The Svedberg Award 2001
Mikael Oliveberg
Department of Biochemistry , Umeå University
Protein folding, misfolding and aggregation
Our research explores the fundamental aspects of protein folding
- the final step in the conversion of genetic information to biological
function.
Background
My undergraduate training was in Engineering Physics at Chalmers
tekniska högskola and Biology at Göteborg University, and my PhD work
(Biophysics) was focused on the redox-linked proton pump cytochrome
oxidase. The latter involved development of new laser techniques for
concomitant detection of electron- and proton-translocation processes
in the ns to ms time regime. In 1992, I started protein-folding studies
as a post-doc at the MRC (Cambridge, UK) and I returned to Sweden in
1996 as “forskarassistent” to build up my own group at Lund University.
In 1999, I was awarded a “särskild forskartjänst” from NFR and shortly
after I was appointed the chair in Biochemistry at Umeå University.
From snap-shot to movie
Based on observations of transition-state changes in the folding
reaction of small proteins (cf. Hammond-postulate behaviour), we have
proposed a folding model based on a broad activation barrier separating
the denatured ensemble from the native state [1, 2]. Based on this
framework, we later used transition state movements to scan the barrier
top, to construct a movie of how a protein folds [3]. The results
support the view that folding is entropically controlled, i.e. contacts
that are close in sequence form first. For a given structure, a simple
rule (of thumb) for predicting the folding sequence would then be: join
nearest element with good contacts through the hydrophobic core.
Fig 1. Schematic illustration of high-energy folding, scanned by transition state movements. The progressively growing transition-state structure (‡) is inferred from effects of mutation on the refolding kinetics at increasing concentrations of denaturant.
Structural gatekeepers
Considering the astronomical number conformations that even a very
short polypeptide may attain, it is remarkable that proteins fold so
fast - often in a fraction of a second. Obviously, the protein somehow
navigates directly to its native state without wasting time by probing
futile misfolds. Recently, we observed that some residues, so called
gatekeepers, help this navigation (keep folding on track) by preventing
the protein from collapsing the wrong way [4]. If we delete these
groups, folding is retarded by detours into misfolded states that need
to entangle to reach the correct physiological structure.
Fig 2. Structural gatekeepers (green) are residues that prevent the protein (S6) from misfolding although they don't “actively” take part in folding of the native structure. Upon deletion of a gatekeeper, the protein displays increased propensity to misfold or aggregate. The residues that form contacts in the misfolded state are shown in blue whereas the residues that are unstructured are coloured red. For clarity, the residues are super-imposed on the native state.
Notably, the presence of folding gatekeepers provides a molecular explanation for the, so called, smooth funnels and Go-behaviour [ 5]: theoretically, folding is remarkably well reproduced by models that consider native contacts only and completely disregard misfolding. Recently, we observed that gatekeepers are also involved in preventing protein aggregation and fibrillation [6].
Protein tetramer zipped up by an Alzheimer sequence: unleashing aggregation by rational design Although protein folding is often utterly robust it sometimes goes wrong. Typically, this leads to uncontrolled accumulation of a generic type of protein aggregates called amyloid fibrils. If the amyloid fibrils deposit in the brain, they cause severe neurological damage as in Alzheimer's disease and in prion diseases. A long-standing problem in dealing with these phenomena is that the structure of the amyloid fibrils has resisted all attempts to be solved, hence obscuring the cause of the fibrils development. We have approached the problem from a folding point of view and transplanted “sticky” parts of the Alzheimer peptide into small model proteins that act as easily soluble scaffolds. As a result, the scaffold proteins assemble into well ordered tetramers whose structure can be determined by x-ray crystallography [ 6]. The interfaces within this tetramer, we believe, represent the first atomic view of interactions between amyloidogenic sequences: the Alzheimer transplant join the tetramer by forming inter-molecular b -sheets - the structural arrangement predicted for amyloid fibrils. To complement the structural data, we can also obtain precise information about the assembly process and the energetic contribution from individual residues from analysis of transient refolding aggregates, see [ 7] .
Fig 3. Designed tetramere zipped up with an Alzheimer sequence (light blue).
Role of topology and protein stability
In the quest to determine the mechanistic factors controlling the
folding rate, considerable interest have been focused on the role of
structural topology versus protein stability. To shed more light on
this issue, we have systematically shuffled the secondary structure
elements of S6 to design circular permutants with altered contact
order. Contact order is a topological parameter that measures the mean
sequence separation between interaction residues in the native state.
The result emphasizes the combined action of topology and stability in
controlling the folding rate constant, and provides, in particular,
information on the link between thermodynamic stability and kinetics
[8].
Fig 5. The circular permutants where constructed by linking the N- and C-termina and placing incisions at suitable sequence positions. Simply cut and paste in the gene.
In essence, the refolding rate constant, k f, is modulated by protein stability ( DG D-N) through the presence of native interactions in the transition state. Comparison across two-state proteins in the literature, yields the relation
where 
is the average 
- value for
stabilising contacts in two-state proteins, i.e. the average Brönsted
slope.
Evolution tunes folding to maximise cooperativity
It is easy to imagine that protein stability needs to be kept within
some limits for biological function. But is there any corresponding
pressure on the folding trajectory and the folding energy landscape? We
have recently demonstrated that reversing the loop entropy of S6 by
circular permutation has a dramatic effect on the transition-state
structure: it changes from uniformly diffuse to discretely polarised.
The phenomenon appears to stem from an evolutionary controlled bias of
the contact energies.
Fig 6. Stability studies of
non-disrupting side-chain truncations of S6, U1A and Ada 2H show that
contacts which anchor largest sequence loops are also the most stable.
There is a clear correlation (R2=0.64
for S6) between the stability loss upon mutation
(∆∆ G ) and
average sequence distance between the lost contacts. The insert
exemplifies one of the largest sequence loops
(45 residues) between contacts lost in the mutation L30A.
Mutant stability data reveal that regions with high loop entropy are balanced by anchoring contacts that are strong. This entropy compensation drives all parts of the protein to fold at the same time, maximises unfolding cooperativity and safeguards the structural integrity of the native state. In some circular permutants, the balance is broken and the globally diffuse transition state collapses into a polarised species with fully formed contacts in one end and none in the other. The unstructured parts of this atypical transition state (the parts that are not captured by the folding nucleus) need then to form on the down-hill side of the folding barrier and display increased propensity to unfold by local fluctuations in the native state. As such local instability could compromise function and even lead to unwanted aggregation [17], it is tempting to suggest that evolution tunes the energy landscape to maximise cooperativity. That is, all parts of the protein are tuned to fold simultaneously by entropy-compensation of the contact energies – the stability is evenly distributed throughout the native structure.
References:
[1] "High-Energy Channeling in Protein Folding"
Maria Silow and Mikael Oliveberg
(1997) Biochemistry , 36, 7633-7637.
[2] "Alternative explanations for multi-state kinetics in protein
folding: transient aggregation and changing transition-state ensambles"
Mikael Oliveberg
(1998) Accounts of Chemical Research , 31, 765-772.
[3] "From snap-shot to movie: phi-value analysis of protein folding
transition states taken one step further"
Tomas Ternström, Ugo Mayor, Mikael Akke and Mikael Oliveberg
(1999) Proc. Natl. Acad. Sci. USA. , 96,
14854-14859.
[4] "Salt-induced detour through compact regions of the protein folding
landscape"
Daniel E. Otzen and Mikael Oliveberg
(1999) Proc. Natl. Acad. Sci. USA. , 96, 11746-11751.
[5] "Characterisation of the transition states for protein folding:
Protein engineering analysis towards a new level of mechanistic detail"
Mikael Oliveberg
(2001) Current opinion in Structural Biology, 11,
94-100.
[6] "Designed protein tetramer zipped together with a hydrophobic
Alzheimer homology: a structural clue to amyloid assembly"
Daniel E. Otzen, Ole Kristensen and
Mikael Oliveberg
(2000) Proc. Natl. Acad. Sci. USA., 97, 9907-9912.
[7] "Transient Aggregates in Protein Folding are Easily Mistaken for
Folding Intermediates"
Maria Silow and Mikael Oliveberg
(1997) Proc. Natl. Acad. Sci. 94, 6084-6086.
[8] "Folding of circular permutants with decreased contact order:
general trend balanced by protein stability"
Magnus O. Lindberg, Jeanette Tångrot, Daniel E. Otzen, Dmitry A.
Golgikh, Alexei V. Finkelstein and Mikael Oliveberg
(2001) J. Mol. Biol. In press.