Complex Multidomain Proteins Research Articles

Ligand-guided selection of aptamers against T-cell Receptor-cluster of differentiation 3 (TCR-CD3) expressed on Jurkat.E6 cells

We recently introduced a screening technology termed ligand-guided selection, (LIGS), to selectively identify target-specific aptamers from an evolved cell-SELEX library. Cell-SELEX utilizes a large combinatorial single-stranded oligonucleotide library and progressively selects DNA ligands against whole cells with variable DNA-binding affinities and specificities by repeated rounds of partition and amplification. LIGS exploits the partition step and introduces a secondary, pre-existing high-affinity monoclonal antibody (mAb) ligand to outcompete and elute specific aptamers towards the binding target of the antibody, not the cell. Here, using anti-CD3ε mAb against the cluster of differentiation 3 (CD3ε), as the guiding ligand against one of the domains of the T-cell Receptor (TCR) complex expressed on Jurkat.E6 cells, we discovered three specific aptamers against TCR complex expressed on an immortalized line of human T lymphocyte cells. In sum, we demonstrate that specific aptamers can be identified utilizing an antibody against a single domain of a multidomain protein complex in their endogenous state with neither post- nor pre-SELEX protein manipulation.

Slowing Translation between Protein Domains by Increasing Affinity between mRNAs and the Ribosomal Anti-Shine-Dalgarno Sequence Improves Solubility.

Recent studies have demonstrated that effective protein production requires coordination of multiple cotranslational cellular processes, which are heavily affected by translation timing. Until recently, protein engineering has focused on codon optimization to maximize protein production rates, mostly considering the effect of tRNA abundance. However, as it relates to complex multidomain proteins, it has been hypothesized that strategic translational pauses between domains and between distinct individual structural motifs can prevent interactions between nascent chain fragments that generate kinetically trapped misfolded peptides and thereby enhance protein yields. In this study, we introduce synthetic transient pauses between structural domains in a heterologous model protein based on designed patterns of affinity between the mRNA and the anti-Shine-Dalgarno (aSD) sequence on the ribosome. We demonstrate that optimizing translation attenuation at domain boundaries can predictably affect solubility patterns in bacteria. Exploration of the affinity space showed that modifying less than 1% of the nucleotides (on a small 12 amino acid linker) can vary soluble protein yields up to ∼7-fold without altering the primary sequence of the protein. In the context of longer linkers, where a larger number of distinct structural motifs can fold outside the ribosome, optimal synonymous codon variations resulted in an additional 2.1-fold increase in solubility, relative to that of nonoptimized linkers of the same length. While rational construction of 54 linkers of various affinities showed a significant correlation between protein solubility and predicted affinity, only weaker correlations were observed between tRNA abundance and protein solubility. We also demonstrate that naturally occurring high-affinity clusters are present between structural domains of β-galactosidase, one of Escherichia coli's largest native proteins. Interdomain ribosomal affinity is an important factor that has not previously been explored in the context of protein engineering.

Efficient segmental isotope labeling of multi-domain proteins using Sortase A.

NMR studies of multi-domain protein complexes provide unique insight into their molecular interactions and dynamics in solution. For large proteins domain-selective isotope labeling is desired to reduce signal overlap, but available methods require extensive optimization and often give poor ligation yields. We present an optimized strategy for segmental labeling of multi-domain proteins using the S. aureus transpeptidase Sortase A. Critical improvements compared to existing protocols are (1) the efficient removal of cleaved peptide fragments by centrifugal filtration and (2) a strategic design of cleavable and non-cleavable affinity tags for purification. Our approach enables routine production of milligram amounts of purified segmentally labeled protein for NMR and other biophysical studies.

Immunogenicity and in vitro and in vivo protective effects of antibodies targeting a recombinant form of the Streptococcus mutans P1 surface protein.

Streptococcus mutans is a major etiologic agent of dental caries, a prevalent worldwide infectious disease and a serious public health concern. The surface-localized S. mutans P1 adhesin contributes to tooth colonization and caries formation. P1 is a large (185-kDa) and complex multidomain protein considered a promising target antigen for anticaries vaccines. Previous observations showed that a recombinant P1 fragment (P1(39-512)), produced in Bacillus subtilis and encompassing a functional domain, induces antibodies that recognize the native protein and interfere with S. mutans adhesion in vitro. In the present study, we further investigated the immunological features of P1(39-512) in combination with the following different adjuvants after parenteral administration to mice: alum, a derivative of the heat-labile toxin (LT), and the phase 1 flagellin of S. Typhimurium LT2 (FliCi). Our results demonstrated that recombinant P1(39-512) preserves relevant conformational epitopes as well as salivary agglutinin (SAG)-binding activity. Coadministration of adjuvants enhanced anti-P1 serum antibody responses and affected both epitope specificity and immunoglobulin subclass switching. Importantly, P1(39-512)-specific antibodies raised in mice immunized with adjuvants showed significantly increased inhibition of S. mutans adhesion to SAG, with less of an effect on SAG-mediated bacterial aggregation, an innate defense mechanism. Oral colonization of mice by S. mutans was impaired in the presence of anti-P1(39-512) antibodies, particularly those raised in combination with adjuvants. In conclusion, our results confirm the utility of P1(39-512) as a potential candidate for the development of anticaries vaccines and as a tool for functional studies of S. mutans P1.

The role of surface electrostatics on the stability, function and regulation of human cystathionine β-synthase, a complex multidomain and oligomeric protein

Human cystathionine β-synthase (hCBS) is a key enzyme of sulfur amino acid metabolism, controlling the commitment of homocysteine to the transsulfuration pathway and antioxidant defense. Mutations in hCBS cause inherited homocystinuria (HCU), a rare inborn error of metabolism characterized by accumulation of toxic homocysteine in blood and urine. hCBS is a complex multidomain and oligomeric protein whose activity and stability are independently regulated by the binding of S-adenosyl-methionine (SAM) to two different types of sites at its C-terminal regulatory domain. Here we study the role of surface electrostatics on the complex regulation and stability of hCBS using biophysical and biochemical procedures. We show that the kinetic stability of the catalytic and regulatory domains is significantly affected by the modulation of surface electrostatics through noticeable structural and energetic changes along their denaturation pathways. We also show that surface electrostatics strongly affect SAM binding properties to those sites responsible for either enzyme activation or kinetic stabilization. Our results provide new insight into the regulation of hCBS activity and stability in vivo with implications for understanding HCU as a conformational disease. We also lend experimental support to the role of electrostatic interactions in the recently proposed binding modes of SAM leading to hCBS activation and kinetic stabilization.

Significance of Unfolding Thermodynamics for Predicting Aggregation Kinetics: A Case Study on High Concentration Solutions of a Multi-Domain Protein

To enable aggregation rate prediction over a broad temperature range for complex multi-domain proteins at high concentrations. Thermal unfolding, non-isothermal kinetics and storage stability studies were conducted on a model multi-domain protein (MDP) at moderate to high concentrations (25-125mg/mL) over a broad temperature range (4-40°C). Storage stability studies indicated the aggregation of MDP in solution to be a second order process. Application of Arrhenius kinetics to accelerated stability data resulted in underestimation of the aggregation rate under refrigerated conditions. Additional studies undertaken to understand the mechanism of the aggregation process highlighted the association of the monomer (or the aggregation competent species) to be the rate-limiting step for aggregation over the temperature range studied. Thermal unfolding studies in the presence of urea were used to calculate the heat capacity change upon unfolding (Δcp,un). The resulting value of Δcp,un when used in the extended Lumry-Eyring model resulted in a more accurate and a conservative estimate of the aggregation rate under refrigerated condition. Some complicating factors for the aggregation rate prediction for multi-domain proteins at high concentration are discussed. The work highlights (i) the significance of incorporating unfolding thermodynamics in protein aggregation rate prediction, (ii) the advantages and challenges associated with the use of the extended Lumry-Eyring (ELE) model for rate prediction and (iii) the utility of using the Arrhenius and the ELE models in tandem during product development.

Exploring a New Role for Molecular Chaperones in Protein Folding

Using force spectroscopy, Mashaghi et al. have demonstrated that the general chaperone trigger factor (TF) promotes the correct folding of the maltose binding protein (MBP) by binding to partially folded states along the folding pathway. Protein folding is a remarkable biological process in which proteins are transformed from linear amino acid sequences into complex three-dimensional structures. Errors during protein folding can lead to neurodegenerative ailments such as Alzheimer's disease and Parkinson's disease. Although the precise in vivo mechanism of this important process remains poorly understood, a network of molecular chaperone proteins are thought to play a crucial role in guiding a protein towards its correct conformation. It has been previously reported that these chaperones assist protein folding by suppressing aggregation, as well as by rescuing misfolded states (Hartl et al., 2011). However, a recent study has shown that molecular chaperones directly influence a protein's conformational search to attain the native structure (Mashaghi et al., 2013). The primary technique used in this study is single- molecule force spectroscopy, in which small amounts of forces (in the piconewton range) are applied to unfold (and refold) single protein molecules, using an instrument called optical tweezers (Jagannathan and Marqusee, 2013). The magnitude of the applied force and the end-to-end extension of the protein molecule are monitored as a function of time. Rare intermediate states can be precisely identified because they have different end-to-end extensions than the native state or the completely unfolded state. Using a combination of force spectroscopy and biochemical essays, Mashaghi and co-workers explored the role of the general chaperone trigger factor (TF) in promoting the folding of maltose binding protein (MBP). In the absence of TF, MBP unfolds predominantly via a single intermediate state, although additional rare protein states are sampled transiently (for less than 1 s). After the addition of TF, the rare, transient states were observed much more frequently and were stable for longer times, suggesting that TF promotes and stabilizes the partially folded states. Furthermore, it was shown that TF does not bind to natively folded MBP, indicating that the molecular chaperone only binds partial folds along the folding pathway, until MBP attains its native state. Because TF association could be particularly beneficial for complex multi-domain proteins, the folding of a construct composed of four MBP repeats (MBP4) was studied. In the absence of TF, MBP4 folds to a compact, non-native state that does not unfold, suggesting the presence of tight misfolding interactions between the domains. The presence of TF significantly restored native folding of MBP4, and the tightly misfolded states were only rarely observed. The work described in this study is of importance to the protein folding field because it demonstrates that molecular chaperones can reshape the folding energy landscape by binding and protecting partially folded states. The stabilization of these states guides the folding pathway towards the correct native structure and away from misfolded conformations.

Divergent α-synuclein solubility and aggregation properties in G2019S LRRK2 Parkinson's disease brains with Lewy Body pathology compared to idiopathic cases

Mutations in LRRK2 are the most common genetic cause of Parkinson's disease (PD). The most prevalent LRRK2 mutation is the G2019S coding change, located in the kinase domain of this complex multi-domain protein. The majority of G2019S autopsy cases feature typical Lewy Body pathology with a clinical phenotype almost indistinguishable from idiopathic PD (iPD). Here we have investigated the biochemical characteristics of α-synuclein in G2019S LRRK2 PD post-mortem material, in comparison to pathology-matched iPD. Immunohistochemistry with pS129 α-synuclein antibody showed that the medulla is heavily affected with pathology in G2019S PD whilst the basal ganglia (BG), limbic and frontal cortical regions demonstrated comparable pathology scores between G2019S PD and iPD. Significantly lower levels of the highly aggregated α-synuclein species in urea–SDS fractions were observed in G2019S cases compared to iPD in the BG and limbic cortex. Our data, albeit from a small number of cases, highlight a difference in the biochemical properties of aggregated α-synuclein in G2019S linked PD compared to iPD, despite a similar histopathological presentation. This divergence in solubility is most notable in the basal ganglia, a region that is affected preclinically and is damaged before overt dopaminergic cell death.

On the supertertiary structure of proteins

Intrinsically disordered proteins and complex multidomain proteins are characterized by a dynamic ensemble of conformations that cannot be unequivocally described by traditional static terms of structural biology. The functional importance of this structural complexity necessitates new standards and protocols for the description and deposition of such 'supertertiary' structural ensembles into structural databases.

A T7 RNA polymerase-based toolkit for the concerted expression of clustered genes

Bacterial genes whose enzymes are either assembled into complex multi-domain proteins or form biosynthetic pathways are frequently organized within large chromosomal clusters. The functional expression of clustered genes, however, remains challenging since it generally requires an expression system that facilitates the coordinated transcription of numerous genes irrespective of their natural promoters and terminators. Here, we report on the development of a novel expression system that is particularly suitable for the homologous expression of multiple genes organized in a contiguous cluster. The new expression toolkit consists of an Ω interposon cassette carrying a T7 RNA polymerase specific promoter which is designed for promoter tagging of clustered genes and a small set of broad-host-range plasmids providing the respective polymerase in different bacteria. The uptake hydrogenase gene locus of the photosynthetic non-sulfur purple bacterium Rhodobacter capsulatus which consists of 16 genes was used as an example to demonstrate functional expression only by T7 RNA polymerase but not by bacterial RNA polymerase. Our findings clearly indicate that due to its unique properties T7 RNA polymerase can be applied for overexpression of large and complex bacterial gene regions.

Do Multiple PHDS Matter? An Analysis of the Dual PHD Domains of Rco1

Post-translational modifications (PTMs) on histone tails are important for many DNA-templated processes, including transcription elongation by RNA polymerase II (RNAPII). Histone PTMs directly alter the accessibility of chromatin and serve as binding sites for effector proteins that mediate numerous downstream biological processes. During transcription elongation in budding yeast (Saccharomyces cerevisiae), the C-terminal domain of RNAPII is hyperphosphorylated at serines 2 and 5 (Ser2 and Ser5) of the YSPTSPS consensus repeat. The histone methyltransferase (HMT) Set2 is recruited to phosphorylated Ser2 and Ser5, promoting the catalysis of lysine 36 (H3K36) methylation on the nucleosomal H3 tail in the bodies of genes. H3K36 methylation is recognized by the Rpd3S histone deacetylase (HDAC) complex, consisting of Rpd3, Sin3, Ume1, Rco1, and Eaf3. Deacetylation of histones behind elongating RNAPII has been shown to suppress cryptic initiation of transcription within the bodies of genes by decreasing chromatin accessibility. While it has been shown that recognition of H3K36 methylation by Rpd3S requires both the chromodomain of Eaf3 and the first PHD domain (PHD1) of Rco1, little is known regarding the importance of other potential histone-interacting domains within the complex. We are specifically interested in the second PHD domain (PHD2) of Rco1 due to its sequence homology to PHD1. We hypothesize that the second PHD domain of Rco1 contributes to the recognition of the Rpd3S complex on nucleosomes. Using a combination of yeast genetics and in vitro biochemical and biophysical assays, we are gaining a better understanding of the functional role of Rco1 PHD2. This study will lend insight into the importance of recognition of histone modifications by multi-domain protein complexes, and will allow us to better understand the function of the Rpd3S complex.

Widespread occurrence of secondary lipid biosynthesis potential in microbial lineages.

Bacterial production of long-chain omega-3 polyunsaturated fatty acids (PUFAs), such as eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3), is constrained to a narrow subset of marine γ-proteobacteria. The genes responsible for de novo bacterial PUFA biosynthesis, designated pfaEABCD, encode large, multi-domain protein complexes akin to type I iterative fatty acid and polyketide synthases, herein referred to as “Pfa synthases”. In addition to the archetypal Pfa synthase gene products from marine bacteria, we have identified homologous type I FAS/PKS gene clusters in diverse microbial lineages spanning 45 genera representing 10 phyla, presumed to be involved in long-chain fatty acid biosynthesis. In total, 20 distinct types of gene clusters were identified. Collectively, we propose the designation of “secondary lipids” to describe these biosynthetic pathways and products, a proposition consistent with the “secondary metabolite” vernacular. Phylogenomic analysis reveals a high degree of functional conservation within distinct biosynthetic pathways. Incongruence between secondary lipid synthase functional clades and taxonomic group membership combined with the lack of orthologous gene clusters in closely related strains suggests horizontal gene transfer has contributed to the dissemination of specialized lipid biosynthetic activities across disparate microbial lineages.

Stability of a guest protein depends on stability of a host protein in insertional fusion

Insertional fusion between host and guest protein domains has been employed to create multi-domain protein complexes displaying integrated and coupled functionalities. The effects of insertional fusion on the stability of a guest protein are however rather controversial. In the study described here, we examined whether the stability of inserted TEM1 beta-lactamase (BLA), as a guest protein, might be affected by the stability of a maltodextrin-binding protein (MBP), as a host protein. Our results indicate that expression levels and in vitro stability of the BLA domain were significantly higher when inserted into thermophilic MBP from Pyrococcus furiosus (PfMBP) compared to mesophilic MBP from Escherichia coli (EcMBP). Moreover, insertion into PfMBP at selected sites was found to improve thermal stability of the BLA domain without compromise in expression levels and BLA activity. Kinetic stabilization during prolonged thermal denaturation of the BLA domain was not guaranteed by insertion into PfMBP, but rather relied on the insertion sites. Taken together, we provide evidence that (i) the stability of the guest protein depended on the stability of the host protein in insertional fusion and (ii) insertion into PfMBP, at least at selected locations, can serve as a novel method of improving protein thermal stability.

Computational Studies of Translocon-Assisted Processes of Membrane Protein Insertion and Translocation

Membrane proteins make up about 30% of all the proteins in the body and represent more than 50% of drug targets. The protein-conducting channel, called translocon, is responsible for protein-membrane integration and the understanding of the mechanism of translocon-associated membrane protein folding has a significant biological and pharmaceutical interest.The translocon is a very large multidomain protein complex and the structural information about the protein-translocon complex is very tentative. Thus brute force all atom MD computer simulations are not expected to be very useful at this stage. In order to advance this challenging direction we introduce a unique coarse grain (CG) method with extended electrostatic treatment. This model allowed us to explore the energetics of the insertion of transmembrane helixes into lipid bilayer trough the translocon and to study the controversial question of the charged residue location in the membrane. More recently we stated to use the CG model in exploring the effect of key mutations that allow the secretion of proteins with defective or absent signal sequences. The preliminary insight provided by this study will also be described.

NMR and small-angle scattering-based structural analysis of protein complexes in solution

Structural analysis of multi-domain protein complexes is a key challenge in current biology and a prerequisite for understanding the molecular basis of essential cellular processes. The use of solution techniques is important for characterizing the quaternary arrangements and dynamics of domains and subunits of these complexes. In this respect solution NMR is the only technique that allows atomic- or residue-resolution structure determination and investigation of dynamic properties of multi-domain proteins and their complexes. As experimental NMR data for large protein complexes are sparse, it is advantageous to combine these data with additional information from other solution techniques. Here, the utility and computational approaches of combining solution state NMR with small-angle X-ray and Neutron scattering (SAXS/SANS) experiments for structural analysis of large protein complexes is reviewed. Recent progress in experimental and computational approaches of combining NMR and SAS are discussed and illustrated with recent examples from the literature. The complementary aspects of combining NMR and SAS data for studying multi-domain proteins, i.e. where weakly interacting domains are connected by flexible linkers, are illustrated with the structural analysis of the tandem RNA recognition motif (RRM) domains (RRM1-RRM2) of the human splicing factor U2AF65 bound to a nine-uridine (U9) RNA oligonucleotide.

Leucine-rich repeat kinase 2-linked Parkinson's disease: clinical and molecular findings.

Mutations in Leucine-rich repeat kinase 2 (LRRK2) gene are the most common cause of sporadic and familial late onset Parkinson’s disease (PD). The G2019S common mutation has been identified about 1% of sporadic cases and 4–7% of familial cases. Over 50 variants have since been identified in LRRK2, and at least 7 of these are confirmed to be pathogenic. In addition to pathogenic mutations, several common polymorphisms in the LRRK2 gene (G2385R and R1628P) have been identified that may explain up to 10% of sporadic PD in Asian populations. LRRK2 is a large complex multidomain protein with 2,527-amino-acid and the molecular weight is 286 kDa. LRRK2 multidomain protein consists of a catalytic core domain, kinase domain and a number of putative protein-protein interaction domains. LRRK2 mutations found in PD families, including the G2019S and I2020T mutations show increased intrinsic kinase activity, when assessed with myelin basic protein as substrate. The modification of LRRK2 GTPase and kinase activity affecting residues in the ROC, COR and mitogen-activated protein kinase kinase kinases domains is believed to lead to neuronal cell death, but the pathways involved remain unclear. A number of in vivo models in C. elegans, D. melanogaster and mice have been developed to study the patho/physiological function of LRRK2. Based on current literature, a toxic gain of function in LRRK2 kinase activity is a possible pathophysiologic mechanism and thus inhibition of kinase activity in experimental models offers a potential therapeutic strategy for LRRK2-linked PD.

Preferential use of protein domain pairs as interaction mediators: order and transitivity

Many protein-protein interactions (PPIs) are mediated by protein domains. The structural data of multi-domain PPIs reveal the domain pair (or pairs) that mediate a PPI, and implicitly also the domain pairs that are not involved in the interaction. By analyzing such data, preference relations between domain pairs as interaction mediators may be revealed. Here, we analyze the differential use of domain pairs as mediators of stable interactions based on structurally solved multi-domain protein complexes. Our analysis revealed domain pairs that are preferentially used as interaction mediators and domain pairs that rarely or never mediate interaction, independent of the proteins' context. Between these extremes, there are domain pairs that mediate protein interaction in some protein contexts, while in other contexts different domain pairs predominate over them. By describing the preference relations between domain pairs as a network, we uncovered partial order and transitivity in these relations, which we further exploited for predicting interaction-mediating domains. The preferred domain pairs and the ones over which they predominate differ in several properties, but these differences cannot yet determine explicitly what underlies the differential use of domain pairs as interaction mediators. One property that stood up was the over-abundance of homotypic interactions among the preferred domain pairs, supporting previous suggestions on the advantages in the use of domain self-interaction for mediating protein interactions. Finally, we show a possible association between the preferred domain pairs and the function of the complex where they reside. hanahm@ekmd.huji.ac.il Supplementary data are available at Bioinformatics online.

Evolved Lactococcus lactis Strains for Enhanced Expression of Recombinant Membrane Proteins

The production of complex multidomain (membrane) proteins is a major hurdle in structural genomics and a generic approach for optimizing membrane protein expression is still lacking. We have devised a selection method to isolate mutant strains with improved functional expression of recombinant membrane proteins. By fusing green fluorescent protein and an erythromycin resistance marker (ErmC) to the C-terminus of a target protein, one simultaneously selects for variants with enhanced expression (increased erythromycin resistance) and correct folding (green fluorescent protein fluorescence). Three evolved hosts, displaying 2- to 8-fold increased expression of a plethora of proteins, were fully sequenced and shown to carry single-site mutations in the nisK gene. NisK is the sensor protein of a two-component regulatory system that directs nisin-A-mediated expression. The levels of recombinant membrane proteins were increased in the evolved strains, and in some cases their folding states were improved. The generality and simplicity of our approach allow rapid improvements of protein production yields by directed evolution in a high-throughput way.

Designer Amphiphilic Short Peptides Enhance Thermal Stability of Isolated Photosystem-I

Stability of membrane protein is crucial during protein purification and crystallization as well as in the fabrication of protein-based devices. Several recent studies have examined how various surfactants can stabilize membrane proteins out of their native membrane environment. However, there is still no single surfactant that can be universally employed for all membrane proteins. Because of the lack of knowledge on the interaction between surfactants and membrane proteins, the choice of a surfactant for a specific membrane protein remains purely empirical. Here we report that a group of short amphiphilic peptides improve the thermal stability of the multi-domain protein complex photosystem-I (PS-I) in aqueous solution and that the peptide surfactants have obvious advantages over other commonly used alkyl chain based surfactants. Of all the short peptides studied, Ac-I5K2-CONH2 (I5K2) showed the best stabilizing effect by enhancing the melting temperature of PS-I from 48.0°C to 53.0°C at concentration of 0.65 mM and extending the half life of isolated PS-I significantly. AFM experiments showed that PS-I/I5K2/Triton X-100 formed large and stable vesicles and thus provide interfacial environment mimicking that of native membranes, which may partly explain why I5K2 enhanced the thermal stability of PS-I. Hydrophobic and hydrophilic group length of IxKy had an important influence on the stabilization of PS-I. Our results showed that longer hydrophobic group was more effective in stabilizing PS-I. These simple short peptides therefore exhibit significant potential for applications in membrane protein studies.

Direct Observation of Tug-of-War during the Folding of a Mutually Exclusive Protein

Although most protein folding studies are carried out on single-domain proteins, over two-thirds of proteins in proteomes are built up from multiple individually folded domains. A significant fraction of these multidomain proteins are domain-insertion proteins, in which one guest domain is inserted into a surface loop of a host protein. Intricate thermodynamic and kinetic coupling between the two domains can have a profound impact on their folding dynamics. Here we use an engineered mutually exclusive protein as a model system to directly illustrate one such complex dynamic process: the "tug-of-war" process during protein folding. By inserting a guest protein I27w34f into a host protein GB1-L5 (GL5), we engineered a novel, mutually exclusive protein, GL5/I27w34f, in which only one domain can remain folded at any given time due to topological constraints imposed by the folded structures. Using stopped-flow techniques, we obtained the first kinetic evidence that the guest and host domains engage in a folding tug-of-war as they attempt to fold, in which the host domain folds rapidly into its three-dimensional structure and is then automatically unfolded, driven by the folding of the guest domain. Our results provided direct evidence that protein folding can generate sufficient mechanical strain to unravel a host protein. Using single-molecule atomic force microscopy, we provide direct evidence for the existence of a conformational equilibrium between the two mutually exclusive conformations. Our results highlight important roles played by the intricate coupling between folding kinetics, thermodynamic stability, and mechanical strain in the folding of complex multidomain proteins, which cannot be addressed in traditional single-domain protein folding studies.