High-throughput beta-galactosidase assay for bacterial cell-based reporter systems.

Abstract

BioTechniquesVol. 36, No. 3 BenchmarksOpen AccessHigh-throughput β-galactosidase assay for bacterial cell-based reporter systemsStacey A. Thibodeau, Rui Fang & J. Keith JoungStacey A. ThibodeauMassachusetts General Hospital, Charlestown, MA, USASearch for more papers by this author, Rui FangMassachusetts General Hospital, Charlestown, MA, USASearch for more papers by this author & J. Keith Joung*Address correspondence to: J. Keith Joung, Molecular Pathology Unit, Department of Pathology, Massachusetts General Hospital, 149 13th Street, 7th floor, Charlestown, MA 02129, USA. e-mail: E-mail Address: jjoung@partners.orgMassachusetts General Hospital, Charlestown, MA, USASearch for more papers by this authorPublished Online:6 Jun 2018https://doi.org/10.2144/04363BM07AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinkedInRedditEmail The Escherichia coli lacZ gene (encoding β-galactosidase) is widely used as a reporter gene in bacteria. We routinely measure lacZ expression in bacterial cell-based two-hybrid assays to determine the relative strengths of various protein-DNA and protein-protein interactions (1,2). A typical experiment in our laboratory requires the simultaneous measurement of many potential interactions in the two-hybrid system under a variety of different conditions. Unfortunately, the classic tube-based, timed end point β-galactosidase assay, originally described by Miller (3), is labor-intensive and requires the undivided attention of a researcher when processing large numbers of samples, thereby limiting the total number of assays that can be performed in a single day. Adaptation of the Miller method to a 96-well plate format and use of a kinetic (rather than a single end point) enzyme assay would significantly reduce the labor required. With such changes, all manipulations could be performed with repeating multichannel pipets, and automated temperature-controlled plate readers could be used to conduct unattended kinetic enzyme assays instead of the timed and monitored measurements required in the traditional Miller method.A complete protocol describing how to adapt the original Miller method to a kinetic assay performed in 96-well format has not yet been described. Various reports in the literature detail the adaptation of some of the steps in the Miller method to 96-well plates (4–8). For example, Griffith and Wolf have provided a comprehensive protocol for performing single end point (i.e., nonkinetic) Miller assays in 96-well format (8). However, none of these previous reports have described how to: (i) optimize the growth of bacteria in a 96-well format; (ii) release β-galactosidase enzyme from bacterial cells without the need for organic solvents (which are difficult to manipulate with multichannel pipets); or (iii) express enzyme activities determined from kinetic assays as traditional Miller units.In this report, we describe a complete protocol for performing kinetic β-galactosidase assays in which all steps have been optimized for a 96-well format (see Figure 1 for overview). We describe and validate novel methods for optimizing growth of bacterial cells in 96-well blocks, for using aqueous detergent solutions to release β-galactosidase enzyme from cells, and for using kinetic assay data to obtain values identical to those derived by the traditional Miller method. Our new high-throughput kinetic protocol is rapid, less labor-intensive than the original Miller method, and expands the possibilities for high-throughput applications requiring large numbers of β-galactosidase assays from bacterial cells.Figure 1. Overview of 96-well format β-galactosidase assay.Flat bottom 96-well plates are used and are made from polystyrene (Fisher Scientific, Pittsburgh, PA, USA). Deep well blocks used are polypropylene Masterblocks (Greiner Bio-One, Monroe, NC, USA). Kinetic assays were performed in a Model 680 Microplate Reader (Bio-Rad Laboratories, Hercules, CA, USA) set to a constant temperature of 28°C. For kinetic assays, absorbance at 415 nm (relative to a blank) was measured at short time intervals of 10–60 s for no more than 60 min. [We limited the length of the assay to ensure that o-nitrophenyl-β-d-galactoside (ONPG) substrate remained in excess and to minimize the effects of reaction evaporation.] For each assay well, a velocity (expressed as A415/time) was calculated by plotting time (x-axis) against A415 measurements between 0.1 and 1.0 (y-axis) and calculating the slope of the line (R2 value of 98% or higher for all experiments) using Microplate Manager™ 5.2 software (Bio-Rad Laboratories). An example of kinetic plots generated by Microplate Manager 5.2 is shown. Velocity and A595 measurements made on the plate reader were exported directly to a Microsoft® Excel® spreadsheet to perform calculations using the equation shown in Figure 2C. Additional protocol details and updates regarding the kinetic microplate assay may be found through a link on our web site at http://www.mgh.harvard.edu/pathology/pathology_research_faculty_joung.htm. β-ME, β-mercaptoethanol.An important requirement for successful adaptation of the Miller method to a 96-well format is the need to optimize growth of logarithmic phase bacterial cultures. Griffith and Wolf previously suggested culturing cells in sterile 96-well blocks on a standard orbital platform shaking at 250 rpm and reported that logarithmic phase growth is similar to that seen in 15-mL tubes aerated under the same conditions (8). However, we have found that logarithmic phase cells do not grow as well in tubes (or in a 96-well block) on a standard orbital shaker as they do in tubes aerated on a roller drum (Figure 2A), a method preferred by many researchers. Optimizing the growth of bacteria in a 96-well format is of critical importance as: (i) we and others have found that the consistency of β-galactosidase assays typically depends upon both the density and the growth rate of cells at the time of assay; and (ii) suboptimal growth conditions are undesirable for slower growing or sicker bacterial strains, a common occurrence when expressing foreign proteins in bacteria (as we frequently do when performing two-hybrid experiments). We hypothesized that the suboptimal growth of cultures in 96-well blocks resulted from inadequate agitation and aeration and thus investigated whether faster rotation using a specialized microplate shaker with a smaller throw radius (Microtitertron; Appropriate Technical Resources, Laurel, MD, USA) might improve aeration and the cell growth rate. As shown in Figure 2A, 96-well subcultures agitated at 900 rpm with a 3-mm throw radius grow essentially indistinguishably from cells grown in glass tubes on a roller drum.Figure 2. Comparison of traditional Miller and optimized 96-well format β-galactosidase assay methods.(A) Comparison of logarithmic phase bacterial cell growth in tubes and 96-well blocks. All bacterial cultures were grown in Luria broth containing antibiotics and isopropyl-β-d-thiogalactopyranoside (IPTG) (50 µM). The bacterial strain used is derived from strain KJ1C (1). A single saturated culture of bacterial cells was diluted 1:40 into subcultures, which were then grown under various conditions at 37°C: (i) in 18 × 150-mm glass tubes on a roller drum (rotating at 100 rpm); (ii) in 96-well Masterblocks with 2.4-mL V-shaped wells on a Microtitertron shaker at 900 rpm (with a 3-mm throw radius and 80% humidity); and (iii) in 96-well blocks or 15-mL tubes on a conventional orbital platform shaker (Multitron; Appropriate Technical Resources) at 250 rpm (with a 25-mm throw radius). Cell densities were determined by measuring the A595 of 200 µL of culture in a well in a polystyrene 96-well plate using a Model 680 Microplate Reader. Each point on the graph represents the average measurement from four independent subcultures. (B) Comparison of lysis and permeabilization methods for releasing β-galactosidase activity. Standard Miller assays were performed on logarithmic phase bacterial cultures expressing various levels of β-galactosidase activity (cultures 1, 2, 3, 4, and 5). (These cells harbor various combinations of plasmids expressing two hybrid proteins that mediate differential activation of a lacZ reporter gene present on a single copy episome stably maintained in the cell.) Samples for assay were prepared from each culture using either cell permeabilization by chloroform/sodium dodecyl sulfate (SDS) treatment or cell lysis by a detergent-based reagent and lysozyme. Each sample was assayed in triplicate with the mean and standard error of measurements shown. (C) Original and modified equations for calculating Miller units of β-galactosidase activity. (D) Direct comparison of traditional and 96-well format β-galactosidase assays. Standard Miller assays were performed according to the original protocol (3,9) except that β-galactosidase was released from cells using Popculture/lysozyme instead of chloroform/SDS. Cultures of bacterial strains expressing low, medium, and high levels of β-galactosidase (strains 1, 2, and 3, respectively) were assayed in parallel using the standard Miller method or the 96-well format protocol. Values are the average of two or three independent measurements with standard errors of the mean shown.We have also developed and validated a novel method for releasing β-galactosidase enzyme from bacterial cultures that is ideally suited for use in standard microplates. The original Miller protocol suggests the use of either toluene or a combination of chloroform and sodium dodecyl sulfate (SDS) to permeabilize the bacteria (3,9). Unfortunately, chloroform and toluene react with and destroy the optical clarity of the polystyrene from which most 96-well plates are made. Griffith and Wolf suggested the alternative method of performing permeabilization in deep well blocks made out of nonreactive polypropylene (8). However, we have found it cumbersome to use volatile organic solvents with repeating multichannel pipets, as they rapidly drip out of pipet tips (presumably due to their low surface tension), and they must be used in chemical fume hoods. Others have suggested the use of infectious, high-titer bacteriophage to lyse bacteria (6,7), a method we also found undesirable, as we wished to avoid the possibility of introducing phage contamination into the laboratory.As an alternative, we use a detergent-based solution to lyse bacteria and thus release β-galactosidase activity from cells. We also note that β-galactosidase expressed in mammalian cells is routinely measured from cell lysates (10). Aqueous solutions that lyse cells directly in culture medium (without the need for centrifugation), which do not react with the polystyrene from which most 96-well plates are made and are easily manipulated using standard multichannel pipets, are commercially available from multiple sources (e.g., PopCulture™; Novagen, Madison, WI, USA). Side-by-side β-galactosidase assays on cells either permeabilized with chloroform and SDS or lysed with PopCulture reagent and lysozyme (according to the manufacturer's instructions) demonstrate that PopCulture/lysozyme-treated cells yielded the same or slightly higher β-galactosidase activity compared to chloroform/SDS-treated cells (Figure 2B). In addition, β-galactosidase is extremely stable in lysates produced with PopCulture/lysozyme, as the enzyme activity remains unchanged even after 18 h of incubation at room temperature (data not shown). Our lysis protocol does not cause excessive bubble formation (which can interfere with plate reader measurements) as has been reported with other detergent-based methods (8,11).To quantify β-galactosidase in our cell lysates, we used a kinetic enzyme assay method instead of the single end point method used in the traditional Miller protocol. Others have previously described kinetic assays for β-galactosidase (4–7), and we prefer this format as it utilizes multiple data points to calculate activity, thereby making the measurements less prone to error than simple end point assays. Kinetic enzyme assays, when performed in a programmable, temperature-regulated microplate reader, do not require monitoring by the researcher as timed absorbance readings can be automatically collected and recorded. In addition, with kinetic assays, there is no requirement for a researcher to monitor and stop reactions at a particular color intensity as there is with an end point protocol. All optical density and absorbance readings can be exported directly from the plate reader to spreadsheet programs to perform calculations, thereby eliminating the possibility of data transcription errors.To our knowledge, none of the kinetic β-galactosidase assay protocols described to date provide guidance on how to express enzyme activity in traditional Miller units. In order to permit the comparison of values, we derived an equation (shown in Figure 2C) designed to express the enzyme activities determined using our kinetic method in Miller units (3,9). We therefore made the following alterations to the original Miller equation (Figure 2C): (i) the correction term (1.75 × A550) for cell debris was eliminated as it is unnecessary since our assays are performed kinetically; (ii) the A420 and time (t) values were replaced by the velocity (V) of the enzyme reaction reported in A415/min; (iii) the A600 term was replaced with an A595 term; and (iv) three conversion factors (CF1, CF2, and CF3) were added. CF1 converts A415 values as measured on a plate reader into A420 values as measured in a spectrophotometer and was obtained by measuring dilutions of an o-nitrophenyl (ONP) solution using the two different machines at the two different wavelengths and then plotting these values against each other (CF1 is the slope of the regression line for this plot). CF2 is a term that accounts for the difference in the yellow color caused by the addition of Na2CO3 (to stop reactions) in the end point assay that does not occur in our kinetic assay and was obtained by plotting A415 values obtained from ONP solutions with and without added Na2CO3 against each other (CF2 is the slope of the regression line fitted to this plot). CF3 converts A595 values into A600 values and was obtained by measuring dilutions of a bacterial culture at A595 in the plate reader and A600 in the spectrophotometer (excluding values that fall outside the linear ranges of the two machines) and then plotting the values against each other (CF3 is the slope of the regression line for this plot). For the machines used in this study, CF1 = 1.86, CF2 = 1.56, and CF3 = 1.91 (data not shown).To test whether our new kinetic microplate-based method yields measurements similar to those obtained with the original, single end point Miller assay, we performed side-by-side assays of cells expressing low, medium, and high levels of β-galactosidase (ranging from as few as 80 to greater than 5000) using the two methods. As shown in Figure 2D, the Miller units obtained using our kinetic data and modified equation were essentially indistinguishable from those obtained using the traditional single end point Miller method. Thus, our new method successfully reproduces the values obtained in the original assay.In this report, we have described a complete and optimized protocol for growing bacterial cultures and performing kinetic β-galactosidase assays in a 96-well format. We validated our new protocol by demonstrating that it yields β-galactosidase values essentially identical to those obtained using the original Miller protocol on matched samples. By adapting the Miller method to a 96-well format, we have achieved significant reductions in the labor required to perform β-galactosidase assays and thus increased the upper limit on the number of samples that can be processed in a single day. The increased throughput afforded by our modified assay has already significantly altered the use of β-galactosidase assays in our laboratory. Experiments involving large combinations of proteins, each performed under a variety of different conditions (e.g., various concentrations of inducers), are now performed in a single day with considerably less effort. The availability of our protocol should stimulate additional high-throughput applications for β-galactosidase assays not feasible or not possible with the current, more labor-intensive method.AcknowledgmentsWe thank Astrid Giesecke for initial feedback on the method, William Lach and Donna Robinson for helpful suggestions on growing bacteria in 96-well blocks, Erica Golemis, Andrew Hirsh, and Scot Wolfe for comments on the manuscript, and David Louis and Robert Colvin for support and encouragement. J.K.J. is partially supported by National Institutes of Health grant no. K08DK02883.