Principal Investigator Robert Sauer
All organisms contain proteolytic machines that consist of an AAA+ ATPase and a compartmentalized peptidase. The AAA+ domains of these multimeric machines form a hexameric ring with a narrow central pore or channel. This hexamer recognizes protein substrates, uses cycles of ATP hydrolysis to unfold these molecules, and translocates the unfolded polypeptide into a sequestered proteolytic chamber for degradation. We study four bacterial AAA+ proteases: ClpXP, HslUV, Lon, and ClpAP. In collaborative studies with Tania Baker’s lab, we are probing the rules of substrate recognition, studying regulatory factors that enhance or inhibit recognition of specific substrates, and dissecting the mechanisms by which these enzymes catalyze protein disassembly and denaturation.
Most if not all substrates of bacterial AAA+ proteases are recognized via short peptide tag sequences, often located at the N- or C-terminus. Some tag sequences bind in the central pore of the AAA+ ring, providing a way for the enzyme to initially grasp the substrate. Other tags tether the substrate to the enzyme by binding to auxiliary domains, either directly or via specialized adaptor proteins. One of our major goals is to identify the peptide sequences that target substrates to different AAA+ proteases, to understand how different types of tags function synergistically, and to determine how these tags mediate highly specific recognition and subsequent protein unfolding/degradation at a detailed structural and mechanistic level.
Unfolding of native proteins by AAA+ ATPases is an active mechanical process. Very stable proteins—those for which spontaneous unfolding can take months or even years—are unfolded in a few seconds by these powerful enzymes. Current evidence suggests that, after binding of the peptide tag, the hexameric enzyme begins to translocate this peptide sequence through a central pore, generating a denaturation force when the attached native protein cannot pass through this narrow aperture. Denaturation of native protein substrates by AAA+ machines depends both on the rate of ATP hydrolysis and on the local stability of the substrate structure immediately adjacent to the degradation tag. By studying the degradation of model substrates varying in stability, we have shown that the ClpXP protease applies an unfolding force iteratively. For some very stable substrates, enzymatic unfolding of a single protein can require hundreds of cycles of ATP hydrolysis. In some cases, the energetic cost of denaturation is so high because substrates undergo multiple rounds of binding, attempted unfolding, and release before denaturation becomes statistically probable. This seemly wasteful mechanism may prevent stalling of the AAA+ motor of ClpXP when a given substrate cannot be readily unfolded and ensure that those substrates in the cell that are easy to unfold are degraded first. Whether similar principles apply to other AAA+ proteases is currently unknown.
Polypeptide translocation, which is needed both for unfolding and for transporting the unfolded chain into the peptidase, is the essential mechanical function of AAA+ proteases. Remarkably, ClpXP and many of its distant cousins can translocate polypeptides from the C-terminus to the N-terminus or in the opposite direction. The step size for ClpXP translocation appears to be roughly four residues per ATP hydrolyzed. Moreover, we found that ClpXP can translocate multiple polypeptides at the same time and can translocate synthetic peptides with d-amino acids, with long stretches of poly-proline, or with as many as ten additional methylene groups between successive peptide bonds. It remains to be determined how the structure of the enzyme allows this remarkable tolerance, whether other AAA+ proteases behave in a similar manner, and whether the translocation process has any chemical specificity or relies on purely physical interactions.
The homohexameric rings of AAA+ proteases contain six potential nucleotide-binding sites. Nevertheless, these enzymes function asymmetrically with different subunits adopting distinct roles during the reaction cycle. For example, no more than four ATP molecules ever bind to the working ClpXP and HslUV enzymes, and states in which all subunits are either ADP bound or nucleotide free are not part of the normal ATPase cycle. By linking wild type and inactive mutant ClpX subunits to form “covalent” hexamers, we have found that ATP hydrolysis in a single subunit can drive the translocation of protein substrates into ClpP. Moreover, ClpX enzymes with just two active subunits unfold native substrates and use ATP as efficiently as the wild-type enzyme to degrade substrates. These and related studies provide evidence for directional communication between ClpX subunits, for additive subunit contributions to overall hexamer activity, and for a probabilistic sequence of ATP hydrolysis in different subunits. We recently solved crystal structures of ClpX hexamers, in different nucleotide states, which shows that asymmetry is generated by distinct rotations between the large and small AAA+ domains. Different rotations in two classes of subunits produce a staggered arrangement of individual large and small domains along the central axis of the hexamer. Deeper understanding will require structures with bound degradation tags or translocating substrates.
Tag recognition is normally viewed as a passive reaction, where only affinity and concentration matter. By contrast, for the Lon protease, we have shown that tags also determine maximal rates of ATP hydrolysis, translocation, unfolding, and proteolysis. Indeed, different tags fused to the same protein change degradation speeds and efficiencies by 10-fold or more. Tag binding to multiple sites in the Lon hexamer differentially stabilizes specific enzyme conformations, including one with high protease and low ATPase activity, and results in positively cooperative degradation. These allosteric mechanisms allow Lon to operate in either a fast or slow proteolysis mode, according to specific physiological needs, and help maximize degradation of misfolded proteins following stress-induced denaturation. Degradation tags also affect the maximum rate of degradation by HslUV by a mechanism that is under active investigation.
For some AAA+ enzymes, the peptidases assemble independently into double-ring multimers (ClpP14, HslV12), which then dock with one or two hexameric ATPases to form the functional ClpXP, ClpAP, and HslUV proteases. For ClpXP, we found that distinct types of static and dynamic interactions are important for binding and for communication between the symmetry mismatched ClpX6 and ClpP7 rings. How these interactions are functionally integrated in the working machine is an active area of investigation.
We have also been developing AAA+ enzymes as tools to study protein function and to analyze macromolecular assemblies. By engineering degradation tags and adaptor proteins, we are creating conditional degradation systems in which a protein bearing a special tag is only degraded when the synthesis of a specific adaptor is induced or when a bipartite adaptor is activated by the presence of a small molecule. In principle, these systems allow specific proteins (even those that are essential) to be removed from the cell in a few minutes and permit studies of the resulting changes in cell physiology on much shorter time scales than are generally possible with other methods. For some macromolecular complexes, certain protein components are required for assembly, and it is not possible to study additional functions of the protein in the complex by deletion of the gene. In a proof-of-principle test case, we have shown that a tagged variant of ribosomal protein L22, which is required for assembly of the 50S subunit, can be pulled out of intact ribosomes by ClpXP without disrupting the integrity of the particle.