2.3.1. Principles of protein folding
Protein structures are based on non-covalent intramolecular interactions within the peptide backbone and between amino acid side chains. The most fundamental insights about protein folding came from early refolding studies (Anfinsen, 1973). Since then, it was shown in numerous in vitro experiments that denatured full-length proteins are able to refold spontaneously into their native state upon removal of the denaturing condition. These findings demonstrated that all the information required for a polypeptide to fold correctly into a specific three-dimensional structure is inherent in its linear amino acid sequence.
Nevertheless, a polypeptide can adopt a tremendous number of theoretical conformations during folding. Therefore, sampling all possible conformations to establish the correct ones would render the folding process impossible to occur on a biologically relevant time scale.
However, proteins typically fold within a few seconds or even microseconds (Buchner et al, 2011; Gruebele, 2005; Kubelka et al, 2004). This discrepancy between theory and experimental observation is described in the so-called Levinthal’s paradox, according to which a protein would never reach its native structure by exploring its entire conformational space (Zwanzig et al, 1992).
Another important discovery was that in most cases, folding is a sequential process rather than a one-step transition. Thereby, proteins fold in a funnel-shaped energy landscape along a pathway of defined intermediates until they reach their free energy minimum in the native state (Levy & Onuchic, 2006). In aqueous solutions water provides the main driving force for folding. In a first phase, a fast hydrophobic collapse of the unfolded peptide gives rise to compact intermediates, or molten globules, that contain native-like secondary structures (e.g.
α-helices and β-sheets) but lack well-packed side chains and a defined tertiary structure
Introduction
peripheral elements and are therefore especially prone to misfold or to aggregate. In this transition state most proteins have to overcome an energetic bottleneck situation upon which the multiple secondary structure elements condense against each other to form the native tertiary structure.
2.3.2. De novo protein folding in the cell
As mentioned above, most of our current knowledge about the driving forces of protein folding is based on the results of in vitro refolding studies. Such experiments are usually performed with denatured full-length proteins under diluted conditions, where the protein concentrations are low. Thus, the entire sequence information is instantaneously available when refolding is initiated. In contrast, protein folding in the cell is considerably different from in vitro refolding. One of the most critical aspects is that the cytosol is a crowded environment composed of organic molecules that reach concentrations between 300 and 400 g/l (Goodsell, 1991; Zimmerman & Trach, 1991). Molecular crowding has been shown to cause excluded volume effects that significantly increase the affinities between interacting proteins (Minton, 2005). Therefore, a multitude of cellular components can potentially interact with nascent polypeptides or partially folded intermediates and influence their folding pathway. In addition, de novo folding goes along with protein synthesis. Protein synthesis by ribosomes, however, is a strict vectorial process that proceeds from N- towards the C-terminus of a polypeptide chain. Since incomplete polypeptide chains cannot fold into their final structures, nascent polypeptides expose hydrophobic sites, which provide a contact surface for unproductive interactions. The probability for the formation of non-native contacts is even enhanced by the fact that translation is slow (~20-80 seconds for a protein of 400 amino acids) compared to average protein folding kinetics, which are usually on the microsecond to second timescale (Jackson, 1998; Zwanzig et al, 1992). This leads to the prolonged exposure of nascent polypeptides in the non-native state. During their synthesis, proteins are therefore especially prone to misfold and to aggregate. In order to prevent folding errors, nature has evolved several strategies, including folding catalysts and molecular chaperones, to keep newly synthesized proteins on the correct folding pathway (Komar, 2009).
2.3.3. Models for de novo protein folding
Different models for de novo protein folding have been proposed (Figure 6). As postulated by Baldwin, protein biosynthesis without folding is energetically unfavorable, because the conformational space and the energy of the growing polypeptide continuously increase with ongoing translation (Baldwin, 1999; Fedorov & Baldwin, 1997). As described above, there is meanwhile substantial experimental support that nascent polypeptides can fold cotranslationally, e.g. during their synthesis. Whereas the ribosomal tunnel only allows the formation of secondary structure elements, such as α-helices, compact three-dimensional folding is supposed to occur outside the ribosome. The compaction of nascent polypeptides on ribosomes was probed with several elegant approaches, including limited proteolysis, analysis of correct disulfide bridge formation, or by conformation-specific antibodies (Hamlin
& Zabin, 1972; Komar, 2009; Land et al, 2003; Netzer & Hartl, 1997). However, the enzymatic activity of full-length nascent chains provided the most solid evidence that proteins can adopt native conformations while being attached to ribosomes (Kudlicki et al, 1995;
Makeyev et al, 1996). Importantly, it was shown that for some proteins cotranslational folding occurs faster (Kolb et al, 2000) and with higher yields of correctly folded species, compared to in vitro refolding (Katranidis et al, 2009; Ugrinov & Clark, 2010). Although cotranslational folding takes place at least for a subset of newly synthesized proteins, the precise mechanism underlying this process and the atomic details of nascent polypeptides on ribosomes remain largely unexplored. Two different scenarios for cotranslational folding can be envisioned (Figure 6). One possibility is that secondary and tertiary structures begin to form gradually as soon as sequence information becomes exposed outside the ribosome. An alternative model suggests that tertiary structure formation could occur in a domainwise manner. In this case, the individual structural units (domains) of a protein may stay largely extended until enough sequence information is available outside the ribosome for productive folding (Zhang & Ignatova, 2011). Thereby, intramolecular interactions would be formed in a hierarchical condensation process, starting with secondary structures based on hydrogen bonds within the peptide backbone, followed by near-range side chain interactions that lead to compact folding of the domain. The long-distance interactions would be established later during synthesis or even posttranslational. This model is especially attractive for larger multi-domain proteins with complex architectures. In both cotranslational folding modes, however, non-native intrachain contacts may become trapped and thereby make folding inefficient and error-prone. It is therefore possible, that although folding may start during synthesis, some proteins acquire their native structures posttranslationally (Figure 6) (Jansens et al, 2002).
Introduction
Figure 6: Models for initial protein folding. In the posttranslational folding mode polypeptides stay unfolded during their synthesis and fold into their native conformation upon release from the ribosome. On the contrary, proteins could fold cotranslationally by forming intermediate structures as soon as sufficient sequence information becomes available outside the ribosome. An alternative model suggests that cotranslational folding of multi-domain proteins could occur in a multi-domainwise manner. In this case, an individual multi-domain remains unfolded until its entire sequence is exposed to allow productive structure formation. The subsequent domains may fold likewise.
Based on (Deuerling & Bukau, 2004).
Nevertheless, it is still challenging to study the details of cotranslational folding mechanisms due to the lack of adequate methods that can reliably detect transient folding intermediates during synthesis. In the present study we applied a new structural approach to visualize the dynamics and conformations of nascent polypeptides by nuclear magnetic resonance (NMR) spectroscopy. This allowed us to obtain detailed insights into their cotranslational structure acquisition pathway.
Posttranslational
N
N C
N C
Cotranslational
N C
Cotranslational domainwise
N N
N N
N