Detection of Cystic Fibrosis Transmembrane Conductance Regulator ∆F508 Gene Mutation Using a Paper-Based Nucleic Acid Hybridization Assay and a Smartphone Camera Karan Malhotraa, M. Omair Noora, and Ulrich J. Krulla*

AbstractDiagnostic technology that makes use of paper platforms in conjunction with the

ubiquitous availability of digital cameras in cellular telephones and personal assistive devices

offers opportunities for development of bioassays that are cost effective and widely distributed.

Assays that operate effectively in aqueous solution require further development for

implementation in paper substrates, overcoming issues associated with surface interactions on

a matrix that offers a large surface-to-volume ratio and constraints on convective mixing. This

report presents and compares two related methods for determination of oligonucleotides that

serve as indicators of Cystic Fibrosis, differentiating between the normal wild-type sequence,

and a mutant-type sequence that has a 3-base replacement. The transduction strategy operates

by selective hybridization of oligonucleotide probes that are conjugated to fluorescent quantum

dots, where hybridization of target sequences causes a molecular fluorophore to approach the

quantum dot and become emissive through fluorescence resonance energy transfer. Detection

can rely on hybridization of target that is labelled with Cy3 fluorophore, or in the presence of

unlabelled target when a sandwich assay format is implemented with a labelled reporter

oligonucleotide. Selectivity to determine the presence of mismatched sequences involves

appropriate selection of nucleotide sequences to set melt temperatures, in conjunction with

control of stringency conditions using formamide as a chaotrope. It was determined that both

direct and sandwich assays on paper substrates are possible to distinguish between wild-type

and mutant-type samples.

Keywords: Fluorescence; quantum dots; oligonucleotides; cystic fibrosis; paper; digital imaging

1

a University of Toronto Mississauga, Department of Chemical and Physical Sciences, 3359 Mississauga Road North, L5L 1C6, Canada, * Corresponding Author, ESI available online

1.IntroductionCystic fibrosis (CF) is a multi-system, fatal autosomal recessive disorder that is

characterized by viscous secretions in the lungs of patients due to mutations in cystic fibrosis

transmembrane conductance regulator protein (CFTR). CF affects 1 in 3000 births with ~70,000

people affected worldwide.1–5 Over 1500 mutations for the CFTR protein have been found but

few are common and fewer result in the disease. Of the small number of mutations,

responsible for the disease state, the deletion of phenylalanine at the 508 position (∆F508) is

responsible for over two-thirds of the cases, while all other mutations account for no more than

5% of the cases individually.2,5,6 Development of sensing technology for detection of ∆F508

would serve to first enable improved screening for identification of heterozygous gene carriers

in untested adults, and tosecond assist clinicians to screen for gene mutations related to CF in

newborns. The current strategies for diagnosing CF are based on newborn screening (NBS)

programs that work via screening for serum markers including the immunoreactive trypsinogen

(IRT) assay.7–9 This assay is typically followed by diagnosis of the genetic basis of disease,

including detection of ∆F508 and related mutations, based on determining the presence of

specific oligonucleotide sequences. Finally, a sweat chloride test is performed to diagnose

patients with CF. All of these techniques require skilled technicians to process samples,

perform, and analyse tests using resource-intensive technologies.10 The aim of theis work

presented herein is to focus on the translation ofexamine a potential point-of-care (POC)

technology that may offer for a low cost, easy to use and portable method for sensing CFTR

2

∆F508 gene mutations, ifbeginning with a focus on a suitable performance from the

transduction strategy can be achieved.

Advances in bioassays and sensing technologies for point-of-care (POC) or resource-

limited settings are guided by recommendations of the World Health Organization, which

indicate that systems must be affordable, sensitive, specific, user-friendly, rapid and robust,

equipment free and deliverable to those who need them (ASSURED).11,12 6,7 Paper is a suitable

substrate for development of analytical devices with microfluidic capabilities (e.g., microfluidic

paper-based analytical devices, PADs), and has been implemented as a platform for screening

and semi-quantitative assays of biomarkers, offering reliable performance at low cost, with

ease of use and disposal.12-14 7–9 As an emerging technology for POC application, PADs are

uniquely poised to function as systems that can process raw samples and then complete an

analysis to yield information regarding the genetic basis of disease.15 0,16 1 Research within the

PAD field has often focused on individual functional components of a complete device

including sample preparation17 2 (i.e. extraction of analytes from complex samples),

amplification of analytes of interest,18 3,19 4 and detection commonly using electrochemical20 15,2 16

or optical (i.e. colorimetric or fluorimetric) techniques.22 17,23 18 For portable or in-field

applications, the preference is isothermal enzymatic amplification, yielding products that are

either labelled or unlabelled with dyes depending on the detection scheme and the desired

analytical figures of merit.24 19–26 1 It is clear that sample processing and gene fragment

amplification can be achieved on paper substrates25 0, providing product for the transduction

step., which is the focus of the work in this investigation.

3

Fluorescence is a high-sensitivity method among oligonucleotide-based detection

strategies.27 2 Labelling of oligonucleotides can be accomplished during the amplification step via

the integration of fluorescently labelled deoxynuclotides,27 2 but is not necessary or desired in

some applications. The performance of fluorescence-based systems can be further improved by

integrating luminescent nanomaterials and adopting a resonance energy transfer (RET) strategy

for application in PADs.28 3 A representation of two analysis formats (based on labelled and

unlabelled amplified oligonucleotide), and immobilization of the sensing platform on paper is

presented in Figure 1.Figure 1 as the basis for the methodology proposed in the work herein.

Figure 1. “Direct” and “Sandwich” FRET assay formats for gQD-probe conjugates. Probe strands (seen in blue) are immobilized on the surface of green-emitting QDs (gQDs) via dithiolane (DTPA) functionality. (A) In direct assay format, the target strand (seen in red) is modified with Cy3 on the proximal end and upon hybridization, is brought to proximity with gQDs allowing for FRET to take place. (B) In sandwich assay format, the probe strand

4

hybridizes with the target strand (seen in red) such that there is an overhang on the distal end. Reporter strand (seen in green) hybridizes with the overhang region of the target strand bringing to proximity the Cy3 label on the proximal end of the reporter.

Quantum dots (QD) have broad absorption wavelengths, narrow and symmetrical

emission photoluminescence (PL) profiles, high quantum yields, and photo/chemical stability.

QDs can serve as efficient donors for RET processes. Integration of biomolecules to surfaces of

QDs is advantageous for a RET based system because the relatively large surface area of QDs

allows for multiple biomolecules to be conjugated. Thus, multiple pathways of RET can take

place, which collectively improves energy transfer efficiency and increases the optical signal.29 4

Of note for signal reproducibility is that a ratiometric data processing approach where acceptor

and QD donor emission are tracked together results in greater precision for transduction of

biological interactions.28 3,30 25

While methods that make use of quantum dotsQDs as FRET donors have been

previously described as a basis for interest in paper-based hybridization bioassays, earlier

investigations have rarely considered the practical challenges required for applications. One

important question is whether sufficient stringency control can be achieved in paper-based

substrates given that such an environment can be different than that experienced when

hybridization takes place in bulk solution. Investigations of the detection of oligonucleotides in a

paper matrix have primarily focused on the hybridization of fully complementary hybrids in the

presence of non-complementary oligonucleotide.22 17,31 26–33 28 The results of these investigations

suggest potential for detection of a three-base pair deletion using a paper-based format to

detect the presence of CFTR ∆F508. The work described herein determined that a paper

5

substrate can serve as a platform for a ratiometric bioassay for detection of nucleic acids using

QDs as RET donors. Green quantum dots (gQDs) and Cy3 dye labelled oligonucleotides were

chosen as the RET pair. Hybridization of complementary strands of oligonucleotides resulted in

proximity of the RET donor and acceptor allowing for the near-field phenomenon to alter the PL

of the FRET pair. Stringency was controlled by addition of formamide to tune selectivity for

wild-type (WT) and mutant-type (MT) targets. Hybridization was conducted in both direct and

sandwich formats with the intention of comparison of analytical performance to guide the

subsequent development of an amplification format in the future. Smartphone imaging was

used to collect PL data. A schematic detailing the operation of the paper-based assay is

presented as Figure 2.Figure 2.

6

Figure 2. Schematics detailing the bioassay of CFTR related genetic mutations. (A) Reaction zones consisted of chemically modified paper that were conjugated with gQD-oligonucleotide probes. Zones contained WT and MT controls, and test zones where unknown samples were spotted and imaged. Detection was based on the principle of RET, with gQDs used as donors and Cy3 labels on oligonucleotide strands as acceptors. (B) Imaging used a smartphone camera, with data processing by ImageJ to split the image to RGB color channels.

2.Experimental MethodsA detailed list of reagents, experimental procedures, and instrumentation can be found

in the supporting information (ESI). The oligonucleotide sequences used in the hybridization

assays are listed below in Table 1 Table 1 (ACGT Corporation, Toronto, ON, Canada; Integrated

DNA Technologies, IDT, Coralville, Iowa, U.S.A). The oligonucleotides were dissolved in

autoclaved deionized water and stored at -20°C.

7

Table 1. Oligonucleotide Sequences used in Hybridization Assays

Name SequenceCFTR WT Probe (WT DTPA) DTPA-5’-AAT ATC ATC TTT GGT GTT-3’CFTR MT Probe (MT DTPA) DTPA-5’-AAT ATC ATT GGT GTT TCC-3’CFTR WT Cy3 TGT Cy3-3’-TTA TAG TAG AAA CCA CAA-5’CFTR MT Cy3 TGT Cy3-3’-TTA TAG TAA CCA CAA AGG-5’CFTR NC Cy3 TGT Cy3-3’-GAA TGA AGG TAC TAA AGA-5’CFTR WT TGT 3’-TTA TAG TAG AAA CCA CAA AGG ATA CTA CTT ATA TCT-5’ CFTR MT TGT 3’-TTA TAG TAA CCA CAA AGG ATA CTA CTT ATA TCT-5’CFTR NC TGT 3’- GAA TGA AGG TAC TAA AGA AAT TGA TAC GGC CCT AGG

TAGCFTR Reporter (REP Cy3) Cy3-5’-TAT GAT GAA TAT AGA-3’

TGT = target; WT = wild type; MT = mutant type; NC = non-complementary; Cy3 = Cyanine 3; DTPA = dithiol phosphoramidite; REP = reporter.

2.1 Materials

2.1.1 Preparation of QD-Probe Oligonucleotide ConjugatesOleic acid capped CdSxSe1-x/ZnS quantum dots with green PL (gQDs, peak PL at 525 nm),

in toluene were made water soluble by a ligand exchange reaction using glutathione (GSH). The

resulting glutathione capped QDs (GSH-QDs) were conjugated with DTPA modified CFTR

oligonucleotide probes (see ESI for details).

2.1.2 Solution-Phase Hybridization AssaysSolution-phase hybridization assays were conducted in direct assay format. For a typical

FRET assay, gQD-probe conjugates were mixed with 3’ Cy3 labelled oligonucleotide targets

(CFTR WT Cy3 TGT and CFTR MT Cy3 TGT) at various concentrations (6.25 – 75 pmol) in 50 mM

borate buffered saline solution (BBS, pH 9.25, 100 mM NaCl), with a reaction time of 15 minutes

prior to sample measurements (See ESI for details).

8

2.1.3 Surface Modification of Paper with Imidazole GroupsEnclosed reaction zones on Whatman cellulose chromatography paper (Grade 1) paper

substrates were prepared by wax printing circular boundaries with a Xerox ColorQube 8570DN

solid ink printer. The paper in the zones was subsequently oxidized to obtain aldehyde

functionalities that were further reacted via reductive amination to obtain imidazole groups on

the paper (see ESI for details regarding modification and fabrication of paper substrates).

2.1.4 Immobilization of QD-Probe Oligonucleotide Conjugates, and Solid-Phase Hybridization Assays

Hybridization assays on paper substrates were conducted using two formats; direct assay

and sandwich assay. For direct assay, first QD-probe conjugates were immobilized on imidazole

modified paper zones, dried in darkness and washed with 50 mM borate buffer (BB, pH 9.25)

for 5 minutes. Next, 3’ Cy3 labelled oligonucleotide targets (CFTR WT Cy3 TGT and CFTR MT Cy3

TGT) were spotted onto paper zones, dried in darkness and washed with BBS for 30 seconds

before drying and imaging. After imaging, papers were washed (if required) using BB containing

formamide for 5 or 10 minutes, and dried before further imaging. For sandwich assay, first QD-

probe conjugates were immobilized onto imidazole modified paper zones, dried in darkness

and washed with BB for 5 minutes. Next, unlabeled oligonucleotide targets (CFTR WT TGT and

CFTR MT TGT) were spotted onto paper zones and were dried in darkness. Next, Cy3 labelled

reporter (CFTR REP Cy3) was spotted and then the paper substrates were washed with BBS

wash for 30 seconds (see ESI for details).

9

2.2 Instrumentation

2.2.1 PL Spectra and Digital Image Acquisition PL spectra for hybridization assays done in solution-phase were acquired using a

QuantaMaster Photon Technology International spectrofluorimeter (London, ON, Canada). The

excitation source was a xenon arc lamp (Ushio, Cypress, CA) and the detector was a red-

sensitive R928P photomultiplier tube (PMT, Hamamatsu, Bridgewater, NJ) (See ESI for further

details). The FRET ratios for the PL spectra were calculated using Equation S3.

Digital color images for paper substrates were acquired using an iPhone SE (Apple,

Cupertino, CA, U.S.A.). A hand-held ultraviolet (UV) lamp (UVGL-58, LW/SW, 6W The Science

Company, Denver, CO, U.S.A.) operated at the long wavelength (365 nm) setting was used to

illuminate paper substrates at a distance 10 cm. Paper substrates were dried in a desiccator

before data collection (dry format). The R/G ratios (ratiometric response) from the digital

images were calculated using Equation S4.

2.2.2 Digital Imaging for Ratiometric Detection of Oligonucleotide Hybridization

Data for a ratiometric format of signal transduction requires simultaneous measurement

of intensity from two wavelength bands associated with the PL of the RET donor and acceptor.

Imaging was done to make use of the R-G-B colour palette of a smartphone camera, with donor

PL associated with the green color channel and acceptor PL was associated with the red color

channel, and dividing the average signal intensity of the red color channel with the green color

channel. Images were processed using ImageJ software to obtain the average signal intensities

10

of the R-G-B color channels in the reaction zones on the paper substrates, with the average

signal obtained via measurement of n = 4 test spots, see Equation S4. Paper zones containing

only gQDs were used as the brightest spots and served as background control. Imaging was

conducted in a dark enclosure, using dried paper which has previously been reported to offer

greater fluorescence intensity.34 29

3.Results and Discussion3.1 FRET Pair Characterization (gQD – Cy3)

The optical signal from the bioassay explored in this investigation was based on the

near-field interaction between gQDs as FRET donors and Cy3 as FRET acceptors. Förster

formalism was used to characterize the FRET pair and the Förster distance was determined to

be 4.7 nm. Detection of target sequences of interest was observed as a decrease in the PL of

the RET donor and an increase in the PL of the RET acceptor. With increasing concentration of

dye labelled target, the fluorescence from the paper zones were observed to change from

green to yellow indicating that RET was occurring (see Figure S5). Details are reported in the

ESI.

3.2 Oligonucleotide Hybridization in Solution Solution-phase assays were conducted to characterize the interaction between probe

and target oligonucleotide strands. It was found that the fully complementary (FC) hybrids,

being gQD-WT probe conjugates with WT Cy3 target strands and gQD-MT probe conjugates

with MT Cy3 target strands, were more stable thermodynamically than the partially

complementary (PC) hybrids, being gQD-WT probe conjugates with MT Cy3 target strands and

11

gQD-MT probe conjugates with WT Cy3 target strands. PC hybrids for gQD-WT probe

conjugates yielded approximately 25% of the absolute FRET intensity of FC hybrids. PC hybrids

for gQD-MT probe conjugates yielded approximately 50% of the absolute FRET intensity of the

FC hybrids (see ESI for details regarding solution-phase assays). The solution-phase experiments

provided guidance for stringency control to distinguish FC from PC targets.

3.3 Oligonucleotide Hybridization in Paper SubstratesSelectivity of base pair hybridization of DNA strands can be controlled by environmental

manipulation for detection of unique oligonucleotide strands of interest. Stringency can be

adjusted by control of the ionic strength, the pH of the hybridization solution, and by altering

the thermodynamic stability of the probe-target conjugate via addition of chaotropes like

formamide. Formamide (HCONH2) can be added to hybridization buffers for lowering

oligonucleotide stability. Chemically, it is a stronger hydrogen bond acceptor than water and

can engage with the hydrogen bond network of oligonucleotides to displace hydration of DNA.

The extent of melt temperature depression caused by addition of formamide is dependent on

factors including GC composition of the oligonucleotide strand, the helical conformation and

the state of hydration. Studies on the denaturing effects from formamide find sensitivity of AT

containing duplexes to be lower than those containing GC, perhaps due to the different

hydration pattern of AT containing oligonucleotides.35 0,36 1 Control of the melting temperature of

hybrids on paper substrates can be achieved by adjusting the ratio (v/v %) of formamide added

to washes and the amount of time that the paper undergoes the wash. A preliminary

12

investigation of the thermodynamic parameters associated with FC and PC oligonucleotide

hybrids was conducted. The nearest-neighbor method was used to determine the

thermodynamic parameters associated with the expected probe – target hybrids used in the

design of this experiment.37 2–39 4 The resulting data was used to interpret the information

produced from the FRET-based system undergoing wash conditions of various stringencies.

3.4 Direct Assay FormatThe direct assay made use of hybridization of probe strands with fluorescently labelled

targets. Imidazole modified paper substrates were used to immobilize either gQD-WT probe

conjugates or gQD-MT probe conjugates. Real samples are expected to be composed of CFTR

WT TGT strands, CFTR MT TGT strands, or a combination of the CFTR TGT strands. Thus, four

different variations of probe and target oligonucleotide conjugates were investigated as

presented in Figure 3

Error: Reference source not found. Hybrids (A) and (B) were fully complementary (FC)

and have ∆Gmax = -30 kcal mol-1 and -31 kcal mol-1 for WT probe – WT Target and MT Probe – MT

Target respectively. Hybrids (C) and (D) were partially complementary (PC) and have ∆Gmax = -14

kcal mol-1 and -14 kcal mol-1 for WT probe – MT Target and MT Probe – WT Target,

respectively.37 2–39 4 These differences in stabilities indicate that careful control of formamide

concentration may be sufficient to distinguish between WT and MT gene fragments at room

temperature.

13

A. WT Probe – WT Target(18 Complementary Base Pairs with Probe)

B. MT Probe – MT Target(18 Complementary Base Pairs with Probe)

∆Gmax = -30 kcal mol-1 ∆Gmax = -31 kcal mol-1

Theoretical Tm = 43C (316 K) Theoretical Tm = 44 C (317 K) C. WT Probe – MT Target(8 Complementary Base Pairs with Probe)

D. MT Probe – WT Target(8 Complementary Base Pairs with Probe)

∆Gmax = -14 kcal mol-1 ∆Gmax = -14 kcal mol-1

Theoretical Tm = 22 C (295 K) Theoretical Tm = 20 C (293 K)3.5

Figure 3. The various probe-target hybrids explored for the direct assay format. (A) refers to WT probe – WT target, (B) refers to MT probe – MT target, (C) refers to WT probe – MT target, (D) refers to MT probe – WT target. Thermodynamic parameters (∆Gmax and Tm) were calculated using the nearest neighbor method.32–34

14

Sandwich Assay FormatA sandwich assay strategy was based on the step-wise hybridization of probe strands

with unlabeled targets and the subsequent hybridization with a fluorescently labelled reporter

sequence. Imidazole modified paper substrates were conjugated with either gQD-WT or gQD-

MT probe systems. Clinical samples are expected to be composed of CFTR WT TGT strands,

CFTR MT TGT strands, or a combination of the CFTR TGT strands. Thus, four different variations

of probe and target oligonucleotide conjugates were investigated as summarized in Figure 4

Error: Reference source not found. In contrast to direct assay, the sandwich assay

consists of two hybridization events. Of the two hybridization events, only the first event was

expected to yield partially complementary (PC) structures while the second event was expected

to always yield fully complementary (FC) structures. For the first hybridization event, hybrids (A)

and (B) are FC and have ∆Gmax = -30 kcal mol-1 and -31 kcal mol-1 for WT probe – WT Target and

MT Probe – MT Target, respectively. Hybrids (C) and (D) are PC and have ∆Gmax = -14 kcal mol-1

and -20 kcal mol-1 for WT probe – MT Target and MT Probe – WT Target, respectively. The

thermodynamic parameters for hybrids (A) and (B) agreed with those determined for the direct

assay, and as expected were higher than the values for hybrids (C) and (D).37 2–39 4 However, the

parameters for hybrid (D) suggest that it was more stable in the sandwich assay format than

the direct assay format. Hybrid (D) in sandwich assay was predicted to form a PC conjugate with

fourteen base pairs, which was six more base pairs than all other PC conjugates (eight). Thus,

discrimination of MT Probe – WT Target and MT Probe – MT Target was predicted to require

wash conditions of greater stringency than other PC conjugates. For the second hybridization

event (target strands and reporter strands), ∆Gmax = -21 kcal mol-1 and Tm = 30C (303 K) were

15

expected. The ∆Gmax for the first hybridization events were higher than the second hybridization

event in FC conjugates. The result was that wash conditions required to achieve the mismatch

discrimination would also result in signal loss for FC conjugates because for a single paper

system, FC hybrids were washed in the same conditions as PC hybrids.

A. WT Probe – WT Target(18 Complementary Base Pairs with Probe)(FC with REP)

B. MT Probe – MT Target(18 Complementary Base Pairs with Probe)(FC with REP)

∆Gmax = -30 kcal mol-1 ∆Gmax = -31 kcal mol-1

Theoretical Tm = 43C (316 K) Theoretical Tm = 44 C (317 K) C. WT Probe – MT Target(8 Complementary Base Pairs with Probe)(FC with REP)

D. MT Probe – WT Target(14 Complementary Base Pairs with Probe)(FC with REP)

∆Gmax = -14 kcal mol-1 ∆Gmax = -20 kcal mol-1

Theoretical Tm = 22 C (295 K) Theoretical Tm = 35 C (308 K)

16

Figure 4. The various probe-target conjugates explored for the sandwich assay format. (A) refers to WT probe – WT target, (B) refers to MT probe – MT target, (C) refers to WT probe – MT target, (D) refers to MT probe – WT target. Thermodynamic parameters (∆Gmax and Tm) were calculated using the nearest neighbor method.32–34

3.6 Optimizing Wash Conditions for SelectivityPaper substrates were washed with borate buffer that contained various content of

formamide from 0 to 10 % v/v ratios for 0, 5, and 10 min for direct assays (up to 20 minutes for

sandwich assays). Following the washes and after drying, the paper substrates were imaged and

the average intensity from reaction zones was measured to calculate a quantitative ratiometric

signal (details in the ESI). A wider range of wash conditions were investigated for the sandwich

assays because the energy associated with the PC hybrid; MT probe – WT Target was larger

than other PC hybrids and could significantly shift conditions for discrimination between FC and

PC hybrids. Of the various conditions investigated, many provided for full discrimination of FC

and PC sequences including all washes done for 10 minutes (see Tables S5 to S8). The optimal

wash conditions for direct assays that provided the best resolution between FC and PC while

minimizing loss of FC hybrid for both WT and MT sequences was BB + 5 % v/v formamide

(BB+5%F) for both WT and MT sequences (presented as Figure 5Figure 5A for WT and Figure

5BFigure 5B for MT). These washing conditions were then further investigated for the analytical

figures of merit, and performance in complex sample matrices.

The optimal wash conditions for the sandwich assay reflected the stability of the PC

hybrids for gQD-MT probe, which was greater than that of all the other sequences (Figure 4

17

Error: Reference source not found). At BB+10%F, complete signal loss for FC gQD-WT

probes was noted. The optimal wash condition for sandwich assays was determined to be

BB+5%F, with WT sequences being washed for 5 minutes while MT sequences were washed for

20 minutes. This is schematically presented in Figure 5CFigure 5C for gQD-WT probe and Figure

5DFigure 5D for gQD-MT probe.

Figure 5. Determination of optimal wash conditions for direct and sandwich assay considered R/G Ratios with variation of formamide concentration for wash times of 0, 5, 10, 15, and 20 min. The optimal wash conditions for direct assay was found to be BB+10%F for 5 minutes for (A) gQD-WT probe sequence and (B) gQD-MT probe sequence. The optimal wash conditions for sandwich assay was found to be BB+5%F for 5 minutes for (C) gQD-WT probe sequence and BB+5%F for 20 minutes for (D) gQD-MT probe sequence.

3.7 Analytical Figures of Merit The performance of the bioassay was investigated in both direct and sandwich assay

formats and concentration-response curves are presented in Figure 6Figure 6. Paper substrates

were washed with BB+10%F or BB+5%F for direct and sandwich assays, respectively, with wash

18

times of 20 minutes for MT sandwich assays, and 5 minutes for direct phase assays and WT

sandwich assays. Performance of the bioassays in the low pmol range is presented as insets for

each of the respective curves. Regression analysis for the dataset was done to obtain the

analytical figures of merit, which are presented in Table 2.Table 2.

The response of the WT and MT direct assays was similar, with sensitivity (slope of

response line), LOD, LOQ and dynamic range being in close agreement. A dynamic range of two

orders of magnitude (0.3 – 15 pmol) was observed, with an LOD (at 3 sigma) of 215 and 285

fmol for WT and MT probes, respectively. This consistency in analytical performance reflects

the similar ∆G and Tm for the two FC and PC hybrids.

The sandwich assay response of WT and MT was found to vary. WT probes had double

the sensitivity in the region of linearity, a lower LOD and LOQ (by three times), and a larger

dynamic range (0.4 – 20 as compared to 1.5 – 20). These differences in analytical performance

are consistent with the thermodynamic stabilities of the various hybrids. MT probes were

required washes of higher stringency, and thus a larger proportion of the FC was lost.

Quantification of the analytical parameters was accomplished using only WT or MT

targets. However, the discrimination of targets in mixtures is also of importance.

Table 2. Analytical Performance Direct and Sandwich Bioassays

Assay Format

Probe Slope of Calibration Curve

r2 LOD LOQ Linear Range (pmol)

Direct Assay

WT 0.3145 0.9857 215 fmol 650 fmol 0.3 – 15 MT 0.3147 0.9680 285 fmol 865 fmol 0.3 – 15

Sandwich WT 0.0486 0.9934 422 fmol 1.28 pmol 0.4 – 20

19

Assay MT 0.0285 0.9779 1.45 pmol 4.38 pmol 1.5 – 20

Figure 6. Concentration-response curves showing the R/G ratiometric response of the direct and sandwich assay formats. (A) gQD-WT probe conjugates were used for determination of Cy3 labelled WT targets, and (B) gQD-MT probe conjugates were used for determination of Cy3 labelled MT targets. (C) gQD-WT probe conjugates were used for determination of unlabelled WT targets with Cy3 labelled reporters, and (D) gQD-MT probe conjugates were used for determination of unlabelled MT targets with Cy3 labelled reporters. Note that the scales are logarithmic. Insets show R/G ratiometric response at low pmol concentration range for respective curves. Each error bar represents one standard deviation for n=4 replicates.

20

3.8 Selectivity for Mixtures of WT and MT TargetsClinical samples of oligonucleotides are expected to be composed of gene sequences of

WT only, MT only (∆F508), or a mixture of WT and MT. The cross-reactivity of WT and MT

sequences must therefore be evaluated. Selectivity assays were determined in direct assay

format, and signal from digital images was measured pre- and post- formamide washing.

Samples of 2.4 pmol of CFTR Cy3 FC targets were spiked with 0 – 2.4 pmol samples of CFTR Cy3

PC targets (e.g., for WT probe, WT targets were spiked with MT targets). Selectivity assays were

also done using the sandwich assay format, and samples of 4.8 pmol of CFTR FC targets were

spiked with 0 – 4.8 pmol samples of CFTR PC targets.

It was found that for both direct and sandwich assays in pre-wash, WT and MT signals

showed linear response with increasing amount of PC targets (Figure 7Figure 7(Ai) and (Bi) for

direct assay and (Ci) and (Di) for sandwich assay, respectively). The signal showed a 25%

increase from 0 to the highest quantity of target that was added. This indicates substantial

contribution of signal from PC hybrids. Post-wash, it was found that there was no contribution

of signal from the addition of PC targets to either WT or MT assays in both direct and sandwich

assays (Figure 7Figure 7(Aii) and (Bii) for direct assay and (Cii) and (Dii) for sandwich assay,

respectively). The results indicate that suitable stringency control can obviate false positives in

mixtures of WT and MT probes.

21

Figure 7. Pre- and post-wash assays for determining the selectivity of (Ai) gQD-WT probes and (Bi) gQD-MT probes in mixtures of WT and MT targets. Selectivity was determined using background corrected R/G ratio plots for hybridization of gQD-probe conjugates with Cy3 labelled targets (for direct assay, A and B), and gQD-probe conjugates with unlabeled targets and Cy3 labelled reporter sequences (for sandwich assay, C and D). Response of the hybridization assay was determined for pre-wash (Ai and Ci) WT probe conjugates, and pre-wash (Bi and Di) MT probe conjugates. Post-wash assays yielded signal response shown in Aii and Cii for WT probe conjugates, and in Bii and Dii for MT probe conjugates. Error bars represent one standard deviation for n = 4 replicates.

3.9 Paper-based Assay Response for Complex Sample MatricesAThe POC bioassaytranslation for nucleic acidof target analysis usingon paper

substrates must berequires a sensing platform that is robust, selective, and insensitivenot

amenable to variations in the sample matrix composition. Fluctuations in the biomolecular

composition and purity of the targets in a sample may arise because of post- sample isolation,

processing, or amplification. Assays performed in solutions containing DNA fragments, proteins,

and serum must demonstrate that performance is not inhibited by the composition of the sample

matrix. A Thus, a robust sensor would experience little signal change in selectivity regardless of

22

the sample matrix and a poor sensor would experience greater signal changes in selectivity and

may fail to achieve enough resolutionwould provide for confidence in determiningfor a true

positive or negative signal.

The performances of the assays were investigated for samples that contained bovine

serum albumin (BSA, 40 mg mL-1), goat serum (GS, 85% v/v) and salmon sperm DNA (SS, 2000

bp fragments, 0.8 mg mL-1). These matrices were used to dilute either, CFTR NC Cy3 TGT, CFTR

WT Cy3 TGT, or CFTR MT Cy3 TGT at 2.4 pmol concentration for direct assay, and at 6 pmol

concentration for sandwich assay. The resulting R/G ratios from direct hybridization assays

(including BBS as a control), are shown in Figure 8Figure 8(A) and (B) for WT and MT probe

conjugates, respectively. R/G ratios from sandwich hybridization assays are shown in Figure

8Figure 8 (C) and (D) for WT and MT probes, respectively.

High selectivity was retained for all hybridization assays in both direct and sandwich

format, with the signal from NC and PC hybrids being within the experimental error. Thus, the

interfering effects of these sample matrices did not compromise the performance of either

direct or sandwich assays.

23

Figure 8. Background corrected R/G ratios for hybridization of gQD-probe conjugates in complex matrices. Direct assays used 2.4 pmol of CFTR Cy3 TGTs in either BBS, GS, SS, or BSA. Response of the hybridization assay was collected for (A) gQD-WT probe conjugates and (B) gQD-MT probe conjugates. Sandwich assays were conducted in similar manner to direct assay and using 6 pmol of CFTR Cy3 TGTs. Response of the hybridization assay was collected for (C) gQD-WT probe conjugates (D) gQD-MT probe conjugates. Error bars represent one standard deviation for n = 4 replicates.

24

4.Conclusions

Fluorescence determination in a paper substrate of a predominant genetic marker for

cystic fibrosis has been explored. This involves distinction between a mutant-type and wild-type

oligonucleotide sequence, either of which could be present individually or in mixture in clinical

samples. A FRET-based hybridization assay using QDs with green emission as donors and a Cy3

molecular fluorophore as an acceptor has provided for two assays methods. One method relied

on labelled oligonucleotide target, as commonly produced during enzyme amplification.

Another method used a sandwich assay format to avoid the need for labelling of

oligonucleotide targets. The inclusion of the direct assay format was intended to examine the

impact that design would have on the performance of the sensing platform, and was not

intended to suggest that the direct assay method was superior in practical applications.

Analytical performance was primarily based on selective melting of undesired hybrids, and

sufficient stringency control was possible to provide reliable detection of targets even in

samples that contained substantial quantities of background protein and nucleic acid. Despite

the performance differences due to thermodynamic stabilities of hybrids formed from two

oligonucleotides compared to sandwich hybrids using 3 oligonucleotides, the data indicates

that both direct and sandwich assays could be implemented to distinguish between wild-type

and mutant-type samples.

AcknowledgementsWe would like to thank Dr. Abootaleb Sedighi for his helpful discussion and assistance with

blind assays. The authors are grateful to the Natural Sciences and Engineering Research Council

of Canada for financial support of this research.

Conflict of InterestThere are no conflicts to declare.

25

References1 B. P. O’Sullivan and S.D. Freedman, Lancet, 2009, 373,1891–1904.12 G.R. Cutting, Nat. Rev. Genet., 2015, 16, 45–56.3 J.R. Riordan, et al., Science, 1989, 245, 1066–1073.24 J. M. Rommens, et al., Science, 1989, 245, 1059–1065.3 B. P. O’Sullivan and S.D. Freedman, Lancet, 2009, 373,1891–1904.5 E. Dequeker, et al., Eur. J. Hum. Genet., 2009, 17, 51–65.4 M. Hodson, A. Bush, and D. Geddes. Cystic Fibrosis, Third Edition; CRC Press, 2012.5 B. Kerem, et al., Science, 2989, 245, 1073–1080. 6 J.R. Riordan, et al., Science, 1989, 245, 1066–1073.7 C. Castellani, et al., Lancet Respir. Med., 2016, 4, 653–661.8 M. Kharrazi, et al., Pediatrics, 2015, 136, 1062–1072.9 L. F. Ross, J. Pediatr., 2008, 153, 308–313. 10 E. Dequeker, et al., Eur. J. Hum. Genet., 2009, 17, 51–65.116 R. W. Peeling, et al., Sex. Transm. Infect., 2006, 82 (Suppl 5), v1–v6.127 A. W. Martinez, et al., Anal. Chem., 2010, 82, 3–10.138 E. Carrilho, A.W. Martinez, and G. M. Whitesides, Anal. Chem., 2009, 81, 7091–7095.149 E. Carrilho, et al., Anal. Chem., 2009, 81, 5990–5998.150 S. C. Fernandes, et al., Anal. Chem., 2017, 89, 5654–5664.161 A. W. Martinez, et al., Anal. Chem., 2008, 80, 3699–3707.172 B. Y. Moghadam, K. T. Connelly, and J. D. Posner, Anal. Chem., 2014, 86, 5829–5837.183 M. O. Noor, et al. Anal. Chim. Acta., 2015, 885, 156–165.194 J. T. Connelly, J. P. Rolland, and G. M. Whitesides, Anal. Chem. 2015, 87, 7595–7601.2015 W. Dungchai, O. Chailapakul, C. S. Henry, Anal. Chem. 2009, 81, 5821–5826.2116 M. N. Tsaloglou, et al., Anal. Biochem., 2018, 543, 116–121.2217 M. O. Noor, A. Shahmuradyan, and U. J. Krull, Anal. Chem. 2013, 85, 1860–1867.2318 M. He, Z. Liu, Anal. Chem. 2013, 85, 11691–11694.2419 A. R. Connolly, and M. Trau, Angew. Chem. Int. Ed., 2010, 49, 2720–2723.250 H. Liu, et al., Anal. Chem. 2013, 85, 7941–7947.261 K. Hsieh, et al., Angew. Chem. Int. Ed. 2012, 51, 4896–4900.272 Q. Ju, M. O. Noor, and U. J. Krull, Analyst, 2016, 141, 2838–2860.283 N. Hildebrandt, et al., Chem. Rev. 2017, 117, 536–711.294 W. R. Algar, et al., Anal. Chem., 2011, 83, 8826–8837.3025 W. R. Algar, A. J. Tavares, and U. J. Krull, Anal. Chim. Acta., 2010, 673, 1–25.3126 P. B. Allen, et al., Lab. Chip., 2012, 12, 2951–2958.3227 Y. Wang, et al., Lab. Chip., 2013, 13, 3945–3955.3328 Y. Song, et al., Anal. Chem. 2014, 86, 1575–1582.3429 M. O. Noor, and U. J. Krull, Anal. Chem., 2014, 86, 10331–10339.350 R. D. Blake, and S. G. Delcourt, Nucleic Acids Res., 1996, 24, 2095–2103.361 J. Fuchs, et al., Anal. Biochem., 2010, 397, 132–134.372 J. SantaLucia, and D. Hicks, Annu. Rev. Biophys. Biomol. Struct., 2004, 33, 415–440.383 J. A. SantaLucia, and P. A. Seneviratne, Biochemistry, 1996, 35, 3555–3562.394 H. T. Allawi, J. A. SantaLucia, Biochemistry, 1997, 36, 10581–10594.

26