and DFHBI fluorescence was subtracted before plotting.

In conclusion we have shown that mutations in the linker region can modulate the stability of the sensors, furthermore, we have created a more sensitive construct, T2. Combining the two constructs confirmed that it was possible to modify the apparent K_d value of the total construct. This is a proof of concept for modulation of the concentration dependency. The dynamic range has been slightly increased, however, additional constructs with markedly changed affinities are needed in order to apply the sensor.

Finding orthogonal optical aptamers

Another aspect of the project is to find an orthogonal optical RNA aptamer, which should function as an internal control for the level of transcription. The aptamer should be orthogonal to the Spinach-SAM-DFHBI complex. This includes the following characteristics:

The excitation or emission wavelengths of the orthogonal aptamers must not overlap.
The emission wavelength of one aptamer must not overlap with the excitation wavelength of the other.
The optical RNA aptamers should not display unspecific binding to non-target ligands.

Synthesis of fluorophores and aptamers

Figure 8: The synthesized fluorophores. The DFHBI fluorophore and the analogues 2-HBI, NBI, DMHBI, DMABI and NI. An analogue of the Mango aptamer fluorophore TO3-PEG2-biotin (TO3).

The four fluorophores, DFHBI, 2-HBI, DMHBI and DMABI from the article by Paige et. al, 2011 were synthesized. In addition to these two new fluorophore analogues were proposed and synthesized. It was anticipated that these would display a bathochromic fluorescence shift due to the extended pi network of the NI fluorophore and the NBI having a nitro-group with a hydroxyl-group acting as an auxochrome. Lastly TO3 from the article by Dolgosheina et. al was synthesized. This fluorophore binds to the optical aptamer called Mango.⁶ The structures of the fluorophores are shown in Figure 8. The corresponding aptamers were produced by assembly PCR, amplified by PCR and transcribed into RNA. The gels showing successful amplification are shown below.¹

Figure 9: left) Assembly PCR visualized in a 1% agarose gel run at 100 Volt/12 cm for 100 minutes. From the left: aptamer 11-3, 2-4, 17-3, Ladder, 3-6, 6-8 and Spinach. right) A 8% polyacrylamide gel run at 400 Volt/19 cm for 75 minutes. From the left: 11-3 (104bp), 2-4 (103bp), 17-3 (104bp), 3-6 (114bp), 6-8 (104bp) and Spinach (95bp). The two gels shows an efficient assembly PCR and transcription of all aptamers.

Evaluation of fluorophore-aptamer pairs

The fluorescense enhancement of the fluorophore upon aptamer binding was initially evaluated (Figure 10). The 2 new fluorophores showed no enhancement and are left out of the graph. As shown in the graph, Spinach and Mango has high fluorescence enhancements.

Figure 10: Experimentally found fluorescence enhancement (Fluorescence of the aptamer-fluorophore complex divided by the fluorescence of the fluorophore) of the different aptamer-fluorophore complexes synthesized as in the article by Paige et. al in 2011.

An experiment was setup to test the orthogonality of Mango, Spinach, and their respective fluorophores. The two constructs are excited at 610 nm for Mango-TO3 and 460 nm for Spinach-DFHBI. The results show that the only significant fluorescence enhancement is achieved for the complexes with the correct fluorophore at the correct excitation wavelength (Figure 11). A small enhancement at 497 nm (maximum emission wavelength for Spinach-DFHBI) is observed for Spinach and TO3 with excitation at 460 nm. The enhancement is 4x the fluorescence of TO3 alone. The orthogonality is still preserved, as there is no fluorescence enhancement for the T1 Spinach-SAM complex. When adding SAM to reach the usual maximum fluorescence the fluorescence enhancement is just 1.03x for T1-TO3. This is shown in Figure 12 (comparison Figure 5).

Figure 11: Fluorescence of Mango & TO3 (top left), Spinach & DFHBI (top right), Mango & DFHBI (bottom left) and Spinach & TO3 (bottom right) measured at exc. 460 nm em. 480-600 nm (red and blue graphs) and exc. 610 nm em. 620-800 nm (purple and green graphs).

Figure 12: Fluorescence of T1 and TO3. From bottom: TO3, TO3 & T1, TO3 & T1 & SAM 100x[T1], TO3 & T1 & SAM 1,000x[T1] and TO3 & T1 & SAM 10,000x[T1].

Based on these data it can be concluded that Mango will be able to serve as an internal control for our sensor.

In vivo setup

The aim was to test the concentration dependent constructs in vivo. The previously published results for in vivo fluorescence of Spinach was reproduced, however our final constructs remain to be tested.

Design and considerations

RNA is difficult to express in vivo as cells have RNases that will degrade it. This is important for the design of RNA constructs for in vivo usage. The RNA can be protected from RNases utilizing tRNA scaffolds as camouflage. The anticodon stem in the tRNA is replaced by the desired RNA insert via restriction sites as seen in Figure 13.

Figure 13: The tRNA scaffold is marked inside the box. The scaffold include the acceptor stem, the D-loop and the TΨC-loop, but not the anticodon. The anticodon is replaced with the desired RNA insert. Fiure adopted from Ponchon et al.⁴

To achieve transcription of the RNA inside cells, DNA plasmids containing the DNA template were transformed into E. coli cells by electroporation. Once inside the cell, the constructs can take advantage of the host cell's transcription and processing machinery. The tRNA scaffold furthermore ensures transcription of the correct length of RNA, which usually is a problem with recombinant RNA expressed in vivo.

A plasmid containing the original Spinach-SAM construct was kindly provided by the Jaffrey lab, at Cornell University. To include a positive control for the experiments a plasmid containing only Spinach was produced. The DNA fragment containing restriction sites was made by assembly PCR using an ultramer (an ultramer a DNA strand of up to 200 bps) and two shorter oligonucleotides. This is shown in Figure 14.

Figure 14: A 1 % agarose gel (11 V/cm) analysis of assembly PCR and restrictions digestion. Lane 1; ladder. Lane 2; the PCR product cleaved with both BglII and XhoI. Lane 3; the uncleaved PCR product (272 bp). Lane 4; the ultramer before PCR.

The DNA was digested with restriction enzymes and ligated into the plasmid. After transformation into E. coli and culturing the plasmid DNA was isolated. From restriction digestion and sequencing data one out of three colonies confirmed the presence of the correct insert.

Confocal fluoroscence microscopy

Three different constructs were tested in E. coli; Spinach, Spinach-SAM and an untransformed control. SAM is endogenously present in the cell. It can be artificially supplemented by adding methionine to the media. The results are shown in Figure 15 and Figure 16. As expected only a small amount of fluorescense was observed for the negative control containing no inserts. The positive Spinach control showed a high level of fluorescence, which was 2.7 times larger than the negative control. When adding methionine, despite our expectations the fluorescence decreased. The Spinach-SAM construct in the presence of DFHBI showed an increase in fluorescence of 1.42. When adding additional methionine the fluorescence decreased again. In the in vitro experiments it was observed that SAM concentrations above a certain treshold had a quenching effect on the fluorescence. It was hypothesized that the reduction in fluorescence might be the result of the same effect.

Figure 15: Fluorescent confocal microscope images of constructs inside E.coli. Upper images) Untransformed negative control (m = 9191.89, n = 9). Middle images) Cells transformed with Spinach in the pressence of DFHBI (m = 24536.92, n = 50)(left), in the pressence of DFHBI and methionine (m = 18462.81, n = 32)(right) (m = 24536.92, n = 50). Lower images) Cells transformed with Spinach-SAM in the pressence of DFHBI (m = 12378.19, n = 16)(lef), in the pressence of DFHBI and methionine (m =10446.45, n = 11)(middle), and without DFHBI and methionine (m =10446.45, n = 11) (left).
m = mean fluorescent intensity per cell, n = cell count

Figure 16: Plot of the mean fluorescent intensity of the Spinach constructs in vivo. The standard deviations are indicated along with the p-value. * indicates that p < 0.0001. There is significant difference between all the samples despite Spinach-SAM w/o DFHBI and the negative control. This is expected.

Paige, J. S.; Wu, K. Y.; and Jaffrey, S. R. (2011) RNA mimics of green fluorescent protein. Science, 333, 642.
Strack, R. L.; Disney, M. D.; and Jaffrey, S. R. (2013) A superfolding Spinach2 reveals the dynamic nature of trinucleotide repeat-containing RNA. Nat. Methods, 10, 1219−1224
Paige, J. S.; Nguyen-Duc, T.; Song, W.; Jaffrey, S. R. (2012) Fluorescence Imaging of Cellular Metabolites with RNA. Science, 335, 1194

Ponchon, L; Dardel, F; (2007) Nat. Methods, 4, 571-576

Vallee-Belisle, A.; Ricci, F.; and Plaxco, K.W. (2012). Engineering biosensors with extended, narrowed, or arbitrarily edited dynamic range. J. Am. Chem. Soc. 134, 2876-2879.

Dolgosheina, E. V; Jeng, S. C. Y.; Panchapakesan, S. S. S.; Cojocaru, R.; Chen, P. S. K.; Wilson, P. D.; Hawkins, N.; Wiggins, P. A.; Unrau, P. J. ACS Chem. Biol., 2014, 9, pp 2412–2420