The KNOTTIN database

Folding may imply complex equilibrium and disulfide reshuffling

  •  Although disulfide bridges are responsible for the high stability of Knottins, they also render the folding process more complex, especially in chemical synthesis.
  • Since Knottins are considered as interesting leads in drug design, it is essential to understand the basic principles that govern the folding process. This would help in rational knottin-based drug-design studies.
  • Main historical and recent efforts along this way are outlined below.
You can toggle between only one (default) or all item display
*All* is currently selected

Squash inhibitors

alpha-Amylase Inhibitor

Carboxypeptidase inhibitor


Spider toxins


General review on the oxidative folding of small disulfide-rich proteins are available [Arolas et al, 2006; Craik 2010].

Squash inhibitors

The Knottins EETI-II and MCoTI-II fold in a simple process

The folding process of the squash inhibitor EETI-II has been studied in several steps

Folding of EETI-II is fast and efficient. Air oxidation of cysteines leads in few hours and with good yield to native disulfide bridges [Le-Nguyen et al., 1989b]

A linear synthetic analog of EETI-II in which all cysteines have been replaced by serines, has been obtained and submitted to careful NMR analyses [Heitz et al., 1995]. Native local structure propensities are clearly detected for the 11-15 310-helix region and for the 22-25 β-turn region. Only very few native non-covalent interactions are detected (proximity between Phe26 and residues Val20)

Mono-disulfide synthetic analogs of EETI-II in which cysteines 2, 9, 19 and 21 or 2, 19, 15 and 27 were replaced by serines, have been synthesized. These compounds do not display more native-like 3D structurations than does the linear analog analog [Heitz et al., 1997]. These results indicate that formation of the first disulfide bridge is non-specific and that at least two disulfides are necessary to stabilize the native fold.

A stable two-disulfide intermediate in the folding process displays most 3D structure of the native compound [Le-Nguyen et al, 1993]. It is shown that the two disulfides II-V and III-VI are necessary AND sufficient to determine the native fold. This corresponds to the elementary two-disulfide Cystine-Stabilized β-sheet (CSB) motif. Similar results have been reported for kalata B1 (see Cyclotides).

The folding pathway of the macrocyclic squash inhibitor MCoTI-II was found to be very similar to that of the linear EETI-II (see above). This study was performed on the natural cyclized compound although it is unknown if in vivo the head-to-tail cyclization occurs prior to oxidative folding. It is likely that oxydative folding of the linear compound precedes cyclization [Cemazar et al., 2006; Cemazar et al., 2008]

The GPNG 22-25 β-turn in EETI-II favors folding and is sequence restricted

The 22-25 GPNG β-turn appeared as the most flexible part in native EETI-II [Chiche et al., 1989; Chiche et al., 1993]. Nevertheless, replacing the corresponding sequence in the homologous inhibitor CMTI-III by the GPNG sequence strongly improved the chemical synthesis yield of the latter compound [Rolka et al., 1991b].

Supporting this observation, a combinatorial library of EETI-II in which the 22-25 GPNG sequence was randomized revealed that the sequence of this turn is significantly constrained and that only the native GPNG sequence provides high yield in the native fold [Wentzel et al., 1999].

α-Amylase Inhibitor

Oxydative folding of AAI involves intermediates with nonnative disulfide bridges between adjacent cycteines

The folding pathway of the α-Amylase Inhibitor AAI from Amaranthus has been analyzed using RP-HPLC, LC-MS, 1H-NMR and photochemically induced dynamic nuclear polarization (photo-CIDNP) experiments [Cemazar et al., 2003].
Folding of AAI proceeds through several fully oxydized intermediates with nonnative disulfide bridges.
The major folding intermediate (MFI) is shown to contain a disulfide bridge between adjacent cysteine residues, i.e. Cys60 and Cys61 according to the knottin unique numbering.

A clear interdependence between the formation of disulfide bridges and conformational folding is demonstrated [Cemazar et al., 2004]. It is proposed that formation of nonnative disulfide bridges facilitates folding by reducing the entropy of the unfolded state. Then the conversion from nonnative disulfide bridges to native disulfide bridges is driven by the formation of native stabilizing non-bonding interactions.

Carboxypeptidase inhibitor

The Knottin PCI folds in a complex process

The folding pathway of the Potato Carboxypeptidase A Inhibitor PCI has been analyzed by structural analysis and stop/go folding experiments [Chang et al., 1994].
Folding of PCI proceeds through an initial stage of nonspecific disulfide formation.
1- and 2- and 3-disulfide intermediates are highly heterogeneous.
Disulfide reshuffling occurs at the final stage which refines and consolidates the scrambled species to acquire the native conformation.

Comparison with other small disulfide-rich proteins show that proteins such as PCI with their native disulfide bonds reduced in a collective and simultaneous manner exhibit both a high degree of heterogeneity of folding intermediates and the accumulation of scrambled isomers along the folding pathway [Chang & Bulychev, 2000a].

Analysis of the unfolding pathway of PCI has revealed the existence of structurally defined unfolding intermediates [Chang et al., 2000b]. It was also shown that the PCI sequence is unable to fold quantitatively into a single native structure, and that, under physiological conditions, the scrambled isomers of PCI that constitute about 4% of the protein are in equilibrium with native PCI.

PCI folding does not depend on the prosequence

Folding of the Potato Carboxypeptidase A Inhibitor PCI has been studied with or without the prosequences, either in vitro or in vivo in Escherichia coli [Bronsoms et al., 2003].
It is shown that the prosequence does not affect folding significantly neither in vitro nor in vivo.


A review on the oxidative folding of conotoxins is available [Bulaj & Olivera, 2008]

Peptidyl prolyl cis-trans isomerases facilitate conotoxin folding

Peptidyl prolyl cis-trans isomerases isolated from the venom gland of Conus novaehollandiae were shown to greatly increase the rate of appearance of hydoxyproline containing conotoxins [Safavi-Hemami et al., 2010].

Hydroxyprolines facilitates folding of ω-conotoxins but do not modify the acitivy

It has been shown that hydroxylation of Prolines improved the in vitro folding yield of ω-MVIIC approximately 2-fold. Conotoxin ω-GVIA displayed similar results. In contrast, hydroxylation of non knottin μ-conotoxins affected activity rather than folding [Lopez-Vera et al., 2008].

Albumin is a redox-active crowding agent that promotes oxidative folding of conotoxins and other cysteine-rich peptides

Oxidative folding reactions were shown to be dramatically accelerated when protein-based crowding agents were present at concentrations lower than those predicted to provide the excluded volume effect. Submillimolar albumin alone appeared as effective as glutathione in promoting the oxidative folding of GI conotoxin [Buczek et al., 2007].

The pro-peptide might play a role in the PDI-catalized folding of conotoxins

The influence of the propeptide on the folding of α-conotoxins has been analyzed [Buczek et al., 2004]. It is shown that the propeptide does not significantly affect oxidative folding of α-conotoxin GI, it does facilitate PDI-assisted folding (PDI: protein disulfide isomerase). It is proposed that the increased length of the precursor may favor binding to PDI. Although α-conotoxins do not share the knottin fold, this result might be of importance for folding of other disulfide-rich proteins.

Non-ionic detergents and low temperature improve folding

Factors that improve oxidative folding of hydrophobic δ-conotoxins have been searched [DeLa Cruz et al., 2003]. It is shown that non-ionic detergents at low temperature (0°C) significantly improves the folding yield of δ-conotoxin PVIA. This effect is attributed to stabilization of native PVIA through non-specific interactions between the hydrophobic part of the detergent and non-polar patches at the surface of PVIA.

Two disufides are necessary for native 3D interaction to occur in omega-conotoxin MVIIA

Three omega-conotoxin MVIIA analogs in which the cysteines of one of the three disulfide bridges are replaced by alanines have been synthesized [Price-Carter et al., 2002; Price-Carter et al., 1998]. For each analog, all mono and two disulfide-bonded species were identified and equilibrium constants evaluated. These studies indicate that, in contrast with results for EETI-II, two-disulfide compounds are not significantly stabilized by non-covalent interactions. However, formation of two native disulfide greatly favor formation of the third disulfide and the native folded structure.

Mature conotoxins are able to refold in vitro

It is shown that in presence of GSSG/GSH, several conotoxins are able to refold to their native conformation with efficiencies ranging from 16% (omega-MVIIC) to 50% (omega-MVIIA, -GVIA and SVIA) [Price-Carter et al., 1996a; Price-Carter et al., 1996b]. Effect of salt and temperature on omega-MVIIC native disulfide brige formation has been reported [Kubo et al., 1996]

The Cys15-Cys26 disulfide is essential for correct folding of omega-conotoxin GVIA

The Cys15-Cys26 disulfide bridge in omega-conotoxin GVIA has been deleted by replacing corresponding cysteines by serines [Flinn et al., 1999]. This analog displayed a gross loss of secondary and tertiary structure, probably responsible for the total absence of activity.

Spider toxins

Mature GsMTx4 is able to fold properly

Oxidative folding of mature GsMTx4 yields the correct folded peptide [Ostrow et al., 2003]. In the absence of GSSG/GSSH, oxidation of the linear precursor yields a mixture of misfolded forms with incorrectly paired disulfides. After addition of GSSG/GSSH, the mixture shifted to the correctly folded state. The propeptides are unnecessary for proper folding. These results are consistent with previous results on ω-conotoxin MVIIA.


Reviews on the oxidative folding of cyclotides are available [Cemazar et al., 2008; Gunasekera et al., 2009]

In vivo Cyclization of cyclotides probably involves an asparaginyl endopeptidase

A highly conserved Asn residue is present at the C-terminus of the mature sequence of most cyclotides, suggesting that an Asn specific enzyme could be involved in the in vivo cyclization. When Oak1, the precursor of kalata B1, is expressed in Nicotiana Benthamiana, a plant which contains well characterized asparaginyl endopeptidases, cyclic oxidized kalata B1 is produced. Interestingly, inhibition or silencing of the asparaginyl endopeptidases in Nicotiana Benthamiana did not affect the production of kalata B1 but significantly decreased the production of the cyclic compound [Saska et al., 2007].

Cyclization favors correct folding of Kalata B1

Oxidation of cysteines during chemical synthesis of Kalata B1 has been performed using two different strategies [Daly et al., 1999]. It is shown that oxidation of the linear precursor, before cyclisation, necessitate an hydrophobic environment. In contrast, if cyclization is performed prior to oxidation, then the hydrphobic environment is unnecessary. This highlights the favorable role of the cyclization in the folding of Kalata B1.

A two-disulfide intermediate accumulates during folding of Kalata B1

Examination of oxidative refolding of kalata B1 shows that a native-like intermediate containing two out of the three native disulfide bridges is accumulated [Daly et al., 2003]. This intermediate contains the II-V and III-VI disufide bridges and is analogous to the intermediate previously observed during folding of EETI-II. In contrast to EETI-II, however, this kalata B1 folding intermediate is a kinetic trap and is not the immediate precursor of the native form.