Principles of microarray analysis
The technical solutions that have been developed for performing microarray analysis, all are miniaturized hybridization assays for studying thousands of nucleic acid fragments simultaneously. All microarray systems share the following key components:
- the array, which contains immobilized nucleic acid sequences, or ‘targets’
- one or more labeled samples or ‘probes’, that are hybridized with the microarray
- a detection system that quantitates the hybridization signal
Manufacturing of microarray slides
Introduction
Microarray analysis is performed on a glass slide. This enables the performance of hybridization assays with fluorescently labeled samples.
Microarray manufacture requires three distinct components:
- production method
- microarray slide
- target genetic content
Production methods
Oligo synthesis
Two parallel approaches have been developed for the production of microarray slides. Nucleic acid targets can either be synthesized directly onto the microarray slide, or purified targets can be deposited onto a solid surface that is capable of binding nucleic acids.
Oligo synthesis is done by attaching chemically modified linker groups onto the glass surface. These linker groups contain photochemically removable protective substances. By masking different predefined positions in different steps, we can synthesize different nucleotides at various regions. Target synthesis proceeds in a step-wise fashion. In each step, the unprotected areas are first activated with light to remove the light-sensitive protective groups.
This method produces arrays of small features that are anchored at their 3′ ends to the array surface. Each feature is made up of oligonucleotides that all have the same nucleotide sequence. These arrays have a high density: an area of 1.6 cm can contain up to 4,00,000 features.
Deposition
Using common deposition methods, purified nucleic acids are attached to a modified glass slide. Typically, small volumes of nucleic acid solution are transferred onto the glass slide. Deposition methods are equally suitable for preparing microarrays containing oligonucleotides, cDNA sequences, as well as genomic DNA.
The deposition involves the formation of covalent bonds between the oligo and the glass surface. It binds the oligonucleotide onto the array. The manufacturing process should meet several criteria. Variation in the quantity of targets deposited, the shape of spots, the regularity of the array pattern, and the carryover of targets could all detrimentally affect the accuracy of microarray data.
Non-contact deposition
The printing heads do not touch the surface of the microarray. Piezoelectric printing and syringe-solenoid methods are the two common variations of this method.
■ In piezoelectric printing, the target solution is drawn into a capillary that is in contact with a piezoelectric crystal. Application of the voltage to the crystal results in a slight conformational change, squeezing the capillary.
■ Syringe-solenoid deposition uses a syringe pump positive displacement method to deposit nano liter volumes onto a slide. A syringe that provides the pressure source is connected to a micro-solenoid valve.
Contact deposition
In contact deposition, solid, hollow, or split-open pen designs are used to transfer target nucleic acid onto the slide surface. These pens are dipped into the target solution, a small volume of which adheres to the pen. When the pen comes into contact with the slide surface, a fraction of the nucleic acid solution on the pen is deposited onto the glass surface.
Slide surface chemistries
Nucleic acids will not attach efficiently to an untreated glass slide. Different chemical treatments have been developed to facilitate the attachment. These treatments have an influence on the density of the molecules that can be attached.Uniformity and thickness of the surface coating are critical quality parameters. These influences spot uniformity, DNA binding, and noise to signal ratio. Commonly used slide surface modifications include the introduction of aldehyde, amino, or poly-lysine groups onto the slide surface.
Common slide types
Aldehyde slides
Amino-modified DNA can be attached to microarray slides that have been modified with aldehyde groups. The amino group can be introduced into DNA in a PCR amplification reaction using amino-modified oligonucleotides. The aliphatic amine on the amino-modified DNA acts as a nucleophile. It attacks the carbon atom aldehydes, which is covalently attached to the surface of the slide. An unstable intermediate is converted to a Schiff base through a dehydration reaction (-H2O), and the DNA is bound to the surface. The unreacted aldehyde groups are reduced to non-reactive primary alcohols by treatment with sodium borohydride (NaBH4).
Amine slides
Amine groups can be introduced onto microarray slides by treating cleaned glass with aminosilane, such as 3-aminopropyltrimethoxysilane. Unmodified DNA can be attached to amine-modified slides, via interactions between negatively charged phosphate groups on the DNA and the positively charged slide surface. This interaction also denatures the DNA which increases its binding affinity. UV treatment is used to further immobilize the DNA onto the slide surface. Attachment via electrostatic interactions is suitable for binding DNA fragments that are longer than 60–70 nucleotides. For attaching oligonucleotides to amine-modified glass, chemical coupling methods must be used.
Poly-lysine slides
Treatment of the slide with poly-lysine creates a positively charged surface to which unmodified DNA can bind via ionic interactions.
Reflective slides
A large proportion of the fluorescent light emitted from the hybridized probe is scattered in all directions when using regular glass arrays. The introduction of a reflective surface below the spotting surface enables a significant amount of this scattered output to be directed towards the detector, hence increasing the amount of signal detected by the system. These reflective slides are constructed by adding a layer of aluminum above the glass surface.
Target nucleic acids
The third critical component in microarray manufacturing is the target nucleic acid. A high concentration of target molecules for deposition. The purity of target solutions is important for both the efficient attachment of nucleic acids to the slide surface and the availability of the immobilized targets for hybridization. The targets are available for hybridization only in single stranded denatured form. This can be achieved by spotting the targets under denaturing conditions. Typically, targets are dissolved in high salt solutions such as 3 × SSC, or in denaturing solvents such as DMSO.
Requirements of labeling methods
Retaining gene expression information
Any labeling methods used in the microarray must cope with the inherent diversity of the transcript sequences. It should contain all the information of the original transcript population. An ideal labeling system is neither biased towards any nucleotide sequences, nor does it label differently transcripts of different sizes or sequences that are expressed at different levels. In reality, existing labeling methods do not convert all information into a labeled form. Enzymatic methods are limited to copying certain nucleic acid sequences, whereas the instability of some transcripts is a general problem for all methods
Length of labeled fragments
Hybridization of nucleic acids and specificity for the complementary targets should be retained for the extraction of accurate information about gene expression. The length of the labeled fragment is an important factor in determining these parameters. For optimal hybridization, probes consisting of fragments of range 200-500 nucleotides sequence are recommended. Longer fragments may not find their targets as efficiently as shorter fragments.
Yield of labeled probe
The amount of labeled probe prepared by the labeling method is important to the sensitivity of microarray experiments. This is because the efficiency of the labeling process is critical in determining the lowest amount of mRNA that can be used to generate detectable signals from microarrays. The efficiency can make the difference between being able to hybridize one or several slides with one probe. Ideally, the labeling method should transform each transcript into a labeled fragment, without any bias towards more highly expressed sequences. If the labeling method results in net amplification of nucleic acid in the labeling process, the amplification process should be linear, i.e. the original ratios of expression levels within the sample should not be changed in the amplification process.
Optimum labeling density
If two or more fluorescent molecules are in close proximity to each other, a significant portion of the absorbed light energy can be spent on interactions between different molecules and dissipated as heat. This will result in the reduction of the amount of fluorescence and it is no longer directly proportional to the number of the fluorophore in the sample. This phenomenon is called “quenching”, and it is an inherent property of fluorescent molecules. Each fluorophore has slightly different quenching properties. In practical terms, this means that for each fluorophore there is an optimal labeling density, or distance between attached labels, which will produce maximum fluorescence from a labeled nucleic acid fragment. Exceeding this optimum labeling density results in a decreased fluorescent signal. Thus optimal optimization density is required in all labeling methods.
Measuring the amount of CyDye in the probe with spectrophotometry
Follow the steps below to measure the amount of incorporated CyDye:
Sample preparation
Dilute an aliquot of a purified probe with water. The required volume depends on the size of available measuring cells. It is recommended to use the smallest cells possible. They minimize cross-contamination from one sample to another. If 100 μl glass cells are used cleaning thoroughly by rinsing with sterile water between samples.
Measurement
- Measure the absorbance spectrum of the sample from 200 to 700 nm against a blank. The absorption spectrum of CyDye contains two peaks, of which the second should be of higher intensity. The first peak indicates the presence of intermolecular interactions between different dye molecules. If both peaks are of nearly similar intensity, this is a sign of quenching. This is usually associated with over labeling of the sample.
- In case of dual labeling, measuring absorbance at 550 nm for Cy3 and at 650 nm for Cy5. The observed absorbance values depend on how much the sample was diluted before measurement, how much RNA was used in the labeling, the labeling method used, and the efficiency of labeling. Typically, values around 0.050 would be expected from first-strand cDNA labeling reactions in which 1 μg of mRNA is used as a template.
- Recover the measured sample for future use. It may be necessary to concentrate the sample before using for microarray hybridization. This can be performed by drying down the sample in a vacuum concentrator or by letting sample that has been heated to 60 °C evaporate to dryness. It is important to protect the sample from light during all handling.
Calculation
The amounts of Cy3 and Cy5 incorporated into probes can be calculated from their respective extinction coefficients and using the following equations:
Extinction coefficients
150 000 M–1 cm–1 at 550 nm for Cy3 and
250 000 M–1 cm–1 at 650 nm for Cy5
Calculation equations
pmol Cy3 in purified sample = (A550 / 150 000) × dilution factor × (z μl) × (w cm) × 1012
where
A550 = absorbance at 550 nm
z μl = volume of the sample after purification
w cm = optical path in cuvette
Cy5 (p mol in sample) = (A650 / 250 000) × dilution factor × (z μl ) × (w cm) × 1012
where
A650 = absorbance at 650 nm
z μl = volume of the sample after purification
w cm = optical path in cuvette
Determination of the amount of labeled nucleic acid in the sample
UV spectrophotometry
Two methods can be used to determine the amount of labeled nucleic acid in the sample: spectrophotometry and radioactive spiking. Nucleic acids absorb at 260 nm, and absorption measurement at this wavelength can be used for quantitative purposes. Distortions may happen due to impurities. Therefore absorption spectra from 200 to 400 nm should be measured, and only if the peak at 260 nm is clearly distinguishable from absorption at near wavelengths, should estimation of nucleic acid amount be made
The following approximations can be used to calculate the amount of nucleic acid in the probe:
1 A260 unit of double-stranded DNA corresponds to 50 μg/ml
1 A260 unit of single-stranded DNA corresponds to 37 μg/ml
1 A260 unit of single-stranded RNA corresponds to 40 μg/ml
Spiking with radioactive nucleotide
For more accurate quantification of labeled nucleic acid labeling can be spiked with a small amount of a radioactive nucleotide. This enables the determination of the amount of nucleic acid synthesized as well as the amount of sample recovered from purification. This approach can be used with all labeling methods in which new nucleic acid is synthesized.
Calculation of labeling density
Labelling density can be defined as the amount of CyDye incorporated into a known amount of nucleic acid.
Labelling density = pmol CyDye in labeled sample
μg nucleic acid in labeled sample
1 μg of cDNA contains approximately 1 × 10-6 /330 = 3030 pmol of nucleotides.
Analyzing CyDye labeled probes with PAGE
In order to analyze CyDye labeled probes using PAGE, follow the procedures detailed below:
Preparation
- PAGE gels suitable for analysis of labeled samples are prepared from standard 6% or 8% (w/v) sequencing gel mix. Slab gel instruments can be used. These provide the added benefit of thicker and deeper wells that simplify sample loading.
- 0.1–1 μl of purified labeling reaction is enough for detection of CyDye labeled nucleic acid by fluorescence scanning. Add 2 μl of formamide and 8 μl of water to each sample. Do not use loading/denaturation buffers which contain dyes such as bromophenol blue or xylene cyanol, as these will interfere with the detection.
- Dilute 4 μl of fluorescent markers with 4 μl of water and 2 μl of formamide. This sizer contains a Cy5-labelled DNA marker ladder.
Denature
Denature all samples by boiling for 2 min at 95 °C. Snap cool on ice before loading on to the gel.
Electrophoresis
- Load samples to a gel that has been pre-electrophoresed for 15–30 min. Perform electrophoresis according to the instructions provided with your equipment. Use 1× TBE as the buffer.
- A marker with 1 μl bromophenol blue is used.It also aids in the runtime of the electrophoresis. Keep the bromophenol blue within the gel.
- Remove and clean the gel plate. Do not let the gel plates to go dry.
Imaging
Scan the gel on an Imager. Detect Cy3 by excitation with 532 nm laser and using emission filter 555 BP 20. Detect Cy5 by excitation with 633 nm laser and using emission filter 670 BP 30. Set PMT to 800 V, focal plane to +3 mm, and use normal sensitivity. The PMT values may need to be adjusted to account for different amounts of sample in the gel.
Hybridization
The process of hybridization is typically performed in order to identify and quantitate nucleic acids within a larger sample. It involves annealing a single-stranded nucleic acid to a target complementary strand. Southern blotting is one well-established hybridization method. The target nucleotide sequence is attached to a membrane and hybridized with a labeled probe. Fluorescence confirms the presence of target sequence and its intensity is proportional to a number of a nucleotide sequence.
There are several critical factors to performing a successful microarray hybridization.
■ Pre-hybridization
■ Hybridization conditions
■ Hybridization buffer
■ Stringency washes
Pre-hybridization
Thi steps consist of incubating the DNA microarray without a probe. Preincubation is done for the following reasons. Badly Adhered DNA sequence is washed away. Pre hybridization ensures the availability of the nucleotide sequence for hybridization. Blocking of non-specific hybridization is done by this step.
Hybridization Conditions
Coverslip Method
Hybridization is done under a coverslip. 30 μl, of the buffer, is used. Dying out of nucleotide around the edges is a limitation. But it can be avoided by maintaining humid experimental conditions. Humidity more than 95 % should be maintained. The hybridization time is 16-hours.
Hybridization buffer
A Hybridization buffer should contain the following components:
- A buffering component that acts to stabilize variations in pH.
- A detergent that acts to lower the surface tension and allow the buffer to flow easily under a coverslip
- A compound that act as rate enhancers, volume excluders, or to speed up the hybridization and lower the melting point(Tm).
Probe concentration
Probe concentration may vary depending on due to the samples used, the slide type and information requirement. For each slide type manufacturers would recommend an optimum concentration of probes.
Probe depletion and target saturation
To provide complete isolation and movement, complete coverage of the array by the buffer solution is mandatory. Once covered, diffusion plays an important role in the movement of these probes to the target nucleotides. A 20 bp oligo nucleotide diffuses to a distance of 3.6 mm during the 18 h period. A cDNA sequence hardly moves throughout the hybridization process. A high concentration of target sequence will lead to a competing inhibition. This phenomenon is known as target depletion. It limits the signal received from the microarray. For a high abundance gene, the amount of probe in solution starts to approach the amount of target present, which can lead to target saturation. Target saturation will be determined by factors such as the amount of target initially spotted on the slide and the amount retained on that slide after pre-treatment, as well as the percentage of that target available for hybridization and the efficiency of hybridization. Together, the sensitivity of detection and target saturation determine the dynamic range of the microarray experiment.
Pre-hybridization
- Store spotted slides in a desiccator until use.
- Read through the pretreatment and hybridization protocols thoroughly before use, as buffers often need preparing and preheating before use.
- Pretreat the number of slides required for the experiment.
- During the pre-treatment stage, prepare the probes. This will often involve drying down equivalent amounts of the two probes of interest together and reconstituting them in the manufacturer’s recommended buffer. Some protocols require heating the probe before use; the probe prepared by reverse transcription will be single-stranded and therefore should not require denaturing before use.
Hybridization
- Once the probe is prepared, lay the spotted slide, DNA side up, on a clean surface. Absorbent tissue that does not release any fibers is a good choice of surface. This is important, as most slides are glass, and dirt on the rear of the slide will affect the result on the front of the slide.
- Using a pipette, transfer the required amount of hybridization buffer/probe mixture onto the slide. Avoid touching the slide surface with the pipette tip. Try to deposit the mixture along with the short side of the slide, away from the spotted area.
- Take a clean coverslip and place it on the slide near the probe mixture, and allow surface tension to speed the buffer along the coverslip. Then gently lower the cover slip, avoiding trapping air bubbles underneath.
- If air bubbles become trapped beneath a coverslip, do not move the coverslip to try and remove them. Movement of the coverslip will result in damage to the targets themselves. Most small air bubbles will disperse once the slide is transferred to hybridization temperature. Larger bubbles can be ‘encouraged’ to move by gently pressing on the surface of the coverslip with a pipette tip.
- The probe mixture is light sensitive, so once the coverslip is on, place the slide in the humid chamber and incubate overnight in the dark.
Stringency washes
The purpose of the post-hybridization washes is to remove all unattached and loosely bound probe molecules. This prevents false positive signals and removes all components of the hybridization buffer, preventing background noise in the form of smearing and speckles. Once the slides have been washed, they should immediately be dried by centrifugation or nitrogen steam to prevent smearing while drying. The slides should then be stored in the dark in a desiccator and scanned as soon as possible. If once scanned, it is found that the slides have a high background or low stringency, it is worth rewashing the slide and re-scanning.
Image analysis
A scanned microarray image records the fluorescent intensities of both the signal and background DNA spots. The first step in data analysis is to identify these spots. Measurement the e intensities of both the spot and background fluorescence. This process is known as gridding. It can be done by many software. It gives an output the size of the spot, the distance between two spots and all other measurable quantities.
Controls
As with all experiments, micro arrays should contain a series of controls to ensure that the data obtained from the arrays is accurate.
Negative controls
Negative controls are spotted DNA sequences that should not hybridize with any labeled probe. The negative controls used should ideally come from organisms that are only distantly related to the organism being studied in the experiment. Under optimal analysis conditions, negative control spots should not give any signal at all.
Poly-adenylated DNA and CotI DNA
When using oligo(dT) to prime first-strand cDNA synthesis, it is possible that the oligo(dT) will prime within the poly-A tail of the mRNA. If this occurs there will be a string of dT bases within the probe. In order to prevent cross reactivity of the poly-dT sequences within the probe with potential poly-dA sequences in the targets, a poly-dA oligo of 80 bases can be included in the hybridization to block the poly-dT.
Positive controls
Labelled DNA
DNA can be labeled with CyDye fluors using polymerase chain reaction (PCR), or in the case of oligos, during the synthesis of the oligos (many oligo manufacturers offer this service). When the labeled DNA is spotted onto the array, the DNA will be fluorescent and serve as a useful positive control for verifying that the target DNA is binding effectively to the slide surface during the hybridization and washes.
Normalization
In order to compare ratio data from one microarray slide to another microarray slide, the ratio data needs to be normalized to correct for experimental variation. The reason for this is that from one slide to another there will be differences between the relative Cy3 and Cy5 signals due to one or more of the following:
- the amounts of mRNA used in the Cy3 and Cy5 labeling reaction
- the efficiency of detection of the Cy3 and Cy5 by the detection system within the scanner
- relative incorporation differences of the Cy3 and Cy5 reverse transcriptases
Visualization and clustering
After microarray data is normalized to account for differences in Cy3 and Cy5 signals, it can subsequently be exported to any number of data visualization software for further analysis. These software products can be used to mine the data for significant changes in gene expression. The process of visualization can significantly enhance data analysis. It can provide helpful features, such as data integration, customized query devices, and pattern recognition. Clustering data points, or genes, that show similar responses on microarray analysis can be used to identify genes that have similar gene expression patterns and which possibly belong to the same pathway.