WO2014130682A2 - Method of analysis using array sensor - Google Patents

Method of analysis using array sensor Download PDF

Info

Publication number
WO2014130682A2
WO2014130682A2 PCT/US2014/017413 US2014017413W WO2014130682A2 WO 2014130682 A2 WO2014130682 A2 WO 2014130682A2 US 2014017413 W US2014017413 W US 2014017413W WO 2014130682 A2 WO2014130682 A2 WO 2014130682A2
Authority
WO
WIPO (PCT)
Prior art keywords
analyte
nanostructures
protein
photoluminescent
sensor
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/US2014/017413
Other languages
French (fr)
Other versions
WO2014130682A3 (en
Inventor
Nigel F. Reuel
Micheal S. STRANO
Ramon WAHL
Bernhard Helk
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Novartis Pharma AG
Massachusetts Institute of Technology
Original Assignee
Novartis Pharma AG
Massachusetts Institute of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Novartis Pharma AG, Massachusetts Institute of Technology filed Critical Novartis Pharma AG
Publication of WO2014130682A2 publication Critical patent/WO2014130682A2/en
Anticipated expiration legal-status Critical
Publication of WO2014130682A3 publication Critical patent/WO2014130682A3/en
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
    • G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
    • G01N33/53—Immunoassay; Biospecific binding assay; Materials therefor
    • G01N33/543—Immunoassay; Biospecific binding assay; Materials therefor with an insoluble carrier for immobilising immunochemicals
    • G01N33/54313—Immunoassay; Biospecific binding assay; Materials therefor with an insoluble carrier for immobilising immunochemicals the carrier being characterised by its particulate form
    • G01N33/54346—Nanoparticles

Definitions

  • the present invention relates to a method of analyzing a product using an array sensor.
  • a method of detecting an analyte can include using an array to provide a plurality of data points that can be used to assess the analyte.
  • a method of detecting an analyte can include contacting a sample including a biologic analyte with a sensor array, the sensor array including a plurality of photoluminescent nanostructures, collecting emissive data from each of the photolummescent nanostructures, and analyzing the emissive data from each nanostructure of the plurality of photolummescent nanostructures to measure a distribution of binding affinities.
  • the method of detecting an analyte can include calculating a calibration curve for the sensor.
  • analyzing the emissive data can include measuring a variance of the emissive data from the plurality of photolummescent nanostructures, measuring a skew of the emissive data from the plurality of photolummescent nanostructures, or measuring a standard deviation of a plume angle of the emissive data from the plurality of photolummescent nanostructures.
  • the method of detecting an analyte can include evaluating a binding heterogeneity of the biologic analyte of the sample from the analyzed emissive data. In some embodiments, the method of detecting an analyte can include evaluating a cell line for production of the biologic analyte of the sample from the analyzed emissive data. In some embodiments, the method of detecting an analyte can include evaluating a glycosylation pattern of the biologic analyte of the sample from the analyzed emissive data.
  • the array can include at least 500 of the photolummescent nanostructures. In some embodiments, the array can include at least 1,000 of the photolummescent nanostructures. In some embodiments, the array can include at least 2,000 of the photolummescent nanostructures. In some embodiments, the array can include at least 5,000 of the photolummescent nanostructures.
  • the senor can include an analyte-binding compound associated with the photolummescent nanostructure.
  • the sensor can include a linker, wherein the analyte-binding compound is associated with the photolummescent nanostructure via the linker.
  • the sensor can include a polymer including clusters of photolummescent nanostructures.
  • the analyte can be a biologic analyte.
  • the biologic analyte can be an antibody.
  • the biologic analyte can be a protein.
  • the antibody can be IgG.
  • the analyte binding compound can be a capture protein comprising Protein A or PSA-lectin.
  • FIGS. 1A-D illustrate nanosensor array fabrication and detection methods.
  • FIG. 1A illustrates a nIR micrograph of nanosensor array - each SWNT-illuminated pixel records changes in intensity upon analyte binding, such as nickel quenching or signal return from IgG adsorption shown.
  • FIG. IB illustrates a multi-layer, poly-acrylamide gel platform used to immobilize the SWNT sensors and provide a porous network for IgG diffusion.
  • SWNT sensors are wrapped in chitosan, nickel-chelated, and functionalized with a capture protein (Protein A or PSA - Lectin).
  • FIG. 1A illustrates a nIR micrograph of nanosensor array - each SWNT-illuminated pixel records changes in intensity upon analyte binding, such as nickel quenching or signal return from IgG adsorption shown.
  • FIG. IB illustrates a multi-layer, poly-acrylamide gel platform used to immobilize the SWNT sensors and provide
  • FIG. 1C illustrates ensemble (i) and all-points histogram (ii) response to lOOmM nickel (1), lmg/ml Protein A (2), 1.5mg/ml Human IgG (3) and 2mg/ml BSA (4) additions.
  • FIG. ID illustrates selectivity of the sensor array (nIR micrograph of starting intensities shown) as demonstrated by a heat map reporting the percent modulation response of the 10,000 top responding pixels after analyte addition.
  • FIGS. 2(a)-(f) illustrate simulation frames.
  • FIGS. 3A-D illustrate modeling a nanosensor array for measuring K D distributions.
  • FIG. 3A illustrates a graph depicting the effect of averaging number of molecules (N) on a single sensor site on measured variance (a m 2 ) assuming a Gaussian shaped K D distribution with known variance (a t 2 ).
  • FIG. 3B illustrates a design regime where a t 2 can be reconstructed showing two spectrum limits (ensemble and single molecule detection) and where the nanosensor platform operates.
  • FIG. 3A illustrates a graph depicting the effect of averaging number of molecules (N) on a single sensor site on measured variance (a m 2 ) assuming a Gaussian shaped K D distribution with known variance (a t 2 ).
  • FIG. 3B illustrates a design regime where a t 2 can be reconstructed showing two spectrum limits (ensemble and single molecule detection) and where the nanosensor platform operates.
  • 3C illustrates modeling nanotube array response - (i) assuming a Weibull (skewed) KD distribution, (ii) Langmuir coverage fraction to determine extent of nanosensor modulation, (iii) resulting simulated response 'plume' normalized by average sensor intensity (I aVg ) where each point represents a sensor site, and (iv) calibration curve from fitting the plume angle mean and standard deviation ( ⁇ ⁇ and ⁇ ⁇ ) with a bivariate Gaussian distribution.
  • FIG. 3D illustrates the effect of changing K D distribution skewness ( ⁇ ) on resulting ⁇ ⁇ and ⁇ ⁇ calibration curves as well as fit parameters 'B' (steepness of calibration curve) and 'C (K D mean).
  • FIGS. 4A-E illustrate experimental results from lyophilized IgG in PBS, murine IgG (TA99) from HEK cells, and human IgG from CHO cells screened on sensor arrays at various concentrations.
  • FIG. 4A are graphs depicting all point sensor response 'plumes' and fit calibration curves.
  • FIG. 4B are graphs depicting the yielding K D 95% confidence intervals, B-fit parameters, and angle saturation values (B SAT ) calculated from the starting intensity distribution of each sensor batch (Eq. 5).
  • FIG. 4C is a graph depicting relation of B-fit parameter and K D skewness parameter ( ⁇ ) as found by simulation results.
  • FIG. 4D is a graph depicting measured K D distributions using assumed Weibull PDF.
  • FIG. 4E is a graph depicting K D histograms directly calculated from sensor response assuming a constant functionalization value - revealing the approximate shape of the true PDF.
  • FIGS. 5A-H illustrate hypermannosylation detection on PSA-lectin conjugated sensor arrays.
  • FIG. 5 A depicts a diagram showing that weaker mannose-PSA lectin interactions can be transduced on independent nanosensors.
  • FIG. 5B are graphs depicting results of testing concept with chicken IgG which has > 40% high mannose glycoforms where human and mouse IgG do not contain these isoforms.
  • FIGS. 5C-E are graphs depicting an all point 'plume' responses and calibration curve (FIG. 5D) as before with corresponding K D distribution (FIG. 5E).
  • FIG. 5F is a graph depicting media compositions used to elicit hypermannosylation in CHO cell culture.
  • FIG. 5G shows IgG titer for each of the culture days.
  • FIG. 5H shows general trends of mannosylation derived from the distribution averages in FIG. 5D.
  • FIGS. 6(a)-(b) illustrate a surface staining experiment to confirm the presence of mannose in the CHO culture samples.
  • 6a) A surface is treated to bind the IgG in the CHO samples and a FITC-labeled PSA lectin is used to stain the mannose content.
  • 6b) Representative FITC fluorescent images of the stained surface showing a marked increase in mannose content as found by the SWNT nanosensors.
  • FIGS. 7A-E illustrate sensor response to local cell production.
  • FIG. 7 A illustrates qualitative images of control and IgG producing cells showing colocalization of SWNT response.
  • FIG. 7B is a graph depicting location of top 1000 SWNT in IgG and control cell images, presented as coverage percentage of cell area (specific) and gel area (non-specific) in each image (FIG. 7C).
  • FIG. 7D are graphs depicting dynamic response of SWNT sensors to IgG producing cells plated on gel presented as distributions of top 1000 SWNT from images and micrographs of scaled SWNT intensity.
  • FIG. 7E are micrographs depicting mapping visible IgG producing cell islands, the co localized SWNT signal underneath, and ranking the islands' productivity based on intensity normalized by island area.
  • FIG. 8 illustrates the experimental domain addressed by arrayed nanosensors in comparison to existing technologies. Also mapped are approximate assay times of each technique.
  • FIG. 9 illustrates that His-tagged sensor protein needed to reduce nonspecific binding - data presented as distribution of all SWNT sensor sites.
  • FIG. 10 illustrates the role of nickel in SWNT sensor modulation.
  • the photolummescent nanostructures are associated with analyte-binding compounds. Binding of the analyte to the analyte-binding compound changes the photolummescent properties of the photolummescent nanostructures. Thus, the presence or absence of the analyte can be determined based on the observed photolummescent properties of the photolummescent nanostructures.
  • the change in the photoluminescence can include a change in photoluminescence intensity, a change in peak wavelength, a Raman shift, or a combination thereof.
  • the plurality of photolummescent nanostructures in the sensor forms an array of nanosensors, each of which provides important data about the analyte, the statistics of which are significant in interpreting the population of analyte species in a sample.
  • the statistical analysis based on a calibration, such as variance, skew, standard deviation and other parameters, can reflect different attributes of the analyte, such as a biologic analyte, for example an antibody or a protein product.
  • the statistical analysis provides information about analytes in the sample, for example, binding heterogeneity of a biologic analyte, cell line for production of a biologic analyte, or glycosylation patterns of a biologic analyte.
  • the array can include more than 100, more than 200, more than 500, more than 1,000, more than 2,000, more than 5,000 nanosensors, or more, each of which has a photoproperty that can be monitored.
  • detection of analytes with photoluminescent nanostructures can be limited by the presence of non-specific background photo luminescence.
  • Background photo luminescence can arise from, among other sources, a support that underlies the photoluminescent nanostructures, e.g., supports including glass or plastic.
  • One way to minimize background photoluminescence is to use a support composed of highly purified materials that are free of contaminants that may be responsible for the background photoluminescence. However, the use of such materials can increase the cost and complexity of analyte detection.
  • Another method to avoid background photoluminescence is to use a detection system that does not detect the background photoluminescence.
  • substrates that suffer from background photoluminescence can nonetheless be used, because the detection system detects only photoluminescence arising from the background photoluminescence. In some cases, this might be achieved with appropriate sets of wavelength filters.
  • filters may not be able to reject a sufficient amount of the background photoluminescence while still allowing a sufficient amount of the photoluminescence arising from the photoluminescent nanostructures to pass.
  • One way to accomplish this is to use an optical detection system that provides a distinct focal plane, where the detected photoluminescence is substantially from sources in the focal plane. Photoluminescent sources outside the focal plane are not detected by the detector. In this way, the detector does not suffer from issues associated with background photoluminescence.
  • a sensor for detecting an analyte can be arranged on a support.
  • Photoluminescence from the sensor is detected by an optical system that detects photoluminescence substantially only from sources in the focal plane. Sources outside the focal plane do not contribute substantially to the detected photoluminescence.
  • the focal plane may extend outside the boundaries of the sensor and include a portion of the support. In this situation, background photoluminescence will be detected.
  • the sensor can be arranged on an intermediate substrate.
  • the intermediate substrate is in contact with the support. The intermediate substrate separates the sensor from the support, such that the focal plane does not include any portion of the support.
  • the intermediate substrate can be advantegously composed of materials that are substantially free of background photoluminescence.
  • the intermediate substrate can be advantageously composed of materials that are compatible with the sensor.
  • the sensor can include a hydrogel, and the intermediate substrate can also include a polymer matrix or a gel.
  • sensor includes (illustrated in lower panel (ii)) support which supports substrate gel, which in turn supports sensor gel.
  • Sensor gel (detailed in upper panel (i)) includes a gel network, a plurality of photoluminescent nanostructures and associated linking polymers, such as a polysaccharide, for example, chitosan.
  • linking polymers such as a polysaccharide, for example, chitosan.
  • Optional crosslinks for example bis- acrylamide crosslinkers can be present between linking polymers, and an analyte-binding compound, such as a nickel chelate.
  • sensor gel is illustrated with analyte being associated with analyte-binding compound, though it will be understood that a sensor gel will typically be formed in the absence of analyte, and analyte may be contacted with sensor gel after formation. Thus in some states, sensor gel is free of analyte and in other states includes analyte (e.g., associated with analyte-binding compound). The presence of the analyte alters the photoluminescent properties of the photoluminescent nanostructures. As also illustrated in FIG. IB, the sensor array may be arranged upon one or more support materials.
  • the solid support may comprise a first gel lacking photoluminescent nanostructures and a second gel comprising photoluminescent nanostructures upon the first gel (FIG. IB).
  • the one or more support material may comprises a solid material (e.g. glass), a first gel lacking photoluminescent nanostructures, and a second gel comprising photoluminescent nanostructures upon the first gel.
  • an array of sensors can reconstruct this important distribution via sampling a large number of independent interactions.
  • Such arrays can also quantify weakly-affined interactions by recording a large number of rare binding events.
  • a nanosensor array can also characterize and differentiate biosynthesis around single cells and colonies, enabling the label- free selection of more productive strains. These properties have the potential to greatly enhance process analytics for biomanufacturing applications.
  • Glycosylation patterns can easily change due to processing conditions (media components, temperature, pH, pC0 2 , dissolved oxygen, cell density, duration, etc.) and the patterns can have a dramatic effect on the pharmacokinetics and immunogenicity of the resulting drug. See, for example, Pacis, E., et al. Biotechnology and Bioengineering 108, 2348-2358 (2011), Gramer, M.J. et al. Biotechnology and Bioengineering 108, 1591-1602 (2011), Lee, S.Y. et al. Process Biochemistry 47, 1411-1418 (2012), Ahn, W.S., et al. Biotechnology and Bioengineering 101, 1234-1244 (2008), Muthing, J. et al.
  • the gel can be a polymer including, but are not limited to, collagen, silicon- containing polymers, polyacrylamides, crosslinked polymers (e.g., polyethylene oxide, polyAMPS and polyvinylpyrrolidone), polyvinyl alcohol, acrylate polymers (e.g., sodium polyacrylate), or copolymers with an abundance of hydrophilic groups.
  • the gel can have a porous structure can be determined by factors including the concentration of polymers and crosslinking agent.
  • the pore size can be in the range of, for example, 10 nm to 1 ,000 nm, 20 nm to 500 nm, 50 nm to 250 nm, or 10 nm to 100 nm.
  • a pore size greater than 10 nm, greater than 20 nm, greater than 30 nm, greater than 40 nm, greater than 50 nm, greater than 60 nm, greater than 70 nm, greater than 80 nm, greater than 90 nm, or 100 nm or greater can be desireable.
  • a nanostructure may comprise different types of gels.
  • the nanostructure may comprise a first gel lacking photolummescent nanostructures and a second gel comprising photolummescent nanostructures positioned upon the first gel.
  • nanostructure refers to articles having at least one cross- sectional dimension of less than about 1 ⁇ , less than about 500 nm, less than about 250 nm, less than about 100 nm, less than about 75 nm, less than about 50 nm, less than about 25 nm, less than about 10 nm, or, in some cases, less than about 1 nm.
  • nanostructures include nanotubes (e.g., carbon nanotubes), nanowires (e.g., carbon nanowires), graphene, and quantum dots, among others.
  • the nanostructures include a fused network of atomic rings.
  • a “photolummescent nanostructure,” as used herein, refers to a class of nanostructures that are capable of exhibiting photoluminescence.
  • photolummescent nanostructures include, but are not limited to, single-walled carbon nanotubes ("SWNT”), double-walled carbon nanotubes, semi-conductor quantum dots, semi-conductor nanowires, and graphene, among others.
  • SWNT single-walled carbon nanotubes
  • photolummescent nanostructures exhibit fluorescence.
  • photolummescent nanostructures exhibit phosphorescence.
  • the carbon nanotube can be classified by its chiral vector (n,m), which can indicate the orientation of the carbon hexagons.
  • the orientation of carbon hexagons can affect interactions of the nanotube with other molecules, which in turn, can affect a property of the nanostructure.
  • a linker can be associated with the nanostructure.
  • the association can be a bond, for example, a covalent, ionic, van der Waals, dipolar or hydrogen bond.
  • the association can be a physical association.
  • at least a portion of the nanostructure can be embedded in the polymer or a portion of the polymer can encompass the nanostructure.
  • a linker can include a polymer.
  • a polymer can include a polypeptide, a polynucleotide or a polysaccharide. Examples of polysaccharides include dextran and chitosan.
  • a polymer can include a plastic, for example, polystyrene, polyamide, polyvinyl chloride, polyethylene, polyester, polypropylene, polycarbonate, polyacrylamide or polyvinyl alcohol.
  • a polymer can be biocompatible, which can mean that the polymer is well tolerated by an organism. More specifically, biocompatibility can mean that a polymer does not elicit an immune response when it is brought in contact with an organism. It can also mean that a polymer can integrate into cell structures, cells, tissues or organs of an organism. The organism can be mammal, in particular, a human.
  • An exemplary polymer can exhibit minimal binding with other molecules.
  • a polymer can have a protein adsorption of less than 5 ⁇ g/cm 2 , less than 1 ⁇ g/cm 2 , less than 0.5 ⁇ g/cm 2 , less than 0.1 ⁇ g/cm 2 , less than 0.05 ⁇ g/cm 2 , or less than 0.01 ⁇ g/cm 2 .
  • the association of a linker with a nanostructure can change a property of the nanostructure.
  • the property can be conductivity, polarity, or resonance.
  • the property can be photoluminescence, including fluorescence or phosphorescence. More specifically, the property can be fluorescence with a wavelength in the near infrared spectrum.
  • the property can be an emission wavelength, an emission intensity, a conductance, an electromagnetic absorbance or an emittance.
  • a linker can be configured to interact with an analyte-binding compound.
  • the analyte- binding compound undergoes a specific and typically reversible binding with an analyte.
  • suitable analyte binding compounds are proteins.
  • proteins that undergo specific and typically reversible binding with an analyte referred to herein as “capture proteins”.
  • a capture protein can include a protein, a polypeptide or a peptide. In some cases, a capture protein can be a complex of proteins. A capture protein can also include a full length protein, a fragment of a protein or a protein domain. A capture protein can be a fusion protein, which can include portions originating from one protein or portions originating from more than one protein. A capture protein can include a protein tag or marker. A capture protein can also be modified, for example, by glycosylation, ubiquitination, PEGylation, SUMOylation or biotinylation. A capture protein can be synthesized from a nucleic acid sequence that was amplified from a cDNA library, genomic DNA, a DNA vector or plasmid, or a DNA fragment.
  • the number of linkers associated with the nanostructure present in the analysis region can exceed the number of capture proteins. More specifically, the number of capture protein binding sites on linkers associated with a nanostructure can exceed the number of capture proteins.
  • the ratio of capture protein binding sites on linkers associated with a nanostructure to capture proteins can be greater than 1.1 to 1, greater than 1.5 to 1, greater than 2 to 1, greater than 5 to 1, or greater than 10 to 1. Having an excess of capture protein binding sites on linkers associated with a nanostructure can minimize the amount of unbound capture protein in a sample. Unbound capture proteins within the sample can compete with capture proteins bound to the composition for binding to the analyte. This can affect the accuracy and/or precision of the analyte detection. Having an excess of capture protein binding sites on linkers associated with a nanostructure can also increase the analyte concentration range over which analyte can be accurately detected because the saturation limit of the binding sites is increased.
  • the interaction between the linker and the capture protein can be binding to a capture protein.
  • the linker can be configured to interact with a capture protein by including a first binding partner in the linker that can interact with the capture protein.
  • the first binding partner can be known binding partner of the capture protein or a portion thereof.
  • the first binding partner can include an ion.
  • the ion can be a metal ion.
  • the metal ion can be a nickel, iron, cadmium, copper, magnesium, calcium, arsenic, lead, mercury or cobalt ion (e.g. Ni 2+ , Fe 2+ , Cd 2+ , Cu 2+ , Mg 2+ , Ca 2+ , As 2+ , Pb 2+ , Hg 2+ or Co 2+ ).
  • the first binding partner can include a protein, a nucleotide, a saccharide, a lipid or combinations thereof. See, for example, U.S. Pat. Pub. 2012/0178640A1 (WO 2012/030961), which is incorporated by reference in its entirety.
  • a linker can further include a chelating region.
  • a chelating region can include a chelator, which can be a polydentate ligand capable of forming two or more bonds with a single central atom.
  • a chelator can include one or more carboxylate ions.
  • a linker can include Na,Na-bis(carboxymethyl)-L-lysine.
  • a chelator can bind to a first binding partner (e.g. a metal ion) in order to incorporate the first binding partner into a linker.
  • the ion can act a proximity quencher of photoluminescent nanostructure.
  • the ion can quench near infrared fluorescence.
  • the quenching can be reversible.
  • the quenching can also depend on the distance between the nanostructure and the ion. In other words, as the distance between the nanostructure and the ion changes, the photoluminescence from the nanostructure can also change. Generally, as the distance between the ion and the nanostructure decreases, the amount of photoluminescence quenching can increase.
  • the capture protein can include a second binding partner, such that the first binding partner and second binding partner can bind together.
  • the second binding partner can be an endogenous motif or endogenous domain within a capture protein.
  • the second binding partner can be added to a capture protein.
  • the second binding partner can be a protein tag.
  • a protein tag can be a peptide sequence grafted onto a protein, which can be used for separating (e.g. using tag affinity techniques), increasing solubility, immobilizing, localizing or detecting a protein.
  • the protein tag can be a histidine tag, chitin binding protein tag, maltose binding protein tag, glutathione-S- transferase tag, c-myc tag, FLAG-tag, V5-tag or HA-tag.
  • One method for adding a second binding partner to a capture protein can include using primers including the sequence encoding for the second binding partner to PCR amplify DNA encoding for the capture protein.
  • a second method can include cloning DNA encoding for the capture protein into an expression vector designed to produce a fusion of the capture protein and the second binding partner.
  • Binding of a first and a second binding partner can be selective binding, which can provide the selectivity needed to bind to the corresponding binding partner (or relatively small group of related molecules or proteins) in a complex mixture.
  • the degree of binding can be less than 100%, less than 90%, less than 80%, less than 70%, less than 60%, less than 50%, less than 40%), less than 30%>, less than 20%> or less than 10%> of a second binding partner present binding to a first binding partner.
  • the degree of binding can be more than 10%, more than 20%, more than 30%, more than 40%, more than 50%, more than 60%, more than 70%, more than 80% or more than 90% of a second binding partner present binding to a first binding partner.
  • a first binding partner and a second binding partner can bind with a dissociation constant less than 1 mM, less than 0.1 mM, less than 0.01 mM, less than 1 ⁇ , less than 0.1 ⁇ , or less than 0.01 ⁇ .
  • a first binding partner and a second binding partner can bind with a dissociation constant greater than lnm, greater than 0.01 ⁇ , greater than 0.1 ⁇ , greater than 1 ⁇ , greater than 0.01 mM, or greater than 0.1 mM.
  • the linker can also be configured to interact with a capture protein by including a region capable of chemically reacting with the capture protein.
  • the chemical reaction can form a covalent, ionic, van der Waals, dipolar or hydrogen bond between the linker and the capture protein.
  • the interaction of a capture protein with a linker associated with a nanostructure can change a property of the nanostructure.
  • the property can be conductivity, polarity, or resonance.
  • the property can be photoluminescence, including fluorescence or phosphorescence.
  • the photoluminescence can be fluorescence with a wavelength within the near infrared spectrum.
  • the property can be an emission wavelength, an emission intensity, a conductance, an electromagnetic absorbance or an emittance.
  • the change in the property can be caused by a change in the distance between an ion in the first binding partner and the nanostructure.
  • a nanostructure property can also change.
  • nanostructure photoluminescence can also change.
  • the capture protein binds to the linker, the distance between the ion and nanostructure can change, which can alter the nanostructure photoluminescence.
  • the amount of photoluminescence quenching can increase.
  • a composition can further include a capture protein, which can be configured to specifically interact with at least one analyte.
  • the capture protein can be configured to specifically bind to at least one analyte.
  • Specific binding can be more limited than selective binding. Specific binding can be used to distinguish a binding partner from most other chemical species except optical isomers, isotopic variants and perhaps certain structural isomers.
  • the degree of binding can be less than 100%, less than 90%>, less than 80%>, less than 70%, less than 60%, less than 50%, less than 40%, less than 30%, less than 20% or less than 10% of an analyte present binding to a capture protein.
  • the degree of binding can be more than 10%, more than 20%, more than 30%, more than 40%, more than 50%, more than 60%, more than 70%, more than 80% or more than 90% of an analyte present binding to a capture protein.
  • An analyte and a capture protein can bind with a dissociation constant less than 1 mM, less than 0.1 mM, less than 0.01 mM, less than 1 ⁇ , less than 0.1 ⁇ , or less than 0.01 ⁇ .
  • An analyte and a capture protein can bind with a dissociation constant greater than lnm, greater than 0.01 ⁇ , greater than 0.1 ⁇ , greater than 1 ⁇ , greater than 0.01 mM, or greater than 0.1 mM.
  • the analyte can be a small molecule, protein, biomolecule, drug, biologic, or a metabolite thereof.
  • the analyte can be a monosaccharide, a polysaccharide, an amino acid, peptide, polypeptide, protein, a nucleotide, an oligonucleotide, a lipid, a polylipid, or a combination thereof.
  • the analyte can be immunoglobulin G (IgG) and the capture protein can be selected to specifically bind IgG.
  • the capture protein can be, for example, protein A or Pisum sativum agglutinin (PSA).
  • exemplary capture proteins may include, for instance, a protein that binds an analyte (e.g., Protein A which binds antibody and / or PSA-Lectin which binds mannose).
  • analyte e.g., Protein A which binds antibody and / or PSA-Lectin which binds mannose.
  • capture proteins may be physically associated with a photolummescent nanostructure and the analyte of interest (e.g., antibody).
  • the physical association may be accomplished using one or more linkers such as a moiety on the capture protein and a second moiety that physically associates with the photolummescent nanostructure or another linker moiety physically associated with the the photolummescent nanostructure.
  • linkers such as a moiety on the capture protein and a second moiety that physically associates with the photolummescent nanostructure or another linker moiety physically associated with the the photolummescent nanostructure.
  • the capture protein may be Protein A comprising a histidine tag that physically associates with nickel which is or becomes physically associated with the photolummescent nanostructure or another linker which is or becomes physically associated with the photolummescent nanostructure (e.g., chitosan).
  • Protein A comprising a histidine tag that physically associates with nickel which is or becomes physically associated with the photolummescent nanostructure or another linker which is or becomes physically associated with the photolummescent nanostructure (e.g., chitosan).
  • Other embodiments are also contemplated by this disclosure as would be understood by those of ordinary skill in the art.
  • the interaction of an analyte with a capture protein that is interacting with a linker associated with a nanostructure can change a property of the nanostructure.
  • the property can be conductivity, polarity, or resonance.
  • the property can be photoluminescence, including fluorescence or phosphorescence. More specifically, the property can be a fluorescent emission within the near infrared spectrum.
  • the property can be an emission wavelength, an emission intensity, a conductance, an electromagnetic absorbance or an emittance.
  • the interaction of an analyte with a capture protein can be reversible, meaning that the analyte can bind to the capture protein and then release and be free of binding.
  • the change in a property of the nanostructure due to the interaction of an analyte with a capture protein can also be reversible.
  • the property of a nanostructure can have a first value, the analyte can bind to the capture protein and alter the property to a second value, then the analyte can release from the capture protein and the property can return to the first value.
  • a method of detecting an analyte can include determining the presence of an analyte in the sample based on the monitored property. Determining the presence of an analyte can include determining the absence of the analyte. In some embodiments, determining the presence of an analyte can include determining the concentration of the analyte, determining the purity of the analyte or determining the quantity of the analyte. In some embodiments, relatively low concentrations or quantities of an analyte can be determined. The ability to determine low concentrations of an analyte may be useful, for example, in detecting trace pollutants or trace amounts of toxins within a subject.
  • analyte concentrations of less than about 100 micromolar, less than about 10 micromolar, less than about 1 micromolar, less than about 100 nanomolar, less than about 10 nanomolar, or less than about 1 nanomolar can be determined.
  • the quantity of the analyte that can be determined can be less than 1 mole, less than 1 millimole, less than 1 micromole, less than 1 nanomole, less than 1 picomole, less than 1 femtomole, less than 1 attomole or less than 1 zeptomole.
  • a single molecule of an analyte can be determined.
  • the purity of the analyte can be greater than 25% pure, greater than 50%, greater than 75% pure, greater than 80%, greater than 85% pure, greater than 90% pure, greater than 95% pure, greater than 99% pure or greater than 99.9% pure.
  • a linker can have a formula: A-L-C, where A can include a polymer, where at least a portion of the nanostructure is embedded in the polymer, L can be a linking moiety including a saturated or unsaturated C4_io hydrocarbon chain optionally containing at least two conjugated double bonds, at least one triple bond, or at least one double bond and one triple bond; said hydrocarbon chain being optionally substituted with Ci_ 4 alkyl, C 2 _ 4 alkenyl, C 2 _ 4 alkynyl, Ci_ 4 alkoxy, hydroxyl, halo, carboxyl, amino, nitro, cyano, C3-6 cycloalkyl, 3-6 membered heterocycloalkyl, unsubstituted monocyclic aryl, 5-6 membered heteroaryl, Ci_ 4 alkylcarbonyloxy, Ci_ 4 alkyloxycarbonyl, Ci_ 4 alkylcarbonyl, or formyl and said hydrocarbon chain being optionally interrupted by O, S, N
  • A can include a polymer [(M) x (N) y (Q) z ] q , where each of M, N and Q, independently, can be selected from the group consisting of a linear or cyclic C3-C8 hydrocarbyl, heterocyclyl, cyclyl, or aryl including one or more amine, alcohol or carboxylic acid group, where each M-N, M-Q or N-Q can include O, S, N(R a ), C(O), N(R a )C(0)0, OC(0)N(R a ), N(R a )C(0)N(R b ), C(0)0, or OC(0)0, each of R a and R b , independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where each of x, y and z can be integers between 0 and 50, 0
  • L can have the formula:
  • each X u X 2 and X 3 can be O, S, N(R a ), C(O), N(R a )C(0)0, OC(0)N(R a ), N(R a )C(0)N(R b ), C(0)0, or OC(0)0, each of R a and R b , independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where the value of n added to o can be 4 to 10.
  • the compound can have the formula,
  • Each of X and X' can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties.
  • the atoms bridging X and X' can be selected to form a 5-membered to 8-membered ring upon coordination to the metal ion.
  • the bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur.
  • the compound can have the formula,
  • Each of X, X' and X" can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties.
  • the atoms bridging X and X', X and X" or X' and X" can be selected to form a 5-membered to 8- membered ring upon coordination to the metal ion.
  • the bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur.
  • C can be derived from HSCH 2 CH 2 CH(SH)(CH 2 ) deliberatelyCOOH, H 2 CH 2 H 2 CH(NH 2 )(CH 2 ) n COOH,
  • the composition can include a nanostructure and a linker having a formula: A-L-C, where A can include the polymer covalently bonded to a portion of the nanostructure, L can be a linking moiety including a saturated or unsaturated C 4 _io hydrocarbon chain optionally containing at least two conjugated double bonds, at least one triple bond, or at least one double bond and one triple bond; said hydrocarbon chain being optionally substituted with Ci_4 alkyl, C 2 _ 4 alkenyl, C 2 _ 4 alkynyl, Ci_ 4 alkoxy, hydroxyl, halo, carboxyl, amino, nitro, cyano, C3-6 cycloalkyl, 3-6 membered heterocycloalkyl, unsubstituted monocyclic aryl, 5-6 membered heteroaryl, Ci_ 4 alkylcarbonyloxy, Ci_ 4 alkyloxycarbonyl, Ci_ 4 alkylcarbonyl, or formyl and said hydrocarbon chain being
  • the composition can include a chelator-containing compound, which can include a chelator region and a non-chelator region.
  • C can be the chelator region.
  • L can include the non- chelator region.
  • A can include a polymer [(M) x (N) y (Q) z ] q , where each of M, N and Q, independently, can be selected from the group consisting of a linear or cyclic C3-C8 hydrocarbyl, heterocyclyl, cyclyl, or aryl including one or more amine, alcohol or carboxylic acid group, where each M-N, M-Q or N-Q can include O, S, N(R a ), C(O), N(R a )C(0)0, OC(0)N(R a ), N(R a )C(0)N(R b ), C(0)0, or OC(0)0, each of R a and R b , independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where each of x, y and z can be integers between 0 and 50, 0
  • L can have the formula:
  • each X u X 2 and X 3 can be O, S, N(R a ), C(O), N(R a )C(0)0, OC(0)N(R a ), N(R a )C(0)N(R b ), C(0)0, or OC(0)0, each of R a and R b , independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where the value of n added to o can be 4 to 10.
  • the compound can have the formula,
  • Each of X and X' can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties.
  • the atoms bridging X and X' can be selected to form a 5-membered to 8-membered ring upon coordination to the metal ion.
  • the bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur.
  • the compound can have the formula,
  • Each of X, X' and X" can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties.
  • the atoms bridging X and X', X and X" or X' and X" can be selected to form a 5-membered to 8- membered ring upon coordination to the metal ion.
  • the bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur.
  • C can be derived from HSCH 2 CH 2 CH(SH)(CH 2 ) deliberatelyCOOH, H 2 CH 2 H 2 CH(NH 2 )(CH 2 ) n COOH,
  • a sensor array can include a plurality of analysis regions on a support.
  • a support can be glass or plastic.
  • a support may also be any material suitable for supporting the nanostructure, as would be understood by those of ordinary skill in the art.
  • An analysis region can be a divot, a tube, a tray, a well, or a similar compartment for suitable for containing a liquid sample.
  • an analysis region can include a droplet or spot on the surface of a support (e.g., a flat support).
  • an analysis region can be formed by spotting the composition on a support.
  • a plurality of analysis regions can be arranged in a pattern on a support.
  • a pattern can include concentric circles, a spiral, a row, a column or a grid.
  • the plurality of analysis regions can include two or more subsets of analysis regions. For example, a first subset of analysis regions can differ from a second subset of analysis regions by including a different nanostructure, a different linker, a different binding partner, a different capture protein, a different analyte or a different sample.
  • a first subset of analysis regions can differ from a second subset of analysis regions by including a different environmental factor including a buffer, a reagent, a nutrient, a serum, an exposure to light, an oxygen concentration, a temperature or a pH.
  • a nanostructure may comprise, for instance, different types of materials.
  • the nanostructure may comprise a first gel lacking photoluminescent nanostructures and a second gel comprising photoluminescent nanostructures positioned upon the first gel.
  • the first and second gel may be positioned upon a solid support material such as, for instance, glass and/or plastic. Any combination of materials may be used as a support material and are contemplated by this disclosure as would be understood by those of ordinary skill in the art.
  • an array of addressable sensors can be multiplexed for the label-free detection of a library of analytes, however such arrays, even when monofunctionalized, have lesser appreciated capabilities that emerge from the ensemble.
  • an array of nanosensors can uniquely monitor weakly-affined analyte interactions via the increased number of observed interactions.
  • One application involves monitoring the metabolically-induced hypermannosylation of human IgG from CHO using PSA-lectin conjugated sensor arrays where temporal glycosylation patterns are measured and compared.
  • the array of sensors can also spatially map the local production of an analyte from cellular biosynthesis, thus providing a powerful approach to clonal selection.
  • productivity of IgG-producing HEK colonies cultured can be ranked directly on the array of nanosensors itself. This study opens new avenues for the use of nanosensor arrays for biomanufacturing applications.
  • this disclosure provides methods for detecting an analyte, comprising: contacting a sample including a biologic analyte with a sensor array, the sensor array including a plurality of photolummescent nanostructures; collecting emissive data from each of the photolummescent nanostructures; and, analyzing the emissive data from each nanostructure of the plurality of photolummescent nanostructures to measure a distribution of binding affinities.
  • the photolummescent nanostructures may be physically associated with a linker (e.g., "wrapped” in chitosan); directly or indirectly “chelated” with a metal ion (e.g., nickel); and / or “functionalized” with a capture protein (e.g., Protein A, PSA-Lectin) (e.g., “functionalized” refers to a direct or indirect physically association providing a functional relationship between the photolummescent nanostructure and the capture protein) (see, e.g., FIGS. IB and 5A).
  • a linker e.g., "wrapped” in chitosan
  • directly or indirectly “chelated” with a metal ion e.g., nickel
  • a capture protein e.g., Protein A, PSA-Lectin
  • the photolummescent nanostructures may be selected from the group consisting of carbon nanotubes, single-walled carbon nanotubes, double-walled carbon nanotubes, semi-conductor quantum dots, semi-conductor nanowires, and graphene.
  • the sensor array is arranged upon one or more support materials which may be at least one gel.
  • the one or more support materials comprises a first gel lacking photolummescent nanostructures and a second gel comprising photolummescent nanostructures upon the first gel. Any number of layers of gels or other support material may be included in the nanostructures.
  • the one or more support materials may also include a solid material (e.g. glass) supporting the photolummescent nanostructures.
  • the one or more support materials comprises a solid material (e.g. glass), a first gel lacking photolummescent nanostructures, and a second gel comprising photolummescent nanostructures upon the first gel.
  • the capture protein is Protein A and/or a derivative thereof (e.g., where the analyte is antibody) or PSA-lectin and/or a derivative thereof (e.g., where the analyte is mannose).
  • the sample may comprises cell that colonize the photolummescent nanostructures and the methods provide for the detection of one or more analytes produced by the cells is detected.
  • the methods described herein may also comprise: calculating a calibration curve for the sensor; analyzing the emissive data by measuring a variance of the emissive data from the plurality of photoluminescent nanostructures (where the emissive data may include measuring a skew of the emissive data from the plurality of photoluminescent nanostructures and/or measuring a standard deviation of a plume angle of the emissive data from the plurality of photoluminescent nanostructures); evaluating a binding heterogeneity of the biologic analyte of the sample from the analyzed emissive data; evaluating a cell line for production of the biologic analyte of the sample from the analyzed emissive data; and/or evaluating a glycosylation pattern of the biologic analyte of the sample from the analyzed emissive data.
  • the arrays described here may include, for instance, at least about any of 10, 20, 30, 40, 50, 60 70, 80, 90, 100, 125, 150, 175, 200, 225, 250, 275, 300, 325, 350, 375, 400, 425, 450, 475, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900, 2000, 2100, 2200, 2300, 2400, 2500, 2600, 2700, 2800, 2900, 3000, 3100, 3200, 3300, 3400, 3500, 3600, 3700, 3800, 3900, 4000, 4100, 4200, 4300, 4400, 4500, 4600, 4700, 4800, 4900, or 5000 photoluminescent nanostructures.
  • these methods may also comprise sensors that include an analyte -binding compound associated with the photoluminescent nanostructure.
  • the sensor may comprises a linker, wherein the analyte-binding compound is associated with the photoluminescent nanostructure via the linker.
  • the sensor may comprise a polymer including clusters of photoluminescent nanostructures.
  • the analyte is a biologic analyte such as a protein (e.g., an antibody such as IgG).
  • the analyte binding compound is a capture protein comprising Protein A or a derivative thereof or PSA-lectin or a derivative thereof.
  • the analyte may be any that specifically interacts with Protein A or PSA-lectin, respectively.
  • the capture protein e.g,. Protein A or a derivative thereof or PSA-lectin or a derivative thereof
  • the capture protein comprises a tag having specificity for a binding partner associated with the photoluminescent nanostructure.
  • the tag is a histidine tag and the binding partner is nickel.
  • Other embodiments are also contemplated by this disclosure as would be understood by those of ordinary skill in the art. A better understanding of the present invention and of its many advantages will be had from the following examples, given by way of illustration. EXAMPLES
  • SWNT Single-walled carbon nanotubes
  • FIG. 1A Single-walled carbon nanotubes
  • a multi-layer, highly- porous (60-90nm) polyacrylamide hydrogel to reduce nonspecific binding and allow for fast diffusion of the IgG analytes to the SWNT sensors (FIG. 1A).
  • a basal substrate gel with no SWNT is used on a glass substrate to position the thin sensor layer in a separate focal plane and eliminate the background fluorescence caused by glass impurities.
  • the platform is excited by a 660nm laser (Crystal Laser) on an inverted microscope (Zeiss D. l) and the nIR emission is collected as an image stack on a 256x320 pixel InGaAs array (Princeton Instruments Acton Array).
  • the SWNT are suspended in chitosan and have been chemically modified as before to display chelated nickel groups that act as both the docking site for a His-tagged capture proteins (Protein A and PSA Lectin in these examples) and as the signal transducer (proximity quencher). See, for example, Reuel, N.F. et al.
  • the nanotube acts as an optical switch, brightening as antibodies bind to the capture protein.
  • the ensemble response of the array is created by averaging the intensity values over the entire array for each time point and is analogous to other bioassay techniques like ELISA and Biacore® SPR measurements.
  • Divalent nickel, protein A, and IgG addition (lOOmM, 1 mg/ml, 1.5 mg/ml) cause a decrease, increase, and additional increase respectively (FIG. lC(i)).
  • Subsequent washing of the gel surface shows negligible effects on the ensemble signal, which can be interpreted as absence of unbinding.
  • Adding a BSA control (2mg/ml) does not elicit a sensor response (FIG. lC(i)).
  • the responses of all SWNT pixels can be monitored as histograms of percent modulation (I 0 -lFinai/I 0 ) (Fig, lC(ii)).
  • Sensor specificity can also be observed qualitatively with a nIR heatmap filtered to the 10,000 most responsive sites (FIG. ID).
  • the sensor protein Protein A or PSA lectin
  • the sensor protein must first be docked to the chelated nickel; otherwise BSA would elicit a response.
  • Example 2(d) with a bivariate Gaussian distribution (FIG. 2(e)).
  • the plume angle mean and standard deviation ( ⁇ ⁇ and ⁇ ⁇ ) are then recorded for each concentration.
  • the program completes 6 iterations to create a smooth calibration curve (FIG. 2(f)).
  • the initial K D skewness parameter ( ⁇ ) is then changed and the program is run again. This was done for 200 different ⁇ values spanning 2 to 45 and the results are shown in FIG. 3D, as described in more detail below.
  • the full code for this simulation can be found in Example 2(K)(1) (Code 3a).
  • nanoscale sensor arrays could be used to measure intrinsic properties of the sample such as affinity distributions.
  • an analyte to have a Gaussian distribution of K D with a set variance (a t 2 ) which is then assayed on a 1 mm 2 array of sensors. If this sensor area is divided in enough independent sensor regions, so that each molecule is assayed independently, the full distribution can be recovered. If each sensor element, however, reports an average of multiple molecule signal transductions the measured sample variance (a m 2 ) becomes much smaller than the actual value (FIG. 3A insert). For a
  • nanotube sensor responds to adsorption is dictated by its starting intensity (Io) and extent of functionalization (PF) which can both be experimentally measured and fit with another Weibull ( ⁇ 2 ) and Gaussian distribution((p 3 ) respectively. It is important to note that the number of binding sites on the senor protein is also imbedded in the functionalization variance (if all nanotubes were uniformly wrapped, functionalized, and had the same number of binding sites, this variable would be a constant).
  • individual sensor responses, R can be obtained as a coverage fraction, intensity, and percent functionalization value ( ⁇ , I 0 , and PF,) randomly generated from their respective PDFs (cpi, cp2, (p 3 ) and subsequently evaluated as:
  • the response is normalized by the average intensity (I aVg ) as each experimental platform's overall intensity may change due to variance in the experimental setup or quality of SWNT. This allows for clear comparison of distinct arrays assayed at different concentrations.
  • I aVg average intensity
  • the 'plume' is the distribution of data points shown in the plot.
  • the experimental 'plumes,' are fit well by a bivariate Gaussian distribution in polar coordinates (6,R) and the mean 'plume' angle ( ⁇ ⁇ ) is used to create the calibration curve (FIG. 2C(iv)).
  • the standard deviation in the plume angle ( ⁇ ⁇ ) also has an interesting dependence on starting K D skew ( ⁇ ) and is reported (FIG. 3C(iv)).
  • the parameter B can be thought of as a measure of 'steepness' and C is equal to the inflection point at K D .
  • the array of nanotube sensors conjugated to His-tagged Protein A was used to assay three different samples of IgG with expected differences in affinity distributions: (1) commercial, lyophilized, polyclonal Human IgG reconstituted in PBS, (2) murine IgG (TA99) from an engineered human embryonic kidney (HEK) cell line, and (3) human IgG (bl2) cultured from Chinese hamster ovary (CHO) cells.
  • K D mean values (95% confidence intervals for each: 11 - 27 ⁇ , 3.3 - 5.3 nM, and 0.6 - 16 nM) comparable to those found in literature for IgG-Protein A interactions (2-50 nM from SPR, 34 nM from acoustic device (see, for example, Nohlden, S. in Department of Physics, Chemistry and Biology, Vol. Masters 68 (Linkopings universitet, Linkoping; 2008), and Saha, K., et al. Analytical Chemistry 75, 835-842 (2003)).
  • the calibration curve also provided 'B' fit parameters (Eq. 4) that are related to the starting distribution skew parameter ( ⁇ ) as solved from the model simulation (FIG.
  • nanosensor arrays Another advantage of nanosensor arrays is their ability to report weak binding events - a greater number of individual sensor sites increases the probability of a detection event and this event is not averaged to null with other non-responsive sites, as in the case for an ensemble sensor.
  • the label-free nature of the platform is also beneficial to detecting weakly-affined ligands since it requires no washing steps.
  • PSA Pisum sativum agglutinin
  • the sensor platform can detect specifically high mannose content IgG (FIG. 5A). Different species of IgG were initially used to test this concept.
  • Chicken IgG contains an appreciable amount of glycoforms with high mannose content (>40% of population) whereas these are virtually absent in human and mouse IgG. See, for example, Raju, T.S., et al. Glycobiology 10, 477-486 (2000), which is incorporated by reference in its entirety.
  • the SWNT sensor responses to human, mouse, and chicken IgG in PBS align with these findings and confirm that the platform can detect mannose species with the PSA lectin specifically (FIG. 5B).
  • a well characterized, high-mannose content IgG sample from a microbial source was then used to further validate detection and obtain a calibration curve and K D distribution as before (Figs. 5C-E).
  • the supematants were collected after each 24 hour period, diluted to a standardized lOng/ml IgG concentration and assayed on our PSA rendered sensor gels.
  • the IgG concentrations were determined with ELISA (FIG. 5G) and if below lOng/ml, the sample was run at stock concentration.
  • the resulting trends (FIG. 5H) determined by the mean percent modulation from the nanosensor array distributions match those found in the previous study: 1) increased mannose content as culture time increases, 2) increased mannose from higher NaCl osmolality, and 3) delayed onset of hypermannosylation from MnCl 2 additive.
  • FIG. 6(a) A surface staining approach was used (FIG. 6(a)) in which Protein A conjugated to a porous resin bead (Pierce 53139) was immobilized to a poly-lysine coated glass surface via glutaraldehyde in 2.5mm diameter Teflon patterned wells.
  • the daily CHO samples diluted to lOng/ml IgG (used in Manuscript Figure 5 assay) were then spotted on the glass slide at 20 ⁇ 1.
  • the IgG was allowed to interact for 2 hours and the chip was then washed in PBS.
  • the chip was then spotted with 300 ⁇ g/ml BSA to block non-specific sites and washed with PBS again.
  • FITC-conjugated PSA (Vector Labs FL-1051) was spotted on the sample at 25ug/ml and allowed to interact overnight. The samples were then washed with PBS and assayed on an inverted microscope reading the visible FITC emission. Qualitatively, the observation of more conjugated FITC-PSA corresponded with the days peak mannose was observed using the nanosensor platform (FIG. 6(b)). This supports the finding of mannose in the IgG samples.
  • a gel with an imbedded array of nanosensors can be used to screen local production of cells.
  • single cells are very difficult to culture for long periods of time on the current porous platform (little indications of healthy, single cell adherence).
  • FOG. 7A Protein A-incubated gel
  • Carbon nanotube-based fluorescent sensor arrays can be used for monitoring distributions of K D , hypermannosylation, and local cellular production of IgG with clear implications in biomanufacturing.
  • the platform was demonstrated with lyophilized IgG in PBS as well as characterizing freshly-expressed IgG in complex media from three different cellular expression systems: HEK, CHO, and a fungal cell line.
  • the sensor array was rendered specific to mannose with PSA-lectin and trends in metabolically induced hypermannosylation from a previous study were confirmed.
  • local production of IgG from HEK cell colonies cultured on sensor arrays was monitored. Better upstream colony selection could be performed with a sensor gel optimized for healthy cell culture.
  • the cell colonies could be exposed to various culture and media conditions and their productivity and glycosylation patterns could be monitored in real time. This platform could lead to more rapid and informed selection of master cell lines and culture conditions based on multiple parameters rather than picking colonies based on static snapshots of productivity provided by current assays. See, for example, Burke, J.F. et al. Genetic Engineering & Biotechnology News 29, 38-39 (2009).
  • sensor arrays in a microfluidic platform can be used to monitor product titer, K D distribution, and glycosylation by periodically sampling the bioreactor, filtering cellular components, diluting to a set level depending on the cell line's average productivity read the fluorescent signal, and then regenerate for the next sample.
  • Protein A gels can be regenerated using a pH 3.0 release wash, similar to regeneration of Protein A purification columns with little loss of sensitivity. Detection of mannose has been validated here but other glycans of interest (galactose, fucose, sialic acids, and non-human, immunogenic glycans like gal-a 1,3-gal (see, for example, Chung, C.H. et al. New England Journal of Medicine 358, 1109-1117 (2008)) could also be detected by multiplexing portions of the nanosensor array with different His-tag lectins.
  • His-tag lectins can be made by chemical conjugation of a His-tag and a target lectin.
  • a kit such as a his-tag linkage kit (Solulink) can be used to form a polyhistidine tagged lectin.
  • suitable starting materials can include a native lechtin (Vector Labs), a linker such as SMCC linker (Pierce), an activating agent such as Traut's reagent (Pierce), and a polyhistidine such as hexahis peptide (Abbiotec).
  • An example protocol can include the following steps:
  • Suitable glycans and lectins can be selected based on the properties to be detected. Examples of glycans or lectins can be identified in the literature. See, for example, van Berkel, P. FL, et al. Biotechnology Progress, Volume: 25, Issue: 1, pages: 244-251 (2009), Hirabayashi, J., J. of ' Biochemistry Volume: 144, Issue: 2, pages: 139-147 (2008), Gabius, H.-J. et al., Trends in Biochem. Sci. Volume: 36, Issue: 6, pages: 298-313 (2011), Dam, T.K.; et al. Glycobio. Volume: 20, Issue: 3, pages: 270-279 (2010), ): Tateno, H.
  • the longstanding goal of nanosensor arrays is to preserve the sensitivity and analytical advantages of single-molecule nanosensors with the multiplexing ability of macroscale techniques, thus filling an untapped analytical regime (FIG. 8).
  • the fast assay time ( ⁇ 5 min) of nanosensor arrays could also provide a disruptive alternative to the more time intensive ELISA and LC/MS analytics that are currently used (FIG. 8).
  • the current limitations of these arrays are the intrinsic variances caused by non-automated production in small batches (16-32 gels per batch). Small variations in polymer casting time, initiator concentration, and washing procedures result in gels with varying levels of functionalization and sensitivity. A standardized gel from an automated printing/production system could reduce this variance and provide a robust tool for biomanufacturing analytics and beyond.
  • SWNT were suspended in chitosan according to Reuel, N.F. et al. J. Am. Chem. Soc. 133, 17923-17933 (2011).
  • 3 mg of purified HiPCO SWNT (Unidym) were added to 20ml of chistosan suspension (0.25 wt% in water containing 1 vol% acetic acid - Sigma).
  • the resulting mixture was tip sonicated (1/4" tip Cole Parmer,Model CV18) at 10W for 45 minutes in an ice bath and table -top centrifuged three times at 13.2 RPM for 90 min each, while collecting the suspended SWNT supernatant and discarding the aggregate pellet after each cycle.
  • the SWNT was then mixed at a 50:50 volume ratio with the polyacrylamide mixture for casting as the top layer.
  • the amount of monomer (acrylamide) and cross-linker ( ⁇ , ⁇ '-Methylenebisacrylamide - both Sigma) are specified using standard %T %C nomenclature, where %T refers to the overall weight % of polymer (monomer and crosslinker) in the solution and %C refers to the wt% of the total polymer that is cross linker.
  • Surface gels with a 3%C and 1%T composition performed well.
  • a substrate gel (6%T, 1%C) was also prepared.
  • TEMED Tetramethylethylenediamine
  • a fresh initiator solution of 1 wt% Ammonium persulfate (APS) was made immediately prior to each gel batch.
  • the APS, bottom and top gel solutions, and substrate chips (8 chamber Lab-Tek by Nunc) were degassed in the glove box antechamber to remove absorbed and dissolved oxygen.
  • a nitrogen controlled glove box (MBraun LABstar)
  • 1 vol% of the APS solution was added to the substrate gel to initiate the polymerization and it was immediately cast (lOOul to each well) and then allowed to cure for one hour.
  • the top gel was then initiated with 1 vol% APS and immediately spotted at 20 ul to each gel surface and allowed to cure for 1 hour.
  • the functionalization steps of the chitosan wrapped SWNT is also similar to that described in Reuel, N.F. et al. J. Am. Chem. Soc. 133, 17923-17933 (2011).
  • the amine groups of the chitosan were reacted with succinic anhydride (133 mM in PBS 7.4 buffer - Sigma) overnight and then washed thoroughly with water.
  • the carboxylic acid functional groups were then activated with lOOmM EDC and 520mM NHS (Sigma) in MES Buffer pH 4.7 (Pierce) for 2 hours.
  • the sensors were manufactured for use with a his-tag sensor protein ((Protein A (Abeam) or PSA lectin).
  • the original SWNT sensors for protein and glycan recognition from our group were cast in a chitosan hydrogel. Not only were the desired attributes of large pore size and local SWNT functionalization hard to control in this gel, but the chitosan proved to exacerbate the non-specific binding of the more 'sticky' IgG and Protein A species. Adding Protein A-Histag followed by Human IgG (Fig. Sla in U.S. Ser. No.
  • the top acrylamide gel layer containing the SWNT sensors was tuned to a maximum pore size to ensure that large antibodies could diffuse to the sensor sites.
  • the effect of crosslinker concentration on pore size has been established qualitatively in literature with TEM imaging (Ruchel, et al. Transmission-Electron Microscopic Observations Of Freeze-Etched Polyacrylamide Gels. Journal of Chromatography 166, 563-575 (1978)). It is noted that these experiments were performed while protecting the radical polymerization from quenching oxygen species in a nitrogen-filled glove box. It was found that solutions with as low as 3 wt% polymer (97% water) were able to crosslink in this environment. It was determined from FITC dextran release profiles (Fig. S2d of U.S. Ser. No.
  • the pore size was also more rigorously probed using a recent hydrogel characterization technique called microscale poroelastic indentation by AFM as explained in the Kalcioglu, Z.I., et al. (From macro- to microscale poroelastic characterization of polymeric hydrogels via indentation. Soft Matter 8, 3393-3398 (2012)) Soft Matter 8, 3393-3398 (2012); see also , which is incorporated by reference in its entirety.
  • a short silicon tip with a 45 um polystyrene sphere (Novascan) was fitted on an AFM (Asylum Research - MFP3D) and the IgorPro software indentation panel was used to drive the tip into the gel at a specified distance and record the force over time.
  • FITC-conjugated dextran particles (Invitrogen) of various sizes (10, 40, 70, and 500 kD) were also absorbed into 150 ⁇ cylindrical gel plugs over 48 hours. The impregnated gels were then removed, washed, and inserted into clean water. The release of the FITC particles was observed by sampling the exterior fluid and assaying the FITC content with a plate reader. Using standard curves, the release was then determined in terms of cumulative mass release over time.
  • the resulting force versus time curve (Fig. S2b of U.S. Ser. No. 61/767,53) exhibited an initial spike (F 0 ) that relates to the shear modulus of the gel (G in Eq 6) and then relaxed to an equilibrium state (F ⁇ ) which relates to the diffusion constant of the displaced fluid.
  • SWNT sensor data was collected on a custom inverted microscope (Zeiss D.l Observer) that was fitted with a 660 nm laser (Crystal Laser, 100 mW).
  • a 20x planar objective (Zeiss) was used and the emission intensities were recorded by a nitrogen-cooled InGaAs array (Princeton Instruments).
  • Win Spec software (Princeton Instruments) was used to collect the SWNT emission and saved as an image stack TIF file. This file was then analyzed using Matlab. Analyte samples were added to the sensor gels by hand, applying the ⁇ sample to the lower right corner of the well, so as not to place the plastic pipette tip in the laser beam path.
  • a sensor gel for testing To prepare a sensor gel for testing, it was first thoroughly washed with PBS to exchange the buffer and then allowed to incubate with the his-tag sensor protein (Protein A (Abeam) or PSA lectin (Vector Labs - conjugated to His-tag peptide (Abbiotec) via Traut's reagent and SMCC linker (Pierce)) at 500 ⁇ / ⁇ 1 overnight. The gel was again washed thoroughly with PBS and then fitted on the microscope for testing.
  • the his-tag sensor protein Protein A (Abeam) or PSA lectin (Vector Labs - conjugated to His-tag peptide (Abbiotec) via Traut's reagent and SMCC linker (Pierce)
  • Divalent nickel cations have been shown to quench anionic surfactant wrapped SWNT 8 and quantum dots (Wu, et al. Ni(2+)- modulated homocysteine-capped CdTe quantum dots as a turn-on photoluminescent sensor for detecting histidine in biological fluids. Biosensors & Bio electronics 26, 485-490 (2010)).
  • the proximity of the chelated nickel groups alter the local electronic environment and offer non- radiative decay pathways for the SWNT exciton. To explore if this is the only mechanism at work, we conceived of an experiment to replace the nickel with another small molecule proxy - biotin.
  • the chitosan wrapped SWNT are biotinylated with a commercially available NHS linker (Pierce EZ-Link NHS-Biotin) and then exposed to nuetravidin and a BSA control (500 ⁇ g/ml).
  • the specific response is again a positive modulation and the control is null (FIG. 10(a)) although the turn-on response is approximately 20% of what is typically seen when nickel is present. This supports two interoperating mechanisms (FIG.
  • the sensor proteins are originally quenched by chelated nickel groups and the return of signal occurs when the nickel is displaced upon binding of the sensor protein or IgG
  • the chitosan wrapped SWNT is originally quenched by aqueous quenching species (water molecules (Strano, M.S. et al. The Role of Surfactant Adsorption during Ultrasonication in the Dispersion of Single- Walled Carbon Nanotubes. Journal of Nanoscience and Nanotechnology 3, 81-86 (2003)), dissolved oxygen, and protons (Blackburn, J.L. et al. Protonation effects on the branching ratio in photoexcited single-walled carbon nanotube dispersions.
  • N number of sites in full distribution
  • Navg number of sites averaged together from full distribution
  • the extent of sensor functionalization was determined by monitoring the extent of quenching caused by nickel addition. This provides a reasonable measure to the accessibility and functionalization of each nanotube.
  • the percent quenching was found to be invariant in SWNT length (banded when plotted vs. I 0 ) and approximated well by a Gaussian distribution (Fig S9c-d of U.S. Ser. No. 61/767,513) with the following PDF (Fig S9d of U.S. Ser. No. 61/767,513):
  • a tricistronic expression cassette pLB2-CMV-GFP-TA99 was created using 2A skip peptides (see Hu, T., et al. Biotechnology Letters 31, 353-359 (2009), which is incorporated by reference in its entirety).
  • the expression cassette was cloned into the lentiviral vector, pLB2 (Stern, P. et al.
  • HEK- 293FT cells (Invitrogen) were transfected with the following plasmids: pLB2-CMV-GFP-TA99, pCMV-dR8.91, and pCMV-VSV-G at a mass ratio of 2: 1 : 1 using PEI (see Zufferey, R., et al. Nature Biotechnology 15, 871-875 (1997); and Dull, T. et al. J. of Virology 72, 8463-8471 (1998)). After 24 hours, fresh media was exchanged. 48 and 72 hours later, supernatant containing lentiviral particles was harvested.
  • HEK-293 cells were transduced twice for 24 hours by incubation with freshly harvested supernatant supplemented with protamine sulfate at 5 ⁇ g/mL.
  • GFP positive cells were selected to a purity of greater than 95% using flow fluorescence activated cell sorting. Details on the generation of the CHO cell line can be found in Hezareh, M., et al. J. Virology 75, 12161-12168 (2001). Additional data is presented in Supplement 2e of of U.S. Ser. No. 61/767,513.
  • the medias used were DMEM (with 4.5 g/L glucose, 10% heat inactivated FBS (Invitrogen), 2 mM L-glutamine, 100 U/ml Penicillin, 100 U/ml Streptomycin - rest Sigma), and GMEM (same additives - Sigma) for the HEK and CHO cultures, respectively.
  • DMEM with 4.5 g/L glucose, 10% heat inactivated FBS (Invitrogen), 2 mM L-glutamine, 100 U/ml Penicillin, 100 U/ml Streptomycin - rest Sigma), and GMEM (same additives - Sigma) for the HEK and CHO cultures, respectively.
  • a serum free media was used for growth and as a buffer in the sensor gel (Invitrogen Freestyle 293). To passage the cells, they were allowed to grow to confluence, washed with PBS, and then released with Trypsin (0.05%> w/0.53 mM EDTA). The cells were then
  • CHO cells were seeded at equal density in small culture flasks (25 cm 2 - Sarstedt) and allowed to grow to confluence with regular GMEM media (overnight). The growth media was then exchanged with 3ml serum free media (Freestlye 293 Invitrogen) and the cells were allowed to produce for 24 hours. The media was then saved and a fresh 3ml of serum-free media was added for the next 24 cycle. This was repeated for 10 days. The IgG content was measured by ELISA (ICL Lab, Inc.) and the samples were diluted to lOng/ml in Freestyle to run on the PSA- incubated SWNT gels.
  • IgG content was measured by ELISA (ICL Lab, Inc.) and the samples were diluted to lOng/ml in Freestyle to run on the PSA- incubated SWNT gels.
  • Npoints 10000 ;
  • Nconc BP* 6+5 ;
  • Kvec zeros (Npoints , 1 ) ;
  • BetaMin 2 ;
  • BetaMax 45 ;
  • nB 200 ; % ## ⁇ Change number of Beta parameters here
  • BetaSpan linspace (BetaMin, BetaMax, nB) ;
  • ThetaSAVE zeros (nB, Nconc) ;
  • ThetaSAVEstd zeros (nB, Nconc) ;
  • K l Cr (wblrnd (KDGlobalMean, Beta) ) ;
  • Kvec ( j , 1) K
  • bins linspace (-11.1, -2.9, 200) ;
  • KDL loglO (1. /Kvec) ;
  • [n x] hist (KDL, bins) ;
  • [n x] hist (Lt ( : , i) , 200) ;
  • %hlegl legend ([ int2str (Cspan ( 1 , 1+BP) ), '
  • RespNorm zeros (NSWNT, Nconc) ;
  • INorm zeros (NSWNT, Nconc) ;
  • Length ( j , i ) wblrnd (6871.7, 2.2)
  • MaxL max (Length) ;
  • MinL min (Length) ;
  • MeanL mean (Length) ;
  • %MeanLNorm (MeanL - MinL) ./ (MaxL - MinL)
  • RespNorm ( j , i ) Theta*L/MeanL (1, i) ;
  • INorm ( j , i ) (L - MinL(l,i)) / (MaxL(l,i) MinL (1, i)
  • %[n x] hist (PM_Ni, 200) ;
  • %Gm mean (PM_Ni) ;
  • %X linspace (-50, 10, 100) ' ;
  • %Y pdf (Gfit, X) ;
  • Theta Lt (j , i) ;
  • Io Length ( j , i ) ;
  • %hleg2 legend ([ int2str (Cspan ( 1 , 1+BP) ), '
  • ThetaMean zeros ( 1 , Nconc) ;
  • ThetaSig zeros ( 1 , Nconc) ;
  • MixGausData zeros ( 3 , 2 , Nconc) ;
  • Cloudl gmdistribution . fit (X, 1 ) ;
  • MixGausData (2 : 3, 1 : 2, i) SIG(:,:,1);
  • Cloud2 gmdistribution . fit (X, 1 ) ;
  • MixGausData (2 : 3, 1 : 2, i) SIG(:,:,1);
  • Cloud3 gmdistribution . fit (X, 1 ) ;
  • MixGausData (2 : 3, 1 : 2, i) SIG(:,:,1);
  • Cloud4 gmdistribution . fit (X, 1 ) ;
  • MixGausData (2 : 3, 1 : 2, i) SIG(:,:,1);
  • Cloud5 gmdistribution . fit (X, 1 ) ;
  • MixGausData (2 : 3, 1 : 2, i) SIG(:,:,1);
  • CloudNull gmdistribution . fit (X, 1 )
  • MixGausData (2 : 3, 1 : 2, i) SIG(:,:,1)
  • h ezcontourf (@ (x, y) pdf (Cloud3, [x y]), [0 .8], [0 .8], 100);
  • h ezcontourf (@ (x, y) pdf (Cloud5, [x y]), [0 .8], [0 .8], 100);
  • MuTheta (Rep, i) MixGausData ( 1 , 1 , i
  • [AX,H1,H2] plotyy (X, mean (MuTheta) , X, mean (S22) , 'plot ') ;
  • figure_name_fig [ int2str (Bloop) , ' . png '] ;
  • Record_temp dlmread (' 3p5p_8um_l . txt ') ;
  • Timel Record_temp ( 5 , 1 ) ;
  • Time2 Record_temp ( 4 , 1 ) ;
  • RD_8um zeros (TI*DR, NF) ;
  • Nrows floor ( TI *DR) ;
  • Data_temp dlmread ([' 3p5p_8um_' , int2str (i) , '. txt ']) ;
  • Data_Temp_Cor ( j , 1) Data_temp ( j , 2 ) - p2*(j-l); end
  • RD_8um(:,i) Data_Temp_Cor ( 1 : Nrows , 1 ) ;
  • CenterD_8um zeros (Nrows , NF) ;
  • CenterD_8um ( j , i) RD_8um(j,i) - RD_8um(l,i);
  • Avg8umCurve zeros (Nrows, 1) ;
  • Avg8umCurve ( i , 1 ) mean (CenterD_8um ( i , SR : ER) ) ;
  • ConfNegCurve (i, 1) mean (CenterD_8um ( i , SR : ER) ) - 2*std (CenterD_8um (i, SR: ER) ) ;
  • Finf mean (Avg8umCurve (Nrows-100 :Nrows, 1) ) ;
  • FinfH mean (ConfPosCurve (Nrows-100 :Nrows, 1) ) ;
  • FH zeros ( 1 : Nrows-indexH+1 , 1 ) ;
  • FtH ConfPosCurve (i, 1) ;
  • FH ( i-indexH+1 , 1 ) (FtH - FinfH) / (FzeroH-FinfH) ;
  • FinfL mean (ConfNegCurve (Nrows-100 :Nrows, 1) ) ;
  • FtL ConfNegCurve (i, 1) ;
  • FL (i-indexL+1, 1) (FtL - FinfL) / (FzeroL-FinfL) ;
  • D fminsearch (@ForceFunc, DO, [], F, a, Xtime) ;
  • Vs 1-Fzero/Finf/2
  • PS 2* (eta/2*D* (l-2*Vs) /G/ (1-Vs) ) ⁇ (1/2) *10 ⁇ 9 %Reported in nanometers! % High error value
  • DH fminsearch (@ForceFunc, DO, [], FH, a, XtimeH) ;
  • VsH l-FzeroH/FinfH/2;
  • VsL l-FzeroL/FinfL/2;
  • ErrorSQR (F(i,l) - ( 0 . 4 91*exp (- .908 * (tau) ⁇ ( 1/2) ) + 0 . 50 9*exp(- 1 . 67 9*tau) ) ) ⁇ 2 + ErrorSQR;
  • the following code was used to analyze the local cell production images. First the images of control cells and IgG producing cells were analyzed to find the top 10000 SWNT and the cell regions were recorded using the ROI matlab function. These locations were then used to determine the coverage of the SWNT in the cell regions (specific) and outside the cell regions (un-specific) using the CellAnalysis Function (below). The time response of the IgG producing cells was determined again using the top 10,000 SWNT pixels and plotting their absolute intensity distributions over time for the series of images collected using the CellTime function (below). Finally, the AnalyzeCellPic function was used to plot the average intensity of cell islands as a contour plot as presented in FIG. 7. function CellAnalysis
  • PNSpec zeros (1 : 21, 1) ;
  • CellLoc csvread( [ ' ⁇ ' , int2str (i) , ' _CellLoc . csv ' ] ) ;
  • SWNTLoc csvread ( [ ' SLoc_Rankl0000_' , int2str (i) , ' . csv ' ] ) ;
  • CellCount sum (sum (CellLoc) ) ;
  • OutCellCount 320*256 - CellCount
  • SWNTinCell 0;
  • SWNToutCell 0;
  • SWNToutCell SWNToutCell + 1 ;
  • PNSpec (i,l) SWNToutCell/OutCellCount*100; % Percent Coverage
  • PLData imread ( [ ' HS ' , int2str (i) , ' . tif ' ] ) ;
  • SWNTLoc csvread ( [ ' SLoc_Rankl0000_' , int2str (i) , ' . csv '
  • PL_0 (countl, 1) PLData (j,k);
  • PL_1 (count2, 1) PLData (j,k);
  • PL_3 (count4, 1) PLData ( j , k) ;
  • count4 count4 + 1 ;
  • X linspace (5000, 30000, 1000) ;
  • n hist (PL_0, X) ;
  • n hist (PL_1, X) ;
  • n hist (PL_2, X) ;
  • n hist (PL_3, X) ;
  • PLData imread ( [ ' S ' , int2str (FN) , ' . tif ' ] , 1) ;
  • AvgSig Stotal/Ccount

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Immunology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Hematology (AREA)
  • Urology & Nephrology (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Microbiology (AREA)
  • Analytical Chemistry (AREA)
  • Biotechnology (AREA)
  • Nanotechnology (AREA)
  • Food Science & Technology (AREA)
  • Medicinal Chemistry (AREA)
  • Physics & Mathematics (AREA)
  • Cell Biology (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Pathology (AREA)
  • Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
  • Investigating Or Analysing Biological Materials (AREA)

Abstract

Arrays can be used to assay binding heterogeneity and other characteristics of biomanufactured products.

Description

METHOD OF ANALYSIS USING ARRAY SENSOR
FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
This invention was made with government support under Grant No. DGE- 1122374 awarded by the National Science Foundation. The government has certain rights in the invention. RELATED APPLICATIONS
This application claims priority to U.S. Ser. No. 61/767,513 filed February 21, 2013, which is hereby incorporated by reference in its entirety into this application.
TECHNICAL FIELD
The present invention relates to a method of analyzing a product using an array sensor. BACKGROUND
Improved analytical technology for increasing production of recombinant proteins is an area of great interest. Cell line generation and selecting culture parameters typically take over a year with current assays and often cell candidates are only picked based on static measurements of productivity. Some post-translational modifications, such as glycosylation patterns, can change in response to changes in process conditions (e.g., media components, temperature, pH, pC02, dissolved oxygen, cell density, duration, and others) and can have a dramatic effect on the properties of the protein product (e.g., pharmacokinetics and immunogenicity of a protein drug). Current analytical technologies for determining titer and glycosylation, such as ELISA and tandem LC/MS systems, respectively, can deliver detailed information but are costly in terms of of time, reagents, and multiplexing capabilities. Furthermore, these methods are incompatible with on-line process use.
SUMMARY
In general, a method of detecting an analyte can include using an array to provide a plurality of data points that can be used to assess the analyte.
In one aspect, a method of detecting an analyte can include contacting a sample including a biologic analyte with a sensor array, the sensor array including a plurality of photoluminescent nanostructures, collecting emissive data from each of the photolummescent nanostructures, and analyzing the emissive data from each nanostructure of the plurality of photolummescent nanostructures to measure a distribution of binding affinities. In some embodiments, the method of detecting an analyte can include calculating a calibration curve for the sensor.
In some embodiments, analyzing the emissive data can include measuring a variance of the emissive data from the plurality of photolummescent nanostructures, measuring a skew of the emissive data from the plurality of photolummescent nanostructures, or measuring a standard deviation of a plume angle of the emissive data from the plurality of photolummescent nanostructures.
In some embodiments, the method of detecting an analyte can include evaluating a binding heterogeneity of the biologic analyte of the sample from the analyzed emissive data. In some embodiments, the method of detecting an analyte can include evaluating a cell line for production of the biologic analyte of the sample from the analyzed emissive data. In some embodiments, the method of detecting an analyte can include evaluating a glycosylation pattern of the biologic analyte of the sample from the analyzed emissive data.
In some embodiments, the array can include at least 500 of the photolummescent nanostructures. In some embodiments, the array can include at least 1,000 of the photolummescent nanostructures. In some embodiments, the array can include at least 2,000 of the photolummescent nanostructures. In some embodiments, the array can include at least 5,000 of the photolummescent nanostructures.
In some embodiments, the sensor can include an analyte-binding compound associated with the photolummescent nanostructure. In some embodiments, the sensor can include a linker, wherein the analyte-binding compound is associated with the photolummescent nanostructure via the linker. In some embodiments, the sensor can include a polymer including clusters of photolummescent nanostructures.
In some embodiments, the analyte can be a biologic analyte. In some embodiments, the biologic analyte can be an antibody. In some embodiments, the biologic analyte can be a protein. In some embodiments, the antibody can be IgG. In some embodiments, the analyte binding compound can be a capture protein comprising Protein A or PSA-lectin.
Other aspects, embodiments, and features will be apparent from the following description, the drawings, and the claims. BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1A-D illustrate nanosensor array fabrication and detection methods. FIG. 1A illustrates a nIR micrograph of nanosensor array - each SWNT-illuminated pixel records changes in intensity upon analyte binding, such as nickel quenching or signal return from IgG adsorption shown. FIG. IB illustrates a multi-layer, poly-acrylamide gel platform used to immobilize the SWNT sensors and provide a porous network for IgG diffusion. SWNT sensors are wrapped in chitosan, nickel-chelated, and functionalized with a capture protein (Protein A or PSA - Lectin). FIG. 1C illustrates ensemble (i) and all-points histogram (ii) response to lOOmM nickel (1), lmg/ml Protein A (2), 1.5mg/ml Human IgG (3) and 2mg/ml BSA (4) additions. FIG. ID illustrates selectivity of the sensor array (nIR micrograph of starting intensities shown) as demonstrated by a heat map reporting the percent modulation response of the 10,000 top responding pixels after analyte addition.
FIGS. 2(a)-(f) illustrate simulation frames. 2a) Starting KD distribution from 10,000 simulated points. 2b) Distributions of Langmuir coverage fraction coefficients (9L). Resulting response of all simulated points in cartesian (2c) and polar coordinates (2d). 2e) Bivariate Gaussian fit of plume angle mean and standard deviation (θμ and θσ). Simulation sensor points at many different concentrations results in calibration curve (2f).
FIGS. 3A-D illustrate modeling a nanosensor array for measuring KD distributions. FIG. 3A illustrates a graph depicting the effect of averaging number of molecules (N) on a single sensor site on measured variance (am 2) assuming a Gaussian shaped KD distribution with known variance (at 2). FIG. 3B illustrates a design regime where at 2 can be reconstructed showing two spectrum limits (ensemble and single molecule detection) and where the nanosensor platform operates. FIG. 3C illustrates modeling nanotube array response - (i) assuming a Weibull (skewed) KD distribution, (ii) Langmuir coverage fraction to determine extent of nanosensor modulation, (iii) resulting simulated response 'plume' normalized by average sensor intensity (IaVg) where each point represents a sensor site, and (iv) calibration curve from fitting the plume angle mean and standard deviation (θμ and θσ) with a bivariate Gaussian distribution. FIG. 3D illustrates the effect of changing KD distribution skewness (β) on resulting θμ and θσ calibration curves as well as fit parameters 'B' (steepness of calibration curve) and 'C (KD mean).
FIGS. 4A-E illustrate experimental results from lyophilized IgG in PBS, murine IgG (TA99) from HEK cells, and human IgG from CHO cells screened on sensor arrays at various concentrations. FIG. 4A are graphs depicting all point sensor response 'plumes' and fit calibration curves. FIG. 4B are graphs depicting the yielding KD 95% confidence intervals, B-fit parameters, and angle saturation values (BSAT) calculated from the starting intensity distribution of each sensor batch (Eq. 5). FIG. 4C is a graph depicting relation of B-fit parameter and KD skewness parameter (β) as found by simulation results. FIG. 4D is a graph depicting measured KD distributions using assumed Weibull PDF. FIG. 4E is a graph depicting KD histograms directly calculated from sensor response assuming a constant functionalization value - revealing the approximate shape of the true PDF.
FIGS. 5A-H illustrate hypermannosylation detection on PSA-lectin conjugated sensor arrays. FIG. 5 A depicts a diagram showing that weaker mannose-PSA lectin interactions can be transduced on independent nanosensors. FIG. 5B are graphs depicting results of testing concept with chicken IgG which has > 40% high mannose glycoforms where human and mouse IgG do not contain these isoforms. FIGS. 5C-E are graphs depicting an all point 'plume' responses and calibration curve (FIG. 5D) as before with corresponding KD distribution (FIG. 5E). FIG. 5F is a graph depicting media compositions used to elicit hypermannosylation in CHO cell culture. Culture was sampled for eight days showing expected changes in IgG titer (ELISA) and mannose content (SWNT sensor array). FIG. 5G shows IgG titer for each of the culture days. FIG. 5H shows general trends of mannosylation derived from the distribution averages in FIG. 5D.
FIGS. 6(a)-(b) illustrate a surface staining experiment to confirm the presence of mannose in the CHO culture samples. 6a) A surface is treated to bind the IgG in the CHO samples and a FITC-labeled PSA lectin is used to stain the mannose content. 6b) Representative FITC fluorescent images of the stained surface showing a marked increase in mannose content as found by the SWNT nanosensors.
FIGS. 7A-E illustrate sensor response to local cell production. FIG. 7 A illustrates qualitative images of control and IgG producing cells showing colocalization of SWNT response. FIG. 7B is a graph depicting location of top 1000 SWNT in IgG and control cell images, presented as coverage percentage of cell area (specific) and gel area (non-specific) in each image (FIG. 7C). FIG. 7D are graphs depicting dynamic response of SWNT sensors to IgG producing cells plated on gel presented as distributions of top 1000 SWNT from images and micrographs of scaled SWNT intensity. FIG. 7E are micrographs depicting mapping visible IgG producing cell islands, the co localized SWNT signal underneath, and ranking the islands' productivity based on intensity normalized by island area.
FIG. 8 illustrates the experimental domain addressed by arrayed nanosensors in comparison to existing technologies. Also mapped are approximate assay times of each technique.
FIG. 9 illustrates that His-tagged sensor protein needed to reduce nonspecific binding - data presented as distribution of all SWNT sensor sites.
FIG. 10 illustrates the role of nickel in SWNT sensor modulation. 10a) Testing the current hypothesis of nickel being the crucial transduction component by swapping it out with another small molecule binding site - biotin. Histograms show sensor responses of the biotinylated chitosan- wrapped SWNT to nuetravidin and BSA. 10b) Two proposed quenching mechanisms at work - 1) the nickel acts as a quencher providing electronic states for non- radiative decay of the SWNT exciton, 2) the bound macromolecule displaces aqueous quenching species such as nickel and water from the SWNT surface. DETAILED DESCRIPTION
Methods of using sensors including plurality of photolummescent nanostructures are described. The photolummescent nanostructures are associated with analyte-binding compounds. Binding of the analyte to the analyte-binding compound changes the photolummescent properties of the photolummescent nanostructures. Thus, the presence or absence of the analyte can be determined based on the observed photolummescent properties of the photolummescent nanostructures. The change in the photoluminescence can include a change in photoluminescence intensity, a change in peak wavelength, a Raman shift, or a combination thereof. Importantly, the plurality of photolummescent nanostructures in the sensor forms an array of nanosensors, each of which provides important data about the analyte, the statistics of which are significant in interpreting the population of analyte species in a sample. Specifically, the statistical analysis based on a calibration, such as variance, skew, standard deviation and other parameters, can reflect different attributes of the analyte, such as a biologic analyte, for example an antibody or a protein product. The statistical analysis provides information about analytes in the sample, for example, binding heterogeneity of a biologic analyte, cell line for production of a biologic analyte, or glycosylation patterns of a biologic analyte. The array can include more than 100, more than 200, more than 500, more than 1,000, more than 2,000, more than 5,000 nanosensors, or more, each of which has a photoproperty that can be monitored.
In some cases, detection of analytes with photoluminescent nanostructures can be limited by the presence of non-specific background photo luminescence. Background photo luminescence can arise from, among other sources, a support that underlies the photoluminescent nanostructures, e.g., supports including glass or plastic. One way to minimize background photoluminescence is to use a support composed of highly purified materials that are free of contaminants that may be responsible for the background photoluminescence. However, the use of such materials can increase the cost and complexity of analyte detection.
Another method to avoid background photoluminescence is to use a detection system that does not detect the background photoluminescence. In such a system, substrates that suffer from background photoluminescence can nonetheless be used, because the detection system detects only photoluminescence arising from the background photoluminescence. In some cases, this might be achieved with appropriate sets of wavelength filters. However, if the background photoluminescence and the photoluminescence arising from the photoluminescent nanostructures occur at overlapping wavelengths, filters may not be able to reject a sufficient amount of the background photoluminescence while still allowing a sufficient amount of the photoluminescence arising from the photoluminescent nanostructures to pass.
It can be advantageous to use a detection system that rejects background photoluminescence on the basis of spatial configuration rather than on the basis of wavelength. One way to accomplish this is to use an optical detection system that provides a distinct focal plane, where the detected photoluminescence is substantially from sources in the focal plane. Photoluminescent sources outside the focal plane are not detected by the detector. In this way, the detector does not suffer from issues associated with background photoluminescence.
Thus, a sensor for detecting an analyte can be arranged on a support. Photoluminescence from the sensor is detected by an optical system that detects photoluminescence substantially only from sources in the focal plane. Sources outside the focal plane do not contribute substantially to the detected photoluminescence. When the sensor is arranged directly on the support, the focal plane may extend outside the boundaries of the sensor and include a portion of the support. In this situation, background photoluminescence will be detected. Rather than being arranged directly on the support, the sensor can be arranged on an intermediate substrate. The intermediate substrate, in turn, is in contact with the support. The intermediate substrate separates the sensor from the support, such that the focal plane does not include any portion of the support. The intermediate substrate can be advantegously composed of materials that are substantially free of background photoluminescence. The intermediate substrate can be advantageously composed of materials that are compatible with the sensor. For example, the sensor can include a hydrogel, and the intermediate substrate can also include a polymer matrix or a gel.
With reference to FIG. IB, sensor includes (illustrated in lower panel (ii)) support which supports substrate gel, which in turn supports sensor gel. Sensor gel (detailed in upper panel (i)) includes a gel network, a plurality of photoluminescent nanostructures and associated linking polymers, such as a polysaccharide, for example, chitosan. Optional crosslinks, for example bis- acrylamide crosslinkers can be present between linking polymers, and an analyte-binding compound, such as a nickel chelate. In FIG. IB, sensor gel is illustrated with analyte being associated with analyte-binding compound, though it will be understood that a sensor gel will typically be formed in the absence of analyte, and analyte may be contacted with sensor gel after formation. Thus in some states, sensor gel is free of analyte and in other states includes analyte (e.g., associated with analyte-binding compound). The presence of the analyte alters the photoluminescent properties of the photoluminescent nanostructures. As also illustrated in FIG. IB, the sensor array may be arranged upon one or more support materials. In some embodiments, the solid support may comprise a first gel lacking photoluminescent nanostructures and a second gel comprising photoluminescent nanostructures upon the first gel (FIG. IB). In some embodiments, the one or more support material may comprises a solid material (e.g. glass), a first gel lacking photoluminescent nanostructures, and a second gel comprising photoluminescent nanostructures upon the first gel.
The concept of vastly multiplexing analyte detection using arrays of independently addressable sensors is ubiquitous in the literature, following successful demonstrations of the DNA and protein microarrays. See, for example, Kingsmore, S.F. Nature Reviews Drug Discovery 5, 310-320 (2006), Nielsen, U.B. & Geierstanger, B.H. Journal of Immunological Methods 290, 107-120 (2004), Epstein, J.R., et al. Analytica Chimica Acta 469, 3-36 (2002), Cheng, M.M.C. et al. Current Opinion in Chemical Biology 10, 11-19 (2006), Schweitzer, B. & Kingsmore, S.F. Current Opinion in Biotechnology 13, 14-19 (2002), and Wilson, D.S. & Nock, S. Angewandte Chemie-International Edition 42, 494-500 (2003). This approach has been a prevailing motivation for further reduction in array size to nanometer dimensions. See, for example, Niemeyer, CM. Nano Today 2, 42-52 (2007), Tarn, J.M., et al. Biosensors & Bioelectronics 24, 2488-2493 (2009), and Tan, CP., et al. Nano Letters 10, 719-725 (2010). Such nanosensor arrays, however, have other important capabilities that are not as well recognized, even when functionalized for just a single analyte. For example, many biological analytes, including antibodies, demonstrate a distribution of dissociation constants even in relatively purified form. See, for example, Werblin, T.P. & Siskind, G.W. Immunochemistry 9, 987-& (1972), Pierson, L., et al. Journal of Immunological Methods 211, 97-109 (1998), and Steensgaard, J., et al. Molecular Immunology 17, 689-698 (1980). Surprisingly, an array of sensors can reconstruct this important distribution via sampling a large number of independent interactions. Such arrays can also quantify weakly-affined interactions by recording a large number of rare binding events. Notably, a nanosensor array can also characterize and differentiate biosynthesis around single cells and colonies, enabling the label- free selection of more productive strains. These properties have the potential to greatly enhance process analytics for biomanufacturing applications.
Improved analytical technology for the rapidly increasing production of clinical recombinant antibodies is an area of great interest. Three apparent needs are better tools for (1) colonal selection, (2) glycan analysis, and (3) determining the affinity distribution or heterogeneity of the expressed product. See, for example, Hacker, D.L. et al. Biotechnology Advances 27, 1023-1027 (2009), Browne, S.M., et al. Trends in Biotechnology 25, 425-432 (2007), Burke, J.F. et al. Genetic Engineering & Biotechnology News 29, 38-39 (2009), Elliott, S. et al. Nature Biotechnology 21 , 414-421 (2003), Hermeling, S. et al. Pharmaceutical Research 21 , 897-903 (2004), Jefferis, R. Biotechnology Progress 21 , 1 1-16 (2005), Jefferis, R. Nature Reviews Drug Discovery 8, 226-234 (2009), Raju, T.S., et al. Glycobiology 10, 477-486 (2000), van Berkel, P.H.C. et al. Biotechnology Progress 25, 244-251 (2009), Werblin, T.P. & Siskind, G.W. Immunochemistry 9, 987-& (1972), Pierson, L., et al. Journal of Immunological Methods 21 1 , 97-109 (1998), and Steensgaard, J., et al. Molecular Immunology 17, 689-698 (1980). Cell line generation and selecting culture parameters typically take over a year with current assays and often cell candidates are only picked based on static measurements of productivity. See, for example,Wurm, F.M. Production of recombinant protein therapeutics in cultivated mammalian cells. Nature Biotechnology 22, 1393-1398 (2004), and Browne, S.M. & Al-Rubeai, M. Trends in Biotechnology 25, 425-432 (2007), and Burke, J.F. et al. Genetic Engineering & Biotechnology News 29, 38-39 (2009).=Glycosylation patterns can easily change due to processing conditions (media components, temperature, pH, pC02, dissolved oxygen, cell density, duration, etc.) and the patterns can have a dramatic effect on the pharmacokinetics and immunogenicity of the resulting drug. See, for example, Pacis, E., et al. Biotechnology and Bioengineering 108, 2348-2358 (2011), Gramer, M.J. et al. Biotechnology and Bioengineering 108, 1591-1602 (2011), Lee, S.Y. et al. Process Biochemistry 47, 1411-1418 (2012), Ahn, W.S., et al. Biotechnology and Bioengineering 101, 1234-1244 (2008), Muthing, J. et al. Biotechnology and Bioengineering 83, 321-334 (2003), Schmelzer, A.E. & Miller, W.M. Hyperosmotic stress and elevated pCO(2) alter monoclonal antibody charge distribution and monosaccharide content. Biotechnology Progress 18, 346-353 (2002), Serrato, J. A., et al. Biotechnology and Bioengineering 88, 176-188 (2004), Majid, F.A.A., et al. Biotechnology and Bioengineering 97, 156-169 (2007), Bumbaca, D., et al. Aaps Journal 14, 554-558 (2012) and Goetze, A.M. et al. Glycobiology 21, 949-959 (2011). Current titer and glycosylation analytical technologies such as ELISA and tandem LC/MS systems respectively can deliver exquisite detail at the expense of much time and reagent. Furthermore their processing steps are prohibitive to any on-line process use. The trend to milliliter sized bioreactors for upstream process optimization will also require platforms that can accurately assay much lower protein and glycan quantities. See, for example, Lee, K.S., et al. Lab on a Chip 11, 1730-1739 (2011). Finally, harsh biomanufacturing process conditions (pH, temperature, mixing) and variability in cell production result in heterogeneous products with a distribution of binding affinities and currently there is no convenient platform on which to measure the dissociation constant (KD) distribution. Emerging nanoengineered sensors fabricated in massive (10,000+ sites) arrays, of up to 10,000 sites or more, could provide solutions to these three areas of biomanufacturing analytics. See, for example, Reuel, N.F., et al. Chemical Society Reviews 41, 5744-5779 (2012). Carbon- nanotube based optical sensors can be used for single protein and single glycan detection. See, for example, Ahn, J.H. et al. Nano Letters 11, 2743-2752 (2011), and Reuel, N.F. et al. Journal of the American Chemical Society 133, 17923-17933 (2011). By creating an array of sensors in a hydrogel platform, binding heterogeneity, weakly-affined hypermannosylation detection, and local cell productivity of biomanufactured products can be assessed.
For example, the gel can be a polymer including, but are not limited to, collagen, silicon- containing polymers, polyacrylamides, crosslinked polymers (e.g., polyethylene oxide, polyAMPS and polyvinylpyrrolidone), polyvinyl alcohol, acrylate polymers (e.g., sodium polyacrylate), or copolymers with an abundance of hydrophilic groups. The gel can have a porous structure can be determined by factors including the concentration of polymers and crosslinking agent. The pore size can be in the range of, for example, 10 nm to 1 ,000 nm, 20 nm to 500 nm, 50 nm to 250 nm, or 10 nm to 100 nm. When the analyte is a macromolecule (e.g., a protein, such as an immunoglobulin), a pore size greater than 10 nm, greater than 20 nm, greater than 30 nm, greater than 40 nm, greater than 50 nm, greater than 60 nm, greater than 70 nm, greater than 80 nm, greater than 90 nm, or 100 nm or greater can be desireable. As described above and illustrated in FIG. IB, a nanostructure may comprise different types of gels. For instance, in some embodiments, the nanostructure may comprise a first gel lacking photolummescent nanostructures and a second gel comprising photolummescent nanostructures positioned upon the first gel.
As used herein, the term "nanostructure" refers to articles having at least one cross- sectional dimension of less than about 1 μιη, less than about 500 nm, less than about 250 nm, less than about 100 nm, less than about 75 nm, less than about 50 nm, less than about 25 nm, less than about 10 nm, or, in some cases, less than about 1 nm. Examples of nanostructures include nanotubes (e.g., carbon nanotubes), nanowires (e.g., carbon nanowires), graphene, and quantum dots, among others. In some embodiments, the nanostructures include a fused network of atomic rings.
A "photolummescent nanostructure," as used herein, refers to a class of nanostructures that are capable of exhibiting photoluminescence. Examples of photolummescent nanostructures include, but are not limited to, single-walled carbon nanotubes ("SWNT"), double-walled carbon nanotubes, semi-conductor quantum dots, semi-conductor nanowires, and graphene, among others. In some embodiments, photolummescent nanostructures exhibit fluorescence. In some instances, photolummescent nanostructures exhibit phosphorescence.
If the nanostructure is a carbon nanotube, the carbon nanotube can be classified by its chiral vector (n,m), which can indicate the orientation of the carbon hexagons. The orientation of carbon hexagons can affect interactions of the nanotube with other molecules, which in turn, can affect a property of the nanostructure.
A linker can be associated with the nanostructure. The association can be a bond, for example, a covalent, ionic, van der Waals, dipolar or hydrogen bond. The association can be a physical association. For example, at least a portion of the nanostructure can be embedded in the polymer or a portion of the polymer can encompass the nanostructure.
A linker can include a polymer. A polymer can include a polypeptide, a polynucleotide or a polysaccharide. Examples of polysaccharides include dextran and chitosan. A polymer can include a plastic, for example, polystyrene, polyamide, polyvinyl chloride, polyethylene, polyester, polypropylene, polycarbonate, polyacrylamide or polyvinyl alcohol.
A polymer can be biocompatible, which can mean that the polymer is well tolerated by an organism. More specifically, biocompatibility can mean that a polymer does not elicit an immune response when it is brought in contact with an organism. It can also mean that a polymer can integrate into cell structures, cells, tissues or organs of an organism. The organism can be mammal, in particular, a human.
An exemplary polymer can exhibit minimal binding with other molecules. In certain circumstances, a polymer can have a protein adsorption of less than 5 μg/cm2, less than 1 μg/cm2, less than 0.5 μg/cm2, less than 0.1 μg/cm2, less than 0.05 μg/cm2, or less than 0.01 μg/cm2.
The association of a linker with a nanostructure can change a property of the nanostructure. The property can be conductivity, polarity, or resonance. The property can be photoluminescence, including fluorescence or phosphorescence. More specifically, the property can be fluorescence with a wavelength in the near infrared spectrum. The property can be an emission wavelength, an emission intensity, a conductance, an electromagnetic absorbance or an emittance.
A linker can be configured to interact with an analyte-binding compound. The analyte- binding compound undergoes a specific and typically reversible binding with an analyte. One class of suitable analyte binding compounds are proteins. In particular, proteins that undergo specific and typically reversible binding with an analyte (referred to herein as "capture proteins") can be suitable.
A capture protein can include a protein, a polypeptide or a peptide. In some cases, a capture protein can be a complex of proteins. A capture protein can also include a full length protein, a fragment of a protein or a protein domain. A capture protein can be a fusion protein, which can include portions originating from one protein or portions originating from more than one protein. A capture protein can include a protein tag or marker. A capture protein can also be modified, for example, by glycosylation, ubiquitination, PEGylation, SUMOylation or biotinylation. A capture protein can be synthesized from a nucleic acid sequence that was amplified from a cDNA library, genomic DNA, a DNA vector or plasmid, or a DNA fragment.
The number of linkers associated with the nanostructure present in the analysis region can exceed the number of capture proteins. More specifically, the number of capture protein binding sites on linkers associated with a nanostructure can exceed the number of capture proteins. The ratio of capture protein binding sites on linkers associated with a nanostructure to capture proteins can be greater than 1.1 to 1, greater than 1.5 to 1, greater than 2 to 1, greater than 5 to 1, or greater than 10 to 1. Having an excess of capture protein binding sites on linkers associated with a nanostructure can minimize the amount of unbound capture protein in a sample. Unbound capture proteins within the sample can compete with capture proteins bound to the composition for binding to the analyte. This can affect the accuracy and/or precision of the analyte detection. Having an excess of capture protein binding sites on linkers associated with a nanostructure can also increase the analyte concentration range over which analyte can be accurately detected because the saturation limit of the binding sites is increased.
The interaction between the linker and the capture protein can be binding to a capture protein. The linker can be configured to interact with a capture protein by including a first binding partner in the linker that can interact with the capture protein. The first binding partner can be known binding partner of the capture protein or a portion thereof. The first binding partner can include an ion. The ion can be a metal ion. The metal ion can be a nickel, iron, cadmium, copper, magnesium, calcium, arsenic, lead, mercury or cobalt ion (e.g. Ni2+, Fe2+, Cd2+, Cu2+, Mg2+, Ca2+, As2+, Pb2+, Hg2+or Co2+). The first binding partner can include a protein, a nucleotide, a saccharide, a lipid or combinations thereof. See, for example, U.S. Pat. Pub. 2012/0178640A1 (WO 2012/030961), which is incorporated by reference in its entirety.
A linker can further include a chelating region. A chelating region can include a chelator, which can be a polydentate ligand capable of forming two or more bonds with a single central atom. A chelator can include one or more carboxylate ions. For example, a linker can include Na,Na-bis(carboxymethyl)-L-lysine. A chelator can bind to a first binding partner (e.g. a metal ion) in order to incorporate the first binding partner into a linker.
The ion can act a proximity quencher of photoluminescent nanostructure. In particular, the ion can quench near infrared fluorescence. The quenching can be reversible. The quenching can also depend on the distance between the nanostructure and the ion. In other words, as the distance between the nanostructure and the ion changes, the photoluminescence from the nanostructure can also change. Generally, as the distance between the ion and the nanostructure decreases, the amount of photoluminescence quenching can increase.
In some embodiments, the capture protein can include a second binding partner, such that the first binding partner and second binding partner can bind together. The second binding partner can be an endogenous motif or endogenous domain within a capture protein. Alternatively, the second binding partner can be added to a capture protein. In some embodiments, the second binding partner can be a protein tag. A protein tag can be a peptide sequence grafted onto a protein, which can be used for separating (e.g. using tag affinity techniques), increasing solubility, immobilizing, localizing or detecting a protein. The protein tag can be a histidine tag, chitin binding protein tag, maltose binding protein tag, glutathione-S- transferase tag, c-myc tag, FLAG-tag, V5-tag or HA-tag. One method for adding a second binding partner to a capture protein can include using primers including the sequence encoding for the second binding partner to PCR amplify DNA encoding for the capture protein. A second method can include cloning DNA encoding for the capture protein into an expression vector designed to produce a fusion of the capture protein and the second binding partner.
Binding of a first and a second binding partner can be selective binding, which can provide the selectivity needed to bind to the corresponding binding partner (or relatively small group of related molecules or proteins) in a complex mixture. The degree of binding can be less than 100%, less than 90%, less than 80%, less than 70%, less than 60%, less than 50%, less than 40%), less than 30%>, less than 20%> or less than 10%> of a second binding partner present binding to a first binding partner. The degree of binding can be more than 10%, more than 20%, more than 30%, more than 40%, more than 50%, more than 60%, more than 70%, more than 80% or more than 90% of a second binding partner present binding to a first binding partner. A first binding partner and a second binding partner can bind with a dissociation constant less than 1 mM, less than 0.1 mM, less than 0.01 mM, less than 1 μΜ, less than 0.1 μΜ, or less than 0.01 μΜ. A first binding partner and a second binding partner can bind with a dissociation constant greater than lnm, greater than 0.01 μΜ, greater than 0.1 μΜ, greater than 1 μΜ, greater than 0.01 mM, or greater than 0.1 mM.
The linker can also be configured to interact with a capture protein by including a region capable of chemically reacting with the capture protein. The chemical reaction can form a covalent, ionic, van der Waals, dipolar or hydrogen bond between the linker and the capture protein.
The interaction of a capture protein with a linker associated with a nanostructure can change a property of the nanostructure. The property can be conductivity, polarity, or resonance. The property can be photoluminescence, including fluorescence or phosphorescence. The photoluminescence can be fluorescence with a wavelength within the near infrared spectrum. The property can be an emission wavelength, an emission intensity, a conductance, an electromagnetic absorbance or an emittance.
Without intending to be bound by any particular mechanism, the change in the property can be caused by a change in the distance between an ion in the first binding partner and the nanostructure. As the distance between the nanostructure and the ion changes, a nanostructure property can also change. For example, as the distance between the nanostructure and the ion changes, nanostructure photoluminescence can also change. When the capture protein binds to the linker, the distance between the ion and nanostructure can change, which can alter the nanostructure photoluminescence. Generally, as the distance between the ion and the nanostructure decreases, the amount of photoluminescence quenching can increase.
In some embodiments, a composition can further include a capture protein, which can be configured to specifically interact with at least one analyte. In particular, the capture protein can be configured to specifically bind to at least one analyte. Specific binding can be more limited than selective binding. Specific binding can be used to distinguish a binding partner from most other chemical species except optical isomers, isotopic variants and perhaps certain structural isomers. The degree of binding can be less than 100%, less than 90%>, less than 80%>, less than 70%, less than 60%, less than 50%, less than 40%, less than 30%, less than 20% or less than 10% of an analyte present binding to a capture protein. The degree of binding can be more than 10%, more than 20%, more than 30%, more than 40%, more than 50%, more than 60%, more than 70%, more than 80% or more than 90% of an analyte present binding to a capture protein. An analyte and a capture protein can bind with a dissociation constant less than 1 mM, less than 0.1 mM, less than 0.01 mM, less than 1 μΜ, less than 0.1 μΜ, or less than 0.01 μΜ. An analyte and a capture protein can bind with a dissociation constant greater than lnm, greater than 0.01 μΜ, greater than 0.1 μΜ, greater than 1 μΜ, greater than 0.01 mM, or greater than 0.1 mM.
The analyte can be a small molecule, protein, biomolecule, drug, biologic, or a metabolite thereof. For example, the analyte can be a monosaccharide, a polysaccharide, an amino acid, peptide, polypeptide, protein, a nucleotide, an oligonucleotide, a lipid, a polylipid, or a combination thereof. As one example, the analyte can be immunoglobulin G (IgG) and the capture protein can be selected to specifically bind IgG. In this case, the capture protein can be, for example, protein A or Pisum sativum agglutinin (PSA). Thus, exemplary capture proteins may include, for instance, a protein that binds an analyte (e.g., Protein A which binds antibody and / or PSA-Lectin which binds mannose). As shown in FIG. IB, such capture proteins may be physically associated with a photolummescent nanostructure and the analyte of interest (e.g., antibody). The physical association may be accomplished using one or more linkers such as a moiety on the capture protein and a second moiety that physically associates with the photolummescent nanostructure or another linker moiety physically associated with the the photolummescent nanostructure. In some embodiments (e.g., as illustrated in FIG. IB), the capture protein may be Protein A comprising a histidine tag that physically associates with nickel which is or becomes physically associated with the photolummescent nanostructure or another linker which is or becomes physically associated with the photolummescent nanostructure (e.g., chitosan). Other embodiments are also contemplated by this disclosure as would be understood by those of ordinary skill in the art.
The interaction of an analyte with a capture protein that is interacting with a linker associated with a nanostructure can change a property of the nanostructure. The property can be conductivity, polarity, or resonance. The property can be photoluminescence, including fluorescence or phosphorescence. More specifically, the property can be a fluorescent emission within the near infrared spectrum. The property can be an emission wavelength, an emission intensity, a conductance, an electromagnetic absorbance or an emittance.
The interaction of an analyte with a capture protein can be reversible, meaning that the analyte can bind to the capture protein and then release and be free of binding. The change in a property of the nanostructure due to the interaction of an analyte with a capture protein can also be reversible. For example, the property of a nanostructure can have a first value, the analyte can bind to the capture protein and alter the property to a second value, then the analyte can release from the capture protein and the property can return to the first value.
A method of detecting an analyte can include determining the presence of an analyte in the sample based on the monitored property. Determining the presence of an analyte can include determining the absence of the analyte. In some embodiments, determining the presence of an analyte can include determining the concentration of the analyte, determining the purity of the analyte or determining the quantity of the analyte. In some embodiments, relatively low concentrations or quantities of an analyte can be determined. The ability to determine low concentrations of an analyte may be useful, for example, in detecting trace pollutants or trace amounts of toxins within a subject. In some embodiments, analyte concentrations of less than about 100 micromolar, less than about 10 micromolar, less than about 1 micromolar, less than about 100 nanomolar, less than about 10 nanomolar, or less than about 1 nanomolar can be determined. The quantity of the analyte that can be determined can be less than 1 mole, less than 1 millimole, less than 1 micromole, less than 1 nanomole, less than 1 picomole, less than 1 femtomole, less than 1 attomole or less than 1 zeptomole. In some cases, a single molecule of an analyte can be determined. The purity of the analyte can be greater than 25% pure, greater than 50%, greater than 75% pure, greater than 80%, greater than 85% pure, greater than 90% pure, greater than 95% pure, greater than 99% pure or greater than 99.9% pure.
A linker can have a formula: A-L-C, where A can include a polymer, where at least a portion of the nanostructure is embedded in the polymer, L can be a linking moiety including a saturated or unsaturated C4_io hydrocarbon chain optionally containing at least two conjugated double bonds, at least one triple bond, or at least one double bond and one triple bond; said hydrocarbon chain being optionally substituted with Ci_4 alkyl, C2_4 alkenyl, C2_4 alkynyl, Ci_4 alkoxy, hydroxyl, halo, carboxyl, amino, nitro, cyano, C3-6 cycloalkyl, 3-6 membered heterocycloalkyl, unsubstituted monocyclic aryl, 5-6 membered heteroaryl, Ci_4 alkylcarbonyloxy, Ci_4 alkyloxycarbonyl, Ci_4 alkylcarbonyl, or formyl and said hydrocarbon chain being optionally interrupted by O, S, N(Ra), C(O), N(Ra)C(0)0, OC(0)N(Ra), N(Ra)C(0)N(Rb), C(0)0, or OC(0)0; each of Ra and Rb, independently, being hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, or L can be a bond, and C can be a metal ion complexing moiety. The composition can include a chelator-containing compound, which can include a chelator region and a non-chelator region. C can be the chelator region. L can include the non- chelator region.
In some circumstances, A can include a polymer [(M)x(N)y(Q)z]q, where each of M, N and Q, independently, can be selected from the group consisting of a linear or cyclic C3-C8 hydrocarbyl, heterocyclyl, cyclyl, or aryl including one or more amine, alcohol or carboxylic acid group, where each M-N, M-Q or N-Q can include O, S, N(Ra), C(O), N(Ra)C(0)0, OC(0)N(Ra), N(Ra)C(0)N(Rb), C(0)0, or OC(0)0, each of Ra and Rb, independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where each of x, y and z can be integers between 0 and 50, 0 and 20 or 0 and 10 and q can be an integer between 1 and 1000, 5 and 500, or 10 and 100.
In some circumstances, L can have the formula:
-[-Xi-(CRaRVX2-(CRaRb)o-X3-]- where each Xu X2 and X3, can be O, S, N(Ra), C(O), N(Ra)C(0)0, OC(0)N(Ra), N(Ra)C(0)N(Rb), C(0)0, or OC(0)0, each of Ra and Rb, independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where the value of n added to o can be 4 to 10.
In some circumstances, C can have the formula, HzG((CH2)nC02H)y and salts thereof, where G can be a bond, C, O, S, P, P=0 or N; n is 0-6; and z and y can be selected to satisfy the valence requirements of G. In other preferred embodiments, the compound can have the formula,
Figure imgf000018_0001
where X and X' can be the same or different and can be metal binding groups including atoms selected from the group of O, S, N, P or 0=P, including carboxyl; Y can be bond, C, O, S, P, P=0 or N; and Z can be a hydrocarbon having a backbone of one to six atoms, such as an alkyl group or alkenyl group. Each of X and X' can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties. In addition, the atoms bridging X and X' can be selected to form a 5-membered to 8-membered ring upon coordination to the metal ion. The bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur.
In yet another preferred embodiment, the compound can have the formula,
Figure imgf000019_0001
where X, X' and X" can be the same or different and can be metal binding groups including atoms selected from the group of O, S, N, P or 0=P, including carboxyl; Y can be a bond, C, O, S, P, P=0 or N; and Z can be a hydrocarbon having a backbone of one to six atoms, such as an alkyl group or alkenyl group. Each of X, X' and X" can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties. In addition, the atoms bridging X and X', X and X" or X' and X" can be selected to form a 5-membered to 8- membered ring upon coordination to the metal ion. The bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur. For example, C can be derived from HSCH2CH2CH(SH)(CH2)„COOH, H2CH2H2CH(NH2)(CH2)nCOOH,
(HOOC(CH2)n)HNCH2CH2NH((CH2)nCOOH), (HOOC(CH2)n)2PCH2CH2P((CH2)n COOH)2, (HOOC(CH2)n)2P(0)CH2CH2 P(0)((CH2)n COOH)2,
HSCH2CH2CH(SH)(CH2)4CONH(CH2)nCOOH, where can is an integer between 1 and 10, or Na,Na-bis(carboxymethyl)-L-lysine.
In some embodiments, the composition can include a nanostructure and a linker having a formula: A-L-C, where A can include the polymer covalently bonded to a portion of the nanostructure, L can be a linking moiety including a saturated or unsaturated C4_io hydrocarbon chain optionally containing at least two conjugated double bonds, at least one triple bond, or at least one double bond and one triple bond; said hydrocarbon chain being optionally substituted with Ci_4 alkyl, C2_4 alkenyl, C2_4 alkynyl, Ci_4 alkoxy, hydroxyl, halo, carboxyl, amino, nitro, cyano, C3-6 cycloalkyl, 3-6 membered heterocycloalkyl, unsubstituted monocyclic aryl, 5-6 membered heteroaryl, Ci_4 alkylcarbonyloxy, Ci_4 alkyloxycarbonyl, Ci_4 alkylcarbonyl, or formyl and said hydrocarbon chain being optionally interrupted by O, S, N(Ra), C(O), N(Ra)C(0)0, OC(0)N(Ra), N(Ra)C(0)N(Rb), C(0)0, or OC(0)0; each of Ra and Rb, independently, being hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, or L can be a bond, and C can be a metal ion complexing moiety.
The composition can include a chelator-containing compound, which can include a chelator region and a non-chelator region. C can be the chelator region. L can include the non- chelator region.
In some circumstances, A can include a polymer [(M)x(N)y(Q)z]q, where each of M, N and Q, independently, can be selected from the group consisting of a linear or cyclic C3-C8 hydrocarbyl, heterocyclyl, cyclyl, or aryl including one or more amine, alcohol or carboxylic acid group, where each M-N, M-Q or N-Q can include O, S, N(Ra), C(O), N(Ra)C(0)0, OC(0)N(Ra), N(Ra)C(0)N(Rb), C(0)0, or OC(0)0, each of Ra and Rb, independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where each of x, y and z can be integers between 0 and 50, 0 and 20 or 0 and 10 and q can be an integer between 1 and 1000, 5 and 500, or 10 and 100.
In some circumstances, L can have the formula:
-[-X1-(CRaRb)„-X2-(CRaRb)0-X3-]- where each Xu X2 and X3, can be O, S, N(Ra), C(O), N(Ra)C(0)0, OC(0)N(Ra), N(Ra)C(0)N(Rb), C(0)0, or OC(0)0, each of Ra and Rb, independently, can be hydrogen, alkyl, alkenyl, alkynyl, alkoxy, hydroxylalkyl, hydroxyl, or haloalkyl, and where the value of n added to o can be 4 to 10.
In some circumstances, C can have the formula, HzG((CH2)nC02H)y and salts thereof, where G can be a bond, C, O, S, P, P=0 or N; n is 0-6; and z and y can be selected to satisfy the valence requirements of G. In other preferred embodiments, the compound can have the formula,
Figure imgf000020_0001
where X and X' can be the same or different and can be metal binding groups including atoms selected from the group of O, S, N, P or 0=P, including carboxyl; Y can be bond, C, O, S, P, P=0 or N; and Z can be a hydrocarbon having a backbone of one to six atoms, such as an alkyl group or alkenyl group. Each of X and X' can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties. In addition, the atoms bridging X and X' can be selected to form a 5-membered to 8-membered ring upon coordination to the metal ion. The bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur.
In yet another preferred embodiment, the compound can have the formula,
Figure imgf000021_0001
where X, X' and X" can be the same or different and can be metal binding groups including atoms selected from the group of O, S, N, P or 0=P, including carboxyl; Y can be a bond, C, O, S, P, P=0 or N; and Z can be a hydrocarbon having a backbone of one to six atoms, such as an alkyl group or alkenyl group. Each of X, X' and X" can include other substituents in order to satisfy the valence requirements, such as for example, amine, thiol, phosphine or phosphine oxide, substituted by hydrogen or other organic moieties. In addition, the atoms bridging X and X', X and X" or X' and X" can be selected to form a 5-membered to 8- membered ring upon coordination to the metal ion. The bridging atoms can typically be carbon, but may be other elements, such as oxygen, nitrogen, or sulfur. For example, C can be derived from HSCH2CH2CH(SH)(CH2)„COOH, H2CH2H2CH(NH2)(CH2)nCOOH,
(HOOC(CH2)n)HNCH2CH2NH((CH2)nCOOH), (HOOC(CH2)n)2PCH2CH2P((CH2)n COOH)2, (HOOC(CH2)n)2P(0)CH2CH2 P(0)((CH2)n COOH)2,
HSCH2CH2CH(SH)(CH2)4CONH(CH2)nCOOH, where can is an integer between 1 and 10, or Na,Na-bis(carboxymethyl)-L-lysine.
Compounds can be prepared according to published procedures such as those described, for example, in Parameswara et al., Synthesis, 815-818 (1980) and Denny et al., J. Org. Chem., 27, 3404 (1962). A sensor array can include a plurality of analysis regions on a support. A support can be glass or plastic. A support may also be any material suitable for supporting the nanostructure, as would be understood by those of ordinary skill in the art. An analysis region can be a divot, a tube, a tray, a well, or a similar compartment for suitable for containing a liquid sample. In some cases, an analysis region can include a droplet or spot on the surface of a support (e.g., a flat support). In those cases, an analysis region can be formed by spotting the composition on a support. A plurality of analysis regions can be arranged in a pattern on a support. A pattern can include concentric circles, a spiral, a row, a column or a grid. In some embodiments, the plurality of analysis regions can include two or more subsets of analysis regions. For example, a first subset of analysis regions can differ from a second subset of analysis regions by including a different nanostructure, a different linker, a different binding partner, a different capture protein, a different analyte or a different sample. Additionally, a first subset of analysis regions can differ from a second subset of analysis regions by including a different environmental factor including a buffer, a reagent, a nutrient, a serum, an exposure to light, an oxygen concentration, a temperature or a pH. As described above and illustrated in FIG. IB, a nanostructure may comprise, for instance, different types of materials. For instance, in some embodiments, the nanostructure may comprise a first gel lacking photoluminescent nanostructures and a second gel comprising photoluminescent nanostructures positioned upon the first gel. In some embodiments, the first and second gel may be positioned upon a solid support material such as, for instance, glass and/or plastic. Any combination of materials may be used as a support material and are contemplated by this disclosure as would be understood by those of ordinary skill in the art.
It is widely recognized that an array of addressable sensors can be multiplexed for the label-free detection of a library of analytes, however such arrays, even when monofunctionalized, have lesser appreciated capabilities that emerge from the ensemble. For example, an array of nanosensors can estimate the mean and variance of the observed dissociation constant (KD), using three different examples of binding IgG with Protein-A as the recognition site, including polyclonal human IgG (Κ0μ = 11 - 27 μΜ), murine IgG (Κ0μ = 3.3 - 5.3 nM), and human IgG from CHO cells (Κομ = 0.6 - 16 nM). Secondly, an array of nanosensors can uniquely monitor weakly-affined analyte interactions via the increased number of observed interactions. One application involves monitoring the metabolically-induced hypermannosylation of human IgG from CHO using PSA-lectin conjugated sensor arrays where temporal glycosylation patterns are measured and compared. Lastly, the array of sensors can also spatially map the local production of an analyte from cellular biosynthesis, thus providing a powerful approach to clonal selection. As an example, productivity of IgG-producing HEK colonies cultured can be ranked directly on the array of nanosensors itself. This study opens new avenues for the use of nanosensor arrays for biomanufacturing applications.
Thus, in some embodiments, this disclosure provides methods for detecting an analyte, comprising: contacting a sample including a biologic analyte with a sensor array, the sensor array including a plurality of photolummescent nanostructures; collecting emissive data from each of the photolummescent nanostructures; and, analyzing the emissive data from each nanostructure of the plurality of photolummescent nanostructures to measure a distribution of binding affinities. In some embodiments, the photolummescent nanostructures may be physically associated with a linker (e.g., "wrapped" in chitosan); directly or indirectly "chelated" with a metal ion (e.g., nickel); and / or "functionalized" with a capture protein (e.g., Protein A, PSA-Lectin) (e.g., "functionalized" refers to a direct or indirect physically association providing a functional relationship between the photolummescent nanostructure and the capture protein) (see, e.g., FIGS. IB and 5A). In some embodiments, the photolummescent nanostructures may be selected from the group consisting of carbon nanotubes, single-walled carbon nanotubes, double-walled carbon nanotubes, semi-conductor quantum dots, semi-conductor nanowires, and graphene. In some embodiments, the sensor array is arranged upon one or more support materials which may be at least one gel. In some embodiments, the one or more support materials comprises a first gel lacking photolummescent nanostructures and a second gel comprising photolummescent nanostructures upon the first gel. Any number of layers of gels or other support material may be included in the nanostructures. The one or more support materials may also include a solid material (e.g. glass) supporting the photolummescent nanostructures. Thus, in some embodiments, the one or more support materials comprises a solid material (e.g. glass), a first gel lacking photolummescent nanostructures, and a second gel comprising photolummescent nanostructures upon the first gel. In some embodiments, the capture protein is Protein A and/or a derivative thereof (e.g., where the analyte is antibody) or PSA-lectin and/or a derivative thereof (e.g., where the analyte is mannose). In some embodiments, the sample may comprises cell that colonize the photolummescent nanostructures and the methods provide for the detection of one or more analytes produced by the cells is detected. In some embodiments, the methods described herein may also comprise: calculating a calibration curve for the sensor; analyzing the emissive data by measuring a variance of the emissive data from the plurality of photoluminescent nanostructures (where the emissive data may include measuring a skew of the emissive data from the plurality of photoluminescent nanostructures and/or measuring a standard deviation of a plume angle of the emissive data from the plurality of photoluminescent nanostructures); evaluating a binding heterogeneity of the biologic analyte of the sample from the analyzed emissive data; evaluating a cell line for production of the biologic analyte of the sample from the analyzed emissive data; and/or evaluating a glycosylation pattern of the biologic analyte of the sample from the analyzed emissive data. The arrays described here may include, for instance, at least about any of 10, 20, 30, 40, 50, 60 70, 80, 90, 100, 125, 150, 175, 200, 225, 250, 275, 300, 325, 350, 375, 400, 425, 450, 475, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900, 2000, 2100, 2200, 2300, 2400, 2500, 2600, 2700, 2800, 2900, 3000, 3100, 3200, 3300, 3400, 3500, 3600, 3700, 3800, 3900, 4000, 4100, 4200, 4300, 4400, 4500, 4600, 4700, 4800, 4900, or 5000 photoluminescent nanostructures. These methods may also comprise sensors that include an analyte -binding compound associated with the photoluminescent nanostructure. In some embodiments, the sensor may comprises a linker, wherein the analyte-binding compound is associated with the photoluminescent nanostructure via the linker. In some embodiments, the sensor may comprise a polymer including clusters of photoluminescent nanostructures. In some embodiments, the analyte is a biologic analyte such as a protein (e.g., an antibody such as IgG). In some embodiments, the analyte binding compound is a capture protein comprising Protein A or a derivative thereof or PSA-lectin or a derivative thereof. Thus, in such embodiments, the analyte may be any that specifically interacts with Protein A or PSA-lectin, respectively. In some embodiments, the capture protein (e.g,. Protein A or a derivative thereof or PSA-lectin or a derivative thereof) comprises a tag having specificity for a binding partner associated with the photoluminescent nanostructure. In some embodiments, the tag is a histidine tag and the binding partner is nickel. Other embodiments are also contemplated by this disclosure as would be understood by those of ordinary skill in the art. A better understanding of the present invention and of its many advantages will be had from the following examples, given by way of illustration. EXAMPLES
Example 1
A. Sensor Fabrication and Detection Method
Single-walled carbon nanotubes (SWNT) are arrayed (FIG. 1A) in a multi-layer, highly- porous (60-90nm) polyacrylamide hydrogel to reduce nonspecific binding and allow for fast diffusion of the IgG analytes to the SWNT sensors (FIG. 1A). See, for example, Ruchel, R., et al. Journal of Chromatography 166, 563-575 (1978), and Wang, J., et al. Advanced Materials 20, 4482-4489 (2008). A basal substrate gel with no SWNT is used on a glass substrate to position the thin sensor layer in a separate focal plane and eliminate the background fluorescence caused by glass impurities. For example, the platform is excited by a 660nm laser (Crystal Laser) on an inverted microscope (Zeiss D. l) and the nIR emission is collected as an image stack on a 256x320 pixel InGaAs array (Princeton Instruments Acton Array). The SWNT are suspended in chitosan and have been chemically modified as before to display chelated nickel groups that act as both the docking site for a His-tagged capture proteins (Protein A and PSA Lectin in these examples) and as the signal transducer (proximity quencher). See, for example, Reuel, N.F. et al. Journal of the American Chemical Society 133, 17923-17933 (2011), which is incorporated by reference in its entirety. In brief, the nanotube acts as an optical switch, brightening as antibodies bind to the capture protein. The ensemble response of the array is created by averaging the intensity values over the entire array for each time point and is analogous to other bioassay techniques like ELISA and Biacore® SPR measurements. Divalent nickel, protein A, and IgG addition (lOOmM, 1 mg/ml, 1.5 mg/ml) cause a decrease, increase, and additional increase respectively (FIG. lC(i)). Subsequent washing of the gel surface shows negligible effects on the ensemble signal, which can be interpreted as absence of unbinding. Adding a BSA control (2mg/ml) does not elicit a sensor response (FIG. lC(i)). Alternatively, the responses of all SWNT pixels can be monitored as histograms of percent modulation (I0-lFinai/I0) (Fig, lC(ii)). Sensor specificity can also be observed qualitatively with a nIR heatmap filtered to the 10,000 most responsive sites (FIG. ID). To render the sensor array specific, the sensor protein (Protein A or PSA lectin) must first be docked to the chelated nickel; otherwise BSA would elicit a response. Although the ensemble response can be efficiently used to construct calibration curves and monitor titer and glycosylation trends, the averaging loses the valuable information about the affinity distribution recorded by each of the individual sensor elements. B. Modeling a Nanosensor Array to Measure KD Distributions
A numerical simulation was also used to determine the analytical relation between KD variance (β - skewness parameter in (pi) and the fit nanosensor response (B-fit parameter). In a single iteration of the program, 10000 sensor sites are generated with random KD, I0, and PF values generated from (pi, φ2, and φ3 respectively. The KD values, centered at ΙΟΟηΜ (FIG. 2(a)), determine the coverage fraction coefficients (9L) (FIG. 2(b)) as specified in the paper (Eq. 2). The response of each sensor is determined for 29 concentration points between InM and 10μΜ (Fig. 2(c)). The responses are then fit in polar coordinates (FIG. 2(d)) with a bivariate Gaussian distribution (FIG. 2(e)). The plume angle mean and standard deviation (θμ and θσ) are then recorded for each concentration. The program completes 6 iterations to create a smooth calibration curve (FIG. 2(f)). The initial KD skewness parameter (β) is then changed and the program is run again. This was done for 200 different β values spanning 2 to 45 and the results are shown in FIG. 3D, as described in more detail below. The full code for this simulation can be found in Example 2(K)(1) (Code 3a).
In addition to the more obvious uses of macroscale sensor arrays (efficient replicates, multiplexed response, spatial mapping of response, etc.), nanoscale sensor arrays could be used to measure intrinsic properties of the sample such as affinity distributions. To mathematically study the emergent properties of monofunctionalized sensor arrays, one can consider an analyte to have a Gaussian distribution of KD with a set variance (at 2) which is then assayed on a 1 mm2 array of sensors. If this sensor area is divided in enough independent sensor regions, so that each molecule is assayed independently, the full distribution can be recovered. If each sensor element, however, reports an average of multiple molecule signal transductions the measured sample variance (am 2) becomes much smaller than the actual value (FIG. 3A insert). For a
-1/2
Gaussian distribution of KD, this decay of (am/ at) appears to be proportional to (N)~ where N is the number of molecules averaged on each sensor site (FIG. 3A). This relation informs the effect of concentration and sensor area and identifies the regime where nanosensor arrays can effectively report the variance of KD (FIG. 3B - assumes lOnm2 area for antibodies). The platform uses nanotubes arranged in 1.4μιη2 pixels in the concentration ranges of nM to μΜ, thus determining distributions of KD among samples is feasible when the intrinsic variance of the sensor response is low. A model is first needed to map our experimental array response back to an initial variance in binding. To simulate the array response, a more realistic distribution of KD for an antibody is used. The Weibull distribution has been proposed for antibody affinity due to its ability to effectively present skewness in the probability distribution function using two shape parameters (a and β). See, for example, Steensgaard, J., et al. Molecular Immunology 17, 689- 698 (1980). Antibody affinity is intuitively non-symmetric as changes in the optimal protein will more often result in reduced binding and rarely enhance the product. Thus an antibody with positive skew in logio(KD) is modeled (tail with higher KD values or analytes with less affinity) and the probability density function (PDF) is expressed as:
Eq.
Figure imgf000027_0001
This distribution is initially centered at KD = 100 nM (can be translated for other values) such that a = 7 and β can vary from 2 to 45+ with 2 being a very large tail (high variance) and 45+ being an equally monodisperse distribution (FIG. 3C(i)). The sensor response is modeled by the Langmuir equation (Eq 2) where the coverage fraction (9L) is determined by 1/KD and sample concentration (C) (FIG. 3C(ii)). In this case, 9L represents the extent to which a nanotube sensor is turned on by a small number of local interacting molecules during the time of light acquisition (Is per frame).
a, /?)
Figure imgf000027_0002
Eq. 2
The extent to which a nanotube sensor responds to adsorption is dictated by its starting intensity (Io) and extent of functionalization (PF) which can both be experimentally measured and fit with another Weibull (φ2) and Gaussian distribution((p3) respectively. It is important to note that the number of binding sites on the senor protein is also imbedded in the functionalization variance (if all nanotubes were uniformly wrapped, functionalized, and had the same number of binding sites, this variable would be a constant). To stochastically simulate the response of an ensemble, for example, individual sensor responses, R, can be obtained as a coverage fraction, intensity, and percent functionalization value (Θ, I0, and PF,) randomly generated from their respective PDFs (cpi, cp2, (p3) and subsequently evaluated as:
Eq.
Figure imgf000028_0001
The response is normalized by the average intensity (IaVg) as each experimental platform's overall intensity may change due to variance in the experimental setup or quality of SWNT. This allows for clear comparison of distinct arrays assayed at different concentrations. By plotting against the normalized starting intensity ((I0 - Imin) / (Imax _ Imin)) characteristic response 'plumes' are observed (FIG. 3C(iii)). The 'plume' is the distribution of data points shown in the plot. The experimental 'plumes,' are fit well by a bivariate Gaussian distribution in polar coordinates (6,R) and the mean 'plume' angle (θμ) is used to create the calibration curve (FIG. 2C(iv)). The standard deviation in the plume angle (θσ) also has an interesting dependence on starting KD skew (β) and is reported (FIG. 3C(iv)). The (θμ) calibration curve is fit by a four parameter logistics curve (used for many other bioassays that exhibit a signal saturation - i.e. ELISA) however in this case two parameters are known from experimental conditions (A = 9min = 0 and
x = log10(C)
Figure imgf000028_0002
Eq. 5
The parameter B can be thought of as a measure of 'steepness' and C is equal to the inflection point at KD. By analyzing the model at many different extents of KD skewness (2 <β <45), the effect of KD variance on the fit parameters were found (FIG. 3D). As the distribution of antibody affinities becomes tighter (β increases) the nanosensor array yields a steeper calibration curve (FIG. 3C fits this relation). This model can be used to analyze experimental data.
C. KD Distributions and Calibration Curves from Experimental Data
The array of nanotube sensors conjugated to His-tagged Protein A (Abeam) was used to assay three different samples of IgG with expected differences in affinity distributions: (1) commercial, lyophilized, polyclonal Human IgG reconstituted in PBS, (2) murine IgG (TA99) from an engineered human embryonic kidney (HEK) cell line, and (3) human IgG (bl2) cultured from Chinese hamster ovary (CHO) cells. The data 'plumes' predicted by the model where observed for each of the systems (FIG. 4A) and the resulting calibration curves (FIG. 4B) yielded KD mean values (95% confidence intervals for each: 11 - 27 μΜ, 3.3 - 5.3 nM, and 0.6 - 16 nM) comparable to those found in literature for IgG-Protein A interactions (2-50 nM from SPR, 34 nM from acoustic device (see, for example, Nohlden, S. in Department of Physics, Chemistry and Biology, Vol. Masters 68 (Linkopings universitet, Linkoping; 2008), and Saha, K., et al. Analytical Chemistry 75, 835-842 (2003)). The calibration curve also provided 'B' fit parameters (Eq. 4) that are related to the starting distribution skew parameter (β) as solved from the model simulation (FIG. 4C). With KD mean (a) and skewness (β), the affinity PDF of each system (Eq. 1) can be determined (FIG. 4D). As expected, the freshly expressed Human IgG has the greatest affinity for Protein A with the least amount of variance. The murine antibody has a comparable KD average, but much more predicted variance. This difference could be due to less efficient binding of murine IgG to Protein A as observed in literature. See, for example, Richman, D.D., et al. Journal of Immunology 128, 2300-2305 (1982). Finally the lyophilized, polyclonal human IgG shows a lOOOx reduction in KD and a much broader distribution of affinities likely due to denaturation damage (see, for example, Schersch, K. et al. Journal of Pharmaceutical Sciences 99, 2256-2278 (2010), which is incorporated by reference in its entirety) or freeze-thaw cycles (vs. the freshly expressed CHO product). One limitation of the modeling approach is the a priori assumption of PDF form (in this case Weibull). The array of nanosensors can be used to report distributions directly, without any assumption of PDF form by calculating the Langmuir coverage from the response of each individual sensor (Eq. 2-3) and creating a histogram of KD values (FIG. 4E). To do this now, however, one assumes a constant percent functionalization value (PF = 0.14, mean of φ3) and chooses a concentration value away from saturation (these assumptions are an approximation as a distribution of functionalization (φ3) exists and each concentration does not align perfectly on the calibration curve, so each 'plume' will yield a slightly different KD histogram). This direct technique for measurement can be improved by better fabrication methods, reducing the intrinsic variance of length and functionalization. The direct reporting approach also presents KD distributions with similar shape and positive skew as the model assumed.
D. Hypermannosylation Detection - Weakly affined interactions on nanosensor arrays
Another advantage of nanosensor arrays is their ability to report weak binding events - a greater number of individual sensor sites increases the probability of a detection event and this event is not averaged to null with other non-responsive sites, as in the case for an ensemble sensor. The label-free nature of the platform is also beneficial to detecting weakly-affined ligands since it requires no washing steps. By swapping out the His-tagged Protein A with a His- tagged, mannose-specific plant lectin, Pisum sativum agglutinin (PSA) the sensor platform can detect specifically high mannose content IgG (FIG. 5A). Different species of IgG were initially used to test this concept. Chicken IgG contains an appreciable amount of glycoforms with high mannose content (>40% of population) whereas these are virtually absent in human and mouse IgG. See, for example, Raju, T.S., et al. Glycobiology 10, 477-486 (2000), which is incorporated by reference in its entirety. The SWNT sensor responses to human, mouse, and chicken IgG in PBS align with these findings and confirm that the platform can detect mannose species with the PSA lectin specifically (FIG. 5B). A well characterized, high-mannose content IgG sample from a microbial source was then used to further validate detection and obtain a calibration curve and KD distribution as before (Figs. 5C-E). The mannose IgG - PSA affinity (KD = 1.3 - 55 μΜ 95% confidence interval) is comparable to literature values (μΜ > KD > mM for lectin-glycan interactions) with a broad distribution (FIG. 5E). See, for example, Hirabayashi, J. Journal of Biochemistry 144, 139-147 (2008). This is expected as 60% of the IgG sample is aglycosylated (verified by MS analysis) and the remaining 40% bear differing lengths and structures of high mannose-type glycans (verified by released glycan analysis).
It has also been shown that changing culture conditions such as media components can affect the onset and extent of IgG hypermannosylation. In order to further validate the platform, an experiment was designed for CHO cultures in which levels of NaCl were increased and a MnCl2 additive was used while monitoring mannose content over time using traditional Peptide- N-Glycosidase F release and capillary electrophoresis. See, for example, Pacis, E., et al. Biotechnology and Bio engineering 108, 2348-2358 (2011). Four identical dishes of CHO cells were cultured and fed media compositions derived from this study (FIG. 5F). The supematants were collected after each 24 hour period, diluted to a standardized lOng/ml IgG concentration and assayed on our PSA rendered sensor gels. The IgG concentrations were determined with ELISA (FIG. 5G) and if below lOng/ml, the sample was run at stock concentration. The resulting trends (FIG. 5H) determined by the mean percent modulation from the nanosensor array distributions match those found in the previous study: 1) increased mannose content as culture time increases, 2) increased mannose from higher NaCl osmolality, and 3) delayed onset of hypermannosylation from MnCl2 additive. The presence of mannose in these samples was confirmed by surface staining with fluorescently tagged PSA and the sensitivity demonstrated by the array exceeded the capabilities of the current protein A capture followed by LC/MS technology. Due to the low volumes of sample and the low level of antibody expression, the antibody capture step was unsuccessful.
An alternative method was therefore used to confirm the presence of mannose in the CHO cultures. A surface staining approach was used (FIG. 6(a)) in which Protein A conjugated to a porous resin bead (Pierce 53139) was immobilized to a poly-lysine coated glass surface via glutaraldehyde in 2.5mm diameter Teflon patterned wells. The daily CHO samples diluted to lOng/ml IgG (used in Manuscript Figure 5 assay) were then spotted on the glass slide at 20μ1. The IgG was allowed to interact for 2 hours and the chip was then washed in PBS. The chip was then spotted with 300μg/ml BSA to block non-specific sites and washed with PBS again. Finally FITC-conjugated PSA (Vector Labs FL-1051) was spotted on the sample at 25ug/ml and allowed to interact overnight. The samples were then washed with PBS and assayed on an inverted microscope reading the visible FITC emission. Qualitatively, the observation of more conjugated FITC-PSA corresponded with the days peak mannose was observed using the nanosensor platform (FIG. 6(b)). This supports the finding of mannose in the IgG samples.
E. Nanosensor Arrays for Monitoring Local Cell Colony Production
A gel with an imbedded array of nanosensors can be used to screen local production of cells. Unfortunately single cells are very difficult to culture for long periods of time on the current porous platform (little indications of healthy, single cell adherence). The single cells that did culture well, however, displayed colocalization of IgG production on a Protein A-incubated gel (FIG. 7A). By seeding a greater number of cells, larger colonies of cells formed on the porous gel surface and were able to produce for longer time periods (24 hours). By analyzing images of control and IgG-producing HEK cells islands after 24 hours of production, a statistical difference between the two production profiles can be seen (see Code 3 c presented in Example 2(K)(3)). By ranking the brightest 1000 SWNT and then querying their location, there is a greater localization of the bright SWNT under IgG producing islands where as they are evenly or randomly distributed within and outside of the control cell islands (FIGS. 7B,C). HEK cells producing IgG were plated on a gel and acquired multiple images of the nIR intensities at 0, 1, 2, and 3 hours. Histograms of the 1000 brightest SWNT pixels in these images show a 'turn-on' trend that is likely due to IgG production (FIG. 7D). Finally, large islands of HEK cells were allowed to grow overnight on a Protein A gel. The nIR response was clearly co-localized under each of the islands, and the response was summed, averaged over the island area and each island was ranked based on productivity (FIG. 7E).
Carbon nanotube-based fluorescent sensor arrays can be used for monitoring distributions of KD, hypermannosylation, and local cellular production of IgG with clear implications in biomanufacturing. The platform was demonstrated with lyophilized IgG in PBS as well as characterizing freshly-expressed IgG in complex media from three different cellular expression systems: HEK, CHO, and a fungal cell line. The sensor array was rendered specific to mannose with PSA-lectin and trends in metabolically induced hypermannosylation from a previous study were confirmed. Finally, local production of IgG from HEK cell colonies cultured on sensor arrays was monitored. Better upstream colony selection could be performed with a sensor gel optimized for healthy cell culture. The cell colonies could be exposed to various culture and media conditions and their productivity and glycosylation patterns could be monitored in real time. This platform could lead to more rapid and informed selection of master cell lines and culture conditions based on multiple parameters rather than picking colonies based on static snapshots of productivity provided by current assays. See, for example, Burke, J.F. et al. Genetic Engineering & Biotechnology News 29, 38-39 (2009). During production, sensor arrays in a microfluidic platform can be used to monitor product titer, KD distribution, and glycosylation by periodically sampling the bioreactor, filtering cellular components, diluting to a set level depending on the cell line's average productivity read the fluorescent signal, and then regenerate for the next sample. Protein A gels can be regenerated using a pH 3.0 release wash, similar to regeneration of Protein A purification columns with little loss of sensitivity. Detection of mannose has been validated here but other glycans of interest (galactose, fucose, sialic acids, and non-human, immunogenic glycans like gal-a 1,3-gal (see, for example, Chung, C.H. et al. New England Journal of Medicine 358, 1109-1117 (2008)) could also be detected by multiplexing portions of the nanosensor array with different His-tag lectins.
His-tag lectins can be made by chemical conjugation of a His-tag and a target lectin. A kit, such as a his-tag linkage kit (Solulink) can be used to form a polyhistidine tagged lectin. Alternatively, for example, suitable starting materials can include a native lechtin (Vector Labs), a linker such as SMCC linker (Pierce), an activating agent such as Traut's reagent (Pierce), and a polyhistidine such as hexahis peptide (Abbiotec). An example protocol can include the following steps:
Prepping SMCC stock
1) Dissolve 6 mg SMCC in 1 mL DMF (the official protocol called for 1.5 mg/mL, but we increased it to avoid adding too much DMF to the protein solution)
Prepping Protein, Adding SMCC
1) Make a 1 mg/mL PBS solution of each lectin you want to tag
2) 25 uL of the SMCC stock was added per mL protein solution (this is assuming the protein is on the order of -50-75 kDa)
3) Incubate this reaction mixture at room temp for 30 min
Prepping Traut stock, Use Traut on His-tag Peptide
1) Make a 14 mM solution of Traut's reagent in H20 (e.g. 2 mg/mL)
2) For a 1 mg/mL solution of HexaHis, add 4.6 uL Traut stock per mL HexaHis solution
3) Incubate at room temp for 1 hour
Tagging the Protein
1) Again assuming the protein is on the order of -50-75 kDa, add 16.8 uL of Traut+His solution per mL of Protein solution (1 mg Protein per mL PBS)
Suitable glycans and lectins can be selected based on the properties to be detected. Examples of glycans or lectins can be identified in the literature. See, for example, van Berkel, P. FL, et al. Biotechnology Progress, Volume: 25, Issue: 1, pages: 244-251 (2009), Hirabayashi, J., J. of 'Biochemistry Volume: 144, Issue: 2, pages: 139-147 (2008), Gabius, H.-J. et al., Trends in Biochem. Sci. Volume: 36, Issue: 6, pages: 298-313 (2011), Dam, T.K.; et al. Glycobio. Volume: 20, Issue: 3, pages: 270-279 (2010), ): Tateno, H. et al. Methods in Enzymology, Volume: 478, pages: 181-195 (2010), and Hsu, K.-L. et al. Current Opinion in Chem. Bio. Volume: 13, Issue: 4, pages: 427-432 (2009), and Krishnamoorthy, L., et al. ACS Chem. Bio. Volume: 4, Issue: 9, pages: 715-732 (2009).
The longstanding goal of nanosensor arrays is to preserve the sensitivity and analytical advantages of single-molecule nanosensors with the multiplexing ability of macroscale techniques, thus filling an untapped analytical regime (FIG. 8). The fast assay time (< 5 min) of nanosensor arrays could also provide a disruptive alternative to the more time intensive ELISA and LC/MS analytics that are currently used (FIG. 8). The current limitations of these arrays are the intrinsic variances caused by non-automated production in small batches (16-32 gels per batch). Small variations in polymer casting time, initiator concentration, and washing procedures result in gels with varying levels of functionalization and sensitivity. A standardized gel from an automated printing/production system could reduce this variance and provide a robust tool for biomanufacturing analytics and beyond.
Example 2
Methods and Materials
A. SWNT Sensor and Gel Platform Fabrication
SWNT were suspended in chitosan according to Reuel, N.F. et al. J. Am. Chem. Soc. 133, 17923-17933 (2011). In brief, 3 mg of purified HiPCO SWNT (Unidym) were added to 20ml of chistosan suspension (0.25 wt% in water containing 1 vol% acetic acid - Sigma). The resulting mixture was tip sonicated (1/4" tip Cole Parmer,Model CV18) at 10W for 45 minutes in an ice bath and table -top centrifuged three times at 13.2 RPM for 90 min each, while collecting the suspended SWNT supernatant and discarding the aggregate pellet after each cycle. The SWNT was then mixed at a 50:50 volume ratio with the polyacrylamide mixture for casting as the top layer. The amount of monomer (acrylamide) and cross-linker (Ν,Ν'-Methylenebisacrylamide - both Sigma) are specified using standard %T %C nomenclature, where %T refers to the overall weight % of polymer (monomer and crosslinker) in the solution and %C refers to the wt% of the total polymer that is cross linker. Surface gels with a 3%C and 1%T composition performed well. A substrate gel (6%T, 1%C) was also prepared. TEMED (Tetramethylethylenediamine) was added in at 0.7 vol% in both the top and bottom gel solutions to stabilize the radical reaction. A fresh initiator solution of 1 wt% Ammonium persulfate (APS) was made immediately prior to each gel batch. The APS, bottom and top gel solutions, and substrate chips (8 chamber Lab-Tek by Nunc) were degassed in the glove box antechamber to remove absorbed and dissolved oxygen. Within a nitrogen controlled glove box (MBraun LABstar), 1 vol% of the APS solution was added to the substrate gel to initiate the polymerization and it was immediately cast (lOOul to each well) and then allowed to cure for one hour. The top gel was then initiated with 1 vol% APS and immediately spotted at 20 ul to each gel surface and allowed to cure for 1 hour.
The functionalization steps of the chitosan wrapped SWNT is also similar to that described in Reuel, N.F. et al. J. Am. Chem. Soc. 133, 17923-17933 (2011). In brief, the amine groups of the chitosan were reacted with succinic anhydride (133 mM in PBS 7.4 buffer - Sigma) overnight and then washed thoroughly with water. The carboxylic acid functional groups were then activated with lOOmM EDC and 520mM NHS (Sigma) in MES Buffer pH 4.7 (Pierce) for 2 hours. After washing with water thoroughly, the gels were then reacted with 34mM Να,Να- Bis(carboxymethyl)-L-lysine hydrate (Sigma) in PBS 7.4 buffer overnight. The gels were then washed and incubated with a lOOmM nickel sulfate solution for 20 minutes. These chips were then washed thoroughly in water and stored in water.
As explained in section C, the sensors were manufactured for use with a his-tag sensor protein ((Protein A (Abeam) or PSA lectin). The original SWNT sensors for protein and glycan recognition from our group were cast in a chitosan hydrogel. Not only were the desired attributes of large pore size and local SWNT functionalization hard to control in this gel, but the chitosan proved to exacerbate the non-specific binding of the more 'sticky' IgG and Protein A species. Adding Protein A-Histag followed by Human IgG (Fig. Sla in U.S. Ser. No. 61/767,513) resulted in the expected loading curves but when other proteins were swapped for Protein A, like EFGP-Histag, we still observed a loading curve from the IgG (Fig. Sib in U.S. Ser. No. 61/767,513). Furthermore, Protein A would also non-specifically bind to the gel, as Protein A without a Histag also resulted in binding (Fig. Sic of U.S. Ser. No. 61/767,513). A set of control experiments show that the docked sensor protein (Protein A) is needed to render the gel response specific, i.e., suppress a response from BSA (FIG. 9). B. Poroelastic Relaxation Indentation and Dextran Release Curves
The top acrylamide gel layer containing the SWNT sensors was tuned to a maximum pore size to ensure that large antibodies could diffuse to the sensor sites. The effect of crosslinker concentration on pore size has been established qualitatively in literature with TEM imaging (Ruchel, et al. Transmission-Electron Microscopic Observations Of Freeze-Etched Polyacrylamide Gels. Journal of Chromatography 166, 563-575 (1978)). It is noted that these experiments were performed while protecting the radical polymerization from quenching oxygen species in a nitrogen-filled glove box. It was found that solutions with as low as 3 wt% polymer (97% water) were able to crosslink in this environment. It was determined from FITC dextran release profiles (Fig. S2d of U.S. Ser. No. 61/767,513) that the largest dextran probe (31.4nm in diameter) was able to freely diffuse into the 3%>T 1%>C optimized gel (%>C is standard nomenclature for the crosslinker weight percent of the total polymer weight percent (%T); thus, in this case, it would be 0.03 wt% of the total mixture).
The pore size was also more rigorously probed using a recent hydrogel characterization technique called microscale poroelastic indentation by AFM as explained in the Kalcioglu, Z.I., et al. (From macro- to microscale poroelastic characterization of polymeric hydrogels via indentation. Soft Matter 8, 3393-3398 (2012)) Soft Matter 8, 3393-3398 (2012); see also , which is incorporated by reference in its entirety. In brief, a short silicon tip with a 45 um polystyrene sphere (Novascan) was fitted on an AFM (Asylum Research - MFP3D) and the IgorPro software indentation panel was used to drive the tip into the gel at a specified distance and record the force over time. This was done in replicates for multiple sites on each gel type. A custom Matlab algorithm was then used to analyze the force relaxation curves and determine the average pore size. FITC-conjugated dextran particles (Invitrogen) of various sizes (10, 40, 70, and 500 kD) were also absorbed into 150 μΐ cylindrical gel plugs over 48 hours. The impregnated gels were then removed, washed, and inserted into clean water. The release of the FITC particles was observed by sampling the exterior fluid and assaying the FITC content with a plate reader. Using standard curves, the release was then determined in terms of cumulative mass release over time.
In more detail, a microsphere that was orders of magnitude larger than the average pore size (R = 22.5 um) is indented into the gel at a specified distance (h) and held while the displaced solvent leaks into surrounding regions. The resulting force versus time curve (Fig. S2b of U.S. Ser. No. 61/767,513) exhibited an initial spike (F0) that relates to the shear modulus of the gel (G in Eq 6) and then relaxed to an equilibrium state (F∞) which relates to the diffusion constant of the displaced fluid. The force relaxation curve was obtained for multiple penetration depths (h = 8, 6, and 4 urn), normalized (Fig S2c of U.S. Ser. No. 61/767,513) and fit by a numerically solved model for an indented sphere geometry3 to determine the diffusion constant (D in Eq 7) and Poisson ratio (vs in Eq 8). These parameters and the viscosity of water (η) were then used to solve for the average pore size of the network (ξρ in Eq 9) (Chan, et al. Spherical indentation testing of poroelastic relaxations in thin hydrogel layers. Soft Matter 8, 1492-1498 (2012)). These measurements yielded an average pore size of 57-93 nm (95% confidence interval) which is 5-9 times the calculated size (-10-1 lnm) of a hydrated IgG antibody (Armstrong, et al. The hydrodynamic radii of macromolecules and their effect on red blood cell aggregation. Biophysical Journal 87, 4259-4270 (2004)) (see Code 3b preseted in Example 2(K)(2)).
Figure imgf000037_0001
Figure imgf000037_0002
= 2(l-vs) Eq. 8
Figure imgf000037_0003
C. Analyzing SWNT Clusters with SEM, Raman Mapping, and TEM
In an attempt to understand the variation between sensor gel batches, substantial characterization of the polyacrylamide hydrogel network was performed. In these experiments, the thin, top polyacrylamide gel entraps the SWNT sensors in a sparsely cross-linked network. If the chitosan wrapped SWNT are added to the solution containing acrylamide monomers and bis- acrylamide crosslinker immediately prior to radical polymerization initiated by ammonium persulfate, the resulting gel has a homogenously distributed SWNT pattern. Conversely, if the SWNT are able to interact with the monomer and crosslinker solution during the bottom gel curing time, the amine groups on the chitosan-wrapped SWNT undergo linear step growth Michael Addition Polymerization with the bisacrylamide cross-linker (forming poly amido amines) leading to larger fluorescent SWNT clusters. The cluster sizes are dependent on the interaction time of these two constituents before starting the radical polymerization that consumes the remaining crosslinkers and immobilizes the clusters. These clusters can be clearly seen in the nIR micrographs (Fig. S2a of U.S. Ser. No. 61/767,513), raman maps, visible micrographs (Fig. S4 of U.S. Ser. No. 61/767,513), SEM (Figs. S2a and S3 of U.S. Ser. No. 61/767,513), and TEM imaging (Fig S5) and were found to be more responsive to IgG than the gels with sparse, evenly distributed SWNT. This is likely due to binding events affecting a greater population of SWNT when in cluster formations versus interacting with only a few SWNT when patterned as more sparse elements. The clusters visible in the top gel nIR micrographs were characterized with SEM (Fig. S3 of U.S. Ser. No. 61/767,513) and the presence of SWNT was confirmed with Raman mapping (Fig. S4 of U.S. Ser. No. 61/767,513) and TEM (Fig. S5 of U.S. Ser. No. 61/767,513). The SEM images were created from flash dried samples as this was found to preserve the structural integrity of the underlying gel and some of the clusters in the top gel. D. Data collection on nIR Inverted Microscope
SWNT sensor data was collected on a custom inverted microscope (Zeiss D.l Observer) that was fitted with a 660 nm laser (Crystal Laser, 100 mW). A 20x planar objective (Zeiss) was used and the emission intensities were recorded by a nitrogen-cooled InGaAs array (Princeton Instruments). Win Spec software (Princeton Instruments) was used to collect the SWNT emission and saved as an image stack TIF file. This file was then analyzed using Matlab. Analyte samples were added to the sensor gels by hand, applying the ΙΟΟμΙ sample to the lower right corner of the well, so as not to place the plastic pipette tip in the laser beam path. To prepare a sensor gel for testing, it was first thoroughly washed with PBS to exchange the buffer and then allowed to incubate with the his-tag sensor protein (Protein A (Abeam) or PSA lectin (Vector Labs - conjugated to His-tag peptide (Abbiotec) via Traut's reagent and SMCC linker (Pierce)) at 500μ§/ηι1 overnight. The gel was again washed thoroughly with PBS and then fitted on the microscope for testing.
E. Quenching and Signal Transduction Studies
Based on previous work (Reuel, N.F. et al. Transduction of Glycan-Lectin Binding Using
Near-Infrared Fluorescent Single-Walled Carbon Nanotubes for Glycan Profiling. Journal of the American Chemical Society 133, 17923-17933 (2011); Ahn, J.H. et al. Label-Free, Single Protein Detection on a Near-Infrared Fluorescent Single -Walled Carbon Nanotube/Protein Microarray Fabricated by Cell-Free Synthesis. Nano Letters 11, 2743-2752 (2011)), it was postulated that the chelated nickel used to bind the His-tag sensor protein was responsible for the modulation of the SWNT fluorescence. Experiments clearly showed that nickel quenches chitosan-wrapped SWNT in solution and immobilized on a surface (Fig. 10). Divalent nickel cations have been shown to quench anionic surfactant wrapped SWNT8 and quantum dots (Wu, et al. Ni(2+)- modulated homocysteine-capped CdTe quantum dots as a turn-on photoluminescent sensor for detecting histidine in biological fluids. Biosensors & Bio electronics 26, 485-490 (2010)). The proximity of the chelated nickel groups alter the local electronic environment and offer non- radiative decay pathways for the SWNT exciton. To explore if this is the only mechanism at work, we conceived of an experiment to replace the nickel with another small molecule proxy - biotin. The chitosan wrapped SWNT are biotinylated with a commercially available NHS linker (Pierce EZ-Link NHS-Biotin) and then exposed to nuetravidin and a BSA control (500μg/ml). The specific response is again a positive modulation and the control is null (FIG. 10(a)) although the turn-on response is approximately 20% of what is typically seen when nickel is present. This supports two interoperating mechanisms (FIG. 10(b)): (1) the sensor proteins are originally quenched by chelated nickel groups and the return of signal occurs when the nickel is displaced upon binding of the sensor protein or IgG, and (2) the chitosan wrapped SWNT is originally quenched by aqueous quenching species (water molecules (Strano, M.S. et al. The Role of Surfactant Adsorption during Ultrasonication in the Dispersion of Single- Walled Carbon Nanotubes. Journal of Nanoscience and Nanotechnology 3, 81-86 (2003)), dissolved oxygen, and protons (Blackburn, J.L. et al. Protonation effects on the branching ratio in photoexcited single-walled carbon nanotube dispersions. Nano Letters 8, 1047-1054 (2008)) have all been shown to quench exposed portions of suspended SWNT) and the bound macromolecule displaces these species to cause a return in signal. A macroscopic swelling mechanism (as in previous work with PVA hydrogels (Barone, P.W. et al. Modulation of Single- Walled Carbon Nanotube Photoluminescence by Hydrogel Swelling. ACS Nano 3, 3869-3877 (2009)) was ruled out as the recorded movies of SWNT arrays during testing show that the SWNT sensors are immobilized and do not move on the pixel resolution recorded (1.2 μιη per pixel).
F. Gaussian distribution (am/ ot) decay is proportional to (Number Sites Averaged)"1 2
To verify Gaussian distribution (am/ at) decay is proportional to (Number Sites
Averaged) -"1/2 , a simulation in which a Gaussian distribution with at is assumed and then a new distribution is created by selecting a number of points (Navg) from the original distribution and then measure this distribution's new standard deviation (am), where:
N = number of sites in full distribution
Navg = number of sites averaged together from full distribution
1 1
o = ^ [Oi - )2 + x2 - μ 2 +■■■ + xN - μ 2] where μ =— χ 1 + - + xN)
1 1 °m = — [(*i - μ)2 + (*2 - μ)2 + - + ON - )21 where μ =—(χ 1 + - + xN)
1
N, avg
Figure imgf000040_0001
By running the simulation multiple times and averaging the simulation results, the relationship found above clearly emerges. Using the simulation of averaging a number of points (Navg) from a given distribution with standard deviation at and determining the ratio of the new standard deviation (am) with the original, the -1/2 power law determined above emerges (as illustrated by FIG. S8 of U.S. Ser. No. 61/767,513). G. Starting Intensity of the SWNT (I0) and the Percent Functionalization (PF)
Two additional inputs to the sensor response are the starting intensity of the SWNT (I0) and the percent functionalization (PF). Both will determine the extent to which a given sensor can modulate. Both can be fit with probability density functions (PDF) from experimental data. An all points histogram of pixel intensity (Fig S9a of U.S. Ser. No. 61/767,513) reveals two distinct populations of pixels - those that are not illuminated by the laser spot and the pixels that contain SWNT emission. The starting SWNT intensities for each sensor gel can be fit by a Weibull distribution (Fig S9b of U.S. Ser. No. 61/767,513) and vary in their mean values (a) and skewness of bright SWNT (β) depending on processing conditions, age of SWNT suspension etc. However, this does not affect the simulation results as variation in the SWNT I0 distribution translates to shifting of the center of the plume distribution in the radial distance when performing the Gaussian bivariate fit in polar coordinates. The plume angle is used to report the calibration curve, and not the radial position, thus any positively skewed Weibull distribution can be used. For the simulation (see Code 3 a presented in Example 2(K)(1)), the following PDF fit (Fig S9b of U.S. Ser. No. 61/767,513) to the first data set (nickel quench in FIG. la,d):
where x = I, 0> a = 6872, and β = 2.2
Figure imgf000041_0001
The extent of sensor functionalization was determined by monitoring the extent of quenching caused by nickel addition. This provides a reasonable measure to the accessibility and functionalization of each nanotube. The percent quenching was found to be invariant in SWNT length (banded when plotted vs. I0) and approximated well by a Gaussian distribution (Fig S9c-d of U.S. Ser. No. 61/767,513) with the following PDF (Fig S9d of U.S. Ser. No. 61/767,513):
1 -(χ-μ)2
φ3 (χ; μ, σ) =— ~ j= e 2cj2 where χ = ΡΡ, μ = —14.2, and σ = 13.9
σ 2π
It was found that the population is fit with a Weibull PDF (Fig. S9b of U.S. Ser. No. 61/767,513) and can be created for each experiment. The population was found to be conserved amongst separated gels and is fit with a Gaussian PDF (Fig. S9d of U.S. Ser. No. 61/767,513) (histogram of quenching percent after nickel addition). H. HEK Cell Line Generation and CHO Origin
A tricistronic expression cassette pLB2-CMV-GFP-TA99 was created using 2A skip peptides (see Hu, T., et al. Biotechnology Letters 31, 353-359 (2009), which is incorporated by reference in its entirety). The light and heavy chain sequences of TA99, a murine IgG2a antibody (Thomson, T.M., et al. J. of Investigative Dermatology 85, 169-174 (1985)), were linked by a T2A sequence. The expression cassette was cloned into the lentiviral vector, pLB2 (Stern, P. et al. PNAS 105, 13895-13900 (2008), which is incorporated by reference in its entirety), modified with a CMV promoter driving GFP-F2A expression, creating the complete plasmid sequence of pLB2-CMV-GFP-F2A-LC-T2A-HC. All cloning was performed using overlap extension PCR. HEK-GFP-TA99 cells were generated using a modified version of a previously described protocol (see Kuroda, H., et al. J. of Virological Methods 157, 113-121 (2009)). Briefly, HEK- 293FT cells (Invitrogen) were transfected with the following plasmids: pLB2-CMV-GFP-TA99, pCMV-dR8.91, and pCMV-VSV-G at a mass ratio of 2: 1 : 1 using PEI (see Zufferey, R., et al. Nature Biotechnology 15, 871-875 (1997); and Dull, T. et al. J. of Virology 72, 8463-8471 (1998)). After 24 hours, fresh media was exchanged. 48 and 72 hours later, supernatant containing lentiviral particles was harvested. HEK-293 cells were transduced twice for 24 hours by incubation with freshly harvested supernatant supplemented with protamine sulfate at 5 μg/mL. GFP positive cells were selected to a purity of greater than 95% using flow fluorescence activated cell sorting. Details on the generation of the CHO cell line can be found in Hezareh, M., et al. J. Virology 75, 12161-12168 (2001). Additional data is presented in Supplement 2e of of U.S. Ser. No. 61/767,513.
I. Cell Passaging and Culture
For standard culture the medias used were DMEM (with 4.5 g/L glucose, 10% heat inactivated FBS (Invitrogen), 2 mM L-glutamine, 100 U/ml Penicillin, 100 U/ml Streptomycin - rest Sigma), and GMEM (same additives - Sigma) for the HEK and CHO cultures, respectively. For cultures used in experiments, a serum free media was used for growth and as a buffer in the sensor gel (Invitrogen Freestyle 293). To passage the cells, they were allowed to grow to confluence, washed with PBS, and then released with Trypsin (0.05%> w/0.53 mM EDTA). The cells were then pelleted, resuspended in fresh media and diluted at a 1 :5 ratio. The cells were passaged every 2-3 days and discarded after the 20th passage. J. Hypermannosylation CHO Culture Experiment
CHO cells were seeded at equal density in small culture flasks (25 cm2 - Sarstedt) and allowed to grow to confluence with regular GMEM media (overnight). The growth media was then exchanged with 3ml serum free media (Freestlye 293 Invitrogen) and the cells were allowed to produce for 24 hours. The media was then saved and a fresh 3ml of serum-free media was added for the next 24 cycle. This was repeated for 10 days. The IgG content was measured by ELISA (ICL Lab, Inc.) and the samples were diluted to lOng/ml in Freestyle to run on the PSA- incubated SWNT gels.
K. Codes Utilized in the Examples
1. Code 3a: Modeling a Nanosensor Array
The following code was used to simulate the effect of the KD distribution skewness parameter (β) on the calibration fit parameter: function Simulation_Effect_of_Beta
% Coded by Nigel F. Reuel on 1 . 8 . 12
% Used to explain:
% 1 . How going from KD --> K --> LangTheta --> Plume of Data
% 2. How to fit the plume of data
% 3. Effect of changing KD distribution skewness parameter (Beta) on fit
% Number of sensor points to simulate:
Npoints = 10000 ;
% Number of concetration points between those that are graphically
represented !
BP = 4 ;
Nconc = BP* 6+5 ;
logC = linspace ( - 9 , - 5 , Nconc) ; % ## < Change Cone Range here!! Lt = zeros (Npoints, Nconc) ;
Kvec = zeros (Npoints , 1 ) ;
KDGlobalMean = 7 ;
% ##### KD Beta Parameter Loop Start
BetaMin = 2 ;
BetaMax = 45 ;
nB = 200 ; % ## < Change number of Beta parameters here
BetaSpan = linspace (BetaMin, BetaMax, nB) ;
ThetaSAVE = zeros (nB, Nconc) ;
SigmaSAVE = zeros (nB, Nconc) ;
ThetaSAVEstd = zeros (nB, Nconc) ;
SigmaSAVEstd = zeros (nB, Nconc) ;
for Bloop = 1 : nB Beta = BetaSpan ( 1 , Bloop) ;
% ##### Replicate Loop START ######
NR = 6; % < Number of Replicates here
MuTheta = zeros (NR, Nconc) ;
S22 = zeros (NR, Nconc) ;
for Rep = 1:NR
for i = 1: Nconc
for j = l:Npoints
K = l Cr (wblrnd (KDGlobalMean, Beta) ) ;
C = 10Λ (logC (1, i) ) ;
Lt ( j , i) = K*C/ (1+K*C) ;
Kvec ( j , 1) = K;
end
if Rep == NR
if i == 1
bins = linspace (-11.1, -2.9, 200) ;
KDL = loglO (1. /Kvec) ;
[n x] = hist (KDL, bins) ;
subplot (2, 3, 1) % < Plot 1 bar (x, n)
xlim ( [-11 -3] )
xlabel ( ' loglO (K_D) ' )
ylabel ( ' requency ' )
title ('K_D Distribution')
text (-7, max (n) * .95, [ ' \leftarrow \beta = ' , int2str (Beta) ] ) end
subplot (2, 3, 2) % < Plot 2
[n x] = hist (Lt ( : , i) , 200) ;
if i == 1+BP
plot (x, n, ' b . ' )
elseif i == 2*BP+2
plot (x, n, ' r . ' )
elseif i == 3*BP+3
plot (x,n, 'g. ' )
elseif i == 4*BP+4
plot (x, n, 'm. ' )
elseif i == 5*BP+5
plot (x, n, ' k. ' )
end
xlim( [0 1] )
hold on
end
end
%Cspan = 10. Λ (logC) *10Λ9; % Reported in nM
if Rep == NR
%hlegl = legend ([ int2str (Cspan ( 1 , 1+BP) ), '
nM' ] , [int2str (Cspan (1, 2*BP+2) ) , ' nM ' ] , [int2str (Cspan (1, 3*BP+3) ) , ' nM' ] , [int2str (Cspan (1, 4*BP+4) ) , ' nM ' ] , [int2str (Cspan (1, 5*BP+5) ) , ' nM ' ] ) ;
%set(hlegl, 'Location', 'Northwest')
title (' Langmuir Coverage')
ylabel ( ' requency ' )
xlabel (' \theta - Coverage Fraction')
hold off end
NSWNT = Npoints;
Length = zeros (NSWNT, Nconc) ;
Response = zeros (NSWNT, Nconc) ;
RespNorm = zeros (NSWNT, Nconc) ;
INorm = zeros (NSWNT, Nconc) ;
% Create matrix of SWNT lengths, then do the calculation:
for i = 1 : Nconc
for j = 1: NSWNT
Length ( j , i ) = wblrnd (6871.7, 2.2)
end
end
% Values to Normalize SWNT Length
MaxL = max (Length) ;
MinL = min (Length) ;
MeanL = mean (Length) ;
%MeanLNorm = (MeanL - MinL) ./ (MaxL - MinL)
for i = 1 : Nconc
for j = 1: NSWNT
Theta = Lt(j,
L = Length(j,
Response ( j , i ) = Theta*L;
RespNorm ( j , i ) = Theta*L/MeanL (1, i) ;
INorm ( j , i ) = (L - MinL(l,i)) / (MaxL(l,i) MinL (1, i)
end
end
% Plot the expected response at different concentrations
%subplot (3, 3, 6) %
%plot (INorm ( : , 1) , RespNorm ( : , 1) , ' .b ' , INorm ( : , 2) , RespNorm ( : , 2) , ' . r ' , INorm ( : , 3) , RespNorm ( : , 3 ) , ' . g ' , INorm ( : , 4 ) , RespNorm ( : , 4 ) , ' . m ' , INorm ( : , 5 ) , RespNorm ( : , 5) , ' . k ' )
%legend ( [int2str (Cspan (1,1)), ' nM ' ] , [int2str (Cspan (1,2)), '
nM' ] , [int2str (Cspan (1, 3) ) , ' nM ' ] , [ int2str (Cspan ( 1 , 4 ) ) , '
nM' ] , [int2str (Cspan (1,5)), ' nM ' ] )
%title (' Simulated Sensor Response - No Functionalization Parameter')
%ylabel ( ' (Io*\theta) /I_a_v_g' )
%xlabel ( ' (Io - I_m_i_n) / ( I_m_a_x-I_m_i_n) ' )
% But the SWNT have been functionalized to a certain extent
PM_Ni = csvread ( ' PM_Nickel . csv ' ) . *100;
%[n x] = hist (PM_Ni, 200) ;
%subplot (3, 3, 7) %
%bar (x, n)
%xlim ( [-35 0] )
%title ( 'Nickel Functionalization' )
%ylabel ( ' requency ' )
%xlabel ( ' % Quench' )
% Read in the nickel quenching data to create your Gaussian fit:
Gfit = gmdistribution . fit (PM_Ni, 1) ;
%GMean = Gfit.mu
%GSigma = Gfit. Sigma;
%Gm = mean (PM_Ni) ;
%Gs = std (PM Ni) ; %Gfit2 = gmdistribution (Gm, Gs , 1 ) ;
%subplot (3, 3, 8) %
%X = linspace (-50, 10, 100) ' ;
%Y = pdf (Gfit, X) ;
%plot(X,Y)
%xlim( [-35 0] )
%xlabel ( ' % Quench' )
%ylabel ( ' f (x; \mu, \sigma) ' )
%title (' Gaussian Fit of Functionalization Extent')
% Now plot what the experimental ouput would look like with
% functionalization distribution
FuncResponse = zeros (NSWNT, Nconc) ;
FuncRespNorm = zeros (NSWNT, Nconc) ;
for i = 1 : Nconc
for j = 1: NSWNT
Theta = Lt (j , i) ;
Io = Length ( j , i ) ;
FuncRange = Io* (1/ (1-random (Gfit) /-100) - 1 ) ;
FuncResponse ( j , i ) = Theta*FuncRange ;
FuncRespNorm (j , i) = Theta*FuncRange/MeanL (1, i) ;
end
end
if Rep == NR
subplot (2, 3, 3) % < Plot 3 plot (INorm ( : , 1+BP) , FuncRespNorm ( : , 1+BP) , ' .b ' , INorm ( : , 2*BP+2) , FuncRespNorm ( : , 2 *BP+2) , ' . r ' , INorm ( : , 3*BP+3) , FuncRespNorm (:,3*BP+3), '.g', INorm ( : , 4*BP+4) , FuncR espNorm (:,4*BP+4), '.m', INorm ( : , 5*BP+5) , FuncRespNorm (:,5*BP+5), '.k')
%hleg2 = legend ([ int2str (Cspan ( 1 , 1+BP) ), '
nM' ] , [int2str (Cspan (1, 2*BP+2) ) , ' nM ' ] , [int2str (Cspan (1, 3*BP+3) ) , '
nM' ] , [int2str (Cspan (1, 4*BP+4) ) , ' nM ' ] , [int2str (Cspan (1, 5*BP+5) ) , ' nM ' ] ) ;
%set(hleg2, 'Location', 'Northwest')
title (' Sensor Response - Cartesian')
ylabel ( ' (Io* (1/ (1-PF) -1) * \theta/ I_a_v_g ' )
xlabel ( ' (Io-I_m_i_n) / ( I_m_a_x-I_m_i_n) ' )
ylim( [0 1.2] )
end
% Now plot what this plume reduces to with your radial analysis
% {
ThetaMean = zeros ( 1 , Nconc) ;
L = zeros ( 1 , Nconc) ;
U = zeros ( 1 , Nconc) ;
ThetaSig = zeros ( 1 , Nconc) ;
Lsig = zeros ( 1 , Nconc) ;
Usig = zeros ( 1 , Nconc) ;
%}
MixGausData = zeros ( 3 , 2 , Nconc) ;
for i = 1 : Nconc
Xn = INorm ( : , i) ;
Yn = FuncRespNorm (:, i ) ;
[THETA1 RHOl] = cart2pol (Xn, Yn) ;
% Plot according to the point density (to see what this looks like) ... %nl = 100;
%n2 = 100;
%[z,C] = hist3([RH01 THETA1 ] , [nl n2]); %xb = C{1,1}; % RHO Values
%yb = C{1,2}; % THETA Values
clear X
X ( : , 1) = THETA1;
X ( : , 2) = RHOl;
if Rep == NR
subplot (2, 3, 4) %< Plot 4 if i == 1+BP
plot ( THETA1 , RH01, '.b')
hold on
Cloudl = gmdistribution . fit (X, 1 ) ;
MU = Cloudl. mu;
SIG = Cloudl . Sigma;
MixGausData (1, 1 : 2, i) = MU;
MixGausData (2 : 3, 1 : 2, i) = SIG(:,:,1);
%ezcontourf (@ (x, y) pdf (Cloudl, [x y] ), [0 .8], [0 1 ] ) ;
elseif i == 2*BP+2
plot ( THETA1 , RHOl, ' .r')
Cloud2 = gmdistribution . fit (X, 1 ) ;
MU = Cloud2.mu;
SIG = Cloud2. Sigma;
MixGausData (1, 1 : 2, i) = MU;
MixGausData (2 : 3, 1 : 2, i) = SIG(:,:,1);
%ezcontourf (@ (x, y) pdf (Cloud2, [x y]), [0 .8], [0 1 ] ) ;
elseif i == 3*BP+3
plot ( THETA1 , RH01, '.g')
Cloud3 = gmdistribution . fit (X, 1 ) ;
MU = Cloud3.mu;
SIG = Cloud3. Sigma;
MixGausData (1, 1 : 2, i) = MU;
MixGausData (2 : 3, 1 : 2, i) = SIG(:,:,1);
%ezcontourf (@ (x, y) pdf (Cloud3, [x y] ) , [0 .8], [0 1 ] ) ;
elseif i == 4*BP+4
plot ( THETA1 , RH01, '.m')
Cloud4 = gmdistribution . fit (X, 1 ) ;
MU = Cloud4.mu;
SIG = Cloud4. Sigma;
MixGausData (1, 1 : 2, i) = MU;
MixGausData (2 : 3, 1 : 2, i) = SIG(:,:,1);
%ezcontourf (@ (x, y) pdf (Cloud4, [x y] ) , [0 .8], [0 1 ] ) ;
elseif i == 5*BP+5
plot ( THETA1 , RH01, '.k')
Cloud5 = gmdistribution . fit (X, 1 ) ;
MU = Cloud5.mu;
SIG = Cloud5. Sigma;
MixGausData (1, 1 : 2, i) = MU;
MixGausData (2 : 3, 1 : 2, i) = SIG(:,:,1);
%ezcontourf (@ (x, y) pdf (Cloud5, [x y] ) , [0 .8], [0 1 ] ) ;
else
CloudNull = gmdistribution . fit (X, 1 ) ; MU = CloudNull . mu;
SIG = CloudNull . Sigma;
MixGausData (1, 1 : 2, i) = MU;
MixGausData (2 : 3, 1 : 2, i) = SIG(:,:,1) end
hold on
else
CloudNull = gmdistribution . fit (X, 1 )
MU = CloudNull .mu;
SIG = CloudNull . Sigma;
MixGausData (1, 1 : 2, i) = MU;
MixGausData (2 : 3, 1 : 2, i) = SIG(:,:,1)
end end
if Rep == NR
xlim([0 .75])
ylim( [0 1] )
title (' Sensor Response - Polar')
hold off
xlabel ( ' \theta ' )
ylabel('R')
% Represent polar fits as bivariate distributions
subplot (2,3,5)
h = ezcontourf (@ (x, y) pdf (Cloudl, [x y]), [0 .8], [0 .8], 100);
set (h, 'LineStyle', 'none');
hold on
h = ezcontourf (@ (x, y) pdf (Cloud3, [x y]), [0 .8], [0 .8], 100);
set(h, 'LineStyle', 'none');
hold on
h = ezcontourf (@ (x, y) pdf (Cloud5, [x y]), [0 .8], [0 .8], 100);
set(h, 'LineStyle', 'none');
hold on
colormap (hot)
title (' Bivariate Gaussian PDF Fit')
xlabel ( ' \theta ' )
ylabel('R')
hold off
end
% Now in your last plot, plot the parameters of interest (Rho-Mu and Rho % sigmall) for i = 1 : Nconc
MuTheta (Rep, i) = MixGausData ( 1 , 1 , i
S22(Rep,i) = MixGausData (2, 1, i) ;
end
Rep
% ##### End for Replicate Loop #####
end X = logC;
subplot (2,3,6)
[AX,H1,H2] = plotyy (X, mean (MuTheta) , X, mean (S22) , 'plot ') ;
set (get (AX (1) , 'Ylabel ' ) , ' String' , ' \theta_\mu ' )
set (get (AX (2) , 'Ylabel ' ) , ' String' , ' \theta_\sigma ' )
set (AX (1) , ' YLim' , [0 .6] )
set (AX (2) , 'YLim' , [0 .1] )
xlabel ( ' loglO (Cone) ' )
title (' Calibration Curve')
set (HI, ' LineStyle ' , '— ')
set (H2, ' LineStyle ', ' : ')
o
% Set up the picture, save and close! set(gcf, 'Position', [5 5 1000 1000])
figure_name_fig = [ int2str (Bloop) , ' . png '] ;
print (gef , ' -dpng ' , figure_name_fig) ;
close
ThetaSAVE (Bloop, : ) = mean (MuTheta) ;
SigmaSAVE (Bloop, : ) = mean(S22);
ThetaSAVEstd (Bloop, : ) = std (MuTheta) ;
SigmaSAVEstd (Bloop, : ) = std(S22);
Bloop
% ###### End for BetaLoop
end
% Save the calibration curve data:
csvwrite ( ' ThetaMu_mean . csv ' , ThetaSAVE) ;
csvwrite ( ' ThetaMu_std. csv ' , ThetaSAVEstd) ;
csvwrite ( ' ThetaSig_mean . csv ' , SigmaSAVE) ;
csvwrite ( ' ThetaSig_std . csv ' , SigmaSAVEstd) ; end
2. Code 3b: Poroelastic Indentation Analysis Code
This code was used to analyze the force relaxation curves presented above to identify the hydrogel pore size: function AFM_PRI_Analysis_v4_8um
% Coded by Nigel F. Reuel on 8.7.2012
% Used to analyze the PRI curves derived from AFM as the Van Vliet group
% has demonstrated - Analyzes the curves to find the following:
% 1. G = shear modulus
% 2. D = diffusion rate
% 3. Vs = Poisson's ratio
% 4. PS = Pore Size
o
% 8.8.2012 v2 provides 95% error limits on the pore calculation
% 8.8.2012 v3 corrects for drift present in low crosslink gels
% 8.27.12 v4 adjusted for new data set (8.23.12 experiments)
8um Indentation Step 30s indentation % (1) Read in the raw data:
% Number of files, time of indentation, and data rate
NF = 18;
TI = 20; % In seconds
% Determine the recording rate:
Record_temp = dlmread (' 3p5p_8um_l . txt ') ;
Timel = Record_temp ( 5 , 1 ) ;
Time2 = Record_temp ( 4 , 1 ) ;
DR = 1/ (Timel-Time2) ; % In Hertz
RD_8um = zeros (TI*DR, NF) ;
Nrows = floor ( TI *DR) ;
for i = 1:NF
Data_temp = dlmread ([' 3p5p_8um_' , int2str (i) , '. txt ']) ;
% Check and correct for drift, using last 5000 time points: L = floor ( TI *DR) ;
L2 = 5000;
yp = Data_temp (L-L2+1 : L, 2) ;
xp = (1 : L2) ' ;
p = polyfit (xp, yp, 1) ;
plot (xp, yp)
%pl = slope and p2 = intercept
% Correct for slope:
p2 = 0;
Data_Temp_Cor = zeros (L,l);
for j = 1 : L
Data_Temp_Cor ( j , 1) = Data_temp ( j , 2 ) - p2*(j-l); end
RD_8um(:,i) = Data_Temp_Cor ( 1 : Nrows , 1 ) ;
end
% Center the values so the initial Force is zero
CenterD_8um = zeros (Nrows , NF) ;
for i = 1:NF
for j = 1 : Nrows
CenterD_8um ( j , i) = RD_8um(j,i) - RD_8um(l,i);
end
end
% Try plotting together to see what they look like:
X = (1 : Nrows) ' ;
plot (X, CenterD_8um)
% Good replication in 1-14
Avg8umCurve = zeros (Nrows, 1) ;
ConfPosCurve = zeros (Nrows, 1) ;
ConfNegCurve = zeros (Nrows, 1) ;
SR = 1;
ER = 14;
for i = l:Nrows
Avg8umCurve ( i , 1 ) = mean (CenterD_8um ( i , SR : ER) ) ;
ConfPosCurve (i, 1) = mean (CenterD_8um ( i , SR : ER) ) +
2*std (CenterD_8um (i, SR: ER) ) ;
ConfNegCurve (i, 1) = mean (CenterD_8um ( i , SR : ER) ) - 2*std (CenterD_8um (i, SR: ER) ) ;
end % Plot the averaged curve with error bars
plot (X, Avg8umCurve , X, ConfPosCurve , ' r-- ' , X, Conf egCurve , 'r--')
csvwrite ( ' AvgF_8um . csv ' , Avg8umCurve )
csvwrite ( ' AvgFP_8um . csv ' , ConfPosCurve )
csvwrite ( ' AvgFN_8um . csv ' , Conf egCurve )
% Plot the force relax curve (w/ error bars)
[Fzero index] = max (Avg8umCurve ) ;
Finf = mean (Avg8umCurve (Nrows-100 :Nrows, 1) ) ;
F = zeros ( 1 : Nrows-index+1 , 1 ) ;
for i = index:Nrows
Ft = Avg8umCurve ( i , 1 ) ;
F ( i-index+1 , 1 ) = (Ft - Finf ) / (Fzero-Finf ) ;
end
Xtime = (( 1 : Nrows-index+1 )* 1 /DR) ' ;
% High error
[FzeroH indexH] = max (ConfPosCurve ) ;
FinfH = mean (ConfPosCurve (Nrows-100 :Nrows, 1) ) ;
FH = zeros ( 1 : Nrows-indexH+1 , 1 ) ;
for i = indexH:Nrows
FtH = ConfPosCurve (i, 1) ;
FH ( i-indexH+1 , 1 ) = (FtH - FinfH) / (FzeroH-FinfH) ;
end
XtimeH = ( (1 :Nrows-indexH+l) *1/DR) ' ;
% Low error
[FzeroL indexL] = max (ConfNegCurve ) ;
FinfL = mean (ConfNegCurve (Nrows-100 :Nrows, 1) ) ;
FL = zeros ( 1 : Nrows-indexL+1 , 1 ) ;
for i = indexL:Nrows
FtL = ConfNegCurve (i, 1) ;
FL (i-indexL+1, 1) = (FtL - FinfL) / (FzeroL-FinfL) ;
end
XtimeL = ( (1 :Nrows-indexL+l) *1/DR) ' ;
% Plot together...
plot (Xtime, F, XtimeH, FH, ' r-- ' , XtimeL, FL, ' r-- ' )
% Save the F and X file
csvwrite ( ' Force_8um . csv ' , F) ;
csvwrite ( ' Time_8um . csv ' , Xtime ) ;
% Solve for tau
R = 22.5*10Λ-6; %Radius of the bead
h = 8*10Λ-6; % Depth of the indentation
a = (R*h) Λ (1/2) ;
eta = 1*10Λ-3; % (N s/m2)
DO = .000000002;
% Calculated value
D = fminsearch (@ForceFunc, DO, [], F, a, Xtime) ;
G = Fzero*3/16/a/h;
Vs = 1-Fzero/Finf/2;
PS = 2* (eta/2*D* (l-2*Vs) /G/ (1-Vs) ) Λ (1/2) *10Λ9 %Reported in nanometers! % High error value
DH = fminsearch (@ForceFunc, DO, [], FH, a, XtimeH) ;
GH = FzeroH*3/16/a/h;
VsH = l-FzeroH/FinfH/2;
PSH = 2* (eta/2*DH* (l-2*VsH) /GH/ (1-VsH) ) Λ (1/2) *10Λ9 %Reported in nanometers! % Low error value
DL = fminsearch (@ForceFunc, DO, [], FL, a, XtimeL) ; GL = FzeroL*3/16/a/h;
VsL = l-FzeroL/FinfL/2;
PSL = 2 * (eta/2*DL* (l-2*VsL) /GL/ (1 -VsL ) ) Λ (1/2) * 10 Λ 9 %Reported in nanometers! end function FZet = ForceFunc ( DO , F, a, Xtime)
rows = length (F);
ErrorSQR = 0 ;
for i = 1 : rows
tau = D0*Xtime (i, 1 ) / a 2 ;
ErrorSQR = (F(i,l) - ( 0 . 4 91*exp (- .908 * (tau) Λ ( 1/2) ) + 0 . 50 9*exp(- 1 . 67 9*tau) ) ) Λ2 + ErrorSQR;
end
FZet = ErrorSQR;
end
3. Code 3c: Local Production Analysis of Cells on Gel Images
The following code was used to analyze the local cell production images. First the images of control cells and IgG producing cells were analyzed to find the top 10000 SWNT and the cell regions were recorded using the ROI matlab function. These locations were then used to determine the coverage of the SWNT in the cell regions (specific) and outside the cell regions (un-specific) using the CellAnalysis Function (below). The time response of the IgG producing cells was determined again using the top 10,000 SWNT pixels and plotting their absolute intensity distributions over time for the series of images collected using the CellTime function (below). Finally, the AnalyzeCellPic function was used to plot the average intensity of cell islands as a contour plot as presented in FIG. 7. function CellAnalysis
% Coded by Nigel F. Reuel on 9.19.2012
% This program does the basic comparison analysis of our control and IgG % producing cells
NFiles = (1:21);
PSpec = zeros (1 : 21, 1) ;
PNSpec = zeros (1 : 21, 1) ;
for i = NFiles
CellLoc = csvread( [ 'Κ' , int2str (i) , ' _CellLoc . csv ' ] ) ;
SWNTLoc = csvread ( [ ' SLoc_Rankl0000_' , int2str (i) , ' . csv ' ] ) ;
CellCount = sum (sum (CellLoc) ) ;
OutCellCount = 320*256 - CellCount;
SWNTinCell = 0;
SWNToutCell = 0;
for j = 1:320
for k = 1:256
if SWNTLoc (j,k) > 0 && CellLoc (j , k) > 0;
SWNTinCell = SWNTinCell + 1; elseif SWNTLoc ( j , k) > 0 && CellLoc ( j , k) == 0;
SWNToutCell = SWNToutCell + 1 ;
end
end
end
PSpec(i,l) = SWNTinCell/CellCount*100; % Percent Coverage
PNSpec (i,l) = SWNToutCell/OutCellCount*100; % Percent Coverage
end
X = (1:21) ';
plot (X, PSpec, X, PNSpec)
Data = zeros (4,1); % 1 = IgG Cell specific, 2 = IgG Cell non specific, 3 = Control Cell Specific, 4 = Control Cell non specific
Error = zeros (4,1);
Data(l,l) = mean (PSpec ( [13: 16 18:21], 1));
Data(2,l) = mean (PNSpec ([ 13 : 16 18:21], 1));
Data(3,l) = mean (PSpec (1 : 7, 1) ) ;
Data(4,l) = mean (PNSpec (1 : 7, 1) ) ;
Error(l,l) = std (PSpec ( [ 13 : 16 18:21], 1));
Error(2,l) = std (PNSpec ([ 13 : 16 18:21], 1));
Error(3,l) = std ( PSpec ( 1 : 7 , 1 ) ) ;
Error(4,l) = std ( PNSpec ( 1 : 7 , 1 )) ;
bar (Data)
hold on
errorbar (Data, Error , 'or')
hold off
ylabel('% Coverage')
end
function CellTime
% Coded by Nigel Reuel on 9.19.2012
% This code analyzes the time traces of the cells
PL_0 = 0;
PL_1 = 0;
PL_2 = 0;
PL_3 = 0;
countl = 1;
count2 = 1;
count3 = 1;
count4 = 1;
for i = 1:19
PLData = imread ( [ ' HS ' , int2str (i) , ' . tif ' ] ) ;
SWNTLoc = csvread ( [ ' SLoc_Rankl0000_' , int2str (i) , ' . csv '
for j = 1:320
for k = 1:256
if SWNTLoc (j,k) > 0
if i < 3 % Time Zero
PL_0 (countl, 1) = PLData (j,k);
countl = countl + 1;
elseif i <10 % Time 1-1.5 hr
PL_1 (count2, 1) = PLData (j,k);
count2 = count2 + 1;
elseif i <18 % Time 2-2.5 hr
PL_2 (count3, 1) = PLData (j,k); count3 = count3 + 1 ;
else % Time 3-3.5 hr
PL_3 (count4, 1) = PLData ( j , k) ;
count4 = count4 + 1 ;
end
end
end
end
end
% Construct histograms for these time points
X = linspace (5000, 30000, 1000) ;
subplot (1,4,1)
n = hist (PL_0, X) ;
bar(X,n)
xlim( [7500 15000] )
subplot (1,4,2)
n = hist (PL_1, X) ;
bar(X,n)
xlim( [7500 15000] )
subplot (1,4,3)
n = hist (PL_2, X) ;
bar(X,n)
xlim( [7500 15000] )
subplot (1,4,4)
n = hist (PL_3, X) ;
bar(X,n)
xlim( [7500 15000] ) function AnalyzeCellPic
% Coded by Nigel F. Reuel on 9.12.2012
% This code will analyze a single cell picture (*.tif)
% For Figure 6 of biomanufacturing paper
%}
% Plot the Data together!!
NC = 10;
% Data file number
FN = 3;
%
CData = zeros (320, 256) ;
PLData = imread ( [ ' S ' , int2str (FN) , ' . tif ' ] , 1) ;
Cell = imread( [ 'C , int2str (FN) , ' .tif ' ] , 1) ;
for i = 1:NC
Ccount = 0;
Stotal = 0;
Cloc = csvread( [ 'C , int2str (FN) , ' Cell Island1 , int2str (i) , 'Loc.csv' ] ) ; for j = 1:320
for k = 1:256
if Cloc(j,k) > 0 Ccount = Ccount + 1 ;
Stotal = Stotal + PLData(j,k)
end
end
end
AvgSig = Stotal/Ccount;
for j = 1:320
for k = 1:256
if Cloc(j,k) > 0
CData(j,k) = AvgSig;
end
end
end
end
contourf (CData)
set (gca, 'Clim' , [0 8000] )
colormap (hot)
m = 10;
imagesc (Cell )
colormap (gray)
m = 10;
imagesc (PLData)
m = 10;
end
For the purposes of interpreting this specification, the following definitions apply and whenever appropriate, terms used in the singular will also include the plural and vice versa. In the event that any definition set forth below conflicts with the usage of that word in any other document, including any document incorporated herein by reference, the definition set forth herein and below shall always control for purposes of interpreting this specification and its associated claims unless a contrary meaning is clearly intended. It is noted that, as used in this specification and the appended claims, the singular forms "a", "an", and "the" include plural referents unless expressly and unequivocally limited to one referent. The use of "or" means "and/or" unless stated otherwise. The use of "comprise", "comprises", "comprising", "include", "includes", and "including" are interchangeable and not intended to be limiting. Furthermore, where the description of one or more embodiments uses the term "comprising" those skilled in the art would understand that, in some instances, the embodiment or embodiments can be alternatively described using the language "consisting essentially of and/or "consisting of." The section headings used herein are for organizational purposes only and are not to be construed as limiting the subject matter described. All documents, or portions of documents, cited in this application, including but not limited to patents, patent applications, articles, books, and treatises are hereby expressly incorporated by reference in their entirety for any purpose. In the event that one or more of the incorporated documents defines a term that contradicts that term's definition in this application, this application controls. While the present invention has been described in terms of the preferred embodiments, it is understood that variations and modifications will occur to those skilled in the art. Therefore, it is intended that the appended claims cover all such equivalent variations that come within the scope of the invention as claimed. Other embodiments are within the scope of the following claims.

Claims

CLAIMS What is claimed is:
1. A method of detecting an analyte, comprising:
contacting a sample including a biologic analyte with a sensor array, the sensor array
including a plurality of photoluminescent nanostructures;
collecting emissive data from each of the photoluminescent nanostructures; and
analyzing the emissive data from each nanostructure of the plurality of photoluminescent nanostructures to measure a distribution of binding affinities.
2. A method of detecting an analyte, comprising:
contacting a sample including a biologic analyte with a sensor array, the sensor array
including a plurality of photoluminescent nanostructures;
collecting emissive data from each of the photoluminescent nanostructures; and
analyzing the emissive data from each nanostructure of the plurality of photoluminescent nanostructures to measure a distribution of binding affinities;
wherein the photoluminescent nanostructures are:
physically associated with a linker;
directly or indirectly chelated with a metal ion; and / or
functionalized with a capture protein.
3. The method of claim 1 or 2 wherein the photoluminescent nanostructures are selected from the group consisting of carbon nanotubes, single-walled carbon nanotubes, double-walled carbon nanotubes, semi-conductor quantum dots, semi-conductor nanowires, and graphene.
4. The method of any one of claims 1-3 wherein the sensor array is arranged upon one or more support materials.
5. The method of claim 4 wherein the one or more support materials comprises a gel.
6. The method of claim 4 or 5 wherein the one or more support materials comprises a first gel lacking photoluminescent nanostructures and a second gel comprising photoluminescent nanostructures upon the first gel.
7. The method of any one of claims 4-6 wherein one or more support materials comprises a solid material (e.g. glass) supporting the photoluminescent nanostructures.
8. The method of any one of claims 4-7 wherein the one or more support materials comprises a solid material (e.g. glass), a first gel lacking photoluminescent nanostructures, and a second gel comprising photoluminescent nanostructures upon the first gel.
9. The method of any one of claims 1-8 wherein the analyte is an antibody.
10. The method of claim 2 wherein the capture protein is Protein A or a derivative thereof.
11. The method of claim 2 wherein the capture protein is PSA-lectin or a derivative thereof.
12. The method of claim 11 wherein the analyte is mannose.
13. The method of any one of claims 1-12 wherein the sample comprises cells that colonize the photoluminescent nanostructures.
14. The method of claim 13 wherein one or more analytes produced by the cells is detected.
15. The method of any one of claims 1-14, further comprising calculating a calibration curve for the sensor.
16. The method of any one of claims 1-15, wherein analyzing the emissive data includes
measuring a variance of the emissive data from the plurality of photoluminescent
nanostructures.
17. The method of any one of claims 1-16, wherein analyzing the emissive data includes
measuring a skew of the emissive data from the plurality of photoluminescent nanostructures.
18. The method of any one of claims 1-17, wherein analyzing the emissive data includes
measuring a standard deviation of a plume angle of the emissive data from the plurality of photoluminescent nanostructures.
19. The method of any one of claims 1-18, further comprising evaluating a binding heterogeneity of the biologic analyte of the sample from the analyzed emissive data.
20. The method of any one of claims 1-19, further comprising evaluating a cell line for
production of the biologic analyte of the sample from the analyzed emissive data.
21. The method of any one of claims 1-20, further comprising evaluating a glycosylation pattern of the biologic analyte of the sample from the analyzed emissive data.
22. The method of any one of claims 1-21, wherein the array includes at least 500 of the
photoluminescent nanostructures.
23. The method any one of claims 1-22, wherein the array includes at least 1,000 of the
photoluminescent nanostructures.
24. The method of any one of claims 1-23, wherein the array includes at least 2,000 of the photo luminescent nanostructures.
25. The method of any one of claims 1-24, wherein the array includes at least 5,000 of the photo luminescent nanostructures.
26. The method of any one of claims 1-25, wherein the sensor includes an analyte-binding compound associated with the photoluminescent nanostructure.
27. The method of any one of claims 1-26, wherein the sensor further comprises a linker, wherein the analyte-binding compound is associated with the photoluminescent nanostructure via the linker.
28. The method of any one of claims 1-27, wherein the sensor includes a polymer including clusters of photoluminescent nanostructures.
29. The method of any one of claims 1-28, wherein the analyte is a biologic analyte.
30. The method of any one of claims 1-29, wherein the biologic analyte is an antibody.
31. The method of any one of claims 1-29, wherein the biologic analyte is a protein.
32. The method of claim 30, wherein the antibody is IgG.
33. The method of any one of claims 1-29, wherein the analyte binding compound is a capture protein comprising Protein A or PSA-lectin.
34. The method of any one of claims 1-33 wherein the capture protein comprises a tag having specificity for a binding partner associated with the photoluminescent nanostructure.
35. The method of claim 34 wherein the tag is a histidine tag and the binding partner is nickel.
PCT/US2014/017413 2013-02-21 2014-02-20 Method of analysis using array sensor Ceased WO2014130682A2 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201361767513P 2013-02-21 2013-02-21
US61/767,513 2013-02-21

Publications (2)

Publication Number Publication Date
WO2014130682A2 true WO2014130682A2 (en) 2014-08-28
WO2014130682A3 WO2014130682A3 (en) 2016-07-21

Family

ID=51391969

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2014/017413 Ceased WO2014130682A2 (en) 2013-02-21 2014-02-20 Method of analysis using array sensor

Country Status (1)

Country Link
WO (1) WO2014130682A2 (en)

Also Published As

Publication number Publication date
WO2014130682A3 (en) 2016-07-21

Similar Documents

Publication Publication Date Title
Ali et al. Molecular imprinted polymer combined with aptamer (MIP-aptamer) as a hybrid dual recognition element for bio (chemical) sensing applications. Review
Li et al. Simple design concept for dual-channel detection of ochratoxin a based on bifunctional metal–organic framework
Jin et al. Semiconductor quantum dots for in vitro diagnostics and cellular imaging
Zhang et al. Amperometric aptasensor for amyloid-β oligomer detection by optimized stem-loop structures with an adjustable detection range
Sapsford et al. Kinetics of metal-affinity driven self-assembly between proteins or peptides and CdSe− ZnS quantum dots
Li et al. Chemical sensors and biosensors for the detection of melamine
US10890582B2 (en) Sensor for detecting analytes
Knezevic et al. Bimodal electrochemiluminescence microscopy of single cells
KR20140143140A (en) Methods and devices for detection and measurement of analytes
CA2643993A1 (en) Methods and arrays for target analyte detection and determination of target analyte concentration in solution
CN106052872B (en) A kind of terramycin SERS detection methods based on nano material self assembly
Liu et al. Nonenzymatic target-driven DNA nanomachine for monitoring malathion contamination in living cells and bioaccumulation in foods
Huang et al. A signal amplification strategy based on peptide self-assembly for the identification of amyloid-β oligomer
Lin et al. Multi-analyte discriminated ECL via photonic crystal-enhanced selectivity
CN107936035A (en) A kind of cysteine-modifying graphene quantum dot GQCY and preparation method are with preparing the application on dopamine luciferase assay reagent
Perlaki et al. Raman spectroscopy based techniques in tissue engineering—An overview
Gagni et al. Supramolecular hydrogels for 3D biosensors and bioassays
Fan et al. Photocurrent switchable dual-target bioassay: Signal distinction and interface reconfiguration via pH-responsive triplex DNA programming
Xiang et al. Light-regulated natural fluorescence of the PCC 6803@ ZIF-8 composite as an encoded microsphere for the detection of multiple biomarkers
CN104391018B (en) Three-dimensional DNA nano-structure, electrochemical biosensor as well as preparation methods and application thereof
Zhang et al. A novel dual recognition sensor based on molecularly imprinted polymers and photochemical aptamer for non-autofluorescence determination of dopamine
Mansouri Recent Advancements in Molecularly Imprinted Polymers Based Aptasensors: Critical Role of Nanomaterials for the Efficient Food Safety Analysis
KR101892668B1 (en) A method for detection of a target analyte using gold nano-probe by copper crystal overgrowth
Zhang et al. Fluorescent metal nanoshells: lifetime-tunable molecular probes in fluorescent cell imaging
WO2014130682A2 (en) Method of analysis using array sensor

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14753819

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase in:

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14753819

Country of ref document: EP

Kind code of ref document: A2