Chemical screening requires methods for the liquid handling of chemical compounds. A cost-effective method was used to verify common pipetting steps in chemical compound screening. The work flow that needs to be verified within the screening process consists of four major steps (Figure 1): 1) rearraying of chemical libraries from different multiple-well plate formats into a standard master (Figure 1, step 1); 2) aliquoting of the master plates into ready-to-use compound plates (Figure 1, step 2); 3) stepwise dilution of compounds in different plates (Figure 1, step 3); and 4) dilution of the compounds in the ready-to-use compound plates with medium for screening purposes (Figure 1, step 4).
Figure 1 - Work flow of the process to be checked for quality.
Figure 2 - Instruments used in the validation process. (Images of Tecan instruments were kindly provided by Tecan.)
Figure 3 - Diagrams used to analyze the 2-μL aliquots 10 times.
The quality control of the rearraying in step 1 during the generation of master plates is described here (Figure 1, step 1). The process was carried out using a Freedom EVO® 1 platform (Tecan, Männedorf, Switzerland) equipped with a Tecan TeMO loaded with disposable 50-μL Axygen (Union City, CA) tips (Tecan Multipipetting Option—pipetting head with 96 channels). It consists of four single pipetting steps aliquoting 25 μL from a 04-072-0000 multiple-well plate (Nerbe, Winsen, Germany) into a X7020 multiple-well plate (Genetix, Boston, MA). The second step (Figure 1, step 2) was carried out with a Biomek FXp and AP384 384-well multipipetting head (Beckman Coulter, Fullerton, CA) using Beckman 384-well 30-μL standard disposable tips (see Figure 2). The instrument pipetted 2 μL 10 times in serial steps from an X7020 multiple-well plate into 10 model 781271 plates (Greiner, Frickenhausen, Germany) (see Figure 3).
The third step (Figure 1, step 3) consisted of several pipetting steps using Axygen 50-μL disposable tips to obtain an eightfold dilution series of compounds in two X7020 plates. First, quadrants 2, 3, and 4 of the first dilution plate and quadrants 1, 2, 3, and 4 of the second dilution plate were prefilled with 30 μL of dimethylsulfoxide (DMSO). Second, 45 μL of a 04-072-0000 multiple-well plate was pipetted into the first quadrant of the first X7020 plate. Lastly, 15 μL was pipetted and mixed in a cascade from quadrant 1 of the first X7020 plate to quadrant 4 of the second X7020 plate to obtain a 1:3 stepwise dilution of the compound solutions. Pipetting was carried out following the standard Z-format. At the end of the process, 15 μL from quadrant 4 of the X7020 plate 2 was discarded to have equal volume in all wells. The fourth and last step to be verified (Figure 1, step 4) consisted of a Freedom EVO robot, using 200-μL Tecan disposable tips, first diluting 2 μL of compounds by adding 198 μL of Dulbecco’s Modified Eagle’s Medium (DMEM). Then, 15 μL of this solution was added to a Greiner 655090 multiple-well plate (clear 96-well cell culture plate for use on microscopes), and 5 μL was added to a 781056 384-well cell culture multiple-well plate (Evotec, Hamburg, Germany).
To determine variations in pipetting processes, the type of errors that need to be recognized have to be determined, and then the precision limits of the quality control experiments must be known. In the present study, the variations that needed to be verified were: 1) variation of the volume that was pipetted into the wells of a multiple-well plate (pipetting accuracy), 2) precision of the pipetting within the plate, 3) precision of the dilution steps, and 4) consistency of pipetting between plates (interplate variation). Based on the quality required and the capabilities of the instruments, the tolerance limits were set to 5% for the intraplate variation (quadrants in the case of the dilution series). For the gravimetric method (see below), the tolerance limit was 5% as well.
Technologies and equipment
To address the above four variations, technologies are required that support the measurement of the values to be monitored. The authors decided on three methods: 1) gravimetric measurement to validate the accuracy of the pipetted volume over a multiple-well plate using a CP 124 S balance (Sartorius, Goettingen, Germany), 2) absorption of OrangeG (Carl Roth GmbH & Co. KG, Karlsruhe, Germany) using a Tecan GENios Plus luminescence reader, and 3) fluorescence emission using Alexa 488 (Invitrogen Corp., Carlsbad, CA) together with a Tecan GENios Pro fluorescence reader. With these simple technologies, it was possible to acquire meaningful validation data for the pipetted volume accuracy, reproducibility of pipetting within the wells, and linearity of dilution over a wide dynamic range.
To verify that the quantity of liquid dispensed over an entire plate was correct, the weight of the plate before and after pipetting was measured. OrangeG was used for this experiment, which allowed the absorbance in each plate to be measured in order to estimate the intraplate variation. To assess interplate variation, the same procedure was repeated several times and the plate-to-plate results were analyzed statistically. These tests could be applied to steps 1, 2, and 4. To verify the dilution procedures, which require a wider dynamic measurement range, an Alexa 488 fluorescent solution was used in combination with weighing.
In order to mimic the screening solutions as closely as possible, DMSO and DMEM solutions were employed when appropriate. When using different solvents or buffers, their effect on the measurement must first be assessed. For the gravimetric measurement, the density of the solutions must to be taken into account for calculating the expected weight increase after pipetting. Furthermore, normalization to blank wells containing either phosphate buffered saline (PBS) only or nothing needs to be carried out. A Microsoft® Excel™ file (Redmond, WA) was created, which enabled the measured data to be added, and the mean, median, standard deviation, minimum, maximum, and relationships of the quadrants to be calculated automatically. Furthermore, the Excel file automatically normalized the data in regard to the blank wells containing media only. To visualize the results, Excel’s conditional formatting was used to highlight values that were out of the tolerance limits, and bar graphs were utilized to visualize the actual results.
Gravimetric and absorbance measurements were easily implemented and required minimal adjustment. For the gravimetric measurement, the expected weight increase was calculated by multiplying the volume pipetted by the density times the number of wells. The net weight was then compared to the expected, calculated value. For the absorbance tests, the OrangeG concentration was adjusted to 0.125 μg/μL in either water or buffer. The pipetting processes were first established with the final plate type to be used in the screening. For the quality control process, the plate types had to be changed from X7020 plates to 781076 and Corning 3675 plates (Corning Incorporated, Corning, NY) to be reader-compatible for absorbance and fluorescence measurements, respectively.
Figure 4 - Relationship between Alexa 488 amounts in DMSO filled with different liquids and their fluorescence measurement values.
It was more difficult than anticipated to set up the fluorescence readout for verification of the dilution steps. First, it was discovered that Alexa 488 dye tends to adhere to the 781076 plates. Plates were routinely reused to reduce costs, and residue of Alexa 488 dye was discovered in the wells. In addition, the fluorescence properties of Alexa 488 were strongly influenced by the concentration and formulation of the DMSO solvent (see Figure 4). Alexa 488 does not fluoresce in pure DMSO. Fluorescence is gradually recovered by dilution with either water or PBS due to its buffering nature, albeit with different intensities. Fluorescence intensity was generally enhanced in PBS/DMSO solutions compared to water/DMSO solutions. Furthermore, DMSO autofluoresces in PBS, and thus the blank wells showed a concentration-dependent readout.
Conclusion
A useful, cost-effective method was developed to verify common pipetting steps in chemical compound screening. These routines employ three different types of commonly available technologies: gravimetric, absorbance, and fluorescence readouts. The statistical treatment of the data was also automated, using Microsoft Excel, which facilitated and increased the robustness of the analysis. Care must be taken to ascertain the compatibility of the multiple-well plate type and the reader. Furthermore, because the fluorescence behavior of chromophores is strongly dependent on the molecular environment, fluorescence emission must be verified in the different solvents to be used.
Mr. Wagner is Head of Automation, HT-Technology Development Studio (TDS); Mr. Gierth is an apprentice; and Dr. Bickle is Service Leader, HT-Technology Development Studio (TDS), Max Planck Institute of Molecular Cell Biology and Genetics, Pfotenhauerstrasse 108/01307 Dresden, Germany; tel.: +49 351 210 2818; fax.: +49 351 210 1349; e-mail: [email protected].