Optofluidic detection of Zika nucleic acid and protein biomarkers using multimode interference multiplexing

: The recent massive Zika virus (ZIKV) outbreak illustrates the need for rapid and specific diagnostic techniques. Detecting ZIKV in biological samples poses unique problems: antibody detection of ZIKV is insufficient due to cross-reactivity of Zika antibodies with other flaviviruses, and nucleic acid and protein biomarkers for ZIKV are detectable at different stages of infection. Here, we describe a new optofluidic approach for the parallel detection of different molecular biomarkers using multimode interference (MMI) waveguides. We report differentiated, multiplex detection of both ZIKV biomarker types using multi-spot excitation at two visible wavelengths with over 98% fidelity by combining several analysis techniques.


Introduction
Zika Virus (ZIKV) infection is caused by a mosquito-borne flavivirus [1,2] and has been strongly correlated with the development of fetal microcephaly and Guillan-Barré syndrome [3][4][5][6][7]. Consequently, the recent 2016 ZIKV outbreak throughout the Americas was declared a public health emergency by the World Health Organization (WHO), and created a pressing need for accelerated development of rapid, specific, and clinically sensitive ZIKV diagnostic techniques [8]. Symptomatic diagnosis of ZIKV in humans is difficult as ZIKV infection can initially present as influenza-like, and can furthermore be easily confused with other flaviviurses, such as Dengue Virus [2]. Currently available tests include qualitative enzymelink immunosorbent assays (ELISA) which target IgM antibodies and qualitative and endpoint polymerase chain reaction (PCR) assays that measure viral RNA concentrations with amplification-based methods. Due to the recorded cross-reactivity with Dengue Virus, Zika Virus antibody detection is not sufficient as a stand-alone diagnostic tool [2,[9][10][11][12], and the involvedness and intricate laboratory setting necessary for PCR tests of nucleic acid targets are non-ideal [13] and limited to nucleic acids. Furthermore, nucleic acid and antibody biomarkers are detectable in different stages of the infection [14]. A biosensor that detects both modalities is, therefore, highly advantageous to accurately diagnose a ZIKV patient and essential to monitoring a possible ZIKV epidemic.
One avenue to accomplish this involves "lab-on-a-chip" devices, which are compact apparatuses that process and analyze clinically relevant chemical or biological samples in small volumes. Optofluidic approaches that integrate photonic principles with biosensing, bioanalysis, a Here, we con Waveguides ( allow for plan 22,[25][26][27]. M preparation st molecular bio In order t photonic appr This device excitation wa the relative ph spot patterns wavelengths enabling on-c detection and nucleic acid m [25,26]. Her scheme that p targets at onc biomarkers is

Principles
We implemen visible light modes with d relative phase spot patterns, we design the corresponds t a scanning el cross section, (bottom left).  uide (LCi-resonant distances [15]. Single mode fiber is butt-coupled to a single mode excitation waveguide, which launches into the MMI excitation waveguide. Fluorescently tagged targets are directed from an input reservoir to an output reservoir through the LC-ARROW microchannel by pressuredriven flow. As these fluorescent targets pass through the excitation cross-section at the intersection of the LC-ARROW and the excitation MMI waveguide, a fluorescence signal is generated. This fluorescence signal propagates down the LC-ARROW, orthogonal to the excitation waveguide, and is coupled to a collection solid-core waveguide. The signal is then transmitted to and analyzed by an avalanche photodiode (APD). The hollow-core ARROW waveguides were 5 μm x 12 μm in cross section geometry and fabricated using a SU8 sacrificial layer technique and standard planar silicon device fabrication methods [27].
The dependence of the spot number (N) on distance from the beginning of the MMI (L) and the excitation wavelength λ can be derived from standard MMI theory [24] and is given by where w is the effective MMI-WG width (here: 75 µm) and n c the effective refractive index of the MMI-WG (here: 1.46). When λ = 633 nm; 7 well-defined spots are produced at the MMI length of L = 2243 µm, where the liquid core ARROW intersects the MMI-WG. At the same intersection, choosing λ = 556 nm produces 8 well-defined spots. This provides a means for spectral multiplexing for our targets. Figure 1(a) also shows both excitation patterns of the MMI waveguide within the liquid-core ARROW filled with quantum dots (right). Target specificity is implemented with a bead-based solid phase extraction method illustrated in Fig. 1(b), which displays the resultant particles for both nucleic acid and protein target isolation. A biotinylated pulldown molecule, either a complimentary oligonucleotide (IDT) specific to a section of the Zika genome or a biotin-functionalized (Roche) ZIKV monoclonal antibody HM333 (East Coast Bio.), was added to streptavidin-coated magnetic microspheres (Thermo) at room temperature for 30 minutes in a medium-high salt solution in accordance with the manufacturer's specifications [12]. The pulldown-functionalized beads were then washed with a low-salt buffer, PBS (Corning). Reporter molecules were then combined with the target nucleic acid (IDT) and target recombinant ZIKV-NS1 protein (East Coast Bio.) and incubated at 37°C for two hours. These reporter molecules were either a complementary nucleic acid probe with TYE665 attached (IDT) or a secondary monoclonal antibody HM332 (East Coast Bio.) functionalized with Cy3 (Lumiprobe). The target and reporter biomarkers were then combined with their respective pulldown beads and left to incubate at room temperature for another 2 hours, at which point they were washed a second time in a low-salt buffer. Each microsphere has the maximum capacity to bind 20 fg of biotinylated antibody or 2.5 x 10 5 molecules of biotinylated ss-oligomer, both of which correspond to a maximum of ~2.4 x 10 5 fluorophores/bead. The isolated target complex is finally pipetted into the input reservoirs of the ARROW chip for detection. Figure 1(c) shows the fluorescence particle trace over time for the protein-bead complex in the absence (top) and presence (bottom) of the NS1 target protein. It is evident from Fig. 1(c) that there is minimal background in the absence of NS1 protein, demonstrating the specificity of this detection scheme for protein detection. This serves as a negative control experiment for our experiments. Specificity for nucleic acid analysis was demonstrated previously in [22,26].

Results and discussion
In Fig. 2, the experimental fluorescence particle trace results for multiplex biomarker detection are shown.   Fig. 3(c)), ating each gures 3(a) ed by the (2) where S(t) is the new, shift-multiplied signal, N is the number of MMI spots, F(t) is the particle fluorescence signal, δt is the corresponding characteristic time difference for each individual nucleic acid (red) or protein (green) particle signals, mentioned above [28]. The result is that S(t) is enhanced when the particle trace is shifted the correct number of times (N) and by the correct δt, and it is quenched when shifted the incorrect number of times and by the incorrect δt. Figure 3(b) shows enhanced signal values for the signals shown in Fig. 2 (b) and (c) as well as Fig. 3(a). When we perform the shift multiply algorithm, we see that the signals are well separated, and the signal with higher N is notably larger as previously reported [28]. Despite a velocity distribution of the ensemble of multichromatic signals, we are able to obtain a velocity-independent confirmation of the fluorescence signal identity using the individual δt of the signal. We also confirmed each peak by vetting with a data parsing scheme that would also mitigate the distribution of velocities in the trace. The spot pattern of the excited particle was directly determined by dividing the total time of the signal by the individual signal's δt, as shown below: where N i is the number of spots created by the MMI excitation of the individual particle, T tot,i is the total time of the individual signal, δt i is the individual signal's characteristic delta t. In order to robustly identify each particle signal, we applied both methods. If the identity of the particle was verified by both methods, then the particle was defined as such. If there was a discrepancy in the two methods, then the particle was left as unidentified. Figure 3(c) shows implementation of this approach for a fluorescence trace where both lasers are exciting the sample in flow. Of the 215 fluorescence signals in the trace, over 98% were identified. We differentiated 134 red nucleic acid signals and 77 protein signals. We repeated this assay three times with comparable results to ensure repeatability of the approach.

Conclusions
We have demonstrated a dual nucleic acid and protein detection scheme using a planar MMI waveguide platform. Zika virus nucleic acid and protein complexes can be excited and detected simultaneously on-chip with high specificity and sensitivity. We demonstrate here that the ARROW optofluidic biosensor is a sample-agnostic platform that is a highly specific and sensitive means of detecting different types of molecular targets at once. The optofluidic platform discussed above can also be incorporated into a dedicated sample preparation and delivery microfluidic layer, as demonstrated in [19,21,25]. These results and the upcoming integrated technology show that this optofluidic approach is capable of detecting different classes of biomarkers down to a single antigen using a simple sample preparation and analysis protocol, in contrast with current gold standard techniques such as RT-PCR and ELISA, which are only suited for one type of target.