Xun Lu

We learned from Biology 101 in college that proteins lose function when they lose their structure. While this is generally true, some proteins, such as heat-shock protein 33 (Hsp33), are actually activated, and thus gain functions, upon losing part of their structures. This “lose-to-gain” trick is intriguing, but why should we care about Hsp33, a bacterial protein? It turns out that Hsp33 is involved in the defense system in bacteria.

When you use bleach, a strong oxidative reagent, to kill bacteria, or when your body fires an oxidative immune response to fight against bacterial infections, you or your body is actually creating an oxidative stressing environment for bacteria. Under such condition, most proteins are unstable and start to unfold, or lose their structures. The unfolding proteins collide with each other, and form solid aggregates that are non-functional. The process of protein aggregation under the oxidative stressing condition is similar to what you would expect to see while boiling an egg (in this case the stress is the elevated temperature). Hsp33 unfolds in the presence of both oxidative stress and heat. However, the unfolding of Hsp33 does not lead it to the tomb, but rather activates this molecular chaperone to bind and protect other proteins from aggregation. In so doing, Hsp33 helps bacteria to survive in the unpleasant environment.

Figure 1. Lose structure to gain function

How does Hsp33 play the “lose (structure)-to-gain (function)” trick, exactly? Under the normal intracellular condition, which is a reducing environment, Hsp33 is inactive and compactly folded. The four strictly conserved cysteine residues of Hsp33 coordinate a zinc ion, forming the zinc redox-center (Figure 1, modified from ref.1). When bacterial cells are subjected to the oxidative stress such as bleach treatment, those four cysteine residues get oxidized and form two pairs of disulfide bonds. This change triggers the release of the zinc ion and a substantial structural rearrangement of the zinc redox-center and the linker region, both of which become less structured. The oxidized Hsp33 form dimers that are fully activated.

Since the activated Hsp33 binds unfolding proteins, and itself is partially unfolded, you may wonder why Hsp33 does not bind to itself (you guessed correctly that Hsp33 does not bind to itself, otherwise what is the point of having Hsp33)? It turns out that Hsp33 preferentially binds early unfolding proteins that are still structured, but not those that are fully unfolded or natively unstructured, as revealed in a recent Cell paper published by Dr. Ursula Jakob’s group (2). They also found that the linker region (green in Figure 1), which undergoes dramatic structural rearrangement upon oxidative activation, plays a direct role in the substrate binding. How the linker region is involved remains to be elucidated. One clue may come from an X-ray crystal structure of an active Hsp33 dimer (pdbid: 1I7F). In the structure, the linker regions from each Hsp33 subunit reaches to and interacts with the partner, forming a domain crossover (3). Perhaps the linker region in the activated Hsp33 is not unfolded (as modeled in Figure 1) but rather flips out from the top of the N-terminal domain.

Once the activated Hsp33 dimer binds its client protein, it holds its client until the non-stressing condition resumes. Then, Hsp33 passes its client to DnaK/J, another class of molecular chaperones that is responsible for assisting the refolding of the client protein. Interestingly, the client protein bound by Hsp33 is not good substrate for DnaK/J, which prefers to bind less structured proteins. As you might imagine, nature must have solved this puzzle. It turns out that the client protein becomes more disordered before it gets released by Hsp33 (ref.2). It is likely that the refolding of Hsp33 under the normal/reducing condition couples the further unfolding of the client protein so that the client is ready to be picked up by DnaK/J.

To end this story, the “lose-to-gain” trick may be played by other proteins as well. The zinc finger is a common motif shared by various proteins across all species (eg., Protein Kinase C in ref.4). It is thus possible that some of them might be redox-regulated similarly as Hsp33.

References

1. Ilbert, M., Horst, J., Ahrens, S., Winter, J., Graf, P.C.F., Lilie, H. & Jakob, U. (2007). The redox-switch domain of Hsp33 functions as dual stress sensor, Nature Structural & Molecular Biology, 14 (6) 563. DOI: 10.1038/nsmb1244

*** Hsp33 is activated by oxidative stress and heat.

2. Reichmann, D., Xu, Y., Cremers, C., Ilbert, M., Mittelman, R., Fitzgerald, M. & Jakob, U. (2012). Order out of Disorder: Working Cycle of an Intrinsically Unfolded Chaperone, Cell, 148 (5) 957. DOI: 10.1016/j.cell.2012.01.045

*** Hsp33’s substrate specificity. The linker region is directly involved in the substrate binding.

3. Kim, S.J., Jeong, D.G., Chi, S.W., Lee, J.S. & Ryu, S.E. (2001). Crystal structure of proteolytic fragments of the redox-sensitive Hsp33 with constitutive chaperone activity, Nature Structural Biology, 8 (5) 466. DOI: 10.1038/87639

*** The X-ray crystal structure of an active Hsp33 dimer revealed the domain crossover by the linker regions.

4. Zhao, F., Ilbert, M., Varadan, R., Cremers, C.M., Hoyos, B., Acin-Perez, R., Vinogradov, V., Cowburn, D., Jakob, U. & Hammerling, U. (2011). Are Zinc-Finger Domains of Protein Kinase C Dynamic Structures That Unfold by Lipid or Redox Activation?, Antioxidants & Redox Signaling, 14 (5) 766. DOI: 10.1089/ars.2010.3773

*** The widely distributed zinc-finger proteins may be redox-regulated similarly as Hsp33.