Screening a library of household substances for inhibitors of phosphatases: An introduction to high-throughput screening


  • Ann T. S. Taylor

    Corresponding author
    1. Chemistry Department, Wabash College, Crawfordsville, Indiana 47933
    • Chemistry Department, Wabash College, Crawfordsville, IN 47933. Tel.: 765-361-6186; Fax: 765-361-6340
    Search for more papers by this author


Library screening methods are commonly used in industry and research. This article describes an experiment that screens a library of household substances for properties that would make a good “drug,” including enzyme inhibition, neutral pH, and nondenaturing to proteins, using wheat germ acid phosphatase as the target protein. An adaptation of the experiment appropriate for lower level biochemistry or outreach is also described.

This work was supported by Wabash College through the Haines Fund for the Study of Biochemistry and the National Science Foundation through Grant DUE 0126242.

Combinatorial procedures have revolutionized chemistry and biochemistry, requiring the use of microscale high-throughput screening techniques to analyze libraries of compounds for specific properties. While these methods are standard in the chemical and pharmaceutical industries [1, 2], this revolution has not yet been incorporated into undergraduate biochemistry curricula.

The target enzyme for this drug screening experiment is wheat germ acid phosphatase. Phosphatases catalyze the hydrolysis of phosphate groups from substrate molecules and are categorized either by their activity conditions (acid or alkaline) or by their substrates (e.g. phosphoinositol phosphatases, protein tyrosine phosphatases, protein serine/threonine phosphatases). They are commonly used in the teaching laboratory in kinetics assays, using p-nitrophenyl phosphate as a substrate [37]. Alkaline phosphatases are commonly used in the research laboratory as antibody conjugates. Protein phosphatases are critically involved in the regulation of many signal transduction processes and are potential drug targets for the treatment of diabetes mellitus, cancer, and inflammation [810].

The goal of the laboratory experiment described here is to introduce students to working with a large number of samples and to simulate a high-throughput screening of a library for a potential drug candidate. The experiment employs strategies similar to those used in a search for inhibitors of protein tyrosine phosphatase 1B to treat diabetes mellitus [1114], though the library size is significantly smaller (21 versus 400,000 [14]) and uses mixtures instead of pure compounds. A library of household substances is screened for inhibition of the enzyme. Because enzyme activity can be altered by changing the pH of the medium or by unfolding the protein, these effects are also evaluated, and any substances that significantly alter either of these parameters are omitted. Students may perform the assays in any order, eliminating substances as soon as they “fail” a test. Finally, students select three substances from the library to determine the type of inhibition produced and Km and Vmax in the presence of the compound.

The experiment occupies a single laboratory period in a one-semester biochemistry course. The students in the class are primarily junior and senior chemistry and biology majors and work as partners. This screening immediately follows an enzyme kinetics experiment, concurrent with lectures on enzyme kinetics and inhibition. Students have prior laboratory experience of pH and protein folding previously in the semester and are expected to design their own procedure in the handout (Fig. 1) and check with the instructor. It reinforces the principles of pH, protein folding, and enzyme kinetics and encourages good laboratory record-keeping skills and experimental design principles. Alternatively, this experiment can also be used by itself, with more detailed directions given. It can easily be adapted for use with lower level biochemistry students or as an outreach experiment with high school students (Fig. 2). Twenty-four-well plates and droppers are used while fewer substances are tested by each partnership, and data from the class can be compiled to provide a larger library. When coupled with a molecular modeling exercise that focuses structural approaches to drug design, the abbreviated experiment is an excellent preface to a tour of a pharmaceutical company.



All chemicals were obtained from Sigma (St. Louis, MO); catalog numbers are indicated. Household chemicals were purchased at local grocery and retail discount stores.

Library Preparation—

The library consists of a wide variety of household substances, ranging from over-the-counter drugs to beauty products to spices (Table I). These substances are prepared as 10% (v/v) aqueous solutions for liquid items and 5% w/v aqueous stock solutions for solids and given an identifying code. The latter solutions are centrifuged to remove any particulate matter. To aid in dispensing the library to students, the samples are distributed both in screw cap tubes and in 96-well dispensing plates, leaving one well in each column blank for controls (Table II). Students use multichannel pipetmen to dispense the samples into the test plates. In the outreach version, disposable plastic droppers are used for all solution additions.

Enzyme Inhibition Assay—

Wheat germ acid phosphatase previously prepared by students in previous weeks or commercially available acid phosphatase (1 U/ml; Sigma P 3627) is used. For each test compound, 20 μl of substance, 12 μl of 5 mm pNPP (0.3 mm final concentration; Sigma N 3254), 100 μl of 100 mm sodium acetate buffer, pH 4.7, and 48 μl of water are mixed in a well of a 96-well plate. Then 20 μl of enzyme solution was added to start the reaction. After 5 min, 100 μl of 1 m NaOH was added to stop the reaction, and the absorbance was measured at 405 nm using a SpectraMax plate reader. The Path Check function was used to reduce error due to variations in volume. A blank reaction with no enzyme is used as a measure of background hydrolysis of the pNPP, and 5 mm sodium phosphate buffer, pH 4.7, was used as a known inhibitor. By simply changing the buffer system to pH 8.0 Tris buffer, the assay is easily adapted for use with alkaline phosphatase as a target enzyme.

In the outreach version of the experiment, five drops of the substance, five drops of phosphatase, and five drops of 1 mm pNPP are mixed. After 5 min, two drops of 0.1 m NaOH are added to the wells. Inhibition is determined visually. If a substance inhibits the enzyme, the sample is colorless; if there is no inhibition, the sample is yellow.

pH Screen—

A universal indicator composed of 0.025 mg/ml thymol blue (Sigma T9887), 0.063 mg/ml methyl red (Sigma M7267), 0.5 g/liter phenolphthalein (Sigma P 9750), and 0.25 mg/ml bromthymol blue (Aldrich 11441-3), in 50% (v/v) ethanol is adjusted to green with 0.5 m NaOH. Alternatively, red cabbage indicator, prepared by mixing chopped red cabbage with hot water, may be used. For each substance, 20 μl of substance, 160 μl of water, and 20 μl of indicator solution are mixed in a well of a 96-well plate. The color is compared with standards [1 m HCl (red), 1 m acetic acid (orange), distilled water (green), 1 m NaOH (blue/purple)].

Folding Assay—

Phycocyanobilin, a fluorescent blue protein from Spirulina, is used to test the effect of the test substances on protein folding. A crude preparation of phycocyanobilin proteins is prepared as described previously [15] in conjunction with the protein folding laboratory exercise. Briefly, each student pair grinds 1 g of Spirulina (available at health food stores or Sigma S 9134) with 0.1 g of sand for 5 min, then adds 25 ml of 0.04 mg/ml lysozyme (Sigma L 6876) in 100 mm sodium phosphate buffer, pH 7.0, and incubates the samples for 15 min at 37 °C. The mixture is centrifuged at 20,000 × g for 10 min. The supernatant is filtered through cheesecloth, and ammonium sulfate is added to give a 50% saturated solution. The solution was centrifuged at 12,000 × g for 15 min to collect the precipitated phycocyanobilin, and the pellet was dissolved in 2 ml of 100 mm potassium phosphate buffer, pH 7.

For each trial, 20 μl of substance, 170 μl of 100 mm sodium phosphate buffer, pH 7.0, and 10 μl of phycocyanobilin extract are mixed in a well of a 96-well plate. The color is compared with a blank in which water was substituted for substance, and to an unfolded control in which 8 m urea in 100 mm sodium phosphate buffer, pH 7.0, is used instead of buffer. Because the urea solution is an irritant, gloves should be worn. Samples are mixed well, and then allowed to equilibrate for 10 min before reading the absorbance at 625 nm. The red fluorescence of the protein is lost upon unfolding, and the intensity of the blue absorption is decreased.


Students used one of two approaches in designing their screening strategy. In the first, they assumed that the most important factor was finding compounds that inhibited the enzyme, so they began with the enzyme assay. In the second approach, students assumed that the enzyme was precious and attempted to remove false positives by conducting the pH and folding tests first. Both strategies gave the same end results, although starting with the pH test reduced the number of samples for further testing more quickly. One challenge was the selection of the concentration of substrate to use in the enzyme assay. After a class discussion, most of the groups selected an amount close to the Km, as all inhibitor types should show a difference in enzyme velocity at this substrate concentration (Fig. 3).

Fig. 4A shows results from the outreach version of the experiment. Column 2, which contains Lemi-Shine, a dishwashing spot reducer, is an example of a false positive. While the enzyme activity was inhibited (row 2), there was a significant pH change (row 3), indicating that this substance would not make a good drug candidate. In contrast, column 5, which contains an extract made from toothpaste, would be a good drug candidate. Because there is no change in the pH (row 3) or protein folding (row 4), but the enzyme activity is reduced, as indicated by the less intense yellow in row 2, this substance merits further studies to determine if it would be a suitable drug.

An important issue in analyzing a library is how to record the results. In the outreach version of the experiment, students are given a template of the 24-well plate on which to record their results (Fig. 2). In the full version of the experiment, most pairs used an Excel spreadsheet to record the results in each well. A color coding or shading scheme was used to keep track of the results (Fig. 4B).

Each pair then selected three substances to characterize more fully. Students attempted to find the Ki and type of inhibition for these substances. The Ki was obtained from the concentration of that inhibited enzyme activity to 50% of the control activity. Ten- to 10,000-fold dilutions of the substances were necessary. Students were given the key to the library (Table I, sorted by code numbers) and looked up the ingredients for these three items on the labels or on the internet and speculated on what could have caused the inhibition. An extension of this laboratory is to have students test the individual ingredients of a substance to determine which component(s) are responsible for the inhibition. This is what medicinal chemists do when they screen natural products for a desired biological property.

Several substances in the library significantly inhibited wheat germ acid phosphatase. Different toothpastes were included in the library, because phosphate and fluoride are common ingredients known to inhibit phosphatases. Compounds containing benzoic acid, a common preservative, also inhibit the enzyme, because the carboxylic acid moiety mimics the charge of phosphate group. This is a real drug candidate, because a benzoic acid derivative has been optimized as a tyrosine phosphatase inhibitor [1113]. The most interesting finding was that an aqueous extract of nutmeg inhibited the phosphatase. The human pharmacological effects of nutmeg are mostly due to myristicin (methoxy-safrole 4%) [16], but the exact species responsible for the inhibition by the aqueous extract has not been identified.

Student response to this laboratory was positive. All student pairs completed the experiment in a three-hour period. They indicated that the laboratory helped them understand the drug development process and how to design an experiment with a large number of variables (Table III). While some would have preferred a step-by-step procedure, others recognized the benefits, as expressed in the following student comment:

For the last three years, I have followed the lab procedure, and there were some days where I hardly remember what I did just that I followed the directions and got something at the end. This lab I will remember because I had to put time into planning as well as a sense of satisfaction that is normally absent when I complete the lab write-up.

Students commented that using procedures with which they had prior experience reinforced the concepts of pH and protein folding, and that the experiment demonstrated the importance of good record-keeping.

This laboratory exercise teaches high-throughput screening techniques while reinforcing basic principles of biochemistry, such as enzyme inhibition, protein folding, and the importance of pH. The use of household substances and a commonly used enzyme increases the relevance and ease of implementation.

Figure FIGURE 1..

Student handout for experiment.

Figure FIGURE 2..

Outreach experiment handout.

Figure FIGURE 3..

Selecting a substrate concentration. A hypothetical graph of velocity as a function of substrate concentration in the presence of no inhibitor, a competitive inhibitor, and an uncompetitive inhibitor shows that selecting a concentration close to the Km (dark line, 0.3 mm) will result in significantly different observed velocities in the presence of both competitive and uncompetitive inhibitors.

Figure FIGURE 4..

Typical student results.A, typical visual plate results from the outreach version of the experiment. Column 1, water; 2, Lemi-shine (dishwasher supplement) (L30); 3, loratadine (L15); 4, nutmeg (N21); 5, toothpaste (C2); 6, almond extract (A10). Row 1, water; 2, enzyme test; 3, pH with red cabbage indicator; 4, protein folding. B, typical student recording grid from a group that began with the pH test. The number at the top of the column is the column number. The numbers in the boxes are the library identification number, followed by the absorbance. The pH test is colored to match the color of the samples: purple, strong base; blue, weak base; green and yellow, neutral; orange, weak base; red, strong acid. Items with lines through them were omitted from further testing.

Table Table I. Components of the household substance library
Type of compoundExamples
Assigned code numbers are indicated in parentheses following the item.
Cleaning productsWindow cleaner (W1)
 Wood oil soap (M16)
 Borax (B20)
 Dishwasher detergent (C27)
 Oxygen bleach (O12)
 Water softener (C28)
 Lemi-shine (L30)
Over-the-counter medicationsTavist (T5)
 Loratadine (L15)
 Bacitracin (B7)
 Immodium AD (123)
SpicesPeppermint extract (P4)
 Vanilla extract (V19)
 Imitation almond flavor (A10)
 Chili powder (C6)
 Ginger (G18)
 Dry mustard (M9)
 Nutmeg (N21)
 Garlic powder (G24)
 Cinnamon (C26)
 Curry powder (C29)
Personal care productsRegular toothpaste (C2)
 Baby toothcleaner (O8)
 Total toothpaste (C14)
 Whitening toothpaste (A22)
 Deep heating rub (M11)
 Mouthwash (M25)
 Hairspray (H13)
Table Table II. Dispensing guide for library of household substances
Table Table III. Student survey responses
QuestionStrongly agree (%)Agree (%)Neutral (%)Disagree (%)Strongly disagree (%)
Students completed and anonymous survey 2 wk after completing the screening laboratory. Fourteen of 17 students completed the survey.
This experiment helped me understand how drug companies develop new drugs35.75014.300
This experiment helped me think about designing an experiment with a large number of variables35.735.728.5700
This exercise helped me better understand the factors that go into designing a drug28.575014.37.10
Future Che 361 students should do this experiment57.135.77.100


—Thanks to Zack Kaur, Tyler Simpson, Galen Collins, Chemistry 361, and Wabash Middle School students for providing data and comments and to Tom Goyne for critical reading of the manuscript.