Defining the condensate landscape of fusion oncoproteins.

Fusion oncoproteins (FOs) arise from chromosomal translocations in ~17% of cancers and are often oncogenic drivers. Although some FOs can promote oncogenesis by undergoing liquid-liquid phase separation (LLPS) to form aberrant biomolecular condensates, the generality of this phenomenon is unknown. We explored this question by testing 166 FOs in HeLa cells and found that 58% formed condensates. The condensate-forming FOs displayed physicochemical features distinct from those of condensate-negative FOs and segregated into distinct feature-based groups that aligned with their sub-cellular localization and biological function. Using Machine Learning, we developed a predictor of FO condensation behavior, and discovered that 67% of ~3000 additional FOs likely form condensates, with 35% of those predicted to function by altering gene expression. 47% of the predicted condensate-negative FOs were associated with cell signaling functions, suggesting a functional dichotomy between condensate-positive and -negative FOs. Our Datasets and reagents are rich resources to interrogate FO condensation in the future.

An Image Analysis Pipeline for Quantifying the Features of Fluorescently-Labeled Biomolecular Condensates in Cells.

Biomolecular condensates are cellular organelles formed through liquid-liquid phase separation (LLPS) that play critical roles in cellular functions including signaling, transcription, translation, and stress response. Importantly, condensate misregulation is associated with human diseases, including neurodegeneration and cancer among others. When condensate-forming biomolecules are fluorescently-labeled and examined with fluorescence microscopy they appear as illuminated foci, or puncta, in cells. Puncta features such as number, volume, shape, location, and concentration of biomolecular species within them are influenced by the thermodynamics of biomolecular interactions that underlie LLPS. Quantification of puncta features enables evaluation of the thermodynamic driving force for LLPS and facilitates quantitative comparisons of puncta formed under different cellular conditions or by different biomolecules. Our work on nucleoporin 98 (NUP98) fusion oncoproteins (FOs) associated with pediatric leukemia inspired us to develop an objective and reliable computational approach for such analyses. The NUP98-HOXA9 FO forms hundreds of punctate transcriptional condensates in cells, leading to hematopoietic cell transformation and leukemogenesis. To quantify the features of these puncta and derive the associated thermodynamic parameters, we developed a live-cell fluorescence microscopy image processing pipeline based on existing methodologies and open-source tools. The pipeline quantifies the numbers and volumes of puncta and fluorescence intensities of the fluorescently-labeled biomolecule(s) within them and generates reports of their features for hundreds of cells. Using a standard curve of fluorescence intensity versus protein concentration, the pipeline determines the apparent molar concentration of fluorescently-labeled biomolecules within and outside of puncta and calculates the partition coefficient (Kp) and Gibbs free energy of transfer (ΔGTr), which quantify the favorability of a labeled biomolecule partitioning into puncta. In addition, we provide a library of R functions for statistical analysis of the extracted measurements for certain experimental designs. The source code, analysis notebooks, and test data for the Punctatools pipeline are available on GitHub: https://github.com/stjude/punctatools. Here, we provide a protocol for applying our Punctatools pipeline to extract puncta features from fluorescence microscopy images of cells.

Gain and loss of TASK3 channel function and its regulation by novel variation cause KCNK9 imprinting syndrome.

BACKGROUND: Genomics enables individualized diagnosis and treatment, but large challenges remain to functionally interpret rare variants. To date, only one causative variant has been described for KCNK9 imprinting syndrome (KIS). The genotypic and phenotypic spectrum of KIS has yet to be described and the precise mechanism of disease fully understood. METHODS: This study discovers mechanisms underlying KCNK9 imprinting syndrome (KIS) by describing 15 novel KCNK9 alterations from 47 KIS-affected individuals. We use clinical genetics and computer-assisted facial phenotyping to describe the phenotypic spectrum of KIS. We then interrogate the functional effects of the variants in the encoded TASK3 channel using sequence-based analysis, 3D molecular mechanic and dynamic protein modeling, and in vitro electrophysiological and functional methodologies. RESULTS: We describe the broader genetic and phenotypic variability for KIS in a cohort of individuals identifying an additional mutational hotspot at p.Arg131 and demonstrating the common features of this neurodevelopmental disorder to include motor and speech delay, intellectual disability, early feeding difficulties, muscular hypotonia, behavioral abnormalities, and dysmorphic features. The computational protein modeling and in vitro electrophysiological studies discover variability of the impact of KCNK9 variants on TASK3 channel function identifying variants causing gain and others causing loss of conductance. The most consistent functional impact of KCNK9 genetic variants, however, was altered channel regulation. CONCLUSIONS: This study extends our understanding of KIS mechanisms demonstrating its complex etiology including gain and loss of channel function and consistent loss of channel regulation. These data are rapidly applicable to diagnostic strategies, as KIS is not identifiable from clinical features alone and thus should be molecularly diagnosed. Furthermore, our data suggests unique therapeutic strategies may be needed to address the specific functional consequences of KCNK9 variation on channel function and regulation.

Enhanced interpretation of 935 hotspot and non-hotspot RAS variants using evidence-based structural bioinformatics.

In the current study, we report computational scores for advancing genomic interpretation of disease-associated genomic variation in members of the RAS family of genes. For this purpose, we applied 31 sequence- and 3D structure-based computational scores, chosen by their breadth of biophysical properties. We parametrized our data by assembling a numerically homogenized experimentally-derived dataset, which when use in our calculations reveal that computational scores using 3D structure highly correlate with experimental measures (e.g., GAP-mediated hydrolysis RSpearman = 0.80 and RAF affinity Rspearman = 0.82), while sequence-based scores are discordant with this data. Performing all-against-all comparisons, we applied this parametrized modeling approach to the study of 935 RAS variants from 7 RAS genes, which led us to identify 4 groups of mutations according to distinct biochemical scores within each group. Each group was comprised of hotspot and non-hotspot KRAS variants, indicating that poorly characterized variants could functionally behave like pathogenic mutations. Combining computational scores using dimensionality reduction indicated that changes to local unfolding propensity associate with changes in enzyme activity by genomic variants. Hence, our systematic approach, combining methodologies from both clinical genomics and 3D structural bioinformatics, represents an expansion for interpreting genomic data, provides information of mechanistic value, and that is transferable to other proteins.

Computational modeling reveals key molecular properties and dynamic behavior of disruptor of telomeric silencing 1-like (DOT1L) and partnering complexes involved in leukemogenesis.

Disruptor of telomeric silencing 1-like (DOT1L) is the only non-SET domain histone lysine methyltransferase (KMT) and writer of H3K79 methylation on nucleosomes marked by H2B ubiquitination. DOT1L has elicited significant attention because of its interaction or fusion with members of the AF protein family in blood cell biology and leukemogenic transformation. Here, our goal was to extend previous structural information by performing a robust molecular dynamic study of DOT1L and its leukemogenic partners combined with mutational analysis. We show that statically and dynamically, D161, G163, E186, and F223 make frequent time-dependent interactions with SAM, while additional residues T139, K187, and N241 interact with SAM only under dynamics. Dynamics models reveal DOT1L, SAM, and H4 moving as one and show that more than twice the number of DOT1L residues interacts with these partners, relative to the static structure. Mutational analyses indicate that six of these residues are intolerant to substitution. We describe the dynamic behavior of DOT1L interacting with AF10 and AF9. Studies on the dynamics of a heterotrimeric complex of DOT1L1-AF10 illuminated describe coordinated motions that impact the relative position of the DOT1L HMT domain to the nucleosome. The molecular motions of the DOT1L-AF9 complex are less extensive and highly dynamic, resembling a swivel-like mechanics. Through molecular dynamics and mutational analysis, we extend the knowledge previous provided by static measurements. These results are important to consider when describing the biochemical properties of DOT1L, under normal and in disease conditions, as well as for the development of novel therapeutic agents.

Phase Separation Mediates NUP98 Fusion Oncoprotein Leukemic Transformation.

NUP98 fusion oncoproteins (FO) are drivers in pediatric leukemias and many transform hematopoietic cells. Most NUP98 FOs harbor an intrinsically disordered region from NUP98 that is prone to liquid-liquid phase separation (LLPS) in vitro. A predominant class of NUP98 FOs, including NUP98-HOXA9 (NHA9), retains a DNA-binding homeodomain, whereas others harbor other types of DNA- or chromatin-binding domains. NUP98 FOs have long been known to form puncta, but long-standing questions are how nuclear puncta form and how they drive leukemogenesis. Here we studied NHA9 condensates and show that homotypic interactions and different types of heterotypic interactions are required to form nuclear puncta, which are associated with aberrant transcriptional activity and transformation of hematopoietic stem and progenitor cells. We also show that three additional leukemia-associated NUP98 FOs (NUP98-PRRX1, NUP98-KDM5A, and NUP98-LNP1) form nuclear puncta and transform hematopoietic cells. These findings indicate that LLPS is critical for leukemogenesis by NUP98 FOs. SIGNIFICANCE: We show that homotypic and heterotypic mechanisms of LLPS control NUP98-HOXA9 puncta formation, modulating transcriptional activity and transforming hematopoietic cells. Importantly, these mechanisms are generalizable to other NUP98 FOs that share similar domain structures. These findings address long-standing questions on how nuclear puncta form and their link to leukemogenesis. This article is highlighted in the In This Issue feature, p. 873.

Pathogenic SPTBN1 variants cause an autosomal dominant neurodevelopmental syndrome.

SPTBN1 encodes ßII-spectrin, the ubiquitously expressed ß-spectrin that forms micrometer-scale networks associated with plasma membranes. Mice deficient in neuronal ßII-spectrin have defects in cortical organization, developmental delay and behavioral deficiencies. These phenotypes, while less severe, are observed in haploinsufficient animals, suggesting that individuals carrying heterozygous SPTBN1 variants may also show measurable compromise of neural development and function. Here we identify heterozygous SPTBN1 variants in 29 individuals with developmental, language and motor delays; mild to severe intellectual disability; autistic features; seizures; behavioral and movement abnormalities; hypotonia; and variable dysmorphic facial features. We show that these SPTBN1 variants lead to effects that affect ßII-spectrin stability, disrupt binding to key molecular partners, and disturb cytoskeleton organization and dynamics. Our studies define SPTBN1 variants as the genetic basis of a neurodevelopmental syndrome, expand the set of spectrinopathies affecting the brain and underscore the critical role of ßII-spectrin in the central nervous system.

Molecular mechanics and dynamic simulations of well-known Kabuki syndrome-associated KDM6A variants reveal putative mechanisms of dysfunction.

BACKGROUND: Kabuki syndrome is a genetic disorder that affects several body systems and presents with variations in symptoms and severity. The syndrome is named for a common phenotype of faces resembling stage makeup used in a Japanese traditional theatrical art named kabuki. The most frequent cause of this syndrome is mutations in the H3K4 family of histone methyltransferases while a smaller percentage results from genetic alterations affecting the histone demethylase, KDM6A. Because of the rare presentation of the latter form of the disease, little is known about how missense changes in the KDM6A protein sequence impact protein function. RESULTS: In this study, we use molecular mechanic and molecular dynamic simulations to enhance the annotation and mechanistic interpretation of the potential impact of eleven KDM6A missense variants found in Kabuki syndrome patients. These variants (N910S, D980V, S1025G, C1153R, C1153Y, P1195L, L1200F, Q1212R, Q1248R, R1255W, and R1351Q) are predicted to be pathogenic, likely pathogenic or of uncertain significance by sequence-based analysis. Here, we demonstrate, for the first time, that although Kabuki syndrome missense variants are found outside the functionally critical regions, they could affect overall function by significantly disrupting global and local conformation (C1153R, C1153Y, P1195L, L1200F, Q1212R, Q1248R, R1255W and R1351Q), chemical environment (C1153R, C1153Y, P1195L, L1200F, Q1212R, Q1248R, R1255W and R1351Q), and/or molecular dynamics of the catalytic domain (all variants). In addition, our approaches predict that many mutations, in particular C1153R, could allosterically disrupt the key enzymatic interactions of KDM6A. CONCLUSIONS: Our study demonstrates that the KDM6A Kabuki syndrome variants may impair histone demethylase function through various mechanisms that include altered protein integrity, local environment, molecular interactions and protein dynamics. Molecular dynamics simulations of the wild type and the variants are critical to gain a better understanding of molecular dysfunction. This type of comprehensive structure- and MD-based analyses should help develop improved impact scoring systems to interpret the damaging effects of variants in this protein and other related proteins as well as provide detailed mechanistic insight that is not currently predictable from sequence alone.

Structural bioinformatics enhances mechanistic interpretation of genomic variation, demonstrated through the analyses of 935 distinct RAS family mutations.

MOTIVATION: Protein-coding genetic alterations are frequently observed in Clinical Genetics, but the high yield of variants of uncertain significance remains a limitation in decision making. RAS-family GTPases are cancer drivers, but only 54 variants, across all family members, fall within well-known hotspots. However, extensive sequencing has identified 881 non-hotspot variants for which significance remains to be investigated. RESULTS: Here, we evaluate 935 missense variants from seven RAS genes, observed in cancer, RASopathies and the healthy adult population. We characterized hotspot variants, previously studied experimentally, using 63 sequence- and 3D structure-based scores, chosen by their breadth of biophysical properties. Applying scores that display best correlation with experimental measures, we report new valuable mechanistic inferences for both hot-spot and non-hotspot variants. Moreover, we demonstrate that 3D scores have little-to-no correlation with those based on DNA sequence, which are commonly used in Clinical Genetics. Thus, combined, these new knowledge bear significant relevance. AVAILABILITY AND IMPLEMENTATION: All genomic and 3D scores, and markdown for generating figures, are provided in our supplemental data. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Novel destabilizing Dynactin variant (DCTN1 p.Tyr78His) in patient with Perry syndrome.

INTRODUCTION: Perry syndrome, also recognized as Perry disease, is a rare autosomal dominant disorder characterized by midlife-onset atypical parkinsonism, apathy or depression, respiratory failure and weight loss caused by a mutation in the Dynactin (DCTN1) gene. CASE DESCRIPTION: A fifty-six years-old adopted male presented with atypical parkinsonism with bradykinesia and postural instability, apathy, weight loss, and recurrent respiratory failure due to central hypoventilation requiring tracheostomy. METHODS AND RESULTS: Clinical workup revealed a novel DCTN1 p.Tyr78His variant. Using bioinformatic protein structure modeling, we compare our patient's variant to known DCTN1 mutations and predict protein stability of each variant at the CAP-Gly domain of p150Glued. All eight variants causing Perry syndrome, as well as Tyr78His, are located at site expected to interact with MAPRE1 tail and are predicted to be destabilizing. Variants causing atypical parkinsonism with incomplete Perry syndrome phenotype (K56R and K68E) are not significantly destabilizing in silico. CONCLUSION: We propose p.Tyr78His as the ninth pathogenic DCTN1 variant causing Perry syndrome. Bioinformatic protein modeling may provide additional window to understand and interpret DCTN1 variants, as we observed non-destabilizing variants to have different phenotype than destabilizing variants.

Novel KLHL26 variant associated with a familial case of Ebstein's anomaly and left ventricular noncompaction.

BACKGROUND: Ebstein's anomaly (EA) is a rare congenital heart disease of the tricuspid valve and right ventricle. Patients with EA often manifest with left ventricular noncompaction (LVNC), a cardiomyopathy. Despite implication of cardiac sarcomere genes in some cases, very little is understood regarding the genetic etiology of EA/LVNC. Our study describes a multigenerational family with at least 10 of 17 members affected by EA/LVNC. METHODS: We performed echocardiography on all family members and conducted exome sequencing of six individuals. After identifying candidate variants using two different bioinformatic strategies, we confirmed segregation with phenotype using Sanger sequencing. We investigated structural implications of candidate variants using protein prediction models. RESULTS: Exome sequencing analysis of four affected and two unaffected members identified a novel, rare, and damaging coding variant in the Kelch-like family member 26 (KLHL26) gene located on chromosome 19 at position 237 of the protein (GRCh37). This variant region was confirmed by Sanger sequencing in the remaining family members. KLHL26 (c.709C > T p.R237C) segregates only with EA/LVNC-affected individuals (FBAT p < .05). Investigating structural implications of the candidate variant using protein prediction models suggested that the KLHL26 variant disrupts electrostatic interactions when binding to part of the ubiquitin proteasome, specifically Cullin3 (CUL3), a component of E3 ubiquitin ligase. CONCLUSION: In this familial case of EA/LVNC, we have identified a candidate gene variant, KLHL26 (p.R237C), which may have an important role in ubiquitin-mediated protein degradation during cardiac development.

Opposing Intermolecular Tuning of Ca2+ Affinity for Calmodulin by Neurogranin and CaMKII Peptides.

We investigated the impact of bound calmodulin (CaM)-target compound structure on the affinity of calcium (Ca2+) by integrating coarse-grained models and all-atomistic simulations with nonequilibrium physics. We focused on binding between CaM and two specific targets, Ca2+/CaM-dependent protein kinase II (CaMKII) and neurogranin (Ng), as they both regulate CaM-dependent Ca2+ signaling pathways in neurons. It was shown experimentally that Ca2+/CaM (holoCaM) binds to the CaMKII peptide with overwhelmingly higher affinity than Ca2+-free CaM (apoCaM); the binding of CaMKII peptide to CaM in return increases the Ca2+ affinity for CaM. However, this reciprocal relation was not observed in the Ng peptide (Ng13-49), which binds to apoCaM or holoCaM with binding affinities of the same order of magnitude. Unlike the holoCaM-CaMKII peptide, whose structure can be determined by crystallography, the structural description of the apoCaM-Ng13-49 is unknown due to low binding affinity, therefore we computationally generated an ensemble of apoCaM-Ng13-49 structures by matching the changes in the chemical shifts of CaM upon Ng13-49 binding from nuclear magnetic resonance experiments. Next, we computed the changes in Ca2+ affinity for CaM with and without binding targets in atomistic models using Jarzynski's equality. We discovered the molecular underpinnings of lowered affinity of Ca2+ for CaM in the presence of Ng13-49 by showing that the N-terminal acidic region of Ng peptide pries open the ß-sheet structure between the Ca2+ binding loops particularly at C-domain of CaM, enabling Ca2+ release. In contrast, CaMKII peptide increases Ca2+ affinity for the C-domain of CaM by stabilizing the two Ca2+ binding loops. We speculate that the distinctive structural difference in the bound complexes of apoCaM-Ng13-49 and holoCaM-CaMKII delineates the importance of CaM's progressive mechanism of target binding on its Ca2+ binding affinities.

Correlation between Gene Variants, Signaling Pathways, and Efficacy of Chemotherapy Drugs against Colon Cancers.

Efficacies, toxicities, and resistance mechanisms of chemotherapy drugs, such as oxaliplatin and 5-fluorouracil (5-FU), vary widely among various categories and subcategories of colon cancers. By understanding the differences in the drug efficacy and resistance at the level of protein-protein networks, we identified the correlation between the drug activity of oxaliplatin/5-FU and gene variations from the US National Cancer Institute-60 human cancer cell lines. The activity of either of these drugs is correlated with specific amino acid variant(s) of KRAS and other genes from the signaling pathways of colon cancer progression. We also discovered that the activity of a non-DNA-binding novel platinum drug, phosphaplatin, is comparable with oxaliplatin and 5-FU when it was tested against colon cancer cell lines. Our strategy that combines the knowledge from pharmacogenomics across cell lines with the molecular information from specific cancer cells is beneficial for predicting the outcome of a possible combination therapy for personalized treatment.

Absence of Activation of DNA Repair Genes and Excellent Efficacy of Phosphaplatins against Human Ovarian Cancers: Implications To Treat Resistant Cancers.

Phosphaplatins, platinum(II) and platinum(IV) complexes coordinated to a pyrophosphate moiety, exhibit excellent antitumor activities against a variety of cancers. To determine whether phosphaplatins trigger resistance to treatment by engaging DNA damage repair genes, a yeast genome-wide fitness assay was used. Treatment of yeast cells with pyrodach-2 (D2) or pyrodach-4 (D4) revealed no particular sensitivity to nucleotide excision repair, homologous recombination repair, or postreplication repair when compared with platin control compounds. Also, TNF receptor superfamily member 6 (FAS) protein was overexpressed in phosphaplatin-treated ovarian tumor cells, and platinum colocalized with FAS protein in lipid rafts. An overactivation of sphingomyelinase (ASMase) was noted in the treated cells, indicating participation of an extrinsic apoptotic mechanism due to increased ceramide release. Our results indicate that DNA is not the target of phosphaplatins and accordingly, that phosphaplatins might not cause resistance to treatment. Activation of ASMase and FAS along with the colocalization of platinum with FAS in lipid rafts support an extrinsic apoptotic signaling mechanism that is mediated by phosphaplatins.

Lessons in Protein Design from Combined Evolution and Conformational Dynamics.

Protein-protein interactions play important roles in the control of every cellular process. How natural selection has optimized protein design to produce molecules capable of binding to many partner proteins is a fascinating problem but not well understood. Here, we performed a combinatorial analysis of protein sequence evolution and conformational dynamics to study how calmodulin (CaM), which plays essential roles in calcium signaling pathways, has adapted to bind to a large number of partner proteins. We discovered that amino acid residues in CaM can be partitioned into unique classes according to their degree of evolutionary conservation and local stability. Holistically, categorization of CaM residues into these classes reveals enriched physico-chemical interactions required for binding to diverse targets, balanced against the need to maintain the folding and structural modularity of CaM to achieve its overall function. The sequence-structure-function relationship of CaM provides a concrete example of the general principle of protein design. We have demonstrated the synergy between the fields of molecular evolution and protein biophysics and created a generalizable framework broadly applicable to the study of protein-protein interactions.

Alterations of Nonconserved Residues Affect Protein Stability and Folding Dynamics through Charge-Charge Interactions.

Charge-charge interactions play an important role in thermal stability of proteins. We employed an all-atom, native-topology-based model with non-native electrostatics to explore the interplay between folding dynamics and stability of TNfn3 (the third fibronectin type III domain from tenascin-C). Our study elucidates the role of charge-charge interactions in modulating the folding energy landscape. In particular, we found that incorporation of explicit charge-charge interactions in the WT TNfn3 induces energetic frustration due to the presence of residual structure in the unfolded state. Moreover, optimization of the surface charge-charge interactions by altering the evolutionarily nonconserved residues not only increases the thermal stability (in agreement with previous experimental study) but also reduces the formation of residual structure and hence minimizes the energetic frustration along the folding route. We concluded that charge-charge interaction in the rationally designed TNfn3 plays an important role not only in enhancing the stability but also in assisting folding.

Conformational frustration in calmodulin-target recognition.

Calmodulin (CaM) is a primary calcium (Ca(2+) )-signaling protein that specifically recognizes and activates highly diverse target proteins. We explored the molecular basis of target recognition of CaM with peptides representing the CaM-binding domains from two Ca(2+) -CaM-dependent kinases, CaMKI and CaMKII, by employing experimentally constrained molecular simulations. Detailed binding route analysis revealed that the two CaM target peptides, although similar in length and net charge, follow distinct routes that lead to a higher binding frustration in the CaM-CaMKII complex than in the CaM-CaMKI complex. We discovered that the molecular origin of the binding frustration is caused by intermolecular contacts formed with the C-domain of CaM that need to be broken before the formation of intermolecular contacts with the N-domain of CaM. We argue that the binding frustration is important for determining the kinetics of the recognition process of proteins involving large structural fluctuations.

Protein recognition and selection through conformational and mutually induced fit.

Protein-protein interactions drive most every biological process, but in many instances the domains mediating recognition are disordered. How specificity in binding is attained in the absence of defined structure contrasts with well-established experimental and theoretical work describing ligand binding to protein. The signaling protein calmodulin presents a unique opportunity to investigate mechanisms for target recognition given that it interacts with several hundred different targets. By advancing coarse-grained computer simulations and experimental techniques, mechanistic insights were gained in defining the pathways leading to recognition and in how target selectivity can be achieved at the molecular level. A model requiring mutually induced conformational changes in both calmodulin and target proteins was necessary and broadly informs how proteins can achieve both high affinity and high specificity.

Allostery and folding of the N-terminal receiver domain of protein NtrC.

The N-terminal receiver domain of protein NtrC (NtrC(r)) exhibits allosteric transitions between the inactive (unphosphorylated) and active (phosphorylated) state on the microsecond time scale. Using a coarse-grained variational model with coupled energy basins, we illustrate that significant loss of conformational flexibility is the key determinant of the inactive (I) → active (A) state transition mechanism of NtrC(r). In particular, our results reveal that the rearrangements of the native contacts involving the regulatory helix-α4 and the flexible ß3-α3 loop upon activation play a crucial role in the activation mechanism. Interestingly, we find that the ß3-α3 loop exhibits a gradual decrease in flexibility throughout the activation transition, while helix-α4, in contrast, becomes more rigid abruptly near the free energy barrier separating the two states. To gain further insight into role these flexible regions play in the transition mechanism, we consider folding of NtrC(r) to both states using a similar model. Our calculated folding routes suggest that helix-α4 becomes structured later when folding to the I state compared to folding of the A state, a result consistent with it is relative conformational flexibility in the two states. Finally, we find a good qualitative agreement between our predicted I → A transition mechanism and the measured backbone dynamics from nuclear magnetic resonance experiments.