Plate assays (1536-well) that require washing (i.e., enzyme-linked immunosorbent assays [ELISAs] and cell-based), are not routinely implemented in high-throughput screening laboratories due to the lack of reliable instrumentation. This became possible with the development of a 1536-well plate washer (GNF Systems, San Diego, CA). An eight-tip angled head dispenses fluid to the wells, and a 32-tip single-column head aspirates fluid from the wells, one column at a time. Typically, this multistep process is the rate-limiting operation of a screening campaign. Assays that contain a wash step can produce 100 plates per 24-hr period, whereas assays without wash steps can produce as many as 500 plates in the same time period. To compensate for these differences, engineers at GNF Systems have designed a 64-tip double-column aspirator head that increases throughput and produces data of equal quality to the original aspirator head.
Figure 1 - 1536-well plate washer on high-throughput screening (HTS) platform, equipped with sonicating wash bath.
The equipment design is featured in Figure 1. There are two dispense heads, each of which can accommodate straight or angled dispensing. Angled dispensing is needed for adherent cells to prevent detachment. The dispense tips are pressure-driven solenoid valves with MINSTAC tips (The Lee Company, Westbrook, CT). The aspirator head is connected to a vacuum, which removes the waste during washing. The washer has been equipped with a sonicating wash bath to provide cleaning for both dispense tips and the aspirator head. There are two tip predispense locations, one of which can be used for recovery of precious reagents, or to keep hazardous material out of the waste stream. The double-column aspirator head can be disassembled into three separate pieces by removing four screws. This allows for easier access during off-line cleaning.
The first experiment to test the double-column aspirator head was the quantification of residual volume. The ability of the aspirator head to consistently and reliably leave behind less than 1 μL/well is especially important for enzymatic assays such as ELISAs. This experiment was performed for two common plate types—1536-well solid-bottom HiBase (Corning Life Sciences, Corning, NY) and 1536-well clear-bottom LoBase (Aurora Biotechnologies, Carlsbad, CA). First an empty assay plate was weighed, and then the plate was filled with 5 μL/well of phosphate-buffered saline (PBS) followed by a 3× wash with PBS. One wash step is defined as filling the well and aspirating to leave behind the desired amount. This was repeated five times for both plate types, and all passed the acceptance criteria, leaving less than 1 μL/well.
Another important experiment was carryover testing to make certain that there was no cross-contamination between the plate wells due to the aspirator head. According to the acceptance criteria, there had to be less than 5% well-to-well crossover of signal. First, 6 μL/well was dispensed to alternating double columns of PBS and PBS + fluorescein. The fluorescence was measured on a Safire II plate reader (Tecan, Männedorf, Switzerland) to determine the average relative fluorescence unit (RFU) of the hot or fluorescent wells. After a 1× wash with PBS, aspirating to 1 μL/well and dispensing to a final volume of 6 μL/ well, the fluorescence was read again to determine the maximum value of the cold or buffer-only wells. Next, a 2× wash (for a total of three washes) was performed, followed by a final read. The percent carryover was then calculated by taking the maximum of any cold well after a wash, divided by the average of all hot wells from the initial read, multiplied by 100. This experiment was repeated five times for both plate types mentioned above. All plates met the acceptance criteria, achieving under 5% crossover of signal.
Figure 2 - HEK cell adhesion experiment. a) Plate prior to washing. b) Plate after 3× wash. No-wash region is columns 45–48, on the right side of the plate. The percent reduction in signal of the washed region compared to the no-wash region is only 5.4% (under the 15% criteria).
Cell adhesion was tested using both human transformed primary embryonic kidney (HEK) and Chinese Hamster Ovary (CHO) cell lines. This was to ensure that the aspirator head could wash effectively without removing any adhered cells. The HEK cells were plated at 4 μL/well, 8000 cells/well, and incubated for approximately 24 hr at 37 °C, 5% CO2, 95% RH. Next, 4 μL/well of Calcein AM fluorescent dye (BD Biosciences, Bedford, MA) at 1 μM was added and incubated for 1 hr at 37 °C, 5% CO2, 95% RH. A fluorescence bottom read was performed on the Safire to get an initial read. After a 1× wash of columns 1–44, leaving columns 45–48 as a no-wash region, the plate was again read. A 2× wash was then performed (for a total of three washes) on columns 1–44, followed by a read. The percent reduction in RFU was calculated for the washed regions based on the initial values of RFU for that plate (refer to Figure 2). The acceptance criteria used were as follows: The percent reduction in signal must be less than 15%, the CVs must be less than 12%, and there could be no drop-out wells (a well with an RFU <0.5× median RFU of the no wash values) or double dispenses (a well with an RFU >1.5× median RFU). An n = 7 was performed, and all plates passed the acceptance criteria.
Figure 3 - CHO cell adhesion experiment. a) Plate after 1× wash. b) Plate after 3× wash. The percent reduction in signal was 3.4 and 4.0% (under the 15% criteria).
The CHO cells were plated at 4 μL/well, 4000 cells/well, and incubated for approximately 24 hr at 37 °C, 5% CO2, 95% RH. A 1× wash (n = 4) or a 3× wash (n = 7) was performed on columns 1–44, leaving columns 45–48 as a no-wash region. With 4 μL/well left in the wells, 4 μL/well of CellTiter-Glo® (Promega Corp., Madison, WI) was then added to columns 1–48. A 10-min room-temperature incubation was followed by a centrifugation step (1000 rpm for 20 sec) to remove bubbles. The luminescence was then detected on a ViewLux™ charge-coupled device (CCD) imager (PerkinElmer, Waltham, MA), and the percent reduction in RLU was calculated based on the no-wash regions of each plate (refer to Figure 3). Using the same criteria as mentioned above with the HEK cells, all plates passed.
To compare the biological results of the double-versus single-column aspirator head, a calcium flux assay was developed using the FLIPRTETRA® reader (Molecular Devices, Sunnyvale, CA), which is equipped with a 1536-well bulk pipettor. CHO cells (4 μL/well) containing the M1 receptor were dispensed at 4000 cells/well and incubated overnight at 37 °C, 5% CO2, 95% RH. A 2× wash with 9 μL/well of assay buffer was performed, leaving the final volume at 4 μL/well. Fluo-4 AM (Invitrogen Corp., Carlsbad, CA), used to measure intracellular calcium flux, was added at 4 μL/well and incubated for 0.5 hr at the above conditions. Test compounds and the control compound were then added using a 50-nL pintool. After an additional 0.5-hr incubation at previous conditions, the plate was transferred to the FLIPRTETRA reader. Two microliters per well of carbachol (a known agonist for the M1 receptor) at a concentration of EC80 was added to columns 1–44, and 2 μL/well assay buffer to columns 45–48. The read involved 3× mixing in the assay plate, 470–495 nM excitation, and 515–575 nM emission (exp. = 0.4 sec/gain = 80), 60 images, 1 per sec.
Figure 4 - Carbachol dose response in CHO M1 cells, comparing single- and double-column aspirator head.
Figure 5 - DMSO plate performed with single-column aspirator head (a) and with double-column aspirator head (b). Carbachol at EC80 was added to columns 1–44; buffer only was added to columns 45–48 (100% inhibition control).
Before any test compound plates were run for the assay, a dose response curve of carbachol was performed for both double- and single-column aspirator heads. This resulted in 286 nM and 359 nM, respectively (Figure 4). When adding carbachol at a concentration of EC80, dimethyl sulfoxide (DMSO) plates were compared for both the single and double-column aspirator heads and achieved very similar biological results (refer to Figure 5). Several known inhibitors were also tested, and yielded IC50 values comparable to those found in the literature.
An ELISA containing a total of nine washes was run successfully on the GNF Systems HTS platform using double-column aspirator heads for all washes. There were no issues affecting biology or throughput, and the sonicating wash baths required very little operator maintenance by preventing any potential clogging of the tips.
Figure 6 - Operational times calculated from ELISA run on HTS system. First four rows are individual operations performed at one washer; in addition to these four operations is cycle time (rate at which plates go into and out of the system). The double-column aspirator gave a 21% increase in throughput for this assay.
The biggest impact of the aspirator is on robotic operational times. The ELISA provided the best example (see Figure 6). One washer on the system performed four different operations in the protocol (which are the first four rows on the chart). The rate-limiting step is the addition of these four operations. This is the cycle time, or the rate at which plates come into and out of the system. HTS was completed in three fewer days with the double-column aspirator head, which was a 21% increase in throughput for this assay.
It has been demonstrated that the double-column aspirator head produces no well-to-well contamination or carryover. It can wash effectively on top of cells, aspirating small volumes to prevent any disturbance of the cells adhered to the bottom of the wells. Bulk amounts of the fluid can be reliably and consistently aspirated from each well, which was essential for the ELISA. The same biological results are achieved when compared to the original head, shown with the calcium flux assay. The aspirator head, which has three separate parts, has made off-line cleaning easier and faster. Furthermore, by sonicating the aspirator tips in a sonicator, less on-line operator maintenance is required over the duration of a screen. The most significant impact of the double aspirator is on operational times, which for the ELISA mentioned, increased throughput by 21%.
Ms. Batchlett is a Chemist, Mr. Cassaday is a Senior Research Engineer, Ms. Ohart is a Staff Biologist, Mr. Berry is a Senior Research Engineer, and Dr. Kornienko is a Senior Research Fellow, Merck & Co., Inc., 140 Wissahickon Ave., North Wales, PA 19454, U.S.A.; tel.: 267-305-3972; e-mail: firstname.lastname@example.org. Mr. Chang is Project Leader, GNF Systems, San Diego, CA, U.S.A.