NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Madame Curie Bioscience Database [Internet]. Austin (TX): Landes Bioscience; 2000-2013.

Cover of Madame Curie Bioscience Database

Madame Curie Bioscience Database [Internet].

Show details

Implications of 3D Domain Swapping for Protein Folding, Misfolding and Function

, , and *.

* Corresponding Author: Laura S. Itzhaki—MRC Cancer Cell Unit, Hutchison-MRC Research Centre, Hills Road, Cambridge, CB2 0XZ,

Three-dimensional domain swapping is the process by which two identical protein chains exchange a part of their structure to form an intertwined dimer or higher-order oligomer. The phenomenon has been observed in the crystal structures of a range of different proteins. In this chapter we review the experiments that have been performed in order to understand the sequence and structural determinants of domain- wapping and these show how the general principles obtained can be used to engineer proteins to domain swap. We discuss the role of domain swapping in regulating protein function and as one possible mechanism of protein misfolding that can lead to aggregation and disease. We also review a number of interesting pathways of macromolecular assembly involving β-strand insertion or complementation that are related to the domain swapping phenomenon.


Three-dimensional domain swapping (referred to subsequently as “domain swapping”) is the process by which two identical protein chains exchange a part of their structure to form an intertwined dimer or higher-order oligomer. The phenomenon was first proposed in the 1960s to explain the behaviour of RNase A1 dimer and somewhat later also tryptophan synthetase2 and tryptophanase3 but the first crystal structures of domain-swapped protein only emerged in the 1980s.4-8 The terminology that is currently used was introduced in 1994 by Eisenberg and colleagues who also put forward a mechanistic framework within which to understand how and why domain swapping occurs.9-11 In this chapter we will: (1) highlight recently determined structures that suggest roles for domain swapping in regulating protein function, (2) discuss quantitative studies of the energetics and kinetics of the domain swapping process and (3) review the evidence for domain swapping as a mechanism of protein misfolding leading to aggregation and disease.

Domain Swapping Terminology

The structure of the subunits within the domain-swapped oligomer is identical to that of the monomer with the exception of the region that connects the exchanging domain with the rest of the protein (Figures 1,2A). In most cases, this so-called “hinge loop” region folds back on itself to form the monomer and adopts an extended conformation in the domain-swapped dimer. Although the process is known as “domain” swapping, proteins are often found to swap only a single secondary structure element such as a β-strand or α-helix rather than a whole domain of structure. The swapped structure can be located in any part of the polypeptide sequence although it is generally at the N- or C-terminus. In some cases, approximately half of the molecule is swapped and it is therefore difficult to define which half constitutes the swapped domain. The interactions made between the swapped domain and the rest of the protein are the same in the oligomer as in the monomer but they are formed in an inter-rather than a intra-molecular fashion. These intermolecular interactions comprise the “primary” interface. Since the subunits are often close to each other in the domain-swapped oligomer, an intermolecular new interface may be created that is not present in the monomeric form and this is known as the “secondary” interface. Domain swapping can potentially occur in a reciprocal manner to form a dimer, in a cyclical manner to form a trimer, tetramer etc., or in an open-ended manner to form an oligomer leaving uncomplemented ends available to assemble further. In order to be classed as a domain swapping protein, both monomer and domain-swapped forms need to have been observed. In some cases however, there is a structure of the domain-swapped form of a protein but no structure of the closed monomer and the protein is therefore considered to be a ‘candidate’ for domain swapping. In other cases, the protein has a homolog that is a closed monomer; these oligomers are classed as “quasi-domain-swapped”

Figure 1. Examples of domain-swapped protein structures.

Figure 1

Examples of domain-swapped protein structures. Monomeric proteins are shown where available. A) α-IPMS (also called LeuA), pdbcode 1sr9, B) RNaseA N-terminal swapped dimer, pdbcode 1a2w, C) RNaseA C-terminal swapped dimer, pdbcode 1f0v, D) RNaseA (more...)

Figure 2. Schematic representation of the key features of domain-swapped structures and of domain-swapping mechanisms.

Figure 2

Schematic representation of the key features of domain-swapped structures and of domain-swapping mechanisms. A) Scheme of domain-swapping showing the hinge loop, primary and secondary interfaces highlighted in red and the formation of cyclical and open-ended (more...)

Domain-Swapped Structures and Regulation of Protein Function

Many of the proteins and protein domains that are commonly used as model systems for studying protein folding and molecular recognition have been crystallised as domain-swapped forms in addition to the monomeric forms, most notably SH212 and SH313domains, staphylococcal nuclease, chymotrypsin inhibitor 214 (all domain-swapped dimers) and barnase (a domain-swapped trimer).15 These examples suggest that, although domain-swapping is relatively rare (there are less than 60 structures of domain-swapped proteins to date), many proteins have regions with features suggesting they could act as hinge loops and it may therefore be possible to induce many proteins to domain swap simply by making a few amino acid substitutions. The potential ease with which a domain-swapped species may become stabilised suggests that domain swapping could be a mechanism for the evolution of larger, complex folds from smaller, simpler ones via a domain-swapped intermediate followed by a gene duplication or fusion process.11 There is a subset of proteins for which evidence suggests that domain swapping regulates function and we shall focus on examples of these proteins here. For other proteins, it remains to be seen whether or not domain swapping is simply an artefact of the high protein concentrations associated with the crystallisation process.

SCAN Domain of HIV-1 Capsid C-Terminal Domain

Formation of the immature retrovirus particle is directed by interactions of the protein Gag with itself, with an arrangement of several thousand copies of the protein. After particle formation, maturation occurs by proteolytic processing of Gag to release the proteins found in the infectious virus. The characteristic conical core structure is formed by the proteolytically-released capsid protein (CA) around the viral genome. The C-terminal domain of CA (CA-CTD) contains the most highly conserved sequence within Gag, a 20-residue sequence known as the major homology region (MHR). Viral assembly and/or infectivity are sensitive to mutations or deletions in this region. Until recently it was thought that CA-CTD functions as a dimerisation domain although the dimer interface does not include the MHR and other CA’s do not dimerise in this way and it was therefore unclear how this interaction could account for the critical role played by CA-CTD in viral assembly. Then two structures emerged: a SCAN domain which is structurally and evolutionary related to the retroviral CA-CTD; and, the other the HIV-1 CA-CTD.16,17 Domain-swapping was induced in the latter by deletion of a single amino acid in the hinge loop region. These structures revealed that the swapped element encompasses the MHR. A functional role of domain-swapping would therefore rationalise the sequence conservation of the MHR sequence and would also explain a number of other features of viral assembly. However, whether or not the CA is domain-swapped in the mature viral core will await the determination of higher resolution structures of this particle.

Forkhead Domain of FOXP2

FOXP2 is a member of newly defined subfamily of the forkhead box (FOX) transcription factors and many disease-causing mutations are found in the forkhead domains of these proteins. The crystal structure of FOXP2 bound to DNA revealed that it forms a domain-swapped dimer.18 In classical FOX protein, a proline is present in a short turn connecting two helices. Replacement of this proline by alanine in FOXP2 results in the formation of a single long helix, thereby rigidifying the polypeptide chain in this region (see also later section on the energetic determinants of domain swapping) and causing it to swap into another molecule. Mutation of the alanine in FOXP2 to proline was found to prevent domain swapping. Alanine is present at this position in all FOXP members, suggesting that domain swapping is a conserved feature and may therefore be required for function. Modelling predicts that, due to the arrangement of the DNA binding surfaces, domain-swapped FOXP dimers can only bind cognate DNA sites that are well separated from each other, or located on different DNA strands, which suggests that these proteins loop DNA or mediate inter-chomosomal associations. Moreover, disease-associated mutations map to the domain-swapped dimer interface.


α-IPMS from mycobacterium tuberculosis is an α-isopropylmalate synthase which catalyses the first committed step of the essential leucine biosynthetic pathway in this organism. The structure of the protein reveals that the 70 kDa momomer folds into two major domains separated by two much smaller linker domains that are joined by a flexible hinge19 (Fig. 1A). The dimer is created by swapping of the major domains between two monomers. The dimer interface is extensive and spans both major domains but not the middle linker domains. 7650 Å of the monomer surface area is buried upon domain swapping, representing a quarter of the total monomer surface. The linker region appears flexible, giving rise to different relative orientations of the N- and C-terminal domains in the two monomers. The catalytic site is located within the N-terminal domain (a TIM barrel). The C-terminal domain is proposed to be the regulatory domain that is required for leucine feedback inhibition. Leucine binds to this domain, with the binding site lying between two helices, one from each of the two monomers. Other enzymes that are subject to end-product regulation bind their product inhibitors on domains remote from the catalytic domain and undergo major conformational changes on product binding; however, this is not the case for α-IPMS. It is thought instead that binding of leucine to the regulatory domain at the dimer interface is communicated to the distant catalytic domain by means of the flexible domain linker.

RNase A

RNase A is probably the best studied domain-swapped protein to date. RNase A is particular interesting because of its ability to swap two different domains, termed “double domain swapping”. RNase A or bovine seminal RNase can swap either an N-terminal (Fig. 1B) or C-terminal region (Fig. 1C and D), with more stringent unfolding conditions favouring the C-terminal swapping.20-24 A trimeric form has also been crystallised in which both N- and C-terminal regions are swapped.25 This feature potentially allows the protein to adopt a variety of differently assembled oligomeric states, including branched structures. Domain swapping in bovine seminal RNase may also be important for function. Domain-swapped dimerisation gives rise to two composite active sites with residues from both subunits contributing to each one and there is cooperativity between the sites.26 Remarkably, the domain-swapped dimer displays selective toxicity for tumour cells whereas the monomer does not.27

Energetic Determinants of Domain Swapping

Hinge Loop Length

The major determinant of domain swapping is the hinge loop, since this is the only region of the protein that adopts a different conformation in monomeric and domain-swapped forms (excluding those dimers in which a secondary interface forms.) The easiest way to increase the domain swapping propensity of a protein is to shorten the hinge loop, thereby making it harder for the polypeptide chain to fold back on itself. Loop deletion is seen in a number of natural proteins and this strategy has been successfully applied to a number of proteins to induce them to domain swap (Fig. 2B). A particularly nice example of this strategy comes from the design of two three-helix bundles of different topologies. Loop deletion in one leads to the formation of a reciprocal domain-swapped dimer whereas loop deletion in the other leads to the formation of fibrils by open-ended domain swapping (Fig. 1F and 2C).28

Hinge Loop Flexibility

A second strategy for engineering domain-swapped dimers has been to introduce greater flexibility into the hinge loop, either by lengthening it or by mutation. A polyglutamine insertion into the long active site loop of chymotrypsin inhibitor 2 resulted in a domain-swapped dimer and higher-order oligomers also.29 The extent of higher-order oligomer formation increased with loop insertions consisting of repeats of glutamine, alanine and glycine, in that order. Likewise for the protein suc1 (Fig. 1G,H), the wild-type protein forms only monomers and domain-swapped dimers at low millimolar concentrations but substitution of both proline residues in the hinge loop for alanines resulted in the formation of trimers and higher-order oligomers and this effect was enhanced when glycine residues were substituted for the prolines.30

Proline Residues in Hinge Loops

The proline residues found in the hinge loop of the protein suc1 appear to play a pivotal role in modulating its domain-swapping propensity. Indeed, Bergdoll and coworkers had observed previously that proline residues are found in the hinge loops of a number of proteins that domain swap.31 The results for suc1 indicated that proline residues create strain in the hinge loop (Fig. 2B). Suc1 is a 113-residue protein comprising a 4-stranded β-sheet that packs against two α-helices.32,33 Domain swapping in suc1 occurs by the exchange of a C-terminal sequence that comprises a central strand of the β-sheet (Fig. 1G,H). Mutation of one or other proline residue in the hinge loop to alanine greatly shifts the monomer-dimer equilibrium, changing the dissociation constant by several orders of magnitude, whereas mutation of other residues in the hinge loop to alanines has very little effect.30 Mutation of the first proline to alanine stabilises the monomer form, whereas mutation of the second proline to alanine stabilises the dimer form. The hinge loop is in a strained conformation in both monomer and dimer forms of the protein, more so in the monomer (Fig. 2B). The strain appears to make the hinge loop act as a loaded molecular spring that releases tension present in the monomer by adopting an alternative conformation in the dimmer.30 The proline residues do not appear to destabilise the hinge loop by making unfavourable interactions with other residues but rather by imposing strain on the backbone. Similar behaviour was observed for protein L.34 Mutations were designed to destabilise the hinge loop conformation of the monomer of protein L by forcing residues into forbidden regions of the Ramachandran plot, thereby inducing an alternative, domain-swapped form.34

Mutation of Residues Distant from the Hinge Loop

The domain swapping potential of the protein CD2 could be altered by making mutations in the secondary interface, the new intermolecular interface that is created in the domain-swapped form. Somewhat surprisingly, however, it has been found for domain-swapped proteins that lack a secondary interface (i.e., proteins for which the interactions made in the domain-swapped structure are identical to those in the monomer with the exception of the hinge loop), that mutations outside of the hinge loop can also alter the domain swapping propensity. These observations suggest signalling, i.e., that sites distant from the hinge loop can sense its conformation. In the case of suc1, mutation of residues in the β-sheet that were distant from the hinge loop was found to alter the equilibrium between monomer and dimer. The effects were not as large as for mutations in the hinge loop, but they were nevertheless significant. Likewise, a site in the β-sheet distant from the hinge loop was found to bind phosphorylated ligands with different affinities in the monomer and the dimer forms. The crystal structures of the monomer and domain-swapped dimer are highlysuperimposable outside of the hinge loop region and therefore they provide no explanation for these effects. The different energetics of the monomer and dimer forms must instead result from a difference in the dynamic properties of the conformational ensemble of the native states that is present in solution. It was proposed that the strain, imposed on the hinge loop by the proline residues, is accommodated differently in the two forms. Consistent with this idea, removing the strain by mutating the proline residues to alanine reduces the difference in binding affinities of the two forms. NMR and molecular dynamics simulations of cks1, the human homolog of suc1, further support the model.35 The experiments revealed conformational heterogeneity in the β-sheet of the monomer; similar behaviour has also been noted for suc1.36 The region of greatest conformational variability mapped to the domain-swapping β-strand. Upon binding of a conserved ligand, cdk2, to one face of the β-sheet containing the hinge loop region, the heterogeneity was frozen out and this complexed form of cks1 bound phosphorylated ligand on the opposite face of the β-sheet with higher affinity than the free form. The observation is consistent with the signal transduction in the β-sheet observed in the domain swapping experiments. It was therefore proposed that conformational heterogeneity is required for cks function and that a side effect of this requirement is that the protein domain has the ability to domain swap.

An example of mutations changing a protein’s tertiary and quaternary structure in a wholly unanticipated way is provided by the immunoglobuliN-binding domain B1 of streptococcal protein G (GB1). GB1 is a 56-residue protein, comprising a 4-stranded β-sheet packing against a single α-helix (Fig. 1I). In the course of screening hydrophobic core mutants for changes in stability some striking results were obtained. A domain-swapped dimer was formed that had the same structural elements as the wild-type monomer but the intertwined nature of the swapping, involving half of the secondary structural elements, was such as to create a significant secondary interface37 (Fig. 1J). A further mutation led to an even more startling result. A double domain-swapped tetramer was formed in which the structural elements were very different from the wild-type monomer38 (Fig. 1K). Closer inspection of the structures led the authors to conclude that destabilisation or opening up of the hydrophobic core upon mutation was compensated by extending the core via oligomerisation. Consistent with the idea that the mutations significantly destabilise the monomer structure, the monomeric form of this mutant variant is only partially folded with characteristics of a molteN-globule state (i.e., secondary structural elements present but not fixed by tertiary contacts).38

Kinetic Mechanisms of Domain Swapping

For domain swapping to occur, it is necessary for the protein to unfold in order for it to release the domain and then reassemble in an intertwined form. The mechanism of how a protein domain swaps and of how monomer and domain-swapped dimer forms interconvert, is therefore essentially a protein folding problem. For many proteins, interconversion is very slow under physiological conditions, consistent with idea of some degree of unfolding being required. The process is speeded up under conditions that favour unfolding. Eisenberg and coworkers proposed that the transition state for the interconversion reaction is an “open” form of the structure, in which the interactions between the exchanging domain and the rest of the protein are broken but the native folds within these two parts are retained. Such a scenario is most likely for proteins that swap a true domain capable of folding independently of the rest of the structure which can then assemble. However, many proteins swap only a small number of secondary structural elements that do not constitute a separate domain. In such cases, it is more likely that folding and association occur in a more concomitant fashion. Suc1 provides an extreme example of this sort of domain swapping, whereby the exchanging “domain” is a single strand of a 4-stranded β-sheet. Although this strand is at the very C-terminus of the polypeptide chain it is a central strand in the β-sheet and therefore integral to its stability. Protein engineering analysis and molecular dynamics simulations of the folding mechanisms of both monomeric and domain-swapped dimeric suc1 revealed that the interactions between this strand and the rest of the protein are formed very early in the folding pathway (and, conversely, very late in the unfolding reaction). Therefore, the entire protein mustunfold substantially in order to interconvert between monomer and dimer. Studies of the folding of the N-terminal, Ig domain of CD2, engineered to domain swap by deletion of residues in the hinge loop, indicate similar behaviour. However, in this case folding proceeds via intermediate in which one half of the domain-swapped dimer is folded and the other half is not.

Domain Swapping and Refolding

The folding reactions of the monomer and the domain-swapped dimer of suc1 are initiated at the same site (or folding nucleus), the only difference being that some of the key interactions of this nucleus are formed intra- vs. inter-molecularly in the two reactions. Consequently, there is ambiguity in the refolding process, which can lead to two different products. The competition between the collapse of a chain on itself to form a monomer and the probability of interacting with another chain before the key interactions are formed is affected by loop lengthening. Loop lengthening slows down the rate of folding of the monomer because of the higher entropic cost of fixing a longer loop than a shorter one and this effect indirectly favours folding to the dimer because it increases the likelihood of association with another chain. An unanticipated result for suc1 was that, whereas mutations in suc1 can alter the domain swapping propensity at equilibrium by many orders of magnitude, the product of refolding of all these mutant variants is overwhelmingly the monomeric form. Thus, the domain-swapped dimer may be the more stable form for certain of the suc1 mutants but it is not kinetically accessible. Evidence suggests that the population of an intermediate state acts to buffer the monomeric folding pathway against the effect of mutations that favour the dimer at equilibrium.

Another effect of loop lengthening is to increase the tendency to aggregate. Many proteins, including CI2, U1A and maltose binding protein, undergo transient association on refolding at moderate to high concentrations and domain-swapping provides a mechanism for such a phenomenon. The formation of higher order oligomers could occur through the same native nucleus by domain swapping in an open-ended, rather than reciprocal manner. These oligomers are only transient because of the uncomplemented ends that are produced by open-ended domain swapping.

Domain Swapping, Protein Misfolding and Aggregation

With the discovery of domain swapping, Eisenberg realised that the phenomenon could be a way for a protein to form aggregates such as those found in disease. Subsequently, two disease-associated proteins, prion protein and cystatin, were crystallised in domain-swapped forms39,40(Fig. 1M-P). There is now increasing evidence that domain swapping is involved in the assembly of some proteins into ordered aggregates. Of the various models for structures of amyloid fibrils that are consistent with its characteristic cross-β diffraction pattern, domain swapping has been implicated in three different ways (reviewed in Nelson and Eisenberg (2006)41). An important common feature of these models, distinct from those models in which the whole protein is envisaged to convert from its native structure to a β-sheet conformation, is that the subunits of the fibril retain much of the native structure of the monomeric form of the protein. This aspect of the domain-swapping models is particularly appealing in view of the fact that amyloid formation in vivo takes place under physiological conditions (albeit facilitated by the likely destabilising nature of disease-associated variants) rather than the harshly denaturing conditions commonly used in vitro to induce fibril formation. Moreover, several disease-associated proteins, such as the prion protein, β2-microglobulin and cystatin, have disulphide bonds and therefore amyloid formation involving large-scale unfolding of the structure would not be possible under nonreducing onditions. According to the cross-β spine model, a small segment of the polypeptide chain has the tendency to stack into a β-sheet.42 The segment may be located at the end of a folded domain or between folded domains and these domains are natively structured in the fibril. Such a model has been proposed for fibrils of β2-microglobulin, in which the β-sheet stacking region is at the N-terminus and the rest of the protein remains natively folded.43 If the stacking region is in the middle of the polypeptide chains in between sub-domains, then these may still be able to reform the native structure by complementing each other via domain swapping. Such a structure may be present in fibrils from a designed variant of RNase A containing a polyglutamine insertion (Q10) in the hinge loop.44 Evidence for the model comes from the observation that fibrils made by mixing two Q10 variants, each of which contains a mutated catalytic residue on either the core domain or the swapped domain, have RNase activity.

In the second model, fibril formation occurs by domain-swapping without formation of a cross-β spine, as has been proposed for fibrils of cystatin and β2-microglobulin.40,45,46 Both proteins have β-sheet structures and reciprocally domain-swapped dimers or oligomers could pack against each other to form a fibril. The potential for domain-swapped dimers to produce high-order assemblies in this way was pointed out by Murray et al for the β-sheet N-terminal domain of the protein CD2.47 Likewise, a crystal-wide β-sheet structure was also observed for the llama VHH-R9 domain built up of stacked domain-swapped dimmers.48 According to the third model, domain swapping could occur in an open-ended or runaway manner to produce high-order oligomers, as has been seen for amyloid fibrils of T7 endonuclease and for trpR, the latter forming a double domain-swapped crystalline array.49,50 Some more detailed examples, including possible pathways by which fibrils are formed from native monomers, are given below.

Prion Protein

The structure of the human prion protein has been solved as a monomer and also as a domain-swapped dimer.39,51(Fig. 1O,P). A disulphide bond links the swapped domain and the remaining protein and therefore monomer-dimer conversion requires reduction of the disulphide bond and then reoxidation. In a set of elegant experiments, Lee and Eisenberg took note of this disulphide bond linkage and were able to develop the first time an in vitro redox method of converting recombinant PrPc (its normal cellular form) to an infectious form they term PrPRDX, which has similar properties to PrPSc. PrPRDX could seed the conversion of PrPC to PrPRDX, and had double the β-sheet content of PrPC.52 They proposed a model for the conversion process involving runaway domain swapping in which the hinge loops of successive subunits form a continuous β-sheet at the centre of the fibril. The domain-swapped PrP globular domains sit on the outside of the β-sheet. The authors highlight the fact that a domain-swapping mechanism for the recruitment of normal protein into the fibrils is able to explain numerous features of prion diseases. In particular, the stability of both forms of the prion protein can be rationalised by the high energy barrier required to break both noncovalent and covalent (disulphide bonds) interactions. Also, self-propagation is able to occur because domain-swapping acts as a templating mechanism.


Strikingly, those mutant variants of GB1 that can exist as domain-swapped dimers (described earlier in this review) form fibrils readily, whereas those that fold only into the monomer form do not.53 A model for the fibrils was constructed involving packing of domain-swapped dimers against each other via an edge β-strand. This arrangement results in a contiguous β-sheet in which the individual β-strands run perpendicular to the long axis consistent with the diffraction pattern of the fibres. Residue Ala34 that switches the protein from a monomer to a domain-swapped dimer was proposed to act as a gatekeeper residue that ensures the protein folds efficiently in the momoeric species and does not misfold into alternative forms. When replaced by larger hydrophobic residues that force the core to expand and be destabilised, alternative conformations become accessible.


Early on in the domain-swapping studies of suc1, a correlation was observed between domain-swapping and aggregation propensities of mutant variants. Electron microscopy revealed that suc1 aggregates have the appearance of a string-of-pearls of dimensions consistent with a runaway domain-swapped arrangement of the molecules. Further experiments verified that the protein in the aggregates was indeed domain-swapped and native-like in structure. These aggregates could be formed by incubating protein at millimolar concentrations and room temperature or by refolding acid-denatured protein at concentrations in the high micromolar range. Interestingly, it was found that mutants with very low domain-swapping propensities and showing negligible aggregationcould be induced to aggregate by refolding in the presence of wild type or domain-swapping prone mutants. The domain-swapped suc1 aggregates were cytotoxic whereas amorphous aggregates, formed by heating the protein, were not. This result appears counterintuitive at first, given the native-like character of the domain-swapped aggregates. However, runaway domain swapping leaves uncomplemented ends and these unfolded elements are likely to be sticky and could potentially interact with various cellular components impairing their function. The observed toxicity is reminiscent of numerous other studies showing that nonfibrillar and prefibrillar assemblies are toxic but the fibril end-products are not. It was proposed that domain-swapping could be one mechanism for the early stages of fibril formation that brings together amyloid-prone segments from different polypeptide chains, thereby facilitating further assembly into fibrils.

Strand Insertion and Complementation: Serpins, Pilus Assembly and Rad51-BRCA2

We conclude this chapter by pointing out that there are other biological assemblies that are reminiscent of domain swapping. In each case, a strand is provided by another molecule of the same protein or of a different protein to complete or expand a β-sheet. Rad51 is a recombinase enzyme that forms nucleofilaments and both Rad51 and its homolog RecA have been shown to form filaments by runaway swapping of a β-strand (Fig. 1E). BRCA2, commonly mutated in breast, ovarian and other cancer types, controls the function of Rad51 in pathways for DNA repair by homologous recombination. The structure of the Rad51-BRCA2 complex recently solved reveals how BRCA2 inhibits the assembly of Rad51 nucleofilaments by mimicking the β-strand exchange process.54

Specific molecular chaperones help to assemble pilin subunits into adhesive pili, which are rod-like structures that allow pathogenic gram-negative bacteria to adhere to and colonize host tissues. Pilus subunits lack the seventh β-strand (G1 strand) that would complete their Ig fold and so expose a hydrophobic groove on their surface. Crystal structures of pilus-chaperone complexes indicate that the chaperone donates a β-strand from its Ig-like domain, thereby completing the Ig fold of the pilus subunit, although interestingly the chaperone G1 strand runs parallel to the F strand of the pilus subunit rather than anti-parallel in the true Ig fold.55,56 A model has been proposed in which the chaperone carries a pilus subunit to the large pore protein (the usher) where the pilus subunit is released and becomes attached to the end of the growing pilus rod (reviewed in 57). This process occurs by an N-terminal extension of the incoming pilus subunit inserting into the end subunit of the rod, replacing the strand temporarily donated by the chaperone. The process is termed “donor strand complementation”.

Serpins (serine protease inhibitors) contain an exposed and mobile reactive loop that presents a peptide sequence as a pseudosubstrate for the target proteinase. Upon docking of the serpin with the protease, a large conformational change occures in the former resulting in the insertion of the cleaved reactive loop as an additional strand into the s-stranded β-sheet of the serpin and consequent inhibition of the protease.58 Disease-associated point mutations act by destabilising the β-sheet allowing the incorporation of the reactive loop of another serpin molecule and thereby leading to the formation of an extended polymer chain.59,60 The Z mutation in α1-antitrypsin, associated with liver disease, is a glutamic acid to lysine substitution at the base of the mobile loop. The Z mutation results in the accumulation of protein in inclusions in the rough endoplasmic reticulum of the liver.61 Z α1-antitrypsin has also been shown to form chains of polymers in vitro when incubated under physiological conditions.59,62 By the same mechanism, inclusions of a mutant neuroserpin are associated with FENIB (familial encephalopathy with neuroserpin inclusion bodies).63-65Interestingly, mutations in the protein antithrombin associated with premature thrombosis, instead of causing polymerisation, result in the formation of inactive 6-stranded monomer that then binds to a normal antithrombin molecule leading to propagation of the conformational inactivation, similar to the proposed prion mechanism.59,66-69 The process has been shown to occur both in vitro and in vivo.


Crestfield AM, Stein WH, Moore S. On the aggregation of bovine pancreatic ribonuclease. Arch Biochem Biophys. 1962;1:217–222. [PubMed: 14023810]
Jackson DA, Yanofsky C. Restoration of enzymatic activity by complementation in vitro between mutant alpha subunits of tryptophan synthetase and between mutant subunits and fragments of the alpha subunit. J Biol Chem. 1969;244:4539–4546. [PubMed: 4897243]
London J, Skrzynia C, Goldberg ME. Renaturation of Escherichia coli tryptophanase after exposure to 8M urea. Evidence for the existence of nucleation centers. Eur J Biochem. 1974;47:409–415. [PubMed: 4607014]
Anderson WF, Ohlendorf DH, Takeda Y. et al. Structure of the cro repressor from bacteriophage lambda and its interaction with DNA. Nature. 1981;290:754–8. [PubMed: 6452580]
Fita I, Rossmann MG. The NADPH binding site on beef liver catalase. Proc Natl Acad Sci. 1985;82:1604–1608. [PMC free article: PMC397320] [PubMed: 3856839]
Parge HE, Arvai AS, Murtari DJ. et al. Human CksHs2 atomic structure: a role for its hexameric assembly in cell cycle control. Science. 1993;262:387–95. [PubMed: 8211159]
Remington S, Wiegand G, Huber R. Crystallographic refinement and atomic models of two different forms of citrate synthase at 2.7 and 1.7 A resolution. J Mol Biol. 1982;158:111–52. [PubMed: 7120407]
Story RM, Weber IT, Steitz TA. The structure of the E. coli recA protein monomer and polymer. Nature. 1992;355:318–25. [PubMed: 1731246]
Bennett MJ, Schlunegger MP, Eisenberg D. 3D domain swapping: A mechanism for oligomer assembly. Prot Sci. 1995;4:2455–2468. [PMC free article: PMC2143041] [PubMed: 8580836]
Bennett MJ, Choe S, Eisenberg D. Domain swapping: entangling alliances between proteins. Proc Nat Acad Sci USA. 1994;91:3127–3131. [PMC free article: PMC43528] [PubMed: 8159715]
Schlunegger MP, Bennett MJ, Eisenberg D. Oligomer formation by 3D domain swapping: a model for protein assembly and misassembly. Adv Protein Chem. 1997;50:61–122. [PubMed: 9338079]
Schiering N, Casale E, Caccia P. et al. Dimer formation through domain swapping in the crystal structure of the Grb2-SH2-Ac-pYVNV complex. Biochemistry. 2000;39:13376–13382. [PubMed: 11063574]
Kishan KVR, Newcomer ME, Rhodes TH. et al. Effect of pH and salt bridges on structural assembly: Molecular structures of the monomer and intertwined dimer of the Eps8 SH3 domain. Protein Science. 2001;10:1046–1055. [PMC free article: PMC2374198] [PubMed: 11316885]
Chen YW, Stott K, Perutz MF. Crystal structure of a dimeric chymotrypsin inhibitor 2 mutant containing an inserted glutamine repeat. Proc Natl Acad Sci USA. 1999;96:1257–61. [PMC free article: PMC15450] [PubMed: 9990011]
Zegers I, Deswarte J, Wyns L. Trimeric domain-swapped barnase. Proc Natl Acad Sci USA. 1999;96:818–822. [PMC free article: PMC15308] [PubMed: 9927651]
Ivanov D, Stone JR, Maki JL. et al. Mammalian SCAN domain dimer is a domain-swapped homolog of the HIV capsid C-terminal domain. Mol Cell. 2005;17:137–43. [PubMed: 15629724]
Sawaya MR, Sambashivan S, Nelson R. et al. Atomic structures of amyloid cross-beta spines reveal varied steric zippers. Nature. 2007;447:453–7. [PubMed: 17468747]
Stroud JC, Wu Y, Bates DL. et al. Structure of the forkhead domain of FOXP2 bound to DNA. Structure. 2006;14:159–66. [PubMed: 16407075]
Koon N, Squire CJ, Baker EN. Crystal structure of LeuA from Mycobacterium tuberculosis, a key enzyme in leucine biosynthesis. Proc Natl Acad Sci USA. 2004;101:8295–300. [PMC free article: PMC420388] [PubMed: 15159544]
Mazzarella L, Capasso S, Demasi D. et al. Bovine seminal ribonuclease—structure at 1.9-Angstrom resolution. Acta Crystallogr D Biol Crystallogr. 1993;49:389–402. [PubMed: 15299514]
Liu Y, Hart PJ, Schlunegger MP. et al. The crystal structure of a 3D domain-swapped dimer of RNase A at a 2.1- A resolution. Proc Natl Acad Sci USA. 1998;95:3437–42. [PMC free article: PMC19854] [PubMed: 9520384]
Liu Y, Gotte G, Libonati M. et al. A domain-swapped RNase A dimer with implications for amyloid formation. Nat Struct Biol. 2001;8:211–4. [PubMed: 11224563]
Esposito L, Daggett V. Insight into ribonuclease A domain swapping by molecular dynamics unfolding simulations. Biochemistry. 2005;44:3358–68. [PubMed: 15736946]
Gotte G, Vottariello F, Libonati M. Thermal aggregation of ribonuclease A. A contribution to the understanding of the role of 3D domain swapping in protein aggregation. J Biol Chem. 2003;278:10763–9. [PubMed: 12533538]
Liu Y, Gotte G, Libonati M. et al. Structures of the two 3D domain-swapped RNase A trimers. Protein Sci. 2002;11:371–80. [PMC free article: PMC2373430] [PubMed: 11790847]
Piccoli R, Di Donato A, D'Alessio G. Co-operativity in seminal ribonuclease function. Kinetic studies. Biochem J. 1988;253:329–36. [PMC free article: PMC1149302] [PubMed: 3178715]
Didonato A, Cafaro V, Romeo I. et al. Hints On the Evolutionary Design of a Dimeric Rnase With Special Bioactions. Protein Science. 1995;4:1470–1477. [PMC free article: PMC2143192] [PubMed: 8520472]
Ogihara NL, Ghirlanda G, Bryson JW. et al. Design of three-dimensional domain-swapped dimers and fibrous oligomers. Proc Natl Acad Sci USA. 2001;98:1404–9. [PMC free article: PMC29269] [PubMed: 11171963]
Gordon-Smith DJ, Carbajo RJ, Stott K. et al. Solution studies of chymotrypsin inhibitor-2 glutamine insertion mutants show no interglutamine interactions. Biochemical and Biophysical Research Communications. 2001;280:855–860. [PubMed: 11162601]
Rousseau F, Schymkowitz JWH, Wilkinson HR. et al. Three-dimensional domain swapping in p13suc1 occurs in the unfolded state and is controlled by conserved proline residues. Procl Natl Acad Sci USA. 2001;98:5596–5601. [PMC free article: PMC33258] [PubMed: 11344301]
Bergdoll M, Eltis LD, Cameron AD. et al. All in the family: Structural and evolutionary relationships among three modular proteins with diverse functions and variable assembly. Protein Science. 1998;7:1661–1670. [PMC free article: PMC2144073] [PubMed: 10082363]
Endicott JA, Noble ME, Garman EF. et al. The crystal structure of p13suc1, a p34cdc2-interacting cell cycle control protein. EMBO J. 1995;14:1004–1014. [PMC free article: PMC398172] [PubMed: 7889931]
Bourne Y, Arvai AS, Bernstein SL. et al. Crystal structure of the cell cycle-regulatory protein suc1 reveals a beta-hinge conformational switch. Procl Natl Acad Sci USA. 1995;92:10232–10236. [PMC free article: PMC40770] [PubMed: 7479758]
O'Neill JW, Kim DE, Johnsen K. et al. Single-Site Mutations Induce 3D Domain Swapping in the B1 Domain of Protein L from Peptostreptococcus magnus. Structure (Camb) 2001;9:1017–1027. [PubMed: 11709166]
Seeliger MA, Spichty M, Kelly SE. et al. Role of conformational heterogeneity in domain swapping and adapter function of the Cks proteins. J Biol Chem. 2005;280:30448–59. [PubMed: 15772084]
Landrieu I, Odaert B, Wieruszeski JM. et al. p13SUC1 and the WW Domain of PIN1 Bind to the Same Phosphothreonine- Proline Epitope. J Biol Chem. 2001;276:1434–1438. [PubMed: 11013245]
Byeon IJ, Louis JM, Gronenborn AM. A protein contortionist: core mutations of GB1 that induce dimerization and domain swapping. J Mol Biol. 2003;333:141–52. [PubMed: 14516749]
Kirsten Frank M, Dyda F, Dobrodumov A. et al. Core mutations switch monomeric protein GB1 into an intertwined tetramer. Nat Struct Biol. 2002;9:877–85. [PubMed: 12379842]
Knaus KJ, Morillas M, Swietnicki W. et al. Crystal structure of the human prion protein reveals a mechanism for oligomerization. Nat Struct Biol. 2001;8:770–4. [PubMed: 11524679]
Janowski R, Kozak M, Jankowska E. et al. Human cystatin C, an amyloidogenic protein, dimerizes through three-dimensional domain swapping. Nat Struct Biol. 2001;8:316–320. [PubMed: 11276250]
Nelson R, Eisenberg D. Recent atomic models of amyloid fibril structure. Curr Opin Struct Biol. 2006;16:260–5. [PubMed: 16563741]
Nelson R, Sawaya MR, Balbirnie M. et al. Structure of the cross-beta spine of amyloid-like fibrils. Nature. 2005;435:773–8. [PMC free article: PMC1479801] [PubMed: 15944695]
Ivanova MI, Sawaya MR, Gingery M. et al. An amyloid-forming segment of beta2-microglobulin suggests a molecular model for the fibril. Proc Natl Acad Sci USA. 2004;101:10584–9. [PMC free article: PMC489978] [PubMed: 15249659]
Sambashivan S, Liu Y, Sawaya MR. et al. Amyloid-like fibrils of ribonuclease A with three-dimensional domain-swapped and native-like structure. Nature. 2005;437:266–9. [PubMed: 16148936]
Staniforth RA, Giannini S, Higgins LD. et al. Three-dimensional domain swapping in the folded and molten-globule states of cystatins, an amyloid-forming structural superfamily. Embo Journal. 2001;20:4774–4781. [PMC free article: PMC125266] [PubMed: 11532941]
Eakin CM, Attenello FJ, Morgan CJ. et al. Oligomeric assembly of native-like precursors precedes amyloid formation by beta-2 microglobulin. Biochemistry. 2004;43:7808–15. [PubMed: 15196023]
Murray AJ, Head JG, Barker JJ. et al. Engineering an intertwined form of CD2 for stability and assembly. Nature Structural Biology. 1998;5:778–782. [PubMed: 9731771]
Spinelli S, Desmyter A, Frenken et al. Domain swapping of a llama VHH domain builds a crystal-wide beta-sheet structure. FEBS Lett. 2004;564:35–40. [PubMed: 15094039]
Guo Z, Eisenberg D. Runaway domain swapping in amyloid-like fibrils of T7 endonuclease I. Proc Natl Acad Sci USA. 2006;103:8042–7. [PMC free article: PMC1472426] [PubMed: 16698921]
Lawson CL, Benoff B, Berger T. et al. coli trp repressor forms a domain-swapped array in aqueous alcohol. Structure. 2004;12:1099–108. [PMC free article: PMC3228604] [PubMed: 15274929]
Zahn R, Liu A, Luhrs T. et al. NMR solution structure of the human prion protein. Proc Natl Acad Sci USA. 2000;97:145–50. [PMC free article: PMC26630] [PubMed: 10618385]
Lee S, Eisenberg D. Seeded conversion of recombinant prion protein to a disulfide-bonded oligomer by a reduction-oxidation process. Nat Struct Biol. 2003;10:725–30. [PubMed: 12897768]
Louis JM, Byeon IJ, Baxa U. et al. The GB1 amyloid fibril: recruitment of the peripheral beta-strands of the domain swapped dimer into the polymeric interface. J Mol Biol. 2005;348:687–98. [PubMed: 15826664]
Rocchi A, Pellegrini S, Siciliano G. et al. Causative and susceptibility genes for Alzheimer's disease: a review. Brain Res Bull. 2003;61:1–24. [PubMed: 12788204]
Choudhury D, Thompson A, Stojanoff V. et al. X-ray structure of the FimC-FimH chaperone-adhesin complex from uropathogenic Escherichia coli. Science. 1999;285:1061–6. [PubMed: 10446051]
Sauer FG, Futterer K, Pinkner JS. et al. Structural basis of chaperone function and pilus biogenesis. Science. 1999;285:1058–61. [PubMed: 10446050]
Remaut H, Waksman G. Structural biology of bacterial pathogenesis. Curr Opin Struct Biol. 2004;14:161–170. [PubMed: 15093830]
Huntington JA, Read RJ, Carrell RW. Structure of a serpin-protease complex shows inhibition by deformation. Nature. 2000;407:923–6. [PubMed: 11057674]
Lomas DA, Evans DL, Finch TJ. et al. The mechanism of Z alpha 1-antitrypsin accumulation in the liver. Nature. 1992;357:605–7. [PubMed: 1608473]
Stein PE, Carrell RW. What do dysfunctional serpins tell us about molecular mobility and disease? Nat Struct Biol. 1995;2:96–113. [PubMed: 7749926]
Sharp HL, Bridges RA, Krivit W. et al. Cirrhosis associated with alpha-1-antitrypsin deficiency: a previously unrecognized inherited disorder. J Lab Clin Med. 1969;73:934–9. [PubMed: 4182334]
Dafforn TR, Mahadeva R, Elliott PR. et al. A kinetic mechanism for the polymerization of alpha1-antitrypsin. J Biol Chem. 1999;274:9548–55. [PubMed: 10092640]
Davis RL, Holohan PD, Shrimpton AE. et al. Familial encephalopathy with neuroserpin inclusion bodies. Am J Pathol. 1999;155:1901–13. [PMC free article: PMC3277299] [PubMed: 10595921]
Davis RL, Shrimpton AE, Holohan PD. et al. Familial dementia caused by polymerization of mutant neuroserpin. Nature. 1999;401:376–9. [PubMed: 10517635]
Belorgey D, Crowther DC, Mahadeva R. et al. Mutant Neuroserpin (S49P) that causes familial encephalopathy with neuroserpin inclusion bodies is a poor proteinase inhibitor and readily forms polymers in vitro. J Biol Chem. 2002;277:17367–73. [PubMed: 11880376]
Alberts B. Molecular biology of the cell. 3rd ed. New York. Garland Pub. 1994
Schreuder HA, de Boer B, Dijkema R. et al. The intact and cleaved human antithrombin III complex as a model for serpin-proteinase interactions. Nat Struct Biol. 1994;1:48–54. [PubMed: 7656006]
Beauchamp NJ, Pike RN, Daly M. et al. Antithrombins Wibble and Wobble (T85M/K): archetypal conformational diseases with in vivo latent-transition, thrombosis and heparin activation. Blood. 1998;92:2696–706. [PubMed: 9763552]
Zhou A, Huntington JA, Carrell RW. Formation of the antithrombin heterodimer in vivo and the onset of thrombosis. Blood. 1999;94:3388–3396. [PubMed: 10552948]
Liu Y, Eisenberg D. 3D domain swapping: As domains continue to swap. Protein Sci. 2002;11:1285–1299. [PMC free article: PMC2373619] [PubMed: 12021428]
Sawaya MR, Guo S, Tabor S. et al. Crystal structure of the helicase domain from the replicative helicase-primase of bacteriophage T7. Cell. 1999;99:167–177. [PubMed: 10535735]
Gallagher T, Alexander P, Bryan P. et al. Two crystal structures of the B1 immunoglobulin-binding domain of streptococcal protein G and comparison with NMR. Biochemistry. 1994;33:4721–4729. [PubMed: 8161530]
Shimba N, Kariya E, Tate S. et al. Structural comparison between wild-type and P25S human cystatin A by NMR spectroscopy. Does this mutation affect the alpha-helix conformation? J Struct Funct Genomics. 2000;1:26–42. [PubMed: 12836678]
Janowski R, Kozak M, Abrahamson M. et al. 3D domain-swapped human cystatin C with amyloid-like intermolecular beta-sheets. Proteins. 2005;61:570–578. [PubMed: 16170782]
Copyright © 2000-2013, Landes Bioscience.
Bookshelf ID: NBK6144


  • PubReader
  • Print View
  • Cite this Page

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...