A Miniaturized Glycan Microarray Assay for Assessing Avidity and Specificity of Influenza A Virus Hemagglutinins

Summary

Using a printed glycan microarray strategy, a conventional 96-well plate assay was miniaturized for analysis of influenza A virus hemagglutinin avidity and specificity for sialic acid containing receptors.

Abstract

Influenza A virus (IAV) hemagglutinins recognize sialic acids on the cell surface as functional receptors to gain entry into cells. Wild waterfowl are the natural reservoir for IAV, but IAV can cross the species barrier to poultry, swine, horses and humans. Avian viruses recognize sialic acid attached to a penultimate galactose by a α2-3 linkage (avian-type receptors) whereas human viruses preferentially recognize sialic acid with a α2-6 linkage (human-type receptors). To monitor if avian viruses are adapting to human type receptors, several methods can be used. Glycan microarrays with diverse libraries of synthetic sialosides are increasingly used to evaluate receptor specificity. However, this technique is not used for measuring avidities. Measurement of avidity is typically achieved by evaluating the binding of serially diluted hemagglutinin or virus to glycans adsorbed to conventional polypropylene 96-well plates. In this assay, glycans with α2-3 or α2-6 sialic acids are coupled to biotin and adsorbed to streptavidin plates, or are coupled to polyacrylamide (PAA) which directly adsorb to the plastic. We have significantly miniaturized this assay by directly printing PAA-linked sialosides and their non PAA-linked counterparts on micro-well glass slides. This set-up, with 48 arrays on a single slide, enables simultaneous assays of 6 glycan binding proteins at 8 dilutions, interrogating 6 different glycans, including two non-sialylated controls. This is equivalent to 18x 96-well plates in the traditional plate assay. The glycan array format decreases consumption of compounds and biologicals and thus greatly enhances efficiency.

Introduction

Wild waterfowl are the natural reservoir for IAV, but IAV is able to cross the species barrier to poultry and mammals, including, humans. Avian IAVs recognize α2-3 linked sialic acids (avian-type receptors), whereas human viruses bind α2-6 linked sialic acids (human-type receptors). To be able to efficiently replicate and transmit between humans an avian IAV needs to bind to human-type receptors ¹.

IAVs are divided based on serology that characterizes the antigenicity of their hemagglutinin (HA) and neuraminidase (NA) envelope glycoproteins. HA binds to sialic acids, whereas NA is the receptor-destroying enzyme at the other end of the viral lifecycle and cleaves sialic acid ². All human infecting viruses, including H1N1, H2N2 and H3N2, have an avian origin ³. Over the last two of decades several avian to human crossovers have occurred, with H5N1, H7N7, and H7N9 being the most well-known; however, other subtypes have infected humans more sporadically (H6N1, H7N1, H7N2, H9N2, H10N7, H10N8) ⁴. Fortunately, it seems that none of these viruses have been able to fully adapt to human-type α2-6-linked sialic acid receptors ^5-8. Adaptation of avian or other zoonotic viruses to accommodate replication and transmission in human hosts could have a devastating effect on human health. Therefore, prior knowledge of how these viruses are evolving to bind human-type receptors will help world-wide surveillance of emerging influenza viruses.

Determination of receptor preference was first elucidated using erythrocytes of different species and remains a favored assay among flu researchers ^9-12. The demonstration that avian viruses recognize α2-3 sialic acid and human viruses α2-6 linked sialic acid was originally based on an assay using hemagglutination of erythrocytes enzymatically engineered to contain each of the linkages ^13,14. Although the readout is hemagglutination, a standard assay for virologists, the underlying glycan structures are not defined, only the terminal linkage. Additionally, the limited availability of the sialyltransferases, used to re-sialylate the cells, have limited the use of this assay ^15-18. Subsequently, other methods of determining receptor-binding preferences were introduced using sialylated glycan structures linked to poly-acrylamide (PAA) or poly-glutamate (PGA) structures in plate-based assays ^19,20. Several variations are possible in coating either the glycans or viruses to microtiter plates, each of which results in a robust, reliable and very sensitive ELISA-type assay ^21-23. Alternatively, biotin-linked glycans can replace PAA/PGA and can be conjugated to streptavidin-coated plates ^2,24. Although some specific sera may be required, ELISAs are standard and several glycans linked to PAA are readily commercially and non-commercially available (Consortium for Functional Glycomics (http://www.functionalglycomics.org)).

Glycan microarray technologies have emerged as an invaluable tool to determine receptor specificity, as multiple different glycans are spotted, and binding to a wide array of different structures can be assessed within a single assay ^25-29. The binding of IAV to these structures provides a better understanding of the glycan structures that IAV preferentially recognizes ^30-33. Glycan microarrays require small amounts of sample volume to perform a binding assay and only uses minute amounts of glycan per spot (2 nl). However, these arrays are typically used only to evaluate specificity of glycans receptors. Analysis of multiple viruses, or hemagglutinin proteins, at multiple concentration ranges can be prohibitive due to the number of slides required. Furthermore, to date, no relative avidity assay has been developed using glycan array techniques.

To combine the low sample requirement afforded by glycan microarray techniques and the sensitivity of PAA-linked glycans in ELISA-based assays, we sought to develop a multi-well glycan array that would allow for high-throughput analysis with similar or better resolution as compared to the traditional ELISA-based assay. Simultaneously, we wanted to minimize the amount of biologicals and reporter chemicals consumed. The ultimate result is a miniaturized avidity assay, specifically developed for monitoring IAV specificity and is equally applicable to assess other glycan binding proteins. Using a glass slide separated into 48 micro-wells by a Teflon mask, 6 different glycans are spotted in 6 replicates per well. The microarray platform affords the same trends in receptor binding seen in the macro ELISA format with several advantages. These include (I) printing of compound in 6 replicates, using minimal sample, versus the coating of several rows in a plate, using 100 µl per well; (II) multiple different compounds analyzed simultaneously in a single well, including controls; (III) a massive decrease in incubation volume and; (IV) a larger dynamic range using a fluorescent readout. A single slide can be calculated to be equivalent to 18x 96-well plates.

With the following protocol, any lab capable of fabricating and analyzing spotted microarrays should be able to manufacture this miniaturized ELISA format.

Protocol

NOTE: All steps are carried out at room temperature unless otherwise indicated. 

1. Array Construction

1. Glycan Preparation and Plate Setup
   1. Prepare a stock of printing buffer. First make 500 ml of a 150 mM stock solution of dibasic sodium phosphate. Also make 50 ml of a 150 mM stock solution of monobasic sodium phosphate solution. In a flask, using a magnetic stir bar and pH meter, slowly titrate the dibasic sodium phosphate solution with the monobasic solution until a pH of 8.5 is reached. Add Tween-20 to 0.005%.
   2. Prepare the glycan samples. Dilute PAA conjugated glycans, (diLacNAc (Galβ1-4GlcNAcβ1-3Galβ1-4GlcNAcβ-PAA), 3SLNLN (Neu5Acα2-3Galβ1-4GlcNAcβ1-3Galβ1-4GlcNAcβ-PAA) and 6SLNLN (Neu5Acα2-6Galβ1-4GlcNAcβ1-3Galβ1-4GlcNAcβ-PAA (Figure 1A)) are to be diluted to 100µg/ml in printing buffer from step1.1.1. Monovalent glycans, diLacNAc (Galβ1-4GlcNAcβ1-3Galβ1-4GlcNAcβ-(CH₂)₃NH₂, 3SLNLN (Neu5Acα2-3Galβ1-4GlcNAcβ1-3Galβ1-4GlcNAcβ--(CH₂)₃NH₂) and 6SLNLN (Neu5Acα2-6Galβ1-4GlcNAcβ1-3Galβ1-4GlcNAcβ--(CH₂)₃NH₂)) to 100µM in printing buffer from step 1.1.1. Finally, prepare the amine-functionalized dyes, to use as grid markers/landing lights. Dye marker spots are printed at 1 µM.
      NOTE: Atto488-NHS dye is used in our prints. However, any amine-functionalized dye, readable by the slide scanner is appropriate for this purpose.
   3. Transfer 10 µl of each glycan sample to 384-well microtiter plates, used for the printing robot. Store unused glycans in printing buffer solution at -20 °C in sealed tubes. Tansfer 10 µl of the dye marker to the printing plate.  
2. Printing
   NOTE: All steps that concern touching or moving the slides must be done with gloves.
   1. Place arrayer pins in the printing head according to the correct configuration of the layout. For this layout use one pin to print all samples and arrays. Print the glycans in blocks of six, skipping one feature (a feature is defined as single spot of a certain compound), on each side, such that spots of the same sample are not printed exactly beside each other.
      NOTE: The end user must decide the configuration.  
   2. Remove slides from bags, open boxes, and remove any dust particles or small pieces of plastic that might be on their surfaces by blowing ultra-high purity Nitrogen gas on each slide. Place the desired number of slides, coated side up, on the arrayer stage.
   3. Program the arrayer, use the parameters of 27 spots per visit to the source plate, using 3 "pre-spots" onto poly-L-lysine coated slides, to remove excess liquid at the tip of the printing pins
      1. Turn on MPU (Main Power Unit) and Climate Control Unit. Open Tiling Analysis software on computer.
      2. Select layout to the print. Select Options → Pin Tool to be used for printing the array 1 x 1 print tool - 1 single pin. Select the Target tab → Tool Array Definition → Print Pattern and add the pitch (spot-to-spot distance), number of rows and columns to accommodate the samples to be printed, and the printing pattern of the samples (for this array an 8 x 8 printing pattern is used at 340 µm pitch for 7 total samples).
      3. Select Target → Slide Layout and program the block-to-block distance to allow printing of each replicate array into the 48-well (4 x 12) slide template. Select Source and add the number of source pickups (1 for each sample when using 1 pin) (for this print 7 source pickups will be used to accommodate 6 sample wells and 1 dye marker well).
         NOTE: The source dialogue should turn GREEN to show that the number of source visits matches the samples to be printed in the print pattern.
      4. Select Target → Adapter Plate and Slide Layout and adjust the slide number (5 slides are printed at a time for this print with 1 pre-spotting slide).
      5. Release the tray by clicking the T1 button to activate the tray and add the desired number of slides to the tray, paying close attention to the locations of the pre-spotting slide (BLUE) and "real" slides.
      6. Place plain glass slides on all remaining vacuum ports to block the vacuum and click OK to activate the vacuum. Check that all slides are held in place and click OK to return the tray to the home position.
      7. Load the printing plate into the plate holder and ensure the rack is set correctly in place. Click GO and allow the printing process to proceed.
   4. Print the remaining 24 spots in replicates of 6 spots per array (4 arrays/source visit).
   5. Load the 384-well plates into the printer and begin printing. Maintain relative humidity, while printing, throughout the print between 55 and 65%. The printer stops after printing 5 slides.  
      NOTE: While the microarray in this assay is printed in a specific 8 x 8 pattern, different microarray printing robots will require equipment specific programming to achieve the desired layout. It is up to the end user to determine the most appropriate pattern given their equipments’ specifications.
3. Humidification/Immobilizing
   NOTE: Once the slides have finished printing, they undergo a temporary immobilization step, also called humidifying. Humidifying post-print helps to homogenize the sample attachment by re-wetting the spots and ensuring complete blocking.
   1. Put slides in a 100% humidity chamber immediately after printing for a period of 30 min. Construct the chamber by placing very wet paper towels flat in the bottom of a large glass dish, placing the slides on some sort of rack, such as a test tube rack, print side up, and sealing with plastic wrap to trap in the moisture.
   2. Number slides with an alcohol/solvent resistant marking pen and store in a desiccation box. Desiccate overnight.
4. Numbering, Blocking, and Storage
   1. Prepare the blocking buffer – 50 mM ethanolamine in 50 mM borate buffer. First, dissolve boric acid, in ddH₂O, to a final concentration of 50 mM in a flask containing a magnetic stir bar. Stir the solution vigorously on a stirring plate until all solid has dissolved. While constantly stirring, add the appropriate amount of ethanolamine from a 16.54 M stock solution to reach the desired 50 mM concentration. Add concentrated sodium hydroxide to adjust to a final pH of 9.2.
   2. Immerse the printed slides in the blocking buffer solution for 1 hr.
   3. Following 1 hr, rinse the slides in ddH₂O to remove any excess traces of blocking buffer. Dispose blocking buffer in an appropriate waste container.
   4. Transfer the slides to glass staining holders and spin dry. Dry arrays by placing them in a centrifuge equipped with swinging plate holders, at a speed of 10 x g for 5 min.
   5. Once the slides are dry, they can be incubated immediately or stored. Storage conditions are -20 °C in a sealed plastic bag.  

2. Analyzing Glycan Binding Proteins

1. Prepare a humidified chamber in which the arrays to be hybridized will fit. A simple chamber consists of a Pyrex dish, layered on the bottom with dampened paper towels and a rack, upon which the slides will rest.
2. Use biotinylated lectins SNA (Sambuca nigra agglutinin) and ECA (Erythrina cristagalli agglutinin) at 10 μg/ml + 2 µg/ml of Streptavidin-555 dye, in probing buffer (PBS + 0.01% Tween-20). For HA (A/Vietnam/1203/04 H5N1 and A/KY/07 H1N1, in this example), use a starting concentration of 20 μg/ml pre-complexed, as previously described ³⁴. Briefly, make a HA:Mouse-α-Strep:Goat-α-Mouse-Alexa647 mixture in blocking buffer (PBS + 3%BSA + 0.01% Tween-20), at a 4:2:1 molar ratio.
   NOTE: The array is probed with lectins to ascertain that glycans are immobilized and detectable (Figure 1B).
3. Transfer the glycan binding protein-antibody mix solution of 20 µl to a 384-well plate and incubate on ice for 30 min. Serially dilute lectin and HA samples 1:1, in their appropriate buffer, 8 times, by mixing 10 µl of sample with 10 µl of buffer.
4. Pipette 8 µl of the sample onto the micro-well surface and incubate in the sealed humidification (100% RH) chamber for 90 min.
5. Wash the slides one at a time by holding the edges and dipping them four times in a mixture of PBS and 0.05% Tween, then four times in PBS, and finally four times in deionized H₂O. Using the same drying protocol described in the blocking step, spin the arrays dry in a centrifuge for 5 min at 10 x g.

3. Scanning and Data Analyses

NOTE: The glycan microarrays can be scanned at different settings, depending on the microarray scanner’s manufacturer. By varying the laser power and resolution, the instrument can produce an image with as many signals as possible within its dynamic range (Figure 1C).

1. Use a fluorescent slide scanner and scan the arrays immediately after incubation with the glycan binding proteins. Take care to ensure the slide is correctly inserted into the scanning instrument.
   1. Open scanning software. Click start to open the program. Connect to scanner: Scanner → Connect or green button in the Scanner navigation panel menu.
   2. Open autoloader door of the scanner. Place slides in slots in autoloader in specific order. Close autoloader door. Click "read loader" in the Tools navigation panel to auto detect slides in the scanner.
   3. Set scan parameter: Scanner → Scan parameters appropriate to the experiment (e.g. 635 laser power high, detection gain 50%). Scan slides: Scanner → scan. Select path by creating a folder to save the scan images and name individual scans. Do this prior to the actual scanning. Repeat for each individual slide. Click OK to start scanning.
2. Use the scanner software to set up the instrument and scan the slides. Set scanner to iteratively scan at low laser power (5 mW) with increasing gain settings of 25%, 50% and 75%.  
   NOTE: The settings can be tested and optimized, but keep in mind that each scan can possibly lower the fluorescence output of the dyes due to photobleaching.
   1. Open data acquisition and analysis software such as Mapix software. Click start to open the program. Open scan image: File → open image. Open GAL file: Analyze → open grid and select the appropriate GAL file.
   2. Enter blocks mode: Image → blocks mode to move the grid. Use the grid markers to set the grid to the appropriate location on the slide. Adjust individual blocks as necessary to match the printed array.
   3. Enter spots mode: Image → spots mode to adjust location and size of individual spots within the block. Once all blocks and spots are adjusted and in place, measure signal intensities by Analyze → photometric calculations. Save the output file.
3. Once the slides have finished scanning, images are analyzed for binding signals with the image-processing program installed (Figure 1D-G). For image analysis, load a GAL (Array List) file over the scanned TIFF image.
   NOTE: The GAL file contains both the spot layout pattern and identities of each spot location. GAL files for arrays are created by the printer software using a tab-delimited text file that contains the identities of the samples and their location in the 384-well source plate.
4. Use the grid marker/landing lights printed on the microarrays to properly place the GAL grid. Individual spots may need to be moved individually, but a well-printed array will generally allow blocks within the grid to be placed with ease. Following grid placement, collect spot intensities by the software and saved as a tab-delimited text file (.txt).  
   1. Using spreadsheet program, open the output file from the scanner. Open the appropriate macro file from the folder location. Click View → Macro → RunMacro.  
      NOTE: The pre-written will copy the raw output data, remove unnecessary rows and columns, sort the data by sample, calculate the average mean signal minus background values and plot the resultant values in to an spread-sheet worksheet containing the graphs for each of the six samples tested on the array.

Results

Printing, Scanning & Data analyses

To ensure proper printing, it is vital to have the correct alignment of the spotted grid within the teflon mask, which delineates each array on the slide. During printing, due to the nature of the teflon coating, spots cannot be seen by the naked eye on the MPX slides. Attention is paid then to the appearance of the pre-spots on the poly-L-Lysine coated slides. Directly after printing, each...

Discussion

Assessing IAV receptor specificity is an important step in analyzing pandemic potential of avian viruses. Sialic acid recognition by the virus is linked to several biological properties such as binding to and release from the cell. Knowledge of which amino-acid mutations are necessary for avian viruses to achieve α2-6 binding and cross the species barrier enables pandemic preparedness. Several assays are used to determine receptor specificity; however, all have their drawbacks, including only measuring avidity and n...

Materials

Name	Company	Catalog Number	Comments
NEXTERION® Slide H MPX-48	Schott	1091525	Microwell slides
ProScanArray Plus	PerkinElmer	discontinued	confocal microarray scanner
Innoscan 1100AL Scanner/Mapix Software	Innopsys	-	confocal microarray scanner
MicroGrid II	Digilab	-	microaray printer
SNA	Vector Labs	B-1305	Plant Lectin
ECA	Vector Labs	B-1145	Plant Lectin
Anti-Strep Antibody	IBA	2-1509-001	Anitbody for HA binding
Anti-Mouse Alexa-647	Life	A-21235	Anitbody for HA binding
Tween-20	Sigma	P2287	detergent
di-basic Sodium Phosphate	Sigma	255793	printing buffer component
mono-basic Sodium phosphate	Sigma	229903	printing buffer component
poly-l-lysine solution	Sigma	P8920	pre-spotting slide component
sodium hydroxide	Sigma	221465	pre-spotting slide component
ethanol	Sigma	493546	pre-spotting slide component
phosphate buffered saline	Corning	46-013-CM	incubation/washing buffer
SMP4B pins	Telechem	SMP4B	printing pin
Compressed Nitrogen (Grade5)	Praxair	UN1066	general dusting/drying tool
Boric Acid	Sigma	B6768-500G	Slide blocking reagent
ethanolamine	Sigma	E9508-500ML	Slide blocking reagent
Atto 488	AttoTec	AD 488-91	Gridmarker on array
PAA-LNLN	Consortium for Functional Glycomics	PA368	Spotted glycans
PAA-3SLNLN	Consortium for Functional Glycomics	PA362	Spotted glycans
PAA-6SLNLN	Consortium for Functional Glycomics	PA343	Spotted glycans
LNLN	Consortium for Functional Glycomics	Te98	Spotted glycans
3SLNLN	Consortium for Functional Glycomics	Te175	Spotted glycans
6SLNLN	Consortium for Functional Glycomics	Te176	Spotted glycans
384-well microtiter plate	Matrix TechCorp	4361	Printing plate
VWR lab marker	VWR	52877-150	Slide Numbering
Wheaton slide staining dish	Sigma	Z103969-1EA	Blocking and Drying