Team:TUDelft/Design

Team

Team Members Collaborations Attributions Sponsors

Project

Project Description Design Engineering Success Results Hardware Contribution Sequencing analysis Partnership Sustainable Development

Lab

Protocols Notebook Parts

Model

Model

Human Practices

Human Practices Integrated Human Practices Education & Public Engagement Proposed Implementation Safety Entrepreneurship

Achievements

Judging Form Medal Criteria Awards and Nominations

Design

The underlying design of AptaVita is based on an in vitro evolution experiment to engineer vitamin-binding aptazymes. These aptazymes are integrated into a genetic circuit containing a reporter gene, which is expressed in a cell-free system. Our designed dedicated hardware enables the interpretation of read-out results.

AptaVita

We aim to develop a quantitative, accessible, and modular biosensor that is able to detect vitamin deficiencies from blood samples to contribute to the fight against hidden hunger. We designed AptaVita through an iterative process based on literature, wet-lab experiments, a dry lab model, and interviews with stakeholders. In the following section, we cover the three main design principles of our project: (i) the engineering of vitamin aptazymes through an in vitro directed evolution method and the technical explanation of this method, (ii) the integration of the evolved aptazymes in a genetic circuit to regulate reporter gene expression using a cell-free system on paper support, and (iii) a dedicated hardware for quantitative and robust read-out of the system (Fig. 1).

Fig. 1 Design modules of AptaVita. From left to right: engineering of vitamin aptazymes through in vitro evolution, integration of aptazymes in a genetic circuit to regulate reporter gene expression using a cell-free system on paper support, and quantitative read-out of the system using dedicated hardware.

DRIVER: De novo Rapid In Vitro Evolution of RNA biosensors

SELEX (systematic evolution of ligands by exponential enrichment) has been the main method used to develop aptamer-based biosensors [1]. However, this technique relies on chemical modifications of the ligand to bind to a stationary phase, which implies several disadvantages: (i) the ligand modifications are specific for every molecule, (ii) the required modification may result in the generation of aptamers that do not bind to the unmodified molecule, (iii) it does not allow for selection against complex ligand mixtures. Other selection methods based on the integration of the aptamer within a ribozyme (aptazyme) have also been developed [2,3]. In those cases, the binding of the molecule modulates cleavage of the aptazyme. However, separation of the cleaved and uncleaved aptazyme fractions depends upon polyacrylamide gel electrophoresis (PAGE), which introduces a manual and time-consuming step in a process where a high number of iterative rounds of selection is needed. In addition, high Mg²⁺ concentrations are required to avoid self-cleavage of the aptazyme during gel separation, which hinders the selection of biosensors that work at biologically relevant low-Mg²⁺ levels [4]. In this project, we used DRIVER, a newly developed method for rapid in vitro evolution of RNA biosensors [4]. In DRIVER, selection rounds can be performed:

• Fully in solution
• Without modification of ligand(s)
• Using biologically relevant Mg²⁺ concentrations
• Targeting a single ligand or ligand mixture

Under these conditions, limitations of the existing methods are addressed. Moreover, although not employed in this project, DRIVER allows for a fully automatable selection using liquid handling robots.

Within our project, we intended to discover aptazymes that bind to specific vitamins, which could be utilized to determine vitamin concentrations using an aptazyme-regulated genetic circuit. In the following sections, we explain how we searched for these vitamin-specific aptazymes by using the DRIVER method.

Directed-evolution principle

Starting with a high-diversity library of potential biosensors, DRIVER makes use of a regeneration strategy within a directed-evolution procedure. This grants the selection of sequences that present a higher level of cleavage in the absence of ligand and a lower level of cleavage in the presence of ligand. During rounds in which the ligand(s) are present, the uncleaved RNA molecules preserve an unmodified 5’ prefix sequence from the previous round. These sequences are PCR amplified using this prefix and a specific suffix as primers. During rounds in which the target ligand(s) are absent, RNA molecules that undergo self-cleavage are regenerated by addition of a different 5’ prefix. Afterwards, these sequences are PCR amplified using the newly added prefix and the specific suffix as primers. By iteratively applying this process, each round selectively enriches the desired subset of sequences from the library. Eventually, RNA sequences whose cleavage is regulated by the target ligand(s) dominate the population.

A simplified animation of the directed-evolution selection principle can be found below:

A more detailed explanation of the library design, vitamin mixtures, and the different steps in DRIVER can be found below.

Library design

The DRIVER method starts with a high-diversity library of approximately 10¹² potential biosensors. For the design of this library, the sequence of the sTRSV hammerhead ribozyme was taken as a template. The two loops of the sTRSV hammerhead ribozyme are randomized and increased in length. It is expected that the binding of the ligand to specific sequences within the large loop, affects the interactions with the smaller loop, thereby hindering self-cleavage [4].

In this project, two different library designs were chosen and tested. The first library consists of sequences with a large loop of 30 random nucleotides (Fig. 2). This choice was motivated by the fact that all but one of the aptazymes found by Townshend et al. had a large loop of 30 nucleotides [4]. The second library consists of sequences with a large loop of 60 random nucleotides as our selected vitamins are larger than the ligands Townshend et al. selected.

Fig. 2 Secondary structure of the RNA library. A large loop with 30 random nucleotides (N30) and a small loop with 6 random nucleotides (N6) are shown. The cleavage site is indicated with a red arrow.

Ligand(s): B vitamins

DRIVER can be performed with either a single ligand or a ligand mixture. We decided to run different selection rounds with a sample containing a folate (vitamin B9) solution and a sample containing a mixture of vitamins. Based on buffer compatibility and vitamin deficiency prevalence, we chose the following water soluble vitamins: thiamine pyrophosphate (B1), riboflavin (B2), pyridoxal 5′-phosphate (B6), and cobalamin (B12). During ligand selection rounds, all vitamins but vitamin B2, were added at a final concentration of 100 µM. Vitamin B2 was added at a final concentration of 43.2 µM due to its lower solubility.

Considering the library designs and ligand mixtures, a total of four samples were run in parallel. Two extra controls were run together with these samples: a positive control with a known theophylline biosensor sequence and a negative control without library.

DRIVER: Step-by-step

In this section, the steps that constitute a DRIVER round will be explained in more detail.

Transcription

Each round begins with a T7 transcription reaction of the original library or the PCR product of the prior round. RNA sequences that can act as biosensors co-transcriptionally self-cleave in the absence of ligands, losing the prefix with which the round starts. On the other hand, RNA sequences that can act as biosensors remain uncleaved when the ligand(s) are present, thereby maintaining the prefix.

In this animation, W is the starting prefix. This prefix is lost when there is self-cleavage (left), and remains when there is no cleavage (right).

During transcription, the use of an excess of rNTPs over standard T7 transcription conditions results in chelation of most of the free Mg²⁺. This allows for selection conditions that are more representative of biologically relevant Mg²⁺ concentrations. Additionally, we carried out transcription at 37 °C, favoring selection of aptazymes that work under this temperature. This temperature is also the target working temperature of our test, where we integrated the evolved aptazymes.

Reverse transcription (RT)

Following transcription, the resulting RNA pool is used as a template for reverse transcription to create the complementary DNA (cDNA). In negative selection rounds (also called uncleavage rounds in our project), the RT primer is a reverse complement to the constant part located at the 3’ end of the RNA. For cleavage selection rounds, the RT primer contains an additional sequence to introduce a different 5’ prefix. The same primer guides the repair of the cleaved molecules in a subsequent ligation step.

In this animation, RNA molecules are reverse transcribed. For cleavage selection rounds, without ligand (left), the RT primer introduces a different prefix (Z prefix) to the starting prefix (W prefix). For uncleavage selection rounds, there is no change of prefix.

Ligation and USER digestion

In cleavage selection rounds, the RT primer partly hybridizes to the nascent cDNA. This way, a partially self-annealing double-stranded hairpin is formed. This structure enhances ligation and therefore circularization of the cDNA. It is important to note that the RT primer should be 5′-phosphorylated for ligation to occur.

The RT primer, additionally contains two uracil bases. A Uracil-Specific Excision Reagent (USER) is then used to cut the cDNA at these two locations. Hence, a linear DNA product with the newly incorporated prefix is released.

In this animation, the cleaved aptazymes are regenerated by circularization of the cDNA to integrate the new prefix. Thereafter, the cDNA is digested to render a linear fragment for the next step. For uncleavage selection rounds, this regeneration step is not required as only uncleaved sequences are to be amplified.

PCR

In both types of selection rounds, with and without ligand, two distinct populations of DNA molecules are present: those corresponding to RNA molecules that did not cleave, containing an unmodified prefix, and those corresponding to RNA molecules that cleaved, containing a new prefix.

As can be observed in the animation below, in selection rounds without ligand(s), the molecules with the newly introduced prefix are to be amplified, corresponding to RNA molecules that underwent self-cleavage. Primers that anneal with the Z prefix and X suffix are used as forward and reverse primers, respectively. In ligand selection rounds, the cDNA corresponding to the uncleaved population is selectively amplified. Primers that anneal with the W prefix and X suffix are used as forward and reverse primers, respectively. Both forward primers include a T7 promoter tail so that the PCR product can be used for transcription in the upcoming round.

The use of the taq error-prone polymerase (error rate 8.0 × 10^-6 mutation frequency/bp/duplication [5]) to amplify the library every round, further increases the diversity of sequences due to mutations throughout the process.

The DRIVER protocol can be found in our Protocol page. Moreover, we created a DRIVER troubleshooting, tips, and tricks guide, which can be found in our Contribution page.

Biosensor hit identification through next-generation sequencing (NGS)

After numerous rounds of selection, DRIVER-evolved libraries are screened to identify sequences that undergo self-cleavage in the absence of the ligand and remain uncleaved in its presence. To do so, a high-throughput assay based on NGS is performed.

In this assay, both the cleaved and uncleaved populations are amplified. After NGS, the relative abundance of cleaved and uncleaved fractions for each distinct sequence in the evolved library is computed. By performing the assay in the presence and absence of the ligand(s), the impact of the ligand(s) on the cleavage fraction can be assessed. Hence, biosensors can be identified and consequently be synthesized and tested out in the lab (Fig. 3).

Fig. 3 Biosensor hit identification scheme (1) Sample preparation: using the same PCR product, two independent rounds are performed, without ligand (top) and with ligand (bottom). (2) NGS: samples are pooled and sequenced with a next-generation sequencing approach. (3) Data analysis: the relative abundance of cleaved and uncleaved fractions for each distinct sequence in the evolved library is computed. Changes in the cleaved:uncleaved ratio in the presence of the ligand indicate that there is a biosensor present.

The library preparation for sequencing can be found at our Protocol page. A more detailed explanation of how the NGS data should be analyzed can be found at NGS pipeline page.

Aptazyme-regulated reporter gene expression using a cell-free system on paper support

Following the cleavage principle explained in the DRIVER: De novo Rapid In Vitro Evolution of RNA biosensors section, an aptazyme-based genetic circuit was designed to regulate the expression of a reporter gene. This expression is facilitated by a cell-free system freeze-dried on a paper support. As proof of concept, we incorporated a known theophylline binding aptazyme [4] in our genetic construct, as the search for vitamin-specific aptazymes through DRIVER was conducted in parallel to the aptazyme-regulated expression experiments.

Genetic circuit design

We designed a genetic circuit where the reporter gene expression depends on the accessibility of the ribosomal binding site (RBS) for translation initiation. This design choice was inspired by the study of Klauser & Hartig in which the exposure of the RBS for translation initiation was controlled using small RNA switches [6]. In our design, the RBS of the reporter gene is also sequestered by an antisense helix. As shown in Fig. 4, the aptazyme sequence is located between the anti-RBS sequence and the RBS. The presence of the aptazyme allows for sequestration or releasing of the RBS in response to the presence or absence of the target ligand.

Fig. 4 Secondary structure of the aptazyme in-between the RBS and an anti-RBS sequence. Upon self-cleavage of the aptazyme, detachment of the antisense strand is expected, liberating the RBS. The cleavage site is indicated with a red arrow.

After transcription of the DNA template comprising the aptazyme and fused reporter gene, the RBS is hidden in the stem of the aptazyme. When the aptazyme is stabilized upon binding of the ligand, the RBS remains entirely sequestered by its antisense strand, and translation is mostly prohibited. Cleavage of the aptazyme in the absence of the ligand liberates the RBS (Fig. 5). As a result, the ribosome can bind to the RBS and translate the downstream reporter gene.

Fig. 5 (A) Aptazyme-regulated gene expression mechanism. Binding of the ligand renders a catalytically inactive aptazyme, the RBS remains entirely sequestered by its antisense strand, repressing translation (left). In the absence of the ligand, self-cleavage of the aptazyme frees the RBS, resulting in the binding of the small ribosomal subunit and initiation of translation (right). (B) Predicted tertiary structure of the uncleaved (left) and cleaved (right) forms of the aptazyme. The RBS is shaded in pink.

As a reporter gene for having a quantitative colorimetric output, we chose lacZ, which encodes for the β-galactosidase enzyme. β-galactosidase produces the chlorophenol red (CPR) indicator upon cleavage of the yellow substrate chlorophenol red-β-D-galactopyranoside (CPRG). CPR is visible to the naked eye and can be measured with our dedicated hardware device by monitoring absorbance. Accordingly, when the ligand is present and the RBS is hidden preventing translation initiation, no color change can be observed. On the contrary, the absence of ligand results in expression of β-galactosidase followed by the colorimetric substrate conversion of CPRG (yellow) to CPR (red).

Improved genetic circuit design

After testing the above-mentioned design, we observed low expression levels. We hypothesized that even if the aptazyme cleaves, the antisense fragment remains attached to the RBS due to the high amount of complementary bases, therefore hindering translation (Fig. 6).

Fig. 6 Schematic representation of the hypothesized scenarios after aptazyme self-cleavage. In the first case, the antisense fragment detaches upon self-cleavage of the aptazyme, liberating the RBS. In the second case, the antisense fragment stays attached to the complementary part of the aptazyme including the RBS. Therefore translation is hindered even when the aptazyme is cleaved.

To further improve the expression of the genetic circuit, we designed a construct in which we modified the sequence of the antisense strand. By doing so, only nine nucleotides, including four nucleotides of the RBS, are complementary to the antisense strand. This is in contrast to our previous design where eleven nucleotides are complementary to the antisense strand, covering the complete RBS (Fig. 7). By lowering the number of complementary base pairs, we expected that the detachment of the antisense strand from the RBS occurs more frequently, increasing translation.

Fig. 7 Secondary structure of the two antisense-aptazyme-RBS designs.(A) Eleven nucleotides are complementary to the antisense strand, covering the complete RBS. (B) Nine nucleotides are complementary to the antisense strand, covering four nucleotides of the RBS. The RBS is shaded in magenta and the two modified nucleotides of the antisense strand are pointed with a black arrow.

Freeze-dried paper-based cell-free expression platform for diagnosis

Our goal is to develop a cold-chain independent vitamin detection kit that can be stored and distributed at room temperature. We aimed to accomplish this by freeze-drying an E. coli based PURE [7] cell-free system onto paper discs. This way, reactions can be assembled and freeze-dried centrally and hydrated with the serum of the patient at the point of need [8,9]. Furthermore, as a result of using a cell-free expression system, where only the necessary machinery for transcription and translation is present: (i) the risk of releasing genetically modified organisms into the environment is eliminated, (ii) a minimum amount of reagents is needed, and (iii) cell wall- impermeable molecules can be detected [8,9].

In order to assemble our genetic circuit, three components, namely (i) the plasmid containing the aptazyme and reporter gene construct, (ii) the PURE cell-free system, and (iii) the yellow substrate CPRG, are embedded and freeze-dried onto paper discs to create a stable and portable expression platform.

Upon rehydration of the paper disc, transcription of the aptazyme and lacZ occurs. The aptazyme-regulated expression of lacZ results in conversion of CPRG (yellow) to CPR (red) depending on the ligand concentration.

By measuring the absorbance through our dedicated hardware, the rate of conversion is related to the vitamin concentrations. For the experimental results of this section, visit our Results page.

In our envisioned test, the vitamin specific aptazymes found with DRIVER will be fused to the lacZ reporter gene. As a positive control to confirm that our vitamin diagnostic test is working, we will use the genetic construct containing the theophylline aptazyme. Upon rehydration with the patient serum, a color change from yellow to red will be observed in the well containing the theophylline construct as this compound cannot be found in human blood.

Other design alternatives were also considered and even tested experimentally. These alternatives include the use of:

• an alternative reporter gene.
• an eukaryotic cell-free system, with the corresponding genetic construct design.
• a microfluidic chip to detect multiple vitamins in a single rehydration step.

For more information on these alternative design considerations check the collapsible below.

Alternative design considerations

Reporter gene alternative: xylE

An alternative reporter gene that we looked into is xylE, which encodes for the catechol-2,3-deoxygenase protein. Similar to lacZ, the expression of xylE results in a colorimetric change, as the colorless substrate pyrocatechol is converted into a yellow 2-hydroxymuconic semialdehyde product [10]. Due to the toxic properties of the pyrocatechol substrate, we decided to continue the project with lacZ as a reporter gene [11].

Cell-free system alternative: Eukaryotic expression system

Townshend et al. [4], applied DRIVER evolved aptazymes to regulate gene expression in vivo using yeast. Inspired by their work, we first set up an in vitro eukaryotic design to regulate expression of lacZ. We chose an in vitro approach as using a cell-free system gives advantages to our project by making our rapid diagnostic test accessible without the use of genetically modified organisms.

We initially chose a wheat germ based cell-free system, which has been proven to show the highest translation efficiency among eukaryotic cell-free protein expression approaches [12]. In this design, the aptazyme sequence was added to the 3’ untranslated region, after the stop codon of the lacZ gene and before the start of the barley yellow dwarf virus 3' untranslated region (BYDV 3’UTR), a known translation enhancer sequence [13]. In this design, cleavage of the aptazyme means that the BYDV 3’UTR enhancer is detached from the mRNA, hindering translation, whereas a stabilized aptazyme results in enhancement of translation upon binding of the ligand (Fig. 8).

Fig. 8 Schematic representation of the eukaryotic aptazyme-regulated expression mechanism. The aptazyme is located between the stop codon of the lacZ gene and the start of the BYDV 3’UTR. Upon binding of the ligand, the aptazyme is stabilized preventing self-cleavage. This way, the BYDV 3’UTR enhancer sequence is preserved, resulting in higher levels of expression (top). Cleavage of the aptazyme means that the BYDV 3’UTR enhancer is detached from the mRNA, resulting in lower levels of expression (bottom).

lacZ was used as the reporter gene, however, background β-galactosidase activity in the wheat germ based cell-free system precluded the use of lacZ with this system. This guided our decision to proceed with a prokaryotic system design. More about our decision making can be found on the Engineering Success page.

Paper-based microfluidic chip

To improve the user-friendliness of the biosensor, we were interested in detecting multiple vitamins by a single rehydration step of the paper disks. For this purpose, we looked into using a microfluidic approach, which has been reported to be a successful strategy by Ghosh et al. [14]. The microfluidics emerges from printing patterns on paper, in our case consisting of multiple wells connected through channels. Subsequently, the paper is heated to melt the ink, which impregnates the paper to create a hydrophobic barrier for liquid flow. We envision a paper chip with a single sample entry point from which the liquid flows to each well containing the biosensor machinery for vitamin detection (Fig. 9). We performed some pilot experiments in which we tested the liquid barrier capacities of the hydrophobic ink. To reach the goal of a single rehydration step in our biosensor, further research on microfluidics regarding flow speed of the liquid, optimal shape, channel thickness, channel length, and patterns has to be done. For the results on microfluidics, visit our Results page.

Fig. 9 Visualization of potential lay-out of the microfluidic chip. Black in the picture represents the ink and red arrow points at the sample entry point. From this entry point the patient sample is able to flow through the channels to the wells containing the biocomponents by microfluidic forces.

Hardware Design

The final step in our design is coupling the colorimetric read-out to the target vitamin concentration resulting from the aptazyme-regulated reporter gene expression. We designed hardware that facilitates the read-out of the paper disc supported expression within a closed box, kept at a constant 37 °C. The box covers the cell-free embedded discs from light pollution. The paper disc is mounted on a test cassette into a small compartment of the box (Fig. 10). The box contains LEDs that emit light at a range of 550-600 nm with a peak at 574 nm, corresponding to the CPR absorbance peak that is read by light sensors inside the hardware. This absorbance is coupled to the vitamin concentration via calibration curves. The test results of the concentration are displayed on top of the box on a user interface. For more information, visit our Hardware page.

Fig. 10 Visualization of our dedicated hardware. Loading of the cassette into the hardware is indicated with the red arrow.

References

1. Ellington, A. D. and Szostak, J. W. (1990). In vitro selection of RNA molecules that bind specific ligands. Nature, 346, 818-822.
2. Koizumi, M., Soukup, G. A., Kerr, J. N. and Breaker, R. R. (1999). Allosteric selection of ribozymes that respond to the second messengers cGMP and cAMP. Nat. Struct. Biol., 6, 1062-1071.
3. Piganeau, N., Jenne, A., Thuillier, V. and Famulok, M. (2001). An allosteric ribozyme regulated by doxycyline. Angew. Chem. Int. Ed. Engl., 40, 3503.
4. Townshend, B., Xiang, J. S., Manzanarez, G., Hayden, E. J. and Smolke, C. (2021). A multiplexed, automated evolution pipeline enables scalable discovery and characterization of biosensors. Nat Commun, 12, 1437.
5. Cline, J., Braman, J. C., & Hogrefe, H. H. (1996). PCR fidelity of pfu DNA polymerase and other thermostable DNA polymerases. Nucleic acids research, 24(18), 3546–3551. https://doi.org/10.1093/nar/24.18.3546
6. Klauser, B., & Hartig, J. S. (2013). An engineered small RNA-mediated genetic switch based on a ribozyme expression platform. Nucleic acids research, 41(10), 5542-5552.
7. Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., Yokogawa, T., Nishikawa, K., & Ueda, T. (2001). Cell-free translation reconstituted with purified components. Nature biotechnology, 19(8), 751-755.
8. Silverman, A. D., Karim, A. S., & Jewett, M. C. (2020). Cell-free gene expression: an expanded repertoire of applications. Nature Reviews Genetics, 21(3), 151-170.
9. Pardee, K., Green, A., Ferrante, T., Cameron, D., DaleyKeyser, A., Yin, P., & Collins, J. (2014). Paper-Based Synthetic Gene Networks. Cell, 159(4), 940–954. https://doi.org/10.1016/j.cell.2014.10.004
10. Lin, X., Li, Y., Li, Z., Hua, R., Xing, Y., & Lu, Y. (2020). Portable environment-signal detection biosensors with cell-free synthetic biosystems. RSC Advances, 10(64), 39261-39265.
11. United States Environmental Protection Agency. Catechol (Pyrocatechol) (2000). Retrieved july 10, 2021, from https://www.epa.gov/sites/default/files/2016-09/documents/catechol-pyrocatechol.pdf
12. Harbers, M. (2014). Wheat germ systems for cell‐free protein expression. FEBS letters, 588(17), 2762-2773.
13. Wang, S., Browning, K. S., & Miller, W. A. (1997). A viral sequence in the 3′‐untranslated region mimics a 5′ cap in facilitating translation of uncapped mRNA. The EMBO journal, 16(13), 4107-4116.
14. Ghosh, R., Gopalakrishnan, S., Savitha, R., Renganathan, T., & Pushpavanam, S. (2019). Fabrication of laser printed microfluidic paper-based analytical devices (LP-µPADs) for point-of-care applications. Scientific reports, 9(1), 1-11.

A big thank you to our sponsors!