An Introduction to printing spotted microarrays
Overview
Microarrays are typically produced by transferring gene or transcript specific PCR amplified cDNA clones or long oligonucleotides in either high salt or other denaturing solutions from 384-well microtitre plates to chemically modified 25 x 75 mm glass microscope slides using robotic contact printing instruments. This process is sometimes termed robotic spotting or just spotting. Although robotic spotters were first described by Schena et al. (1995) they are also available commercially.
Printing microarrays
The spotting pins are a vital component because they are the only part of the instrument to contact the probe DNA and the slides. Spotting pins draw fluid into the pins by capillary action when a pin loading is performed. The pins enter adjacent wells of the 384-well microtitre plate because they have a fixed pitch of 450 µm in the print-head. The probe DNA within the microtitre plates therefore needs to be arranged accordingly for any given miroarray layout to be printed correctly.
Example source visit layouts
A) A 384-well Genetix plate. B) 12x4 pin-tool source visit layout. C) 4x4 pin-tool source visit layout
When depositing a spot on the slide, surface tension interactions between the substrate surface and spotting buffer lead to spot formation when the pins pull away. Spot diameter is determined by a variety of parameters including the spotting buffer, substrate slide, temperature, relative humidity and the pin itself. Most spotting pins are blunt ended (50 to 100 µm) with a capillary or split and contain a storage reservoir. An example of one such pin is shown in the figure below.
Electron micrograph of a BioRobotics MicroSpot 2500 pin
Spot density is determined by the diameter of the printed spots because this dictates the centre-to-centre spot distance that can be used during printing. Additionally, the accuracy and reliability of the robotic spotter will determine the reproducibility of the spot positioning within each microarray. This is sometimes referred to as pin or spot wobble. An imprecise instrument with lots of pin or spot wobble will require a greater centre-to-centre spot distance, which will reduce the maximum spot density that can be achieved.
Sub-grids and meta-grids
Each pin within the print-head will produce a single pin-patch or a sub-grid of spots. The sum total of sub-grids printed by one pin-tool is called a meta-grid. A single microarray can consist of one or more meta-grids. The number of meta-grids that be printed per microarray is limited by the space available. FlyChip typically spots microarrays using 48 pins and therefore has 48 sub-grids per microarray but just one meta-grid. Other formats are also possible with our set-up and have indeed been produced.
Microarray meta-grid and sub-grid organisation
Rush spots
The first few spots to be printed after a pin loading, the rush spots, are larger than any subsequently printed spots, as when a pin loading is performed some spotting solution will coat the outer surfaces of each pin. Until printing has exhausted this additional spotting solution the printed spots will not have a regular diameter. This problem can be overcome by discarding the first few slides after each pin loading or by using blotting slides that can then be discarded. The later option is preferable when printing arrays with an inter-spot distance that is less than the maximum spot size.
Rush spots
The above image is a close-up of 1 sub-grid printed with Cy3-labelled sonicated salmon sperm DNA using a BioRobotics MicroGrid II 600 spotter MicroSpot 2500 pins. The spotter started to print spots in the bottom left-hand corner and moved to the right and then up the image. 35 spots were printed per pin loading. This image clearly shows how the spot size initially decreases in diameter before reaching a uniform size. The number of slides to be discarded at the start of each print-run or the number of blotting spots to be printed should be defined in advance.
Spot morphology
Key determinants of spot morphology have previously been described in the literature. These include substrate chemistry and hydrophobicity, spotter calibration and print settings, spotting buffer viscosity, pH and evaporation, probe DNA concentration, room temperature and relative humidity. Because of these complex interactions it is not possible to predict in advance which combination of conditions should be used for any given clone-set or spotter. For this reason, we routinely check and then optimise the spotting conditions for each of our clone-sets.
Wash cycles
Wash cycles are performed between each pin loading to ensure that the spotting solution from one loading does not contaminate any subsequent pin loadings. The number and length of wash cycles is a key determinant of how long any given print run will last for because the time taken to perform the pin loadings and spot depositions is minimal. Care must therefore be taken to ensure that that wash cycle is optimised for a fast print-run time and low probe DNA carry-over.
Cleanliness
Microarrays should be printed, stored, hybridised and scanned in a clean dust-free environment to ensure that arrays of the highest possible quality are produced. Spotting pins contain small capillaries and reservoirs that can be easily blocked by dust. Such blockages can lead to pins performing badly or being predisposed to further blockages. Poorly performing pins will either print inconsistently or not at all and a significant amount of potentially valuable data will be lost.
| A clean pin | A dirty pin |
|---|---|
|
|
The biological samples being co-hybridised to the microarray are labeled with the Cy3 and Cy5 dyes respectively. Unfortunately, dust and other airborne contaminants are fluorescent under the Cy3 and/or Cy5 wavelengths. The complexities of image analysis mean that any fluorescent contaminant that lies within the spot will adversely affect the measured spot signal and hence, interpretation of the experimental result.
Dust particles
Spot tracking
Each microarray spot has a unique position and each spot corresponds to a specific probe DNA from a specific well of a microtitre plate. There are typically thousands to tens of thousands of probe DNAs distributed between tens of source plates and thousands to tens of thousands of spots on each microarray. A single microarray experiment can consist of a few or a hundred microarray hybridisations. Tracking where and what each probe DNA is on each microarray is therefore an important issue.
Most robotic spotters are supplied with a data tracking program that uses an input file to describe the positions of each probe DNA within the microtitre plates and another file that defines how the microarray was printed to produce a description of where each probe DNA is within each microarray. These spot identities can then be imported into a spot finding and quantification tool that will 'append' the fluorescence spot signal. These data are then analysed to determine what affect any given experimental condition or treatment has had on the gene expression of the samples being compared.
Summary
Many disparate factors need to be considered when printing microarrays. These affect the quality of the microarrays produced, the rate of production and the ease with which any downstream data analysis may be performed. Failure to fully appreciate these limitations will result in sub-standard microarrays being produced and experimental data that may be unreliable. Extreme care and consideration must therefore be taken to ensure that the best quality microarrays are produced.
R. Auburn (17-02-2006).
