This web page was produced as an assignment for an undergraduate course at Davidson College.

Single Molecule Protein Interactions

What do proteins do?

Proteins are central to the fundamental chemistry of life. Cells are a buzzing milieu of diverse proteins performing a wide range of functions. Proteins cooperate to execute many complex tasks, such as transmitting information along signaling pathways and assembling and disassembling molecules during metabolism.

Where is the protein located?

Our current understanding of fundamental cellular processes has emerged from a combination of methods, including direct observation and indirect experiments. Direct observation is appealing, because it enables researchers to see how proteins fit within the spatial context of the cell. Viewing the locations of proteins in a live cell often relies on the small protein GFP, which remarkably fluoresces green without any need for added substrates. By genetically fusing GFP to a protein of interest, we can identify the subcellular localization of the protein using fluorescence microscopy. Portions of the cell containing the protein fluoresce green.

In most applications, the resolution available to conventional fluorescence microscopes is insufficient to track single proteins (reviewed in Huang et al., 2010). Inevitable slight imprecisions in focus are problematic at high magnification. Despite the limitation, GFP tags provide excellent information about the general location of proteins within the cell.

What is the protein’s structure?

We can, however, resolve the structure of a protein in atomic detail using X-ray crystallography. Growing protein crystals is challenging, and success requires good fortune as much as it does technical expertise, but the results are stunning. The structure of GFP (see the image) reveals a barrel shape made of beta sheets, called a “beta barrel.” Inside the beta barrel is the chromophore, responsible for the fluorescence that makes GFP useful to researchers.

Crystal structures can both test and generate hypotheses about the activities performed by the many proteins in a cell. Additional experiments are required to interrogate proteins in their native physiological contexts, such as how often another protein is bound and what consequences result from the binding event.

Green fluorescent protein, GFP. The chromophore, responsible for the fluorescence mechanism of GFP, is colored green and not to scale. Coordinate data obtained from PDB ID 1EMA and rendered with Jmol.

What interacts with the protein?

Proteins do not act in isolation. Interactions between proteins are key to many biological processes. Immunoprecipitation and Western blotting have been used to map out many biological pathways where protein-protein interactions serve an important role. Western blots use an antibody against a protein of interest to test for its presence in a sample, and immunoprecipitation generates samples containing the binding partners of a protein. Together, they can tell us if two proteins interact in a particular organism, tissue, or physiological state.

But, ultimately, Western blots are averages. Retrieving enough protein to see a band in the blot requires pooling the lysates from many cells. Through this averaging process, Western blots hide the biological variation that exists from cell to cell, and from protein to protein within each cell. The nuance between the two questions, 'Do proteins interact?' and 'How do proteins interact?' is not trivial. Biological systems are dynamic, complex, and sometimes chaotic. A mechanistic and quantitative understanding requires precision in details.

How do proteins perform their functions?

Fluorescence microscopy, X-ray crystallography, and immunoprecipitation can provide fundamental information about what a protein does. However, studying proteins in bulk can hide insights that arise from appreciating the variation that exists or examining cells at a very small scale. In contrast, through examining proteins individually, single molecule experiments can reveal unexpected behaviors and the diversity that exists in cells.

How does an enzyme interact with its substrate?

Cholesterol oxidase, as the name suggests, oxidizes cholesterol (Lu et al., 1998). The net reaction is loss of electrons by cholesterol and gain of electrons by oxygen, with hydrogen balancing the charges. The reaction mechanism requires that the active site of cholesterol oxidase bind the coenzyme FAD. Electrons are transferred from cholesterol to FAD in the first step of the reaction, and subsequently passed along to oxygen.

Cholesterol oxidase. The active site of cholesterol oxidase binds the coenzyme FAD (highlighted in light blue) to catalyze the oxidation of cholesterol. During the reaction, FAD accepts electrons from cholesterol prior to transfer to oxygen. In its electron-bound state, FAD cannot fluoresce. Coordinate data obtained from PDB ID 1B4V and rendered with Jmol.

FAD is fluorescent, but in an electron-bound state its ability to fluoresce is suppressed. Therefore when bound to cholesterol oxidase you would expect each FAD molecule to transition back and forth between fluorescent and non-fluorescent states, depending on the progress of the reaction. In other works, FAD should blink on and off: on by default, but switched off in the intermediate reaction state where electrons have been taken from cholesterol but have not yet been handed off to oxygen.

Flashes of FAD fluorescence are precisely what Lu et al. found. In their experiment, the authors used purified cholesterol oxidase embedded in a gel containing the substrate cholesterol. The surface of the gel was illuminated with light to excite the FAD molecules, and a camera detected any light emitted. 

By combining data obtained from observing many individual enzymes in the gel matrix, they found that the reaction followed unexpected kinetics. The rates of FAD flashes did not fit the original prediction. By generating a mathematical model and testing the model against their experimental data and against computer simulations, they concluded that the best explanation for the data was that reaction rate depended on which of two major conformations the enzyme exhibited at a given time. In other words, the efficiency of the enzyme depended on its ability to change conformations between individual reactions. Once the reaction was complete, the protein was not necessarily ready to accept more cholesterol. A conformation change must occur first, a step ignored by conventional models of enzyme kinetics. Proteins are more flexible than static images tempt us to believe, and enzyme efficiency depends on the individual history of each molecule.

Motor proteins: the ultimate single molecule experiments

Experiments like observing cholesterol oxidase can help us understand how single enzymes behave, but no field has benefited more strongly from single molecule experiments than study of the cell biology of motor proteins.

The eukaryotic cytoskeleton not only provides structural rigidity. It is also a dynamic transportation network for organelles, particles of mRNAs bound to protein, and other cellular cargo. The motor proteins dynein and kinesin walk (literally!) along microtubules. Members of the dynein protein family walk in the opposite direction as those of the kinesin family. While the proteins are quite different in structure, they both contain two domains that can bind tubulin, the monomer used to construct microtubules. Alternate binding of the two domains causes the motor protein to walk or slide along the microtubule, dragging attached cargo along with it.

If microtubules are organized like a highway system, we might expect particular motor proteins to travel only along a subset of microtubules. It would be much more efficient for a motor protein destined for the cell membrane to travel along the most direct route possible, rather than detouring halfway to visit the Golgi.

We know that the microtubules in a cell are not all the same; there are subpopulations. But how these differences translate into specialized transport of cargo has not been as clear. Through single molecule experiments, Cai et al. (2009) found that the Kinesin-1 motor localizes to a subset of microtubules (see the figure). In the image, microtubules are labeled with the fluorescent protein mCherry (shown in red), and Kinesin-1 by the fluorescent protein mCitrine (shown in green). Rather than seeing spots of mCitrine-labeled Kinesin-1 scattered across the mCherry-labeled microtubules, they found that the proteins are limited to a relatively small fraction of microtubules. This suggests that Kinesin-1 does not bind to microtubules randomly, but the experiment does not exclude the possibility of a burst of activity, such as many vesicles being trafficked in a similar way at one time.

An additional experiment clarified the ambiguity, showing that the microtubules ‘preferred’ by Kinesin-1 are marked with post-translational modifications, which is associated with stable microtubules. It appears that Kinesin-1 is specialized to the 'tried and true' part of the cytoskeleton, while other motor proteins walk along parts of the cytoskeleton being actively assembled and disassembled.

High resolution microscopy of Kinesin-1 on microtubules. Kinesin-1 proteins are shown in green, and microtubules in red. Kinesin-1 localizes to a subpopulation of microtubules, rather than being broadly distributed across the cell. Scale bar indicates 4 m. Image obtained from Cai et al. (2009), Figure 1G (reproduced under terms of the CC BY license).


The question 'What does this protein do?' is complex and multifaceted. Immunoprecipitation, Western blotting, and X-ray crystallography are central to gathering fundamental information about proteins, but they do not provide a complete story. Proteins are autonomous machines, and experiments that look at proteins in aggregate will inevitably miss crucial parts of the story. Single molecule experiments reveal the underlying variation and unexpected behaviors that would otherwise be lost to an average.


Huang B, Babcock H, Zhuang X. 2010. Breaking the diffraction barrier: super-resolution imaging of cells. Cell 143:1047-1058.

Lu HP, Xun L, Xie XS. 1998. Single-molecule enzymatic dynamics. Science 282:1877-1882.

Cai D, McEwen DP, Martens JR, Meyhofer E, Verhey KJ. 2009. Single molecule imaging reveals differences in microtubule track selection between kinesin motors. PLoS Biol 7:e1000216.

Eric Sawyer's Home Page

Genomics Page
Biology Home Page

Email Questions or Comments.

Copyright 2014 Department of Biology, Davidson College, Davidson, NC 28035