Crowding and function reunite

The insides of cells are crowded. Fig. 1A is a dramatic representation of the fact that macromolecules occupy 30% of the cellular volume, with their concentrations reaching 300 g/L (1). For comparison, the concentration of protein in an egg white is only one third of that value. However, almost everything we know about biological macromolecules in solution comes from data acquired under conditions in which the concentration of macromolecules rarely exceeds 10 g/L. In PNAS, Dhar et al. (2) report the broadest study to date about the effects of macromolecular crowding on the properties of a globular protein.

Fig. 1.

(A) The Escherichia coli cytoplasm, as modeled by McGuffee and Elcock (1). (B) Open form of phosphoglycerate kinase (PDB ID code 1QPG), visualized with visual molecular dynamics (20).

A powerful combination of experiment and theory is brought to bear here because the work is a collaboration between a team of experimentalists, the Gruebele group, and a team of theoreticians, the Cheung group. Their data run the gamut of crowding effects, covering folding kinetics, equilibrium stability, structure, and function. Before describing the system and the results, we consider what is expected when a globular protein in solution is crowded by other macromolecules.

An elevator ride provides a simple analogy. If you are alone, it is easy to raise your wrist to eye level to check your watch. If the elevator is crowded, you are more likely to keep your hands to your sides. In terms of equilibrium, this is just a restatement of Le Chatelier’s principle: crowding favors species that require the least space.

In the late 1950s, Ogston (3) suggested that volume exclusion by large molecules would affect equilibria. Three years later, Ogston and Phelps (4) proved it. Minton and Wilf (5) put the field on thermodynamic terra firma in terms of theory and experiment, and they coined the phrase “macromolecular crowding” in 1981. The most recent crowding paper to evoke intense interest among biochemists is from the groups of Wittung-Stafshede and Cheung, showing that crowding can deform the native state of a distinctly nonspherical globular protein, because a nonnative state takes up less space than its native state (6). A thorough and thought-provoking review by Elcock (7) of crowding research showcases the current state of the field.

Despite early interest (5), more effort has been focused on crowding and protein stability than on crowding and protein function. Dhar et al. (2) bring function to the forefront again with their study of the enzyme phosphoglycerate kinase, which catalyzes the initial ATP-generating step of glycolysis. Most importantly, the enzyme is stunningly bilobal (Fig. 1B), which makes its open and unbound state distinctly noncompact. One substrate, ADP, binds an inner face, whereas the other substrate, 1,3-bisphosphoglycerate, binds the opposing inner face. Bringing the lobes together is thought to be involved in catalysis (8, 9). As such, crowding would be predicted to increase the enzyme’s activity.

The authors tested this hypothesis by crowding a solution containing phosphoglycerate kinase with the highly soluble, uncharged, 70-kDa synthetic polymer Ficoll. The polymer is made by cross-linking sucrose molecules and is most well known for its ability to separate the components of blood serum. To detect changes in distance between the lobes, the Gruebele group covalently attached a different fluorescent label to each lobe and monitored Ficoll’s effect on the fluorescence [i.e., fluorescence resonance energy transfer (FRET)]. Cheung’s group focused on the computer simulations that provide molecular-level interpretations of the data. The simulations start with the crystal structure of the enzyme (PDB ID code 1QBG) and modeled Ficoll as a hard sphere with a radius of 5.5 nm.

First, the effect of Ficoll on protein stability was measured by assessing the melting temperature. This task was accomplished by monitoring the FRET signal while raising the temperature. Like many proteins, this one denatures cooperatively, and the authors find that the midpoint of the transition increases with Ficoll concentration. It is well known that denatured states are less compact than native states; thus, the knee-jerk conclusion is that macromolecular crowding stabilizes the protein because the native state takes up less space.

Not simply resting at this result, Dhar et al. (2) performed an important control to determine whether the macromolecular nature of Ficoll causes the stability increase. The fundamental component of Ficoll, sucrose, was used as a comparable small molecule crowder. The result was revealing; sucrose alone increases the melting temperature. Thus, the stabilization afforded by Ficoll is from favorable interactions between the polymer and the native protein (10) or a solvophobic effect (11) rather than a result of macromolecular crowding.

Crowding effects on folding rates were examined by jumping the temperature and looking at relaxation of the FRET response. The interesting observation is that a minimum at 100 g/L is observed in a plot of the time constant for relaxation vs. Ficoll concentration. Results from simulations provide an explanation. The relaxation rate increase from 0 to 100 g/L is interpreted as decompaction of the denatured state ensemble. This interpretation stands in contrast to work recently described by the Gierasch group (12), suggesting a decrease in folding rates from crowding-induced collapse of the denatured state. The decreasing relaxation rate above 100 g/L Ficoll observed by Dhar et al. (2) is explained by the general increase in viscosity. Both the Gierasch group (12) and the Fersht group (13) report similar results. The sucrose control supports the conclusion that increasing the viscosity slows folding and proves that it is the macromolecular nature of Ficoll that causes the increased rate below 100 g/L.

The centerpiece of the work addresses crowding-induced conformation changes and their relationship to enzyme activity. The results are simple to state; the more that volume is excluded by Ficoll, the closer the lobes approach each other. Sucrose had almost no effect, again showing that compaction arises from the polymeric nature of Ficoll. In terms of function, 200 g/L Ficoll has a large effect on the enzymatic activity, increasing it by an order of magnitude. Correcting for the increased viscosity rockets this value to nearly two orders of magnitude, but this correction is not relevant if we want to make a comparison with what happens in cells, because the cytosol is viscous. Sucrose, conversely, reduces the activity, showing that the effect comes from macromolecular crowding.

Simulations confirm these results and identify three forms of the enzyme at normal temperatures. The first is an open form, like the one in Fig. 1B. The second form is a closed state with the lobes pressed together. The third form is more spherical and arises from a twist on compaction. It is unclear whether the closed form or the spherical form is the one with more activity, but it is more likely to be the closed form because of the active-site distortions observed in its spherical relative. The main conclusion arising from this clever combination of wet biochemistry and computer simulation is that crowding increases the enzyme’s activity by bringing together the two halves of the active site.

Dhar et al. bring function to the forefront again with their study of the enzyme phosphoglycerate kinase.

Three potential caveats should be borne in mind for all crowding studies using uncharged synthetic polymeric crowders such as Ficoll. First, such molecules are not completely impenetrable, nor are they perfect spheres (14, 15). Second, the properties of their solutions change with concentration (16). At a low concentration, these crowders act as individual molecules. As their concentrations increase, the single molecules entangle, like spaghetti in a bowl. Whether the crowders act as single molecules or as a collection of tangled ones should make a difference to the protein being studied. Importantly, the concentration at which this change occurs is often within the range of crowder concentrations used in studies of crowding effects.

Finally, and most importantly from the standpoint of biological relevance, cells are crowded by proteins (and nucleic acids) rather than by synthetic polymers. Unlike synthetic polymers, globular proteins can be treated more like spheres. Furthermore, proteins have charged surfaces that are studded with the same hydrogen bond donors and acceptors that stabilize the protein under investigation. Returning to the elevator analogy, imagine now that you and your fellow riders are adorned with Velcro, which would make it difficult to lower your wrist should you get it raised. These weak nonspecific interactions have been identified in crowded protein solutions (17). In summary, there is a lot to be learned from using synthetic crowders, but we must recognize that their solutions are not the same as intracellular conditions.

The most physiologically relevant way to study the effects of crowding on proteins is to study them in cells. Such endeavors have been called postreductionist protein science (18). There remains, however, a need for what might be called neoreductionist approaches like those of Dhar et al. (2). Such studies will serve as a bridge, at least until synthetic biology (19) matures and makes it possible to manipulate, and thereby disentangle, all the effects within the intracellular eye candy shown in Fig. 1A.