By Hazel Richards, Robotics Project Manager, Genpak Ltd.
This article is the second in the series on bringing microarray technology into practice. In this article, Hazel Richards of Genpak Ltd. (Brighton, UK) steps you through two ways of creating your own array, manually or with the use of a robot.
Introduction (Back to Top)
At the simplest level, a microarrayer is a system for transferring material from one location, the source, to a second location, the target. In practice, source material is usually contained in a microtitre plate, and the target is either a glass microscope slide or nylon filter.
It is possible to produce microarrays using either an automated system or by hand. Robotic systems have several advantages: they are fast and accurate, produce large numbers of identical arrays, and don't get bored or want time off. They are also relatively expensive from the point of view of initial purchase costs.
Hand-Spotted Arrays (Back to Top)
It is perfectly feasible to produce an array by hand. Obviously, the scale is much smaller than an automated system, but it is a reasonable alternative if only a few slides, each with 200 different samples or less, are required.
The center to center distance between adjacent spots is known as the feature pitch. It is reasonably easy to produce hand-spotted arrays with a feature pitch of 0.5 or 1.0 mm. Although it is possible to "handspot" at pitches of about 150–200 nm, it's not recommended.
Using a pitch of 1.0 mm and an 8 x 12 grid layout, which corresponds to a standard 96 well microtitre plate, means that up to 6 grids will fit on a standard microscope slide. However, this doesn't leave much of a handling margin; two grids per slide is much easier to work with. Each grid can be different, or they can be duplicates. Some degree of magnification is necessary—two-fold is adequate. Because the most time-consuming element of producing an array is the cleaning cycle, working on several slides at once is recommended. Three slides, for example, fit nicely into the field of view of a low-power magnifying lamp.
Stronger magnification is required to produce arrays at 0.5 mm pitch. Using a stereo zoom microscope at six fold magnification works well, but obviously only one slide can be produced at a time. Although each 0.5 mm format grid is quite small, it is best to keep the total number of grids per slide down to eight or less.
Grids are laid with the aid of a simple template placed underneath the slides. This can be a piece of graph paper taped to the underside of the slide (very cheap, but awkward to use); an indexed printed grid (still cheap, and much easier use); or an engraved stage similar to a microscope stage (much more expensive). An easy and sensible solution is to use an indexed adhesive label for each slide. This provides a permanent means of identifying each sample within a grid. Genpak can supply suitable indexed and gridded adhesive labels to fit standard microscope slides in either 1.0 or 0.5 mm format.
Sample material is colorless and dries almost instantly when spotted out, making it extremely easy to lose track. Hence, it is necessary to keep a careful record of which samples have been spotted out and their locations within an array. Scientists at Genpak are working on incorporating a dye into sample solutions that won't interfere with either hybridization reactions or affect laser scanning.
Samples are spotted out using a printing pin identical to those used in the automated version. Spotting is not difficult—it only requires a steady hand and lots of patience. Spots are simply produced by gently touching the pin down onto the slide surface. The first few spots should be made onto a "waste" area, as they are inevitably uneven until excess sample material on the outside of the pin has drained away.
To produce a circular spot the pin should be held as vertically as possible, if the angle is much less than 80°, the spots will be oval. Pins should be touched down gently but firmly; if not, the spots end up comma shaped. The dwell time (how long the pin actually touches the surface) also affects the volume that is deposited. Too much sample results in an uneven "cracked" spot or a doughnut effect. Neither is good for reading the array following hybridisation.
Practice really does make perfect with hand spotting. Tyro spotters might want to practice producing dummy arrays, using DNA and a suitable dye. Practice arrays can be viewed under ultraviolet light or with a stereo microscope. The effectiveness of the cleaning protocols can also be checked using fluorescent dyes.
The pin must be cleaned between successive samples. We find that a simple 4-stage process is adequate. The pin is first touched down onto a blotting pad to remove sample material, then washed in distilled water and rinsed in 95% ethanol. Finally, the pin is dried by blotting onto a second pad. This is the same protocol as we use in our genArray 21 automated system. Following the completion of each set of slides, the pin is cleaned using ultrasound.
genStation 3 (Back to Top)
Genpak has developed a bench-top workstation, the genStation 3, to facilitate the production of hand spotted arrays. The workstation is designed to bring together all the components, including a cleaning station, required to produce an array. It is available in two versions. The basic version is most suited to producing 1.0mm arrays, and incorporates a low power magnification system and halogen light source. The enhanced version incorporates a stereo zoom microscope with an integral light source. It is equally suitable for producing arrays pitched at 1.0 mm, 0.5 mm, or less. Using a 0.5 x objective increases the working distance. This system has a magnification range of x5 to x40.
To eliminate the spotters dilemma of "Where was I?" or "Which well is next?", both systems incorporate a sample tracking device and counter. This reduces the likelihood that samples will be over-spotted, omitted, or duplicated. Separate sheltered storage areas for clean and arrayed slides, as well as for spotting pins, are also provided.
The genStation 3 system retails from $3000.
Motion Control System (Back to Top)
It is important that the X-Y positional accuracy of the robot is as high as possible. Most systems on the market have an X-Y positional accuracy of ±10 µm, a few of ±5 µm. The genArray 21 system has a positional accuracy of ±1 µm. An X-Y accuracy of only ±10 µm means that the pitch between adjacent spots, or features, could be as much as 20 µm "out" from that intended. Minimizing positioning errors is of particular importance when it comes to reading the array, since individual features may be only 50 µm in diameter.
Positional accuracy of the Z-axis is usually an order of magnitude greater than X-Y axes; the genArray 21 system is accurate to within ±0.125 µm.
Printhead (Back to Top)
Printheads maintain a constant positional relationship between the robot and the printing pins. They must have a footprint that is compatible with both source and target surfaces. The number of and spacing between printing pin positions is determined by the geometries of source and target surfaces.
The center to center distance (pitch) between adjacent wells of a microtitre plate varies with the format of the plate; 96 well plates have a 9.0 mm pitch, for 384 plates it is 4.5 mm, and for 1536 plates it is 2.25 mm. The diameter of the individual wells also reduces, from 7.0 mm to 1.5 mm.
In practice this means that pins are positioned a multiple of 1.5mm apart. Most commonly, the pitch is set at 4.5 mm or 9.0 mm. A 9.0 mm pitch is most suited to sampling from 96 well plates. Size limitations imposed by the target microscope slides means that a maximum of 12 pins, in a 3 x 4, array can be accommodated. Such an array would pick up samples from a block of 12 wells (e.g., A1 to C4) from a 96 well plate, or alternate wells in a block 24 wells, A1 to E8, from a 384 well plate.
A pitch of 4.5 mm is best suited for 384 well plates. A maximum of 48 pins, in a 4 x 12 array, is compatible with microscope slide geometry. A 4.5 mm printhead can be used to sample 96 well plates if alternate pin positions are left vacant.
Generally speaking, if the footprint of the printhead is n.p, where n = the number of pins on any axis, and p = the pitch, then the minimum size of the target surface, excluding a handling margin, is (n+1) p.
Printing Pins (Back to Top)
The standard of any array is a direct reflection of the quality of the printing pins used to produce it. Pins must conform to an exact specification to ensure that uptake and deposit volumes for any given set of samples is constant. The uptake and deposit volumes will vary slightly between different samples due to the effects of density and viscosity on capillary action. The diameter of any printed feature depends on the volume of material deposited as well as the dimensions of the pin tip where it touches the target surface. Within any one set of arrays, the printed features must be uniform in size. Features should be circular in shape (greatest possible surface area within the smallest perimeter) to maximize the number of significant pixels when reading the array. Pins must also be robust and long lasting, and as easy to clean as possible.
Cleaning Station (Back to Top)
Pins must be cleaned between consecutive sampling operations. Cleaning protocols vary greatly in their complexity, depending on the degree of paranoia in the operator. However, an increase in complexity does not necessarily mean that such a protocol is more efficient than a simple one. The standard genArray cleaning protocol is a 4-stage process described above. Additionally, pins are cleaned between successive runs using ultrasound, which is essential to maintain optimum pin performance.
Workstation (Back to Top)
The robot must be supported, and isolated from the external environment. Any vibration experienced by the robot will be transmitted to the printhead and pins. This will make the controller "hunt" for its correct position, and reduce the overall performance of the system. Robots are usually supported on a workstation that is made up of a frame that isolates the robot from external vibration sources and a "breadboard" type tabletop. The tabletop must be capable of quickly damping down any vibration generated by the operation of the robot itself, and also provide a very flat surface on which to place source and target material, as well as any other components required by the system.
PC Control (Back to Top)
The robot is controlled through a PC, using Windows based software. Our software offers the user a choice of options: they can choose a "ready-to-go" program from a pre-set menu, or can design a custom program to suit their individual requirements. Variables include: the size (96, 384, 786, or 1536 wells) and number of source plates, type of printing pin, the number of copies of each sample to be printed, feature size and spacing (pitch), and the number of target slides or filters.
It is possible to build your own arrayer; there is lots of information available on the web. However, designing, sourcing, and constructing a system involves a lot of time and effort. Commercially available systems eliminate such problems, and some, are very reasonably priced. The genArray 21 system costs from $65,000.
For more information: Hazel Richards, Genpak Ltd., Sussex Innovation Centre, Science Park Square, Brighton BN1 9SB UK, Tel: +44 1273 704470. Fax: +44 1273 626213.