Separation Of Binding Molecules

  • Published: Dec 5, 2013
  • Earliest Priority: May 31 2012
  • Family: 1
  • Cited Works: 41
  • Cited by: 3
  • Cites: 1
  • Sequences: 87
  • Additional Info: Cited Works Full text

SEPARATION OF BINDING MOLECULES

RELATED APPLICATION

This application claims the benefit of U.S. Provisional Application No. 61/653,472, filed on May 31, 2012. The entire teachings of the above application are incorporated herein by reference.

GOVERNMENT SUPPORT

This invention was made with government support under Grant No. DGE- 1122374 awarded by the National Science Foundation. The government has certain rights in this invention.

BACKGROUND OF THE INVENTION

Nucleic acid aptamers are used widely in biological applications such as gene regulation, cell surface biomarker profiling, and as therapeutics. New aptamers are identified from highly diverse libraries using the Systematic Evolution of Ligands by Exponential enrichment (SELEX). See Stolteburg R, Reinemann C, Strehlitz B (2007) Biomolecular Engineering 24:381-403. An important limiting step in this process is the efficiency and reliability with which relatively rare but desirable aptamers are enriched from a large background of non-binding library members. A need exists for a general, widely accessible and inexpensive partitioning strategy.

SUMMARY OF THE INVENTION

In a particular aspect, the invention is directed to a method of separating one or more binding molecules that are specifically bound to one or more binding targets from a mixture of bound and unbound molecules. The method includes introducing the mixture into at least one inlet of a microfiuidic device at a sample flow rate, the microfluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate the one or more molecules bound to the binding target along portions of the cross-section of the channel based on size, wherein bound molecules flow along a radially innermost portion of the channel to a first outlet, and unbound molecules flow along one or more other portions of the channel to at least one other outlet, thereby separating the one or more binding molecules that are specifically bound to the one or more binding targets from the mixture of bound and unbound molecules.

In another aspect, the invention is directed to a method of identifying one or more binding molecules that specifically bind to one or more binding targets. The method includes introducing a mixture of one or more binding molecules and the one or more binding targets into at least one inlet of a microfluidic device at a sample flow rate, the microfluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate the one or more molecules bound to the binding target along portions of the cross-section of the channel based on size, wherein bound molecules flow along a radially innermost portion of the channel to a first outlet, and unbound molecules flow along one or more other portions of the channel to at least one other outlet, and determining the binding molecule that is bound to the binding target, thereby identifying the binding molecule that specifically binds to the binding target.

This invention has many advantages, including a high partitioning efficiency and continuous resolution of a well-mixed suspension of particles (e.g., nucleic acid library or a library of other affinity molecules, such as proteins (e.g., antibodies), nucleic acids (e.g., aptamers), or synthetic (e.g., small molecules, peptoids) libraries) into at least two distinct flow streams that can be collected separately.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing will be apparent from the following more particular description of example embodiments of the invention, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating embodiments of the present invention.

FIGS. 1A-1C are schematic illustrations of the I-SELEX microfluidic device and its use in aptamer selection. FIG. 1 A: The microchannel design consists of a bi- loop spiral of radius ~lcm with dual inlets and outlets. Pre-incubated bead/cell target-aptamer library mixtures and sheath buffer are pumped through the right and left inlets of the device, respectively. Under the influence of Dean drag forces (FD), unbound aptamers migrate along Dean vortices towards the outer wall and are diverted to the waste (bottom) outlet. Target beads/cells (and any bound aptamers) experience additional strong inertial lift forces (FL) and are focused along the inner microchannel wall and collected in the product (top) outlet. FIG. IB: a schematic of the I-SELEX procedure showing: (1) negative selection of random library against sRBCs; (2) positive selection of surviving library on tRBCs; (3) partitioning of iRBC-bound aptamers from unbound library in the I-SELEX device; (4) recovery, RT-PCR amplification and in vitro transcription to enrich tRBC-bound aptamers. FIG. 1C: optical cross sections through the device were taken at positions 1-5 along the length of the channel using high-speed confocal microscopy. Fluorescently labeled CRP aptamers injected via the sample inlet (right) enter the microchannel nearest the inner wall (positions 1 and 2). Dean forces focus free aptamers along the channel midline (position 3). Once focused along the midline (position 4), the aptamer stream migrates toward the outer wall (positions 4 and 5) before exiting the device.

FIGS. 2A-2C are summaries of the synthesis and quantitative

characterization of a thrombin-displaying model system. FIG. 2A: human red blood cells (PvBCs) are surface biotinylated then coated with streptavidin to produce scaffold-only RBCs (sRBCs). Biotinylated human a-thrombin is added to sRBCs in titrated quantities to create thrombin-displaying RBCs (^RBCs). Excess biotin is added to both sRBCs and iRBCs to cap residual open sites on streptavidin. FIG. 2B: quantitative control over the amount of thrombin displayed from tRBCs can be achieved by titrating the concentration of biotinylated human a-thrombin used is step 3 above. iRBCs used in these experiments typically displayed ~103-104 thrombin molecules/cell. The plot shows efficient binding of fluorescently labeled thrombin aptamer to ^RBCs but not sRBCs. FIG. 2C: fluorescently-labeled Toggle- 25 aptamer selectively binds iRBCs with an apparent IQ ~ 34 nM by FACS analysis. No significant binding to sRBCs is detected.

FIGS. 3A-3D are illustrations of the optimization of the operating flow rates within the I-SELEX device to achieve stringent separation of a model non- interacting aptamer from tRBCs, while maximizing tRBC recovery. FIG. 3A:

average fluorescence intensity line scans showing the normalized distribution of unbound FITC-labeled CRP aptamers (200 nM) across the channel width at increasing flow rates. Approximate positions of the product and waste outlets are indicated. Corresponding fluorescence images illustrating flow positions of unbound aptamers are also shown as an inset (dashed lines indicate the approximate position of the microchannel walls). FIG. 3B: average composite images indicate unbound aptamers move to the outer wall and are diverted into the waste (bottom) outlet. FIG. 3C: average composite images indicate efficient tRBC focusing to the inner microchannel wall and diversion into the product (top) outlet. In both FIG. 3B and FIG. 3C, the sample input and sheath buffer flow rates are 150 μΐ, min"1 and 1500 μΐ, min"1, respectively, and the dashed lines indicate approximate positions of the microchannel walls and bifurcation. FIG. 3D: recovery of tRBCs at the product outlet as a percentage of the cells loaded into the device when operated at different sample input flow rates. In all cases, the sheath buffer flow rate was 10-fold higher than the sample input flow rate.

FIGS. 4A-4B are illustrations of the I-SELEX device's high partitioning efficiency and use to selectively recover and enrich target aptamers from mock libraries. FIG. 4A: near quantitative recovery of Toggle-25 bound to tRBCs was reproducibly achieved, whereas scrToggle-25 recovery using tRBC targets was below the limit of detection. These data are consistent with a device partitioning efficiency that is >106. FIG. 4B: mock SELEX libraries containing Toggle- 25 :i'crToggle-25 in 1 :10 and 1 : 100 ratios, respectively, were incubated with tRBCs and partitioned in a single pass through the I-SELEX device. Toggle-25 is selectively enriched and becomes the dominant species post-selection as determined by quantitative RT-PCR. FIGS. 5A-5D are illustrations of the use of the I-SELEX device for de novo discovery of high affinity aptamers. FIG. 5 A: bio-layer interferometry kinetic binding data on the interaction between probe-immobilized thrombin and the selected RNA pools from each round are shown. Toggle-25 and scrToggle-25 were included as positive and negative controls, respectively. High affinity (~ 40 nM) binding of the pool from the fifth selection round was observed. FIG. 5B: 5-12 and Toggle-25 show enhanced binding to fRBCs but not ^RBCs. scr Toggle-25 and the captured oligonucleotides used for aptamer fluorescent labeling do not exhibit binding to either cell type. FIG. 5C: clones sequenced from the Rounds 3 and 5 selected pools were collectively analyzed using MEME (Bailey TL, Elkan C (1994) Proc Int Conf Intell Syst Mol Biol 2: 28-36). A conserved heptanucleotide motif (Motif 1), distinct from that present in Toggle-25, was detected in 12/38 unique clones (Cluster I sequences). By mfold, Motif 1 is predicted to be part of a stem- loop secondary structure element, as shown for 5-12. See Zuker M (2003) Nucleic Acids Res 31 : 3406-3415. FIG. 5D: using 5-12 as an example, the conserved motif was shown to contribute significantly to high affinity binding to thrombin.

Truncated 5-12 {5-Ylmini), in which the conserved motif is retained in the predicted parental stem-loop arrangement, bound thrombin with a ¾ ~ 5 nM, similar to 5-12 (Kd ~ 2 nM). However, 5-\2mut, which is the full-length aptamer containing mutations to the motif indicated in FIG. 5(B), bound thrombin with significantly lowered thrombin affinity (¾ > μΜ).

DETAILED DESCRIPTION OF THE INVENTION

A description of example embodiments of the invention follows.

Described herein is a micro fluidics approach based on inertial focusing and Dean flow phenomena in curved (e.g., spiral) channels to achieve high stringency separation of one or more binding molecules that are specifically bound to one or more binding targets from, for example, a library of unbound nucleic acids. A high partitioning efficiency (e.g., about 1 ,000,000 or greater) can be achieved in a single pass of a selection mixture through this device. The device facilitates continuous resolution of a well-mixed suspension of binding molecules (e.g., particles) and unbound molecules (e.g., from a nucleic acid library) into two distinct flow streams that can be collected separately. Subsequent processing can be performed as in existing separation protocols (e.g., SELEX protocols, refered to herein as inertial SELEX (I-SELEX)). Using a whole cell target, the efficiency of this device is shown herein. Specifically, the suitability of the device was established by its ability to identify high affinity aptamers within 3-5 rounds of de novo selection for a high diversity library. The methods described herein offer a generic strategy for improved partitioning when working e.g., with bead-immobilized and unmodified whole cell targets, and can be seamlessly integrated into any existing binding molecule (e.g., aptamer) selection workflow. To validate this strategy, a model system was used for an I-SELEX procedure in which thrombin was displayed from the surface of intact human red blood cells (RBCs). This permitted characterization of the high partitioning efficiency achievable with this device, and the ability to selectively enrich a known aptamer from mock libraries. Finally, successful de novo I-SELEX was demonstrated using a library containing ~1014 unique sequences. High affinity (low nM Kd) aptamers were recovered containing a conserved motif distinct from previously described thrombin aptamers by the third round of selection, and these were dominant by the fifth round of selection. The methods herein provide a simple and inexpensive microfluidic strategy to enable the efficient discovery of new binding molecules (e.g., aptamers).

Accordingly, in a particular aspect, the invention is directed to a method of separating one or more binding molecules that are specifically bound to one or more binding targets from a mixture of bound and unbound molecules. The method includes introducing the mixture into at least one inlet of a microfluidic device at a sample flow rate, the microfluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate the one or more molecules bound to the binding target along portions of the cross-section of the channel based on size, wherein bound molecules flow along a radially innermost portion of the channel to a first outlet, and unbound molecules flow along one or more other portions of the channel to at least one other outlet, thereby separating the one or more binding molecules that are specifically bound to the one or more binding targets from the mixture of bound and unbound molecules. In another aspect, the invention is directed to a method of identifying one or more binding molecules that specifically bind to one or more binding targets. The method includes introducing a mixture of one or more binding molecules and the one or more binding targets into at least one inlet of a microfluidic device at a sample flow rate, the microfluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate the one or more molecules bound to the binding target along portions of the cross-section of the channel based on size, wherein bound molecules flow along a radially innermost portion of the channel to a first outlet, and unbound molecules flow along one or more other portions of the channel to at least one other outlet. The binding molecule that is bound to the binding target is determined (identified), thereby identifying the binding molecule that specifically binds to the binding target.

A variety of methods of determining the identity of the one or more binding molecules that are bound to the one or more target molecules can be used and will depend upon the characteristics of the binding molecule(s). Examples of methods include amplification methods (e.g., polymerase chain reaction (PCR)),

sequencing/cloning methods, electrophoretic (e.g., gel, capillary) methods, chromatographic methods and the like.

The methods can further include contacting the bound and unbound binding molecules with the binding target prior to introducing the mixture into the microfluidic device for an amount of time in a range of between about 1 second and about 1 week, such as about 1 hour, or about 1 day, or about 2 days, or about 3 days. For example, the methods can further comprise, prior to introducing the mixture into the device, contacting the one or more binding molecules with the one or more binding targets, thereby producing the mixture, and the mixture is maintained under conditions in which the one or more binding molecules can bind the one or more binding targets, contacting the one or more binding targets with the one or more binding molecules being assessed for binding to the binding target, thereby producing the mixture, and maintaining the mixture under conditions in which the one or more binding molecules can bind to to the one or more binding targets. In a particular aspect, these steps are performed prior to (e.g., immediately prior or after a suitable period of time (e.g., minutes, hours, days, weeks, etc.)) introducing the mixture into the device.

In yet another aspect, the invention is directed to a method of separating a binding molecule that specifically binds to a binding target. The method includes contacting the binding target with one or more binding molecules being assessed for binding to the binding target, thereby producing a combination. The method further includes maintaining the combination under conditions in which the one or more binding molecule can bind the binding target if the one or more binding molecule specifically binds the binding target, and introducing the combination into at least one inlet of a micro fluidic device at a sample flow rate, the micro fluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate combinations along portions of the cross-section of the channel based on size, wherein combinations flow along a radially innermost portion of the channel to a first outlet, and other binding molecules flow along other portions of the channel to at least one other outlet, thereby separating the binding molecule that specifically binds to the binding target.

In still another aspect, the invention is directed to a method of identifying a binding molecule that specifically binds to a binding target. The method includes contacting the binding target with one or more binding molecules being assessed for binding to the binding target, thereby producing a combination. The method further includes maintaining the combination under conditions in which the one or more binding molecule can bind the binding target if the one or more binding molecule specifically binds the binding target, and introducing the combination into at least one inlet of a microfluidic device at a sample flow rate, the microfluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate combinations along portions of the cross-section of the channel based on size, wherein combinations flow along a radially innermost portion of the channel to a first outlet, and other binding molecules flow along other portions of the channel to at least one other outlet, thereby identifying the binding molecule that specifically binds to the binding target. As used herein, a binding molecule includes any molecule that specifically binds to a binding target. Binding molecules include nucleic acids (e.g.,

oligonucleotides), peptides (e.g, proteins), organic small molecules (e.g, antibiotics, fluorophores) and variants thereof (e.g., chemically modified; mutants). Examples of nucleic acids include DNA, RNA, ribozymes, DNAzymes, and variants thereof. The nucleic acid can be single stranded or double stranded, and can be naturally occuring and/or chemically synthesized, and/or chemically modified (e.g., can include one or more modified bases). For example, chemically modified nucleic acids can introduce new features such as functional groups (e.g., to provide enhanced binding and/or stability to the binding target). Such modifications include modification(s) at the second position of the sugar in a nucleic acid (e.g., replace ribose 2 ΌΗ with a 2 -NH2 or 2 -F group), inclusion of one or more 2 -O-methyl substituted nucleotides, modification of nucleic bases at the C-5 position of pyrimidines and/or the C-8 position of purines, modification of the phosphate backbone (e.g., replacement of non-binding oxygen in the phoshodiester linkage with sulfur), the inclusion of a label (e.g., radioactive labeled nucleotides or fluorescent molecules), and combinations thereof. Examples of peptides include antibodies, enzymes, and ligands (e.g., growth factors). Examples of small molecules (e.g., organic or inorganic) include drugs (e.g., antibiotics), fluorophores, and dyes. In a particular aspect, the one or more binding molecules is an aptamer.

In another aspect, the one or more binding molecules are present in a library (e.g., a nucleic acid library; a peptide library; combinatorial library). The library can be a completely or partially random, and/or a completely or partially structured library. In one aspect, the one or more binding molecules are present in a nucleic acid library. The nucleic acid library can comprise any number of nucleic acids, such as about 105, 106, 107, 108, 109, 1010, 1011, 1012, 1013, 1014, 1015, 1016, 1017, 1018, 10!9, 1020 or more nucleic acids. In other aspects, the one or more binding molecules are present in a peptide library (e.g., a bacterial display, a phage display, a

1 2 yeast display, a ribosome display). The peptide library can comprise about 10 , 10 , 103, 104, 105, 106, 107, 108, 109, or more peptides. In other aspects, the one or more binding molecules are present in a combinatorial library. The combinatorial library can comprise 101, 102, 103, 104, 105, 106 or more particles. The binding affinity of the bound molecules can be in a range of between about 10 pM and about 1 μΜ, such between about 2 nM and about 200 nM, between about 20 nM and about 100 nM, between about 50 nM and about 150 nM, and between about 100 nM and about 200 nM.

A wide variety of binding targets can be used in the methods described herein, from single molecules to complex binding targets and/or whole organisms (e.g., single cell or multi-cell organism). The dimensions of the device determine the size of binding targets that can be separated (i.e., focused or resolved). The limit of resolution is determined by limitations on the fabrication of device height dimensions and large pressure drops at operating flow rate conditions in smaller channels. In one aspect, the diameter of the binding targets can be in a range of between about 1 μπι and about 20 μηι, such as about 2 μηι, about 3 μτη, about 4 μιη, about 5 μπι, about 6 μτη, about 7 μηι, about 8 μπι, about 9 μηι, about 10 μηι, and about 15 μιη. In another aspect, smaller diameter binding targets (e.g., nucleic acids, small molecules) can be focused if they are attached to supramolecular structures, such as beads.

Examples of the one or more binding targets can include nucleic acids, proteins, whole cells, cell membranes, viruses (e.g., bacteriophages), bacteria, cell surface molecules, receptors, ligands, tissues, organisms, proteins, and

supramolecular structures. Examples of cells are red blood cells, such as thrombin- displaying red blood cells, mammalian cells, eukaryotic cells, and prokaryotic cells. Examples of supramolecular structures are scaffolds, beads, liposomes, and organelles. Specific examples of binding targets include inorganic components (e.g., Zn2+, Ni2+), small organic molecules (e.g., ethanolamine, theophylline), nucleotides and derivatives, cofactors, nucleic acids, amino acids, carbohydrates, antibiotics, peptides and proteins, and complex structures. As used herein, a bound binding molecule is a molecule that is specifically bound to the binding target.

In some aspects, the binding molecules and/or binding target can comprise a solid phase to which it can be attached. Examples of a suitable solid phase include a bead (e.g., a magnetic bead), a scaffold and the like. Any suitable microfluidic device can be used (e.g., see PCT Application PCT/US2011/027276 published as WO 2011/109762 Al on September 9, 2011, the contents of which are incorporated by reference in their entirety). Micfofluidic devices for use in the methods provide herein can be fabricated as described in the exemplification.

As further described herein, the microfluidic device can have one or more (at least one) inlet for introduction of the sample into the device. For example, the device can have one, two, three, four, five, six, seven, eight, nine, ten, etc., inlets.

The sample can be introduced into the device using a variety of techniques known to those of ordinary skill in the art. For example, the sample can be introduced using a syringe and/or a pump.

The sample flow rate can be in a range of between about 50 μΤ/ηιίη and about 200 such as about 60 μΤ/min, about 70 μΤ/min, about 80 μΤ/min, about 90 μΤ/min, about 100 μΤ/ητϊη, about 110 μΤ/ηιΐη, about 120 μΤ/ηώι, about 130 μΙ7ιηίη, about 140 μΙ7ιηϊη, about 150 μΤ/min, about 160 μΤ/η ίη, about 170 μί/ηιήι, about 180 μΙ7ηιΐη, about 190 μΙ7ηιίη.

Similarly, the microfluidic device can have one or more outlets. In some aspects, the device can have one, two, three, four, five, six, seven, eight, nine, ten, etc., outlets. In a particular aspect, the device has at least 2 outlets. In another aspect, the device has 3 outlets. In yet another aspect, the device has 4 outlets. In still another aspect, the device has 8 outlets. In a particular aspect, the more than one other outlet can have a diameter that is greater than the diameter of the first outlet.

The device also comprises one or more channels, for example one, two, three, four, five, six, seven, eight, nine, ten, etc., channels connecting the one or more inlets to the one or more outlets. The channel(s) comprise a cross section of a height and a width defining an aspect ratio that enables separation of the binding target(s) from the remainder of the molecules in the sample. As used herein, an aspect ratio is the ratio of a channel's height divided by its width and provides the appropriate cross section of the channel to allow the binding targets to flow along at least one portion of the cross section of the channel to a first outlet, and the remaining molecules to flow along a different (e.g., second, third, fourth, etc.) part or cross section of the channel and not to the same outlet as the binding targets, such as to a distinct (e.g., second, third, fourth, etc.) outlet. The appropriate aspect ratio causes the binding targets to flow along a distinct portion of the channel based on a difference in a structural characteristic of the binding targets in the sample, compared to the same or similar structural characteristic of the remaining molecules in the sample. Examples of such structural characteristics include cell size and the like.

As will be appreciated by those of ordinary skill in the art, the channel can have a variety of shapes (e.g., curved, spiral, multiloop, s-shaped, linear) provided that the dimensions of the channel are adapted to isolate the one or more molecules bound to the binding target along portions of the cross-section of the channel based on size.

In one aspect, the channel is curved. In a particular aspect the channel is a spiral. The height of the spiral channel can be in a range of between about 10 μηι and about 200 μηι, such as about 100 μηι and about 140 μπι. The width of the spiral channel can be in a range of between about 100 μηι and about 500 μηι. The length of the spiral channel can be in a range of between about 1 cm and about 10 cm.

In one aspect, the spiral channel can be a bi-loop spiral channel. In another aspect, the spiral channel can be 2-loop spiral channel. In yet another aspect, the spiral channel can be 3 -loop spiral channel. In still another aspect, the spiral channel can be 4-loop spiral channel. In another aspect, the spiral channel can be 5-loop spiral channel, etc.

The radius of the spiral channel can be adapted to yield a Dean number in a range of between about 1 and about 10, such as a radius of about 1 cm that yields a Dean number equal to about 5. The length of the spiral channel can be equal to or greater than about 3 cm, such as about 9 cm, about 10 cm, about 15 cm, and about 20 cm. The width of the spiral channel can be in a range of between about 100 μπι and about 1,000 μπι, such as about 200 μηι, about 300 μηι, about 400 μιη, about 500 μηι, about 600 μιη, about 700 μιη, about 800 μηι, and about 900 μηι. The height of the spiral channel can be in a range of between about 20 μη and about 200 μιη, such as about 30 μηι, about 40 μηι, about 50 μτη, about 60 μηι, about 70 μηι, about 80 μτη, about 90 μηι, about 100 μηι, about 110 μτη, about 120 μηι, about 130 μτη, about 140 μηι, about 150 μπι, about 160 μηι, about 170 μηι, about 180 μηι, and about 190 μηι. The aspect ratio of the channel can be in a range of between about 0.1 and about 1, such as about 0.12, about 0.2, about 0.3, about 0.4, about 0.5, about 0.6, about 0.7, about 0.8, and about 0.9.

In a particular aspect, the method can further include introducing a buffer into one or more inlets (e.g., the same inlet or a different inlet, such as a second inlet) of the micro fluidic device at a buffer flow rate. A variety of buffers can be used in the methods described herein. In one aspect, a buffer with a similar viscosity as the sample is used. In one aspect, the buffer can be saline, such as phosphate buffered saline supplemented with calcium. In another aspect, the buffer can be culture media, such as mamallian cell culture media (e.g., HEPES-buffered RPMI media). In a particular aspect, the buffer can be a buffer that facilitates binding of a binding molecule to a binding target, such as a thrombin binding buffer. The buffer is also sometimes referred to herein as a sheath buffer. The buffer flow rate can be in a range of between about 500 μΐνπώι and about 2,000 μΤ/ηιίη, such as about 600 μΕ/πιϊη, about 700 μΐνιηϊη, about 800 μΤ/min, about 900 μΤ/min, about 1,000 μΤ/πώι, about 1 ,100 μΕ/ηιϊη, about 1,200 μΤ/ηώι, about 1,300 μΕ/ηήη, about 1,400 μΙ7ηιϊη, about 1,500 μΕ/πώι, about 1,600 μΐ,/ηιΐη, about 1,700 μΤ/ηιίη, about 1,800 μΙ7ηιϊη, and about 1,900 μΤ/min.

In some aspects, the sum of the buffer flow rate and the sample flow rate can be equal to an overall flow rate that yields a Reynolds number equal to or greater than about 50. The buffer flow rate can be less than or equal to the sample flow rate, although focusing and thereby isolating the binding targets in the device is increasingly improved with a buffer flow rate higher than the sample flow rate, such as at least about 2 times greater than the sample flow rate, such as about 3 times, about 4 times, about 5 times, about 6 times, about 7 times, about 8 times, about 9 times, and about ten times greater than the sample flow rate.

The microfluidic device described herein can be used to separate one or more binding molecules that are specifically bound to one or more binding targets from a mixture of bound and unbound molecules. In some aspects, one or more

microfluidic devices (e.g., a cascade of microfluidic devices, e.g., in parallel or in sequence) described herein can be used to separate one or more binding molecules that are specifically bound to one or more binding targets from the mixture of bound and unbound molecules.

As will be appreciated by those of skill in the art, the methods can also further comprise enriching the binding targets, for example, by passing the mixture through one or more devices one or more times. The mixture can be passed through the device one or more times, such as 2 times, 3 times, 4 times, or 5 times, thereby enriching the sample of the one or more binding molecules that are specifically bound to the one or more binding targets. Additionally, the method can include reintroducing the bound molecules into the inlet of the microfluidic device, or into the inlet of another microfluidic device.

In some aspects, the method can further comprise collecting the one or more binding molecules that are specifically bound to the one or more target molecules from the first outlet (e.g, for further analysis and/or characterization such as determination of binding affinity, cloning, sequencing). In some aspects, the methods can further comprise separating (isolating) the one or more binding molecules from the one or more target molecules. For example, in one aspect, in which the one or more binding targets, or if the one or more binding molecules are nucleic acids, the binding molecule(s) can collected and/or separated from the target molecule and amplified by, for example, polymerase chain reaction (PCR), reverse transcriptase PCR, or a combination thereof. In addition, or in the alternative, the one or more binding molecules are nucleic acids, the binding molecule(s) can collected and/or separated from the target molecule and cloned, sequenced, subjected to binding studies, subjected to chromatographic studies, nuclear magnetic resonance studies and the like.

The performance of this device can be characterized by its partitioning efficiency, defined as the ratio of the number of input sequences to the number of non-binding sequences recovered in the product output after partitioning. The partitioning efficiency of isolating bound molecules from unbound molecules can be greater than or equal to about 100,000, such as about 200,000, about 300,000, about 400,000, about 500,000, about 600,000, about 700,000, about 800,000, about 900,000, about 1,000,000, about 1,100,000, about 1,200,000, about 1,300,000, about 1,400,000, about 1,500,000, and about 2,000,000. In a particular aspect, the partitioning efficiency of this device is equal to or greater than 1,000,000.

The need for a partitioning strategy has been addressed by developing a continuous partitioning device based on inertial microfluidics in curvilinear channels, which, in some aspects is referred to as inertial SELEX or I-SELEX. Inertial microfluidics is a powerful technique for high throughput particle (e.g., cell) separation, as previously demonstrated in applications such as blood fractionation, circulating tumor cell (CTC) enrichment from blood, and stem cell synchronization. See Di Carlo D, et al. (2008) Analytical chemistry 80: 2204-221 1); Bhagat AAS, et al. (20U) Lab on a Chip 1 1 : 1870-1878; Lee WC, et al. (20U) Lab on a Chip 11 : 1359. In this work, a curved device that imparts differential inertial and Dean migratory effects on micron-sized particles (e.g. beads and cells) versus

macromolecules (e.g. nucleic acid libraries) within a mixture is used to stringently separate these components.

EXEMPLIFICATION RESULTS

I-SELEX microfluidic device overview

The device used for I-SELEX is shown schematically in FIG. \A. The user interface is simple, requiring only a pair of syringe pumps for routine operation. Importantly, the current device is used strictly as a generic strategy for achieving the partitioning step in SELEX (FIG. IB), and so it is easily integrated with existing sample preparation, and library preparation, manipulation and deconvolution procedures. Operationally, the nucleic acid library-target mixture and a sheath buffer are introduced via the right 110 and left 120 inlets of the device 100 (FIG. IA), respectively, at an empirically determined appropriate flow rate. After a brief period during which inertial forces within the device stabilize, particles (and any bound aptamers) are recovered at the product 130 (top) outlet, while the non-binding fraction of the nucleic acid library is diverted to the waste 140 (bottom) outlet (FIG. IA). This flow pattern in the device is stable indefinitely once established, and permits continuous input mixture fractionation and sampling of complex libraries. Under typical operating conditions, a library size of ~1014 can be sampled and partitioned in -10 minutes at a sample flow rate of 150 μΐ, min"1 and ~ 2 xlO6 cells min"1.

Fluid mechanics principles governing the design and operation of the I-SELEX device

Due to centrifugal acceleration in curvilinear channels 125, an embodiment of which is shown in FIG. 1A, faster-moving fluid at the channel center moves towards the outer wall in a radial direction from the channel midline. Conservation of mass principles dictate that fluid near the channel walls circulates inwardly.

Consequently, two symmetrical and counter-rotating Dean vortices perpendicular to the main axial flow in the channel are established. See Di Carlo D (2009) Lab Chip 9: 3038-3046. The magnitude of these Dean vortices is determined by the dimensionless Dean number parameter (De), which relates channel dimensions, curvature, and flow rate as described in Equation (1): μ 2R 2R

where p is fluid density (kg m"3), Uf is the average primary channel velocity (m s"1), Dh is the microchannel hydraulic diameter defined as 2w x h/(w + h), μ is fluid viscosity (kg m'V1), R is the radius of curvature, and Re is Reynolds number. See Berger SA, Talbot L, Yao LS (1983) Annual Review of Fluid Mechanics 15: 461- 512. Ookawara et al. later formulated an empirical expression for the average Dean velocity (U) for a given De as:

UDe = 1.84 x lO"4 De a (m s'1) (2)

See Ookawara S, Street D, Ogawa K (2006) Chemical engineering science

61 : 3714-3724. Due to transverse Dean flows, particles flowing in a curvilinear channel experience lateral drag forces (/¾), which allow them to migrate across streamlines. Dean drag increases in magnitude with particle size and channel width. See Martel JM, Toner M (2012) Physics of Fluids 24: 032001. The lateral distance traversed by a particle due to Dean flow can be defined in terms of 'Dean Cycle' (DC). A particle that travels across the entire channel width (x-axis) has completed half a Dean Cycle (DC 0.5), and a full Dean cycle (DC 1) upon returning to its starting x-coordinate. The path length of a full Dean Cycle (LDC) is approximated by:

LDC ~ 2w + h (3)

where w and h are the channel width and height, respectively. Particles may undergo multiple Dean Cycle migrations, the number of which increase directly with channel length (Z), and flow rate (Re). In addition to Dean drag forces, particles in curvilinear microchannels experience an appreciable inertial lift force (F , which is the combination of a shear gradient lift force (directed toward the channel walls) and a wall-induced lift force (directed away from the channel walls). The superposition of competing inertial lift (F£) and Dean drag (FD) forces results in particle focusing at two equilibrium positions-one within each Dean vortex-provided the particle size (dp) and channel height (h) satisfy the following criterion:

aj/h > 0.07 (4)

See Di Carlo D, et al. (2008) Analytical chemistry 80: 2204-221 1 ; Di Carlo D, Irimia D, Tompkins G, Toner M (2007) Proceedings of the National Academy of Sciences of the United States of America 104: 18892. With optimized channel dimensions and flow conditions, these hydrodynamic forces act differentially on particles to achieve highly efficient size-based separation. If the force ratio, Rf, defined as FJJFD is≥ 1 , lift forces dominate, and inertial focusing at distinct equilibrium positions near the inner channel wall based on particle size is attained. As the equilibrium position of a particle is strongly dependent on Rf, which varies with the third power of particle diameter, this principle has been used to stringently separate mixtures of micron-sized particles or cells. See Kuntaegowdanahalli SS, Bhagat AAS, Kumar G, Papautsky I (2009) Lab on a Chip 9: 2973; Di Carlo D, Irimia D, Tompkins RG, Toner M (2007) Proceedings of the National Academy of Sciences of the United States of America 104: 18892, Kuntaegowdanahalli SS, Bhagat AAS, Kumar G, Papautsky I (2009) Lab on a Chip 9: 2973 and Bhagat AAS, Kuntaegowdanahalli SS, Papautsky I (2008) Lab on a Chip 8: 1906; Di Carlo D, et al. (2008) Analytical chemistry 80: 2204-221 1. However, inertial focusing of sub-micron sized particles (diameter < 1 μιη) and macromolecules is not possible in these devices due to the negligible inertial forces exerted on nanoscale species (Rf < 1). Instead, Dean drag forces dominate, and continuously drive these small species along circulating secondary flows to induce homogeneous mixing. This

phenomenon would be undesirable in a SELEX application, as it severely reduces partitioning of the 'free' nucleic acid library away from the aptamers bound to the micron-sized bead/cell target. Currently, inertial focusing of molecules is not possible due to difficulties in fabricating sufficiently small microchannels to satisfy the focusing criterion (<¾/¾ > 0.07), while tolerating the large pressure drop inherent at high flow rate conditions in these devices.

This limitation has been addressed by using a two-inlet, spiral channel design (FIG. I A) in which the sample stream is introduced via the right inlet 1 10 while a sheath buffer is pumped via the left inlet 120 at a higher flow rate to form a tight sample stream at the inner wall. If the channel 125 is truncated at L = n x (DC 0.5) for « e N, circulating macromolecules migrate along the midline (dictated by Dean flow) as a focused band towards the outer channel wall, and are maximally spatially resolved from inertially-focused particles migrating near the inner channel wall (Fig. 1 C). Additionally, this microchannel has a low aspect ratio (h/w « 1) as Dean drag forces become stronger with increasing channel width. See Martel JM, Toner M (2012) Physics of Fluids 24: 032001. An exemplary device 100 according to this invention measures 9 cm (/) x 500 μπι (w) x 60 μηι (h) with a dual-inlet and asymmetric dual-outlet spiral microchannel. The product 130 and waste 140 outlet diameters are 150 μηι and 350 μηι, respectively. The channel height was selected such that particles > 6 μηι in diameter (e.g., human RBCs) predominantly experience inertial forces (a ¾ ~ 0.1) and focus to the inner channel wall, while

macromolecules (e.g., nucleic acid SELEX libraries) experience Dean drag forces (aplh « 1) and are transported to the outer channel wall by the time they reach the device outlet (DC 0.5). This allows tightly focused particles/cells (and any bound macromolecules) to be efficiently collected at the product outlet 130, while unbound macromolecules are diverted to the waste outlet 140. Primarily, channel height determines particle focusing, and this parameter can easily be varied to

accommodate beads/cells of different sizes.

Whole cell target system used to quantitatively validate the I-SELEX device A synthetic whole cell model was used to quantitatively characterize the performance of this device during SELEX. Human red blood cells (RBCs) were selected as a model cell type as they present two extreme yet realistic challenges faced in whole cell-SELEX. First, the cell surface is dominated by a single glycoprotein, glycophorin A, which is present at -0.5-1 x 106 molecules per cell. See Chasis JA, Mohandas N (1992) Blood 80: 1869-1879. This can be used to recapitulate the scenario in complex target whole cell-SELEX where relatively rare surface targets may be occluded by proteins of significantly higher abundance. Second, RBCs have very high cell surface glycan content, and the majority of these terminate in negatively charged sialic acid residues. Thus, RBCs naturally display abundant glycan targets that favor recovery of low affinity (¾ ~ μΜ affinity) aptamers. See Masud MM, Kuwahara M, Ozaki H, Sawai H (2004) Bioorg Med Chem 12: 1111-1 120. While preserving the above characteristics, the RBC surface was modified to display human -thrombin as a target protein. This allowed one to take advantage of the previously described 'Toggle-25' thrombin aptamer to characterize the I-SELEX device, and to stringently exclude unfavorable target characteristics as the primary reason for potential failure of a de novo SELEX experiment. See White R et al. (2001) Mol Ther 4: 567-573.

First, RBCs were lightly biotinylated using NHS ester chemistry and coated with streptavidin to generate 'scaffold' RBCs (sRBCs) (FIG. 2A). Biotinylated thrombin was then bound to the cell surface to produce 'thrombin-displaying' RBCs (tRBCs). The remaining biotin-binding sites on streptavidin were capped using an excess of free biotin. The final amount of thrombin displayed on iRBCs was determined for each new batch prepared, and this parameter was controlled by titrating the concentration of biotinylated thrombin used (FIG. 2B). tRBCs used in these experiments typically displayed ~103-104 molecules/cell, which is -50- to 1000- fold lower than glycophorin A. Using 3'-FITC labeled Toggle-25 aptamer, it was confirmed that high affinity and specific binding to iRBCs occurs (apparent ¾ ~ 34 nM), while no binding to sRBCs is observed, as expected (FIG. 2C). This established that the displayed thrombin is accessible and selectively recognized by a cognate nucleic acid aptamer, confirming the suitability of this model system for device characterization.

The I-SELEX device facilitates resolution of tRBCs and aptamers into distinct streams

tRBCs and the 5'-FITC labeled C-reactive protein (CRP) DNA aptamer (no binding to tRBCs) were used to empirically determine the optimal input flow rates needed to simultaneously: (i) focus the /RBC stream at the inner wall of the device and ultimately into the product outlet; (ii) stringently divert the non-interacting CRP aptamer to the outer wall of the device and into the waste outlet; while (in) minimizing tRBCs entering the waste outlet. In the limit, these boundary criteria define perfect stringency in the ideal SELEX experiment, where the non-interacting nucleic acid library is completely excluded from the product outlet, and all tRBCs with aptamer bound are collected at the product outlet. See Huang C-J, Lin H-I, Shiesh S-C, Lee G-B (2010) Biosensors and Bioelectronics 25 : 1761-1766.

Predictable control over the free aptamer stream was empirically determined to occur when the sheath buffer flow rate is ten-fold greater than the sample input flow rate. A minimum overall flow rate (sheath buffer + sample) must be maintained {Re > 50 or U/ ~ 900 μΐ min"1) to ensure inertial focusing. Simultaneously, the flow rate-induced increase in Fi from the sheath stream results in better partitioning of the unbound nucleic acid library to the outer channel wall. At sample input flow rates of 50 \xL min"1 (total flow rate = 550 μΐ^ min"1), unbound aptamers are collected in the product outlet with tRBCs. As the sample input flow rate is progressively increased to the 180 μΕ min"1 maximum tested (total flow rate = 1980 μΐ, min"1), the unbound aptamer is increasingly recovered in the waste outlet (FIG. 3/4). Since the diameter of the waste outlet is greater than that of the product outlet, near-complete collection of unbound aptamers in the waste channel can be achieved while inertially focused RBCs remain at the inner wall during their short (< 1 s) passage through the device (FIG. 3B). tRBC recovery at the product outlet is inversely dependent on sample input flow rate (Fig. 3Q, as increased channel velocity [//increases the magnitude of Dean drag. Based on these data, a sample input flow rate = 150 μΕ min"1 and sheath buffer flow rate = 1500 μΕ min"1 were selected as standard operating conditions for achieving high partitioning between tRBCs and unbound aptamer while maintaining high (-85%) tRBC recovery at the product outlet (FIG. 3Q. The I-SELEX device has a high partitioning efficiency and permits selective enrichment of aptamers from a mock library

First, whether selective recovery of aptamers bound to tRBCs can be achieved in this device was tested. The Toggle-25 thrombin aptamer and a scrambled version (scr Toggle-25) with no affinity for thrombin or fRBCs were used. These were incubated separately at 100 nM with tRBCs collectively presenting an ~10-fold excess of thrombin binding sites prior to partitioning using the I-SELEX device. By quantitative RT-PCR, Toggle-25 and i"crToggle-25 levels were determined at the sample input and product outlet (tRBC-bound fraction). As shown in FIG. 4A, Toggle-25 was quantitatively recovered, while the amount of scrToggle- 25 collected at the product outlet was below the limit of detection.

From these data, the partitioning efficiency (PE) attainable with the I- SELEX device was determined. PE is a common metric used to evaluate SELEX washing methods, and is a measure of the device's ability to reject non-binding sequences from the recovered pool containing true aptamers. See Berezovski M, et al. (2005) Journal of the American Chemical Society 127: 3165-3171 and Lou X, et al. (2009) Proceedings of the National Academy of Sciences of the United States of America 106: 2989-2994. PE is defined by the ratio of the number of input sequences to the number of non-binding sequences recovered in the product output after partitioning. A single pass through this device reproducibly removed > 106 non-binding sequences (Fig. 4A), establishing this as the lower limit of PE for the I- SELEX device. This PE is similar to or exceeds that attained in NECEEM methods, and is consistent with the reproducibly high PEs accessible using microfluidics.

Whether riRJBC targets could significantly enrich Toggle-25 thrombin aptamers from a mock library containing excess scrToggle-25 was empirically tested in preparation for conducting a de novo SELEX experiment. Two artificial SELEX libraries were prepared, each containing ~1014 total molecules, and either 1% or 10% Toggle-25. Each library was incubated with ^RBCs (107 cells; 104 thrombin molecules/cell) and this mixture was subjected to a single partitioning step through the I-SELEX device. As shown in FIG. 4B, significant and preferential enrichment of Toggle-25 was achieved and the recovered pools from both mock libraries were dominated by Toggle-25.

Successful de novo selection using the I-SELEX device

Using a randomized library containing ~1014 different sequences, de novo SELEX was conducted using ^RBCs as a whole cell target. For each round, the partitioning step was achieved in a single pass of the tRBC-library mixture through the I-SELEX device. iRBCs were collected at the product outlet, and the bound RNA recovered, amplified by RT-PCR then transcribed in vitro using standard procedures in preparation for the next round. A total of five rounds of selection were completed. The bulk initial library and the Rounds 1-5 selected pools were evaluated for thrombin binding by bio-layer interferometry using biotinylated thrombin immobilized on streptavidin probes. The Round 5 selected pool exhibited thrombin binding (Kd ~ 4 nM), similar to Toggle-25 (FIG. 5A).

26 and 17 clones were sequenced from the Rounds 3 and 5 selected pools, respectively (Table 1), and these sequences were analyzed using MEME. See Bailey TL, Elkan C (1994) Proc Int Conflntell Syst Mol Biol 2: 28-36. The sequence of variable regions from clones isolated after three and five rounds of iRBC I-SELEX are shown in Table 1 below. By MEME analysis, these are grouped into three clusters based on the presence of a GUUACUG (SEQ ID NO: 1) motif (Cluster I), a UUACCCAAG (SEQ ID NO:2) motif also present in Toggle-25 (Cluster II) or absence of a recognizable conserved motif (Cluster III). The number of times a clone was represented in the sequenced pool is indicated in parentheses following the clone ID. Dissociation constants for binding to thrombin are shown for tested clones (Clone ID in bold text).

Table 1 Sequence of variable region

Cluster I

GUUACUG AUCUUCCUGCAGCGCGAAUCACAUGUAUGAAGCCGGAUCGACG

(SEQ ID NO:3)

A GUUACUG CGCUCUUACGAGGUAACUACUUAGUUGGCAUUACGUAGUACU

(SEQ ID NO:4)

A GUUACUG AGCUCUUGUGUGUUACAGUUGAGAAUCACAACGAUUCCCUG

(SEQ ID NO: 5)

A GUUACUG GGUUCCCCUUCCCACACGCCAUCAUCGUAUGCCGGCAAACGA

(SEQ ID NO:6)

A GUUACUG AGUUCCGGGGAGAGGGGUAGAGCUCUACCGCCCAAAUAGCGA

(SEQ ID NO:7)

GUUACUG GGUUCCGCACAAAUGAGAUUUAUGUUUUUCUAAUCUGCCUCA

(SEQ ID NO: 8)

GUUACUG AGAGCCCUUGACCUCUGGAGCCCACGACGUCGUGAAUAUGAGG

(SEQ ID NO:9)

GUUACUG ACACUCCCCGUUGGUGCGAAGCACCAAGUAACGACAGACUCAG

(SEQ ID NO: 10)

GUUACUG AGCUCAGUCGGGGUGACGCGUCACCCUCUAGGAGAGACUCUGU

(SEQ ID NO: 11)

A GUUACUG AUGUCCCCGAACUGGUGUGGCAAUGGCAGACAUUGCAGCAGG

(SEQ ID NO: 12)

A GUUACUG CGCUCCUAUACUGGCUCAUUGUGCCCGGCAUCGAGGACAUUG

(SEQ ID NO: 13)

GUUACUG AACUCUCGCCGUGGCGCCACGUUGAUAUGCAGUCGACUUCAAG

(SEQ ID NO: 14)

Cluster II

GAACAAAGCUGAAGUAC UUACCCAAG AUCAUCCCGA (SEQ

ID NO: 15)

UGAUGUAAGACUGUUACUUGUGUGUUA UUGCCCAAG UUUGGUCUGUAUUG

(SEQ ID NO: 16)

UGGUGACUGGGUGAGAAUAGCAUU UUGCCCUAG

UCAUUCGAGUCUACAAC (SEQ ID NO : 17)

CUGCCUGAGACUGCACUUUUUCGCUCC UUCCCCAAG UCUUUUUGCGAGUU

(SEQ ID NO: 18)

CGAAUGCCGAAAUUAGAGCGUG UUCUCCAAG CCCGACUCGUACGGCUCGA (SEQ ID NO: 19)

Cluster III

UCAAAUUAGUGCGCGAGGACUAUUAGGACGUCCAGUACAUAAUAUCAGAA

(SEQ ID NO:20)

C G AAAAAC AGUAUUAC G AUG C AUC CUUC GUAAUC C AC CUUG AAAUAAAAC

(SEQ ID NO:21 )

UAAGUUGCAGUAAUGGAGUUC CUGAAAUGGACGUACGGUC CUAUU (SEQ ID

NO:22)

UAGUCAUUGGUUAAAAACCAGUCUCCAAUCGAAUAUUACCGGUUAUGUUU

(SEQ ID NO:23)

CAACGGAUCACACGCAAGCGAAUCUAUGUCCUUCGGGUCUAAGGGUCGCU

(SEQ ID NO:24)

UCCAUUUCUUCUGAAACGAUGAGCAC (SEQ ID NO:25)

-9 ACGAACGUACAAGUUAUGAGAACUGACCUCCCUAUUGUCAUCCACACUCA

(SEQ ID NO:26)

-1 1 ACUUAAGAAUCGACCUGGGGUGACAAUUAAUCCUUCUUUGUUUUGUUAAC (SEQ ID NO:27)

3- -12 200 GAUCAAAUUUAAUAUAUCCGAUCUGGAGAUCUAAUAUCAUAUUGUAGUGG

(SEQ ID NO:28)

13 ACAGGUUUAUAAAGUUUUCGUUUCGAUGGAUUUUAAAACAAUCUAGAUCC

(SEQ ID O:29)

3- -14 AUAAGCUGAUUUGUUCAGAACAUCUUAAUAUACAGUCACACAGGCUGCAG

(SEQ ID NO:30)

3- -15 CCAUUUUAAGGAAUCACGCUGCCCCUACCCUGCGGUUCAGCCGACCUUU (SEQ

ID NO:31)

3- -16 GGUUUUAAUUUAGUACUUCUCUAUAGUUUUACUAGCCUUGAACAUUGGGA

(SEQ ID NO:32)

3- -17 CUUUGCCUUACAUCAAUAUAGGUCAACAUAAGUAGUGUAACGUUUGGUCA

(SEQ ID NO:33)

3- -18 UACCUGUUAAGAAUGAGUUCAAGAGUAAAAGUUAUGUCCUGCAUGUCUAG

(SEQ ID NO:34)

3- -24 UGGGAGAC CG AUAUAAGCAAUGUAUC AC CUGUGCGC AUUGAAGUCGGACU

(SEQ ID NO:35)

3· -26 CAUUUGCGCGAAUCAGGCUUGCUUGUGUUAACACUGCAACGGUUUUAUCG

(SEQ ID NO:36)

3- -27 CCGCGGACUUGCACUGUACCGCGUACCUUGCCUAAUUGUUCUGUAAAGAA

(SEQ ID NO:37)

3- -28 AAGGUUGAUGGCGUUGAUUCUUAGUCAUUAAUCGUUGUCGUUAAAGACGU

(SEQ ID NO:38)

3- -29 AUGAUUGAUUAUCAUGAAAAUGGCUGGUGGUAAACAUCACCUGAUAUUG (SEQ

ID NO:39)

5- -7 UUGCACCAAGGUUCGGAAAUUGAGCGGGCGACGCGUCGAAUGCCGCUAAU

(SEQ ID NO:40)

The majority (86%) of the sequences were unique. One clone (3-19) was duplicated, and another was represented four times (3-7, 5-3, 5-9 and 5-16). Clones could be broadly grouped into three clusters based on the presence of a MEME- identified motif or lack thereof. Cluster I sequences contained a conserved

GUUACUG{AIGI (SEQ ID NO:41) motif (Motif 1), and 2/25 and 10/13 (31% overall) of the unique clones recovered from rounds 3 and 5, respectively, fell into this group. Clones 5-5 and 5-12, arbitrarily chosen as representatives from this cluster, were identified as high affinity thrombin aptamers by bio-layer

interferometry (IQ ~ 2 nM for both). 5-12 binding to ^RBCs and sRBCs was tested. Similar to Toggle-25, 5-12 exhibited significant binding to tRBCs but not sRBCs as expected based on their thrombin-specific binding properties (FIG. 5B). To understand whether Motif 1 is involved in binding, secondary structures predicted by mfold were used to guide the design of truncated or mutated versions of 5-12 (FIG. 5Q. See Zuker M (2003) Nucleic Acids Res 31 : 3406-3415. Motif 1 is predicted to lie within a stem-loop region of these aptamers. 5-12mini was generated by truncating the parent aptamer while preserving the predicted stem-loop structure containing the conserved motif, and high-affinity thrombin binding was determined to be retained (¾ ~ 5 nM). In the context of the full-length aptamer, the motif was mutated while preserving the stem-loop element to produce 5-12mui. This mutated aptamer exhibited significantly reduced thrombin binding (¾ > 3 μΜ) (FIG. 5D). Taken together, these data indicate that the conserved motif present in Cluster I aptamers may be required for high affinity binding to thrombin.

A sequence identical to Toggle-25 was not present amongst the clones sampled, and Motif I is distinct from that implicated in Toggle-25 binding to thrombin. See White R, et al. (2001) Mol Ther 4: 567-573. However, Cluster II clones contained a UU[G/C/A]YCC[A/U]AG (SEQ ID NO:42) motif (Motif 2) that is shared between Toggle-25, 3-4, 3-10, 3-19 and 5-6. The representative 3-19 exhibited very low binding affinity to thrombin (> 1 μΜ). Thus, while this conserved feature may confer some degree of binding to thrombin, it appears insufficient for mediating the high affinity binding observed for Toggle-25. Cluster III consists of sequences that do not contain the above motifs or another sequence feature that is strongly conserved amongst the group. 3-7, 3-9 and 3-12 were tested for thrombin binding. Although 3-7 was represented by four separate clones, it did not detectably bind thrombin. It is possible that this sequence persisted due to a replication advantage rather than target-binding affinity. Clone 3-12 bound thrombin with modest affinity (Kd ~ 200 nM ), while 3-9 bound with high affinity (¾ ~ 4 nM). Altogether, using I-SELEX with a complex whole cell target, a new class of high-affinity binding thrombin aptamers that are sufficiently enriched from a starting library of ~1014 sequences was identified within three rounds of selection. It is also worth noting that, while the I-SELEX partitioning step is stringent, aptamers with affinities varying by 100-fold (-2-200 nM) can be recovered. Thus, without having to devise alternative partitioning protocols, I-SELEX may be broadly useful in recovering both high and modest affinity aptamers to a given target, thus expanding the option space for identifying the most suitable aptamer for the intended application. Discussion

The application of inertial micro fluidics principles was demonstrated in a spiral device capable of rapidly and stringently resolving micron-sized particles (RBCs) from macromolecules (a nucleic acid library) to successfully perform whole cell-SELEX. This device functions with very high partitioning efficiency (> 106) and affords high affinity aptamers in as few as three rounds of selection. This device provides a generic strategy for effectively completing the critical partitioning step in SELEX, and has the advantage of being equally applicable to bead-immobilized targets and directly to whole cells without the need for modification. The theoretical principles guiding device design and operation are sufficiently well understood such that new designs to accommodate particles of different sizes can be achieved.

Alternatively, standardization of the partitioning step using a single device design is feasible under optimal operating conditions as the device has sufficient tolerance to accommodate a range of particle sizes with no adverse impact on resolving particles and macromolecules into distinct output streams.

The microfluidics partitioning strategy described herein takes advantage of the combined effects of inertial focusing of ^RBCs and well-controlled Dean migration of unbound nucleic acid library along the channel midline in spiral microchannels. By choosing a low-aspect ratio channel design, the resolving power of this device is significantly enhanced such that highly efficient separation of micron-sized particles (VRBCs) from macromolecules (nucleic acid library) is attainable. Remarkably, such stringent partitioning occurs extremely rapidly, as a particle or cell spends less than a second within the device under standard operating conditions. Additional particle/cell rotation due to high-shear gradients and secondary Dean flows near the channel wall might play a role in enhancing removal of weakly-bound aptamers from target cells. While tRBCs remained focused near the inner wall, Dean vortices establish a lateral shear gradient at their focusing positions, which may cause iRBCs to rotate as they flow along the channel. This hypothesis is supported by a recent study showing that the combination of cell rotation and transverse motion in a spiral channel enhanced transfection efficiency via more homogenous electroporation of individual cells. See Wang J, Zhan Y, Ugaz VM, Lu C (2010) Lab on a Chip 10: 2057. By varying these parameters, the degree of Dean flow-induced target cell/particle surface "washing" achieved during I-SELEX can potentially be tuned. This could allow selection stringency to be modulated with greater flexibility than is typical in other microfluidics-based aptamer selection devices.

Attaining a generally applicable and highly efficient partitioning strategy can be readily and synergistically integrated with other recent advances aimed at improving the efficiency, reproducibility and likelihood of routinely discovering target aptamers with the desired target-interaction profiles. For example,

improvements in library design and chemical diversity can enhance discovering high affinity aptamers to a more diverse set of targets. See Ruff KM, Snyder TM, Liu DR (2010) J Am Chem Soc and Davis JH, Szostak JW (2002) Proc Natl Acad Sci USA 99: 1 1616-1 1621 ; Keefe AD, Cload ST (2008) Current opinion in chemical biology 12: 448-456 and Vaught JD, et al. (2010) J Am Chem Soc 132: 4141 -4151 ; Gold L, et al. (2010) PLoS ONE 5: el 5004. Similarly, high-throughput DNA sequencing holds promise for more comprehensively deconvoluting selected libraries 'in real time' . Through systematic round-by-round analysis of selected libraries, it has been observed that high affinity aptamers are frequently sequences exhibiting the highest enrichment between consecutive rounds. See Zimmermann B, et al. (2010) PLoS ONE 5: e9169, Gold L, et al. (2010) PLoS ONE 5: el 5004 and Hoon S, et al. (201 1) BioTechniques 51 : 413-416. The option to deep sequence a selected library may be viewed as a simple solution to circumvent developing highly efficient partitioning strategies for SELEX. See Hoon S, et al. (201 1)

BioTechniques 51 : 413-416 and Kupakuwana GV, Crill JE, McPike MP, Borer PN (201 1) PLoS ONE 6: el 9395. However, as observed by Kupakwana et al, when partitioning was achieved using a low efficiency resin-based strategy, a substantial number of recovered sequences reflected inefficiency during the partitioning step. Furthermore, deep sequencing and follow-up characterization of putative aptamers require investing significant resources. Therefore, a highly efficient partitioning strategy significantly increases the probability that identified sequences are indeed aptamers. This strongly favors integrating high-efficiency partitioning with deep sequencing to improve the success rate achieved in de novo aptamer discovery. A simple, reusable and broadly applicable inertial microfluidics device has been designed and validated as an efficient way of achieving stringent, single pass library partitioning during SELEX. The device can be inexpensively fabricated and routine operation is straightforward. While using a single device for a single target has been demonstrated above, this approach is amenable to both automation and multiplexing. Combined with improvements in starting library design and synthesis, and selected library deconvolution procedures, this facilitates more robust and higher throughput selections in the future. METHODS

Device Fabrication & Flow Conditions

Microfluidic devices were fabricated in polydimethylsiloxane polymer (PDMS, Sylgard 184, Dow Corning, USA) using the double molding process reported previously. See Hou HW, et al. (2010) Lab Chip 10: 2605-2613. Briefly, the patterned silicon wafers were silanized with trichloro(lH,lH,2H,2H- perfluorooctyl)silane (Sigma Aldrich, USA) for 1 hr and PDMS prepolymer mixed in 10: 1 (w/w) ratio with curing agent was poured onto the silanized wafer and baked at 80 °C for 1 hr. The cured PDMS mold then acted as a template for subsequent PDMS casting (negative replica). The PDMS master template was silanized for 1 hr before use to aid release of subsequent PDMS microchannels. Finally, holes (1.5 mm) for inlets and outlets were punched and the PDMS microchannels were irreversibly bonded to microscopic glass slides using an air plasma machine

(Harrick Plasma Cleaner, USA) and left for 2 hr at 70 °C to complete the bonding.

Each sample mixture was pumped into the inner inlet at 150 μΐ, min"1 while sheath buffer (Thrombin Binding Buffer or TBB = 20 mM HEPES pH 7.4, 150 mM NaCl, 2 mM CaCl2) was pumped through the outer inlet at 1500 yL min"1. Sample at the product outlet (cells and aptamers) was collected after 1.5 minutes of run time to allow for establishment of Dean vortices along the channel length.

Microscopy Samples (107 cells/mL) and sheath buffer (TBB) were pumped through the microfluidic device using two syringe pumps (NE-1000, New Era Pump Systems Inc., USA) and the ratio between the sample and sheath buffer flow rates was fixed at 1 : 10. For image acquisition, the microchannels were mounted on an inverted phase contrast microscope (Olympus 1X71) equipped with a Hamamatsu Model C4742-80-12AG CCD camera (Hamamatsu Photonics). IPLab (Scanalytics) software was used for video acquisition and captured videos were analyzed using ImageJ® software. For confocal microscopy, images were acquired at 10X magnification (488 nm excitation) at an interval of 2 μπι in z-axis (vertical axis) using LSM 510 meta laser scanning confocal (Carl Zeiss, Jena, Germany) mounted on an Axiovert 100M Inverted Microscope (Carl Zeiss, Microimaging).

Cell Synthesis and Quality Control

Human PvBCs were washed twice with 10 mL PBS, pH 8, and 2 x 109 were resuspended to 5% hematocrit in 4 mL PBS, pH 8 with 1 mg EZLink Sulfo-NHS- Biotin (Thermo Scientific). Cells were incubated for 30 minutes rotating at room temperature (25 °C) then washed twice with 10 mL PBS, pH 7.4. Biotinylated RBCs were resuspended to 5% hematocrit in PBS, pH 7.4 and incubated with 200 μg streptavidin (ThermoScientific) for 30 minutes at room temperature. Half of this cell suspension was directly transferred to a fresh tube and incubated with 100 of biotin-saturated PBS, pH 7.4 for 15 minutes at room temperature to cap free biotin- binding sites and create sRBCs. The other half of the cell suspension was incubated with thrombin-BFPRck (Haematologic Technologies Inc.) for 30 minutes prior to biotin capping to create tRBCs. tRBCs were synthesized with surface thrombin concentrations spanning 0.1-1000 nM. All cells were stored at 4 °C between use in Gibco RPMI 1640 media (Life Technologies) supplemented with 24 mM HEPES pH 7.4, 2 g/L sodium bicarbonate, 0.25% Albumax II (Life Technologies), 0.1 mM hypoxanthine and 50 mg/L gentamicin. Effective surface thrombin concentrations were determined by nonlinear regression analysis of titrated monoclonal anti- thrombin antibody (Haematologic Technologies) labeled with a 1 : 1000 dilution of Alexa Fluor 488 goat anti-mouse secondary antibody (Life Technologies) before FACS analysis (FL1 channel). Binding data were fit to the following model: Y = Bmax*X/(X + Kd) + NS*X + YO, where Y is fluorescence signal (RFU), Bmax is the total number of antibody-accessible binding sites, X is the number of antibody molecules added to the solution, NS is the non-specific binding component, and YQ is the background fluorescence.

Aptamer Design & Synthesis

The DNA template for the thrombin-binding Toggle-25 aptamer was assembled by PCR using primers CMB49, CMB50, CMB51 and CMB56 (Table 2). See White R, et al. (2001) Mo I Ther 4: 567-573.

Table 2. List of primers used in this work

CMB49 CCGCTCGAGTAATACGACTCACTATAGGGAGAGAGGAA

(SEQ ID NO:43)

CMB50 CTTCAGCTTTGTTCCCCATCCCTCTTCCTCTCTCCCTATAGT

GAGTC (SEQ ID NO:44)

CMB51 GATGGGGAACAAAGCTGAAGTACTTACCCAAGATCATCCC

GAACGA (SEQ ID NO:45)

CMB56 TCGACCTCTGGGTTATGTCGTTCGGGATGATCTTGGG (SEQ

ID NO:46)

CMB77 GGGAGAGAGGAAGAGGGATGGG (SEQ ID NO:47)

CMB94 TCGACCTCTGGGTTATGTCGTTC (SEQ ID NO:48)

CMB95 CCGCTCGAGTAATACGACTCACTATAGGGAGACAAGAT

(SEQ ID NO:49)

CMB96 GCAGCGTTCCTCGATGGCCTTGATCTTGTCTCCCTATAGTG

AGTCG (SEQ ID NO:50)

CMB97 CATCGAGGAACGCTGCCTACACACATGGACATAAACAAGA

GCCGA) (SEQ ID NO:51)

CMB98 CCATTTTCTCGCCCTCTTCGGCTCTTGTTTATGTCCATG

(SEQ ID NO:52)

CMB104 TTCTCGCCCTCTTCGGCTCTTGTTTA (SEQ ID NO:53)

CMB106 GAGACAAGATCAAGGCCATCGAGGAA (SEQ ID NO:54)

CMB107 CACTATAGGGAGCTCAGAATAAACGCTCAAACGAACGTAC

AAGTTATGAGAACTG (SEQ ID NO:55)

CMB108 CGAATGAGTGTGGATGACAATAGGGAGGTCAGTTCTCATA

ACTTGTACGTTCGT (SEQ ID NO:56)

CMB109 GCCGGATCCGGGCCTCATGTCGAATGAGTGTGGATGACAA

TAGG (SEQ ID NO:57)

CMB1 12 CACTATAGGGAGCTCAGAATAAACGCTCAACAACGGATCA

CACGCAAGCGAA (SEQ ID NO:58)

CMB1 13 GCGACCCTTAGACCCGAAGGACATAGATTCGCTTGCGTGT

GATCCGTTGTTG (SEQ ID NO:59)

CMB114 GCCGGATCCGGGCCTCATGTCGAAAGCGACCCTTAGACCC GAAGGACATAGA (SEQ ID NO:60)

CMB116 AGTCTCTCCTAGAGGGTGACGCGTCACCCCGACTGAGCTCA

GTAACTTGA (SEQ ID N0:61)

CMB1 17 GCCGGATCCGGGCCTCATGTCGAAACAGAGTCTCTCCTAG

AGGGTGACGCGT (SEQ ID NO:62)

CMB1 19 CGAACCACTACAATATGATATTAGATCTCCAGATCGGATAT

ATTAAATTTGATCTTG (SEQ ID NO:63)

CMB120 GCCGGATCCGGGCCTCATGTCGAACCACTACAATATGATAT

TAGATCT (SEQ ID NO:64)

CMB122 CTCGCAAAAAGACTTGGGGAAGGAGCGAAAAAGTGCAGTC

TCAGGC (SEQ ID NO:65)

CMB123 GCCGGATCCGGGCCTCATGTCGAAAACTCGCAAAAAGACT

TGGGGAAGGA (SEQ ID NO:66)

CMB 125 TCGCTATTTGGGCGGTAGAGCTCTACCCCTCTCCCCGGAAC

TCAGTAACTTTGAG (SEQ ID NO:67)

CMB126 CACTATAGGGAGCTCAGAATAAACGCTCAAGTTACTGAGC

TCAGTCGGGGT (SEQ ID NO:68)

CMB127 GCCGGATCCGGGCCTCATGCTGAGCTCAGTAACTTGAGCGT

TTATTCTGAGCTCCCTAT (SEQ ID NO:69)

CMB130 CACTATAGGGAGCTCAGAATAAACGCTCAAACCGATGAGC

TCAGTCGGGGT (SEQ ID NO:70)

CMB131 AGTCTCTCCTAGAGGGTGACGCGTCACCCCGACTGAGCTCA

TCGGTTTGA (SEQ ID N0:71)

CMB132 GCCGGATCCGGGCCTCATGTCGAAACAGAGTCTCTCCTAG

AGGGTGACGCGT (SEQ ID NO :72)

CMB133 ATAGGGAGCTCAGAATAAACGCTCAAGATCAAATTTAATA

TATCCGATCTGGAGATCTAA (SEQ ID NO:73)

CMB134 CTATAGGGAGCTCAGAATAAACGCTCAACTGCCTGAGACT

GCACTTTTTCG (SEQ ID NO:74)

CMB 139 CTCTCCCCGGAACTCATCGGTTTTGAGCGTTTATTCTGAGC

TCCCTATAGTG (SEQ ID NO:75)

CMB140 TCGCTATTTGGGCGGTAGAGCTCTACCCCTCTCCCCGGAAC

TCATCGGTTTTGAG (SEQ ID NO:76)

CMB141 GCCGGATCCGGGCCTCATGTCGAATCGCTATTTGGGCGGTA

GAGCTC (SEQ ID NO:77)

CMB142 CACTATAGGGAGCTCAGAATAAACGCTCAAAGTTACTGAG

TTCCGGGGAGAG (SEQ ID NO:78)

CMB152 CTGGTCATGGCGGGCATTTAATTCGGGCCTCATGTCGAAAC

AGAG (SEQ ID NO: 79)

CMB153 CTGGTCATGGCGGGCATTTAATTCTCGACCTCTGGGTTATG

TCGTT (SEQ ID NO:80)

CMB154 CTGGTCATGGCGGGCATTTAATTCCCATTTTCTCGCCCTCTT

CG (SEQ ID NO:81)

SELEX CCGAAGCTTAATACGACTCACTATAGGGAGCTCAGAATAA

Library ACGCTCAA- [N50] -TTCGACATGAGGCCCGGATCCGGC (SEQ template ID NO:82) SELEX CCGAAGCTTAATACGACTCACTATAGGGAGCTCAGAATAA

Forward ACGCTCAA (SEQ ID NO: 83)

Primer

SELEX GCCGGATCCGGGCCTCATGTCGAA (SEQ ID NO: 84)

Reverse

Primer

Capture Biotin-CTGGTCATGGCGGGCATTTAATTC (SEQ ID NO:85) Oligo scr Toggle-25, a non-binding variant was prepared by scrambling the thrombin-binding region of Toggle-25 and installing unique primer binding sites while retaining overall length and (G+C) content. The scrToggle-25 DNA template was PCR-assembled using primers CMB95, CMB96, CMB97 and CMB98. Four aptamer clones from Round 3 were PCR-assembled using the SELEX Forward Primer and the following sequence-specific primers: 3-7 (identical to 5-3), CMBl 12, CMB l 13, CMB l 14; 3-9, CMB107, CMB108, CMB 109; 3-12, CMB133, CMBl 19, CMB 120; and 3-19, CMB134, CMB 122, CMB123. Three aptamer clones from Round 5 were PCR assembled using the SELEX Forward Primer and the following sequence-specific primers: 5-3 (see above); 5-5, CMB142, CMB125, CMB141 ; 5- 12, CMB126, CMBl 16, CMBl 17. Mutant 5-12 {S-Umut) templates were synthesized with SELEX Forward Primer and sequence-specific primers CMBl 30, CMB 131 and CMB l 32. The minimized aptamer 5-12mini) was synthesized from the DNA template assembled from the SELEX Forward and CMBl 27 primers. All PCR reactions were performed identically with Phusion High-Fidelity PCR polymerase (New England BioLabs) according to manufacturer's instructions, with the exception that internal primers were used at 10-fold lower concentrations than the external primers.

Partitioning Efficiency & Enrichment Experiments

For partition efficiency experiments, 100 nM of either Toggle-25 or scrToggle-25 was incubated in 1 mL TBB with 107 tRBCs for 20 minutes at room temperature with continuous gentle inversion before being passed through the I- SELEX device. tRBCs were recovered directly from the sample outlet onto a vacuum filter plate membrane (Millipore MSHVS4510). Bound aptamers were eluted (off-vacuum) by a five minute incubation with 1 mM EDTA in PBS, pH 7.4 followed by plate centrifugation. Quantitative polymerase chain reaction (qPCR) with SYBR Green was used to determine absolute levels of Toggle-25 and •scrToggle-25 recovered after device partitioning. Recovered aptamers were ethanol precipitated and reverse transcribed using Superscript III Reverse Transcriptase (Life Technologies) with either Toggle-25 reverse primer CMB94 or scrToggle-25 reverse primer CMB104. The partner forward primers (CMB77 and CMB106, respectively) were added with 1 x SYBR Green for qPCR quantitation using 40 cycles of amplification on a Light Cycler 480 (Roche Applied Science). For enrichment experiments, 100 nM total aptamer library (either 90% or 99% scrToggle-25) was incubated with 107 ^RBCs for 20 minutes at room temperature with continuous gentle inversion before being passed through the device. In these enrichment experiments, where the recovered aptamer sample contained two species, samples were divided in half after recovery and precipitation, followed by aptamer-specific reverse transcription and qPCR, as described above.

I-SELEX

The DNA template for the random library containing a 50 nucleotide randomized region was chemically synthesized (Integrated DNA Technologies). The primers used for RT and PCR amplification of the library are summarized in Table 2. A 2'-fluoro-pyrimidine RNA library containing approximately 3xl014 unique sequences was synthesized using the DuraScribe T7 Transcription Kit (Epicentre Biotechnologies). RNA was denatured at 65 °C for 5 minutes, followed by refolding for 10 minutes at room temperature. The aptamer library was then incubated with 108 sRBCs in 1 mL TBB for 30 minutes at room temperature with continuous gentle mixing. sRBCs and the aptamers bound to them were removed by centrifugation. The supernatant was incubated with 106 /RBCs for 60 minutes at room temperature with continuous gentle mixing. After incubation, the entire binding reaction was partitioned in a single pass through the I-SELEX device at a flow rate of 150 μΕ min"1 with sheath buffer (TBB) pumped at 1500 μΕ min"1. Cells were recovered and aptamers eluted as described above for enrichment experiments, above. The recovered aptamers were RT-PCR amplified using an empirically determined minimum number of PCR cycles, minimized to prevent formation of truncation or PCR giant products (14 cycles on average). Half of the recovered PCR product (~1 μg) was stored, and the remainder used for in vitro transcription (37 °C for 6- 15 hours) to prepare RNA library for the subsequent round. Remaining template DNA was removed by Turbo DNase (Life Technologies). RNA was purified by phenol- chloroform extraction and ethanol precipitation, then re-dissolved in ddH20. Five rounds of selection were performed.

Flow Cytometry

To test binding of Toggle-25 to iRBCs versus sRBCs, Toggle-25 was fluorescently labeled by addition of a single 3'-amino-2',3'-ddATP (TriLink

Biotechnologies) by Poly(A) Polymerase (New England BioLabs) followed by incubation with DyLight 488 NHS ester (Pierce). Toggle-25 was titrated against 106 cells in 200 μΕ TBB containing 0.1% BSA for 3.5 hours at room temperature. Cells were washed twice with TBB + 0.1% BSA prior to resuspension in 90 iL TBB followed by FACS analysis (BD Accuri C6 flow cytometer, FL1 channel).

To test binding of I-SELEX aptamer 5-12 to iRBCs, Toggle-25, scr Toggle- 25 and 5-12 were synthesized with a 3' 24 nucleotide

(GAAUUAAAUGCCCGCCAUGACCAG) (SEQ ID NO: 86) using PCR extension oligos CMB153, CMB 154 and CMB 152, respectively. See Li N, et al. (2009) J Proteome Res 8: 2438-2448. A capture oligonucleotide (5 '-biotin-

CTGGTCATGGCGGGCATTTAATTC) (SEQ ID NO: 87) complementary to the aptamer extension was synthesized and fluorescently labeled with streptavidin- phycoerythrin (SP). Aptamers were incubated with equimolar concentrations of capture oligonucleotide for 5 minutes at 75 °C then cooled slowly to 4 °C. Aptamer- capture oligonucleotide and SP were pre-complexed by incubating them in a 1 :2 molar ratio for 15 minutes at room temperature. Excess biotin (20 μΕ biotin- saturated PBS, pH 7.4) was added for 15 min at room temperature to cap any free biotin-binding sites. The aptamer-capture oligonucleotide-SP complex (100 nM) or capture oligonucleotide-SP complex (100 nM) was incubated with either 106 tRBCs or i'RBCs in 100 μΕ TBB containing 1 μg BSA and 0.25 μg yeast tRNA for 1 hour at room temperature. Cells were washed with TBB and resuspended in 50 μΐ, TBB prior to FACS analysis (FL2 channel). Bio-Layer interferometry binding studies

Aptamer binding kinetics were determined using a BLItz machine and Streptavidin (SA) Dip and Read biosensors (Forte Bio). TBB was used for probe hydration (10 min), baseline readings, aptamer dilutions, and dissociation steps.

Thrombin-BFPRck (Haematologic Technologies Inc.) at 10 nM was immobilized on a SA biosensor via a 300 s loading incubation. Aptamers were in vitro transcribed, DNase treated, and refolded as described above, then incubated with thrombin- loaded SA biosensors (association: 180 s; dissociation: 180 s) at various

concentrations. Probes were regenerated using the following protocol: 60 s incubation in regeneration buffer (1 M NaCl, 1 niM NaOH) followed by 3 x 60 s incubations in TBB. The probe was regenerated at least once prior to collection of binding data using BLItz software for analysis. The relevant teachings of all patents, published applications and references cited herein are incorporated by reference in their entirety.

While this invention has been particularly shown and described with references to example embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.

CLAIMS

What is claimed is:

1. A method of separating one or more binding molecules that are specifically bound to one or more binding targets from a mixture of bound and unbound molecules, comprising introducing the mixture into at least one inlet of a microfluidic device at a sample flow rate, the microfluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate the one or more molecules bound to the binding target along portions of the cross-section of the channel based on size, wherein bound molecules flow along a radially innermost portion of the channel to a first outlet, and unbound molecules flow along one or more other portions of the channel to at least one other outlet, thereby separating the one or more binding molecules that are specifically bound to the one or more binding targets from the mixture of bound and unbound molecules.

2. The method of claim 1, wherein the one or more binding molecules comprise one or more nucleic acids, peptides, one or more organic small molecules, or a combination thereof.

3. The method of claim 2, wherein the one or more organic small molecules comprises an antibiotic, a fluorophore or a combination thereof.

4. The method of any one of claims 1 -3, wherein the one or more binding targets comprise all or a portion of a cell, a cell membrane, a receptor, a protein, an antibody, a supramolecular structure, or a combination thereof.

The method of claim 4, wherein the cell is a red blood cell, a mammalian cell, an eukaryotic cell, or a prokaryotic cell.

6. The method of any one of claims 4 or 5, wherein the cell is a thrombin- displaying red blood cell.

7. The method of claim 4, wherein the supramolecular structure is a scaffold, a bead, a liposome, an organelle, or a combination thereof.

8. The method of any one of the preceding claims, further comprising

introducing buffer into one or more inlets of the microfluidic device at a buffer flow rate.

9. The method of claim 8, wherein the buffer flow rate is higher than the

sample flow rate.

10. The method of any one of claims 8 or 9, wherein the buffer flow rate is at least about three times greater than the sample flow rate. 1 1. The method of any one of claims 8-10, wherein the buffer flow rate is in a range of between about 500 μϋ/ηύη and about 2,000 μΤ/πιϊη.

12. The method of any one of claims 1-11, wherein the sample flow rate is in a range of between about 50 μί/ηιίη and about 200 μΤ/τηϊη.

13. The method of any one of claims 8-12, wherein the sum of the buffer flow rate and the sample flow rate is equal to an overall flow rate that yields a

Reynolds number equal to or greater than about 50.

14. The method of any one of claims 8-13, wherein the buffer flow rate is about ten times greater than the sample flow rate.

15. The method of any one of the preceding claims, wherein the curved channel is a spiral channel.

16. The method of claim 15, wherein the spiral channel is a bi-loop spiral

channel.

17. The method of any one of claims 15 or 16, wherein the radius of the spiral channel is adapted to yield a Dean number in a range of between about 1 and about 10.

18. The method of any one of claims 15-17, wherein the length of the spiral channel is equal to or greater than about 3 cm.

19. The method of any one of claims 15-18, wherein the width of the spiral channel is in a range of between about 100 μηι and about 1,000 μπι. 20. The method of any one of claims 15-19, wherein the height of the spiral channel is in a range of between about 20 μη and about 200 μιη.

21. The method of any one of the preceding claims, wherein the aspect ratio of the channel is in a range of between about 0.1 and about 1.

22. The method of any one of the preceding claims, wherein the one other outlet has a diameter that is greater than the diameter of the first outlet.

23. The method of any one of the preceding claims, further comprising

collecting the one or more binding molecules that are specifically bound to the one or more binding targets from the first outlet.

24. The method of any one of the preceding claims, wherein a binding affinity of the bound molecules is in a range of between about 2 nM and about 1 μΜ.

25. The method of any one of the preceding claims, wherein the mixture is

passed through the device one or more times, thereby enriching the sample of the one or more binding molecules that are specifically bound to the one or more binding targets. 26. The method of claim 25, wherein the mixture is passed through the device one or more times by reintroducing the bound molecules into the inlet of the microfluidic device.

27. The method of any one of the preceding claims, wherein a partitioning

efficiency of isolating bound molecules from unbound molecules is at least about 1 ,000,000. The method of any one of the preceding claims, wherein the other portions of the channel are the radially outermost portions of the channel.

The method of any one of the preceding claims, further comprising separating the one or more bound molecules from the one or more binding targets.

The method of claim 29, wherein the one or more binding molecules are nucleic acids, and the method further comprises amplifying the one or more bound molecules.

The method of claim 30, wherein the one or more bound molecules is amplified by polymerase chain reaction (PCR), reverse transcriptase PCR, or a combination thereof.

The method of any one of claims 30 or 31, further comprising sequencing the one or more bound molecules.

The method of any one of the preceding claims, further comprising contacting the bound and unbound binding molecules with the binding target prior to introducing the mixture into the microfluidic device.

The method of any one of the preceding claims, wherein, prior to introducing the mixture into the device, the one or more binding molecules are contacted with the one or more binding targets, thereby producing the mixture, and the mixture is maintained under conditions in which the one or more binding molecules can bind the one or more binding targets.

A method of identifying one or more binding molecules that specifically bind to one or more binding targets, the method comprising:

a) introducing a mixture of one or more binding molecules and the one or more binding targets into at least one inlet of a microfluidic device at a sample flow rate, the microfluidic device comprising at least one curved channel, wherein each curved channel has a length, a radius, and a cross-section of a height and a width defining an aspect ratio adapted to isolate the one or more molecules bound to the binding target along portions of the cross-section of the channel based on size, wherein bound molecules flow along a radially innermost portion of the channel to a first outlet, and unbound molecules flow along one or moreother portions of the channel to at least one other outlet; and

b) determining the binding molecule that is bound to the binding target, thereby identifying the binding molecule that specifically binds to the binding target. 36. The method of claim 35, wherein the one or more binding molecules

comprise one or more nucleic acids, peptides, organic small molecules, or a combination thereof.

37. The method of claim 36, wherein the one or more organic small molecules comprise an antibiotic, a fluorophore or a combination thereof. 38. The method of any one of claims 35-37, wherein the one or more binding targets comprise all or a portion of a cell, a cell membrane, a receptor, a protein, an antibody, a supramolecular structure, or a combination thereof.

39. The method of claim 38, wherein the cell is a red blood cell, a mammalian cell, an eukaryotic cell, or a prokaryotic cell. 40. The method of any one of claims 38 or 39, wherein the cell is a thrombin- displaying red blood cell.

41. The method of claim 38, wherein the supramolecular structure is a scaffold, a bead, a liposome, an organelle, or a combination thereof.

42. The method of any one of claims 35-41 , further comprising introducing

buffer into one or more inlets of the microfluidic device at a buffer flow rate.

43. The method of claim 42, wherein the buffer flow rate is higher than the

sample flow rate. 44 The method of any one of claims 42 or 43, wherein the buffer flow rate is at least about three times greater than the sample flow rate.

45. The method of any one of claims 42-44, wherein the buffer flow rate is in a range of between about 500 μΙ7ηιίη and about 2,000 μΙ7ηιΐη. 46. The method of any one of claims 35-45, wherein the sample flow rate is in a range of between about 50 μΙ7ηπη and about 200 μΤ/min.

47. The method of any one of claims 42-46, wherein the sum of the buffer flow rate and the sample flow rate is equal to an overall flow rate that yields a Reynolds number equal to or greater than about 50. 48. The method of any one of claims 42-47, wherein the buffer flow rate is about ten times greater than the sample flow rate.

49. The method of any one of claims 35-48, wherein the curved channel is a spiral channel.

50. The method of claim 49, wherein the spiral channel is a bi-loop spiral

channel.

51. The method of any one of claims 49 or 50, wherein the radius of the spiral channel is adapted to yield a Dean number in a range of between about 1 and about 10.

52. The method of any one of claims 49-51 , wherein the length of the spiral channel is equal to or greater than about 3 cm.

53. The method of any one of claims 49-52, wherein the width of the spiral channel is in a range of between about 100 μηι and about 1,000 μηι.

54. The method of any one of claims 49-53, wherein the height of the spiral channel is in a range of between about 20 μηι and about 200 μηι.

55. The method of any one of claims 35-54, wherein the aspect ratio of the channel is in a range of between about 0.1 and about 1.

56. The method of any one of claims 35-55, wherein the one other outlet has a diameter that is greater than the diameter of the first outlet. 57. The method of any one of claims 35-56, further comprising collecting the one or more binding molecules that are specifically bound to the one or more binding targets from the first outlet.

58. The method of any one of claims 35-57, wherein a binding affinity of the bound molecules is in a range of between about 2 nM and about 1 μΜ. 59. The method of any one of claims 35-58, wherein the mixture is passed

through the device one or more times, thereby enriching the sample of the one or more binding molecules that are specifically bound to the one or more binding targets.

60. The method of claim 59, wherein the mixture is passed through the device one or more times by reintroducing the bound molecules into the inlet of the microfluidic device.

61. The method of any one of claims 35-60, wherein a partitioning efficiency of isolating bound molecules from unbound molecules is at least about 1,000,000. 62. The method of any one of claims 35-61, wherein the other portions of the channel are the radially outermost portions of the channel.

63. The method of any one of claims 35-62, further comprising separating the one or more bound molecules from the one or more binding targets.

64. The method of claim 63, wherein the one or more binding molecule are nucleic acids, and the method further comprising amplifying the one or more bound molecules. The method of claim 64, wherein the one or more bound molecules is amplified by polymerase chain reaction (PCR), reverse transcriptase PCR, or a combination thereof.

The method of any one of claims 64 or 65, further comprising sequencing the one or more bound molecules.

The method of any one of claims 35-66, further comprising contacting the bound and unbound binding molecules with the binding target prior to introducing the mixture into the microfluidic device.

The method of any one of claims 35-67, wherein prior to introducing the mixture into the device the one or more binding targets are contacted with the one or more binding molecules being assessed for binding to the binding target, thereby producing the mixture, and the mixture is maintained under conditions in which the one or more binding molecules can bind the one or more binding targets.

Download Citation


Feedback