Using Molecular Models To Assess Agonists and ... - ACS Publications


Using Molecular Models To Assess Agonists and...

0 downloads 69 Views 2MB Size

Activity Cite This: J. Chem. Educ. XXXX, XXX, XXX−XXX

pubs.acs.org/jchemeduc

Using Molecular Models To Assess Agonists and Antagonists for CellSurface Receptors Daniel D. Schwert* and Scott M. Gruenbaum Department of Natural and Health Sciences, University of St. Francis, 500 Wilcox Street, Joliet, Illinois 60435, United States S Supporting Information *

ABSTRACT: A series of tabletop models was built to demonstrate the interactions between a fictional cell-surface receptor, the receptor’s natural ligand, and eight potential therapeutic agents. Students assessed the eight therapeutic agents as agonists, antagonists, or neither agonists nor antagonists, first as a pencil-and-paper exercise and then subsequently using the three-dimensional models. Students found the activity to be engaging and were more accurately able to identify the correct roles of the agents when using the models. KEYWORDS: First-Year Undergraduate/General, Upper-Division Undergraduate, Biochemistry, Hands-On Learning/Manipulatives, Drugs/Pharmaceuticals, Molecular Recognition, Medicinal Chemistry, Receptors



BACKGROUND

Recognizing the important structural features of a natural ligand and its corresponding binding site is an important step in the development of agonists and antagonists for pharmaceutical use.1 As both the ligand and the receptor binding site are threedimensional structures capable of motion, studying ligand binding using static structures drawn on paper is limited at best. Conveying the details of three-dimensional structures using two-dimensional delivery methods (e.g., paper, computer screens) is a challenge faced by instructors in numerous fields.6 To combat this issue, instructors often to turn to models, such as anatomical models in biology and topographical models in geography courses. Organic chemists often utilize molecular model kits to demonstrate the shapes of compounds. Researchers have hypothesized that such models increase intelligence as a result of sensory−motor activity.7 A student’s ability to understand the three-dimensional structures of molecules is important because, in chemistry, structure determines function. Structure is especially important when studying molecular interactions, as both steric and electronic effects can determine how or even if two compounds will associate. For example, a recent review of the binding interactions between the active site of phosphodiesterase (PDE), PDE substrates, and PDE inhibitors showed that a variety of noncovalent interactions are possible.8 Extending the use of model kits to biomolecules becomes problematic due to the sizes of many biomolecules and the costs associated with obtaining the needed model pieces. Potential solutions include modeling only portions of proteins9 or having students pool their model kits to build larger structures.10 Another way to work around the cost/size problem is by using inexpensive alternatives as stand-ins for plastic models. For example, paper coils can demonstrate the

Cell surfaces contain protein receptors that allow cells to sense and respond to changes in the environment.1 Ligands bind to these receptors, triggering conformational changes. These changes in receptor shape lead to signal transduction events: communication of a message from outside to inside the cell. Our sense of taste and smell, as well as some hormonal regulation, occurs through the mechanism of such receptors.2−4 An agonist is a ligand with a structure similar to the natural ligand that binds to the receptor and triggers the same response as the natural ligand.1 Agonists are used as drugs when the receptor is not being triggered enough under “normal” conditions. An antagonist is a ligand that binds to the receptor but does not trigger a response. Antagonists, sometimes called blocking agents, are used as drugs when the receptor is being overstimulated. An agonist and antagonist for the adrenergic receptors5 are shown in Figure 1.

Figure 1. (A) Epinephrine (adrenaline), the natural ligand for the adrenergic receptors. (B) Isoproterenol, an agonist for the adrenergic receptors, is used to treat bronchospasms. (C) Propranolol, an antagonist, is used to treat hypertension, among other disorders.5 © XXXX American Chemical Society and Division of Chemical Education, Inc.

Received: December 21, 2017 Revised: May 1, 2018

A

DOI: 10.1021/acs.jchemed.7b00980 J. Chem. Educ. XXXX, XXX, XXX−XXX

Journal of Chemical Education

Activity

common secondary structures of protein molecules,11 and aluminum foil can demonstrate the contours of a choline transporter.12 While these models can be effective, they are limited in that they do not easily illustrate the concept of induced fit and/or conformational selection. (While evidence supports the theory of conformational selection,13 a survey of recent textbooks indicates that induced fit is the term in greater current usage.) Induced fit occurs when the structure(s) of one or both of the interacting compounds changes to maximize intermolecular noncovalent interactions.1 While both the paper and foil models discussed above have inherent flexibility, over time repeated flexing may destroy the models. Alternatively, other groups have recommended the use of digital models to demonstrate interactions between compounds. The use of touch-screen devices in conjunction with plastic models to explore molecular geometries has been investigated.14 It was found that the ability to rotate structures on-screen was a more effective learning tool than having the students study the molecular structures on paper. Another study found that using software to illustrate the development of drugs was effective, though time-intensive.15 Finally, digital models were combined with touch-based feedback (haptics) to allow students to explore the forces and stochastic nature of ligand−protein interactions.16 Regardless of whether the models are digital or hand-held, the pedagogical value of models is well-established.17 In this paper, a tabletop, manipulable model is outlined that demonstrates interactions between a fictional cell-surface receptor, its natural ligand, and potential therapeutic agents. The principal pedagogical goal of this activity was to determine whether students using these models can more accurately identify the role of therapeutic agents (agonist, antagonist, or neither) versus solely viewing the agents on paper. We were also interested to see if the models increased student engagement and interest in the activity.



Figure 2. (A) Model of the cell-surface receptor. (B) The binding site undergoes a conformational change upon binding of the natural ligand.

THE MODELS

A model of a fictional cell-surface receptor was constructed using wood and a molecular model kit. An image of the completed model is shown in Figure 2, and construction details are provided in the Supporting Information, including alternative models built from foam board or corrugated plastic. The model is hinged, allowing for a conformational change upon binding of the natural ligand. Models of the natural ligand (Figure 3, structure N) and eight potential agonists/antagonists (Figure 3, structures A−H) were built using molecular model kits. Noncovalent interactions between the receptor and ligands were simulated using magnets and Velcro. Weaker magnets represented hydrogen bonds, with the poles of the magnets oriented so a hydrogen bond donor interacts with a hydrogen bond acceptor but is repelled by another donor. Similarly, stronger magnets were used to represent ionic groups. London force interactions between nonpolar regions were represented using Velcro.

Figure 3. Natural ligand (N) and potential therapeutic agents (A−H).

prerequisite of organic chemistry), the activity was spread over two 50 min class periods. Initially, pairs of students were given a handout (Supporting Information) describing the binding site of the receptor and diagrams showing how the natural ligand interacts with the binding site and causes a conformational change upon binding (see Figure 4 for details). In this activity, in order for the signal to be transduced, all three noncovalent interactions (ion−ion interaction, hydrogen bond, and London forces) between the ligand and the receptor must occur, and the receptor must undergo a complete 90° conformational change. The students then had 30 min to assess the eight potential therapeutic agents by comparing their line structures to the natural ligand. Students categorized each agent as “An agonist”, “An antagonist”, or “Neither an agonist nor an antagonist” and then explained their observations. The following class period, each pair of students was given a receptor model and molecular model kit parts. The students built agents N and A−H and then used the models to categorize the agents and explain their observations. The



THE ACTIVITY The cell-surface receptor activity was performed in two separate courses: a medicinal chemistry course and an introductory-level general−organic−biochemistry (GOB) course. In the medicinal chemistry course (upper-division undergraduate science majors, B

DOI: 10.1021/acs.jchemed.7b00980 J. Chem. Educ. XXXX, XXX, XXX−XXX

Journal of Chemical Education

Activity

Table 1. Therapeutic Agent Role Assignments by Students in an Upper-Level Medicinal Chemistry Coursea Students Selecting a Role Using Line Structures

Students Selecting a Role Using Molecular Models

Agentb

Agonist

Antagonist

Neither

Agonist

Antagonist

Neither

A B C D E F G H

14c 12c 9c 0 12 0c 0 0

2 2 7 14c 2c 14 0 8c

0 2 0 2 2 2 8c 0

12c 10c 14c 0 2 5c 0 0

2 0 2 14c 7c 9 0 6c

2 6 0 2 7 2 8c 2

a

All students performed the line structure exercises one class period prior to using the molecular models. bAgents A−F were evaluated by 16 students; agents G and H were evaluated by 8 students. Most students worked in pairs. cCorrect roles.

Table 2. Therapeutic Agent Role Assignments by Students in an Introductory-Level GOB Coursea Students Selecting a Role Using Line Structures

Students Selecting a Role Using Molecular Models

Agentb

Agonist

Antagonist

Neither

Agonist

Antagonist

Neither

A B C D E F G H

21c 5c 15c 11 5 17c 0 6

4 18 6 6c 18c 8 3 15c

0 2 4 8 2 0 22c 4

23c 11c 23c 2 6 16c 0 2

2 12 2 21c 17c 9 0 16c

0 2 0 2 2 0 23c 7

a

All students performed the line structure exercises prior to using the molecular models. bMost students worked in pairs; one pair of students left blanks for agent G, resulting in only 23 responses. c Correct roles.

Figure 4. Binding site of the cell-surface receptor shown from the side (A) and end (B) views. (C) The natural ligand in the binding site (this representation corresponds to Figure 2A). (D) Upon binding, the receptor undergoes a conformational change driven by noncovalent interactions between the ligand and the binding site (this representation corresponds to Figure 2B).

“neither” for therapeutic agents A−H after performing the pencil-and-paper exercise (“Line Structures” columns) and then using the tabletop models (“Molecular Models”). As indicated by footnote c, agents A, B, C, and F are agonists for this model receptor, agents D, E, and H are antagonists, and agent G is neither. Table 2 similarly includes student data for the introductory-level course. In Figure 5, the percentage of students who correctly identified each agent using line drawings (gray bars) and the three-dimensional models (black bars) is shown for both the upper-level (top) and introductory-level course (bottom). For most agents, the black bar is longer than or similar to the gray bar. In all 16 cases, the use of the models either increased or did not substantially change (within two students or one student pair) the number of students who correctly identified the role of the agent. Averaging both courses together, students correctly identified the role of the agent 60% of the time using only line drawings and 73% of the time when using the models. For the upper-level (introductory-level) course, these percentages are 60% (60%) for the line drawings and 68% (76%) using the models. Therefore, our data suggest that, on average, the models better enabled students to visualize and “fit” the ligands into the receptor binding site. Several agents proved to be generally problematic. For example, many students in both courses incorrectly identified agent F even when using the model, and students in the

original handouts were returned, and the students were given 15 min to compare the results of their two assessments. In the GOB course (mostly first-year undergraduate prenursing majors, no prerequisite), the activity proceeded as in the medicinal chemistry course. However, both the line drawing and model components were performed over the course of 60 min rather than two class periods. Given the introductory nature of this course, the instructor also spent significant time discussing protein structure, cell-surface receptors, and ligand binding prior to the start of the activity. Discussions after the activity focused primarily on the role of intermolecular forces in binding rather than the more complicated biochemistry of protein receptors.



RESULTS This activity was tested in two separate courses: an upper-level medicinal chemistry course (2 sections, 16 students total), and an introductory-level GOB course (3 sections, 25 students total). In all cases, students responded positively to the models and remained engaged throughout the activity. The results of the student responses are summarized in Tables 1 and 2. In Table 1, each entry indicates the number of students in the upper-level course who selected “agonist”, “antagonist”, or C

DOI: 10.1021/acs.jchemed.7b00980 J. Chem. Educ. XXXX, XXX, XXX−XXX

Journal of Chemical Education

Activity

example, several groups were observed to bind a ligand to the model receptor, make all three connections, observe a 90° angle, and then mark “neither” or “antagonist”. While these students were not corrected (as we did not want to affect the data), this is an error that would be relatively easy to fix through instructor intervention during the activity. The other major error was not exploring the full conformational space of each ligand. Several ligands (e.g., agent F) required the students to rotate a number of bonds in order to make all three noncovalent connections. Some students never tried to “wiggle” the molecule and just decided it did not fit. Again, this is something that could be corrected during the activity or discussed afterward as a group. Several of the ligands also present opportunities for postactivity discussion. For example, agents A and N are enantiomers. In many real receptors, enantiomers do not bind with the same affinity as the natural ligand. In this activity, the all-or-nothing nature of the binding interactions demonstrates a limitation of the model. Other such limitations (e.g., mechanical force is needed for the conformation change, agonists/antagonists may work by a variety of mechanisms) may also be discussed after the activity.



FUTURE IMPROVEMENTS As a result of student feedback and instructor observations, there are several improvements that could be considered. Additional binding groups or steric blocks can be secured to the model. It may also be instructive for students to propose their own ligands. To do this, they could design their therapeutic agents on paper; build each agent using a model kit, magnets, and Velcro; and then assess their agents in the model receptor. For example, they could see what differences might result from replacing the hydroxyl group with an amino group (no change expected) or a methyl group (no binding results). Alternatively, a model of the binding site of a real receptor (or the active site of an enzyme) could be constructed using the models presented in this paper as a template. Modeling an actual receptor would add a layer of realism to the activity, as the students could assess therapeutic agents currently in development and/or clinical use. The only changes that would need to be made to the model are the addition or subtraction of various binding groups, magnets, and patches of Velcro. In summary, the development of a hands-on, generalizable model of a cell-surface receptor was presented. This model was found to be an effective and engaging way to teach students about ligand binding and noncovalent interactions that can be used in both introductory and upper-level undergraduate courses.

Figure 5. Graphical depiction of the percentage of students who correctly identified the role of each potential agent using line drawing (gray bars) and the three-dimensional models (black bars). The results from the upper-level medicinal chemistry course are shown at the top, and the results from the introductory-level GOB chemistry course are shown at the bottom.

introductory course had trouble with agent B. In the introductory-level course, it also at first appears as if the model did not significantly change the number of students who correctly identified agent E (18 vs 17 students correct). However, a closer look at the data reveals that actually 8 students changed their answer from correct to incorrect, while 7 changed from incorrect to correct. This large degree of uncertainty suggests that there was something about certain agents that confused students. In observing the students, three primary causes of error/ confusion were noted. First, some of the agents are borderline agonists or antagonists. For example, agent E is meant to be an antagonist, as it is just slightly too large to make all three noncovalent connections due to the nonplanar cyclohexane ring. However, if the model measurements are a bit off, it is possible for agent E to be (barely) an agonist. It is thus important for instructors to carefully test each model prior to class to ensure that agents actually fulfill their intended role. Small adjustments (e.g., moving the Velcro or bending the 90° angle bracket) can be made to improve the fit of the ligands. For students, one common error was not paying careful attention to the definition of “agonist” or “antagonist”. For



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available on the ACS Publications website at DOI: 10.1021/acs.jchemed.7b00980. Building instructions (PDF, DOCX) Line drawing activity for medicinal chemistry (PDF, DOCX) Model activity for medicinal chemistry (PDF, DOCX) Comparison activity for medicinal chemistry (PDF, DOCX) Activity worksheet for GOB course (PDF, DOCX) D

DOI: 10.1021/acs.jchemed.7b00980 J. Chem. Educ. XXXX, XXX, XXX−XXX

Journal of Chemical Education



Activity

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Daniel D. Schwert: 0000-0001-8110-0959 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS D.D.S. thanks Valerie O’Loughlin for the introduction to the Scholarship of Teaching and Learning. The assistance of Susan Renner and David Hatz was invaluable in the construction of the models (and ensuring that we finished with as many attached fingers as when we started).



REFERENCES

(1) Patrick, G. L. An Introduction to Medicinal Chemistry, 5th ed.; Oxford University Press: Oxford, UK, 2013; pp 42−46, 106−111, 215−244. (2) Neer, E. J. Heterotrimeric G proteins: organizers of transmembrane signals. Cell 1995, 80, 249−257. (3) Schlessinger, J. Cell signaling by receptor tyrosine kinases. Cell 2000, 103, 211−225. (4) Rang, H. P. The receptor concept: pharmacology’s big idea. Br. J. Pharmacol. 2006, 147, S9−S16. (5) Nursing 2016 Drug Handbook; Comerford, K. C., Ed.; Wolter Kluwer: Philadelphia, 2016; pp 1207−1209 and 1614−1615. (6) Harle, M.; Towns, M. A review of spatial ability literature, its connection to chemistry, and implications for instruction. J. Chem. Educ. 2011, 88, 351−360. (7) Smith, L. B. Cognition as a dynamic system: principles from embodiment. Dev. Rev. 2005, 25, 278−298. (8) Jansen, C.; Kooistra, A. J.; Kanev, G. K.; Leurs, R.; de Esch, I. J. P.; de Graaf, C. PDEStrIAn: a phosphodiesterase structure and ligand interaction annotated database as a tool for structure-based drug design. J. Med. Chem. 2016, 59, 7029−7065. (9) Halkides, C. J. Using molecular models to show steric clash in peptides: an illustration of two disallowed regions in the Ramachandran diagram. J. Chem. Educ. 2013, 90, 760−762. (10) Hageman, J. H. Use of molecular models for active learning in biochemistry lecture courses. J. Chem. Educ. 2010, 87, 291−293. (11) Pikaart, M. The turn of the screw: an exercise in protein secondary structure. Biochem. Mol. Biol. Educ. 2011, 39, 221−225. (12) Geldenhuys, W. J.; Hayes, M.; Van der Schyf, C. J.; Allen, D. D.; Malan, S. F. Receptor surface models in the classroom: introducing molecular modeling to students in a 3-D world. J. Chem. Educ. 2007, 84, 979−982. (13) Changeux, J.-P.; Edelstein, S. Conformational selection or induced fit? 50 years of debate resolved. F1000 Reports 2011, 3. DOI: 10.3410/B3-19. (14) McCollum, B. M.; Regier, L.; Leong, J.; Simpson, S.; Sterner, S. The effects of using touch-screen devices on students’ molecular visualization and representational competence skills. J. Chem. Educ. 2014, 91, 1810−1817. (15) Rodrigues, R. P.; Andrade, S. F.; Mantoani, S. P.; Eifler-Lima, V. L.; Silva, V. B.; Kawano, D. F. Using free computational resources to illustrate the drug design process in an undergraduate medicinal chemistry course. J. Chem. Educ. 2015, 92, 827−835. (16) Bivall, P.; Ainsworth, S.; Tibell, L. A. E. Do haptic representations help complex molecular learning? Sci. Educ. 2011, 95, 700−719. (17) Stull, A. T.; Gainer, M.; Padalkar, S.; Hegarty, M. Promoting representational competence with molecular models in organic chemistry. J. Chem. Educ. 2016, 93, 994−1001.

E

DOI: 10.1021/acs.jchemed.7b00980 J. Chem. Educ. XXXX, XXX, XXX−XXX