Journal of Molecular Biology

KdpFABC is a high-affinity prokaryotic K+ uptake system that forms a functional chimera between a channel-like subunit (KdpA) and a P-type ATPase (KdpB). At high K+ levels, KdpFABC needs to be inhibited to prevent excessive K+ accumulation to the point of toxicity. This is achieved by a phosphorylation of the serine residue in the TGES162 motif in the A domain of the pump subunit KdpB (KdpBS162-P). Here, we explore the structural basis of inhibition by KdpBS162 phosphorylation by determining the conformational landscape of KdpFABC under inhibiting and non-inhibiting conditions. Under turnover conditions, we identified a new inhibited KdpFABC conformation that we termed E1-P tight, which is not part of the canonical Post-Albers transport cycle of P-type ATPases. It likely represents the biochemically described stalled E1-P state adopted by KdpFABC upon KdpBS162 phosphorylation. The E1-P tight state exhibits a compact fold of the three cytoplasmic domains and is likely adopted when the transition from high-energy E1-P states to E2-P states is unsuccessful. This study represents a structural characterization of a biologically relevant off-cycle state in the P-type ATPase family and supports the emerging discussion of P-type ATPase regulation by such conformations.

Proteins targeted to the secretory pathway are involved in a myriad of biological processes but can only do so when properly folded. Within the endoplasmic reticulum, glycoprotein folding is regulated by the enzyme UDP-glucose:glycoprotein glucosyltransferase (UGGT) and its oxidoreductase partner, the 15-kDa selenoprotein (SEP15 aka SELENOF). The interaction between these two chaperones is poorly understood, limiting understanding of their function. SEP15 is comprised of two domains, a C-terminal thioredoxin-like domain, the structure of which has been reported (PDB 2A4H), and an approximately 50-residue long N-terminal cysteine-rich domain (CRD), of unknown structure. Here, we use a combination of AlphaFold structural predictions and NMR spectroscopy to elucidate the structure of the SEP15 CRD, which mediates the interaction with UGGT. These data reveal that this domain forms a previously undescribed helical fold stabilized by three disulfide bonds between residues C10-C42, C21-C43, and C24-C39. Furthermore, our results validate our reported model of the UGGT/SEP15 complex and lay the foundation for future studies of its interaction with glycoprotein substrates.

We employed NMR spectroscopy to investigate the structure and dynamics of the class B J domain protein (JDP) of S. cerevisiae (Sis1) complexed with an EEVD peptide of HSP70. It is widely recognized that the interactions between the EEVD motif and Sis1 play a crucial role in the chaperone activity. Notably, the deletion of the EEVD impairs the ability of Sis1 to bind with HSP70, while leaving the interaction between the class A JDP Ydj1 and HSP70 unaffected. Leveraging the advantages of NMR, which is particularly suitable for studying transient interactions, we provide compelling evidence that the EEVD motif transiently engages multiple sites on Sis1. Our findings revealed that EEVD binds to two distinct sites within the C-terminal domain I (CTDI) of Sis1. The interaction at these sites plays a crucial role in anchoring HSP70 to Sis1 at site I, as well as displacing the client protein at site II. Notably, site II is also the binding site for the client protein, and its displacement occurs through competition with the binding to site II. In addition to these interactions, we observed that EEVD, as a transient electrostatic binder, also interacts with the J domain and the GF-rich loop located between the J domain and -helix 6. We propose that the interaction between EEVD and Sis1 facilitates the dissociation of -helix 6, promoting a conformational state that is more favorable for interaction with HSP70 at the nucleotide-binding domain (NBD) and substrate-binding domain (SBD) interface. We also employed -synuclein as a substrate to investigate the competitive nature between EEVD and the client protein. Our experimental findings provided evidence supporting the interaction of EEVD with the client protein at multiple sites. Our findings contribute essential insights into the mechanistic cycle of class B JDPs, paving the way toward a more complete understanding of the primary function of Sis1, which is the transfer of the client protein to HSP70, where multiple site transient interactions play a collective role.

Eukaryotic ribosomes are enriched with pseudouridine, particularly at the functional centers targeted by antibiotics. Here we investigated the roles of pseudouridine in aminoglycoside-mediated translation inhibition by comparing the structural and functional properties of the wild-type ribosomes and those lacking pseudouridine (cbf5-D95A). We showed that the cbf5-D95A ribosomes have decreased thermostability and high sensitivity to aminoglycosides. When presented with an internal ribosome entry site (IRES) RNA, elongation factor eEF2, GTP, sordarin, hygromycin B preferentially binds to the cbf5-D95A ribosomes during initiation by blocking eEF2 binding and stalls the ribosomes in a non-rotated conformation, further hindering translocation. Hygromycin B binds to the inter-subunit bridge B2a that is known to be sensitive to pseudouridine, revealing a functional link between pseudouridine and aminoglycoside inhibition. Our results suggest that pseudouridine enhances both thermostability and conformational fitness of the ribosomes, thereby influencing their susceptibility to aminoglycosides. HighlightsO_LILoss of pseudouridine increases cell sensitivity to aminoglycosides C_LIO_LIPseudouridine enhances ribosome thermostability C_LIO_LIHygromycin B competes with eEF2 for the non-rotated ribosome C_LIO_LIHygromycin B deforms the codon-anticodon duplex C_LI

The E. coli cytidine repressor (CytR) is a member of the LacR family of bacterial repressors that regulates nine operons with distinct spacing and orientations of recognition sites. Understanding the structural features of the CytR DNA-binding domain (DBD) when bound to DNA is critical to understanding differential mechanisms of gene regulation. We previously reported the structure of the CytR DBD monomer bound specifically to half-site DNA and found that the DBD exists as a three-helix bundle containing a canonical helix-turn-helix motif, similar to other proteins that interact with DNA [Moody, et al (2011), Biochemistry 50:6622-32]. We also studied the free state of the monomer and found that since NMR spectra show it populates up to four distinct conformations, the free state exists as an intrinsically disordered protein (IDP). Here, we present further analysis of the DBD structure and dynamics in the context of full-site operator or nonspecific DNA. DBDs bound to full-site DNA show one set of NMR signals, consistent with fast exchange between the two binding sites. When bound to full-length DNA, we observed only slight changes in structure compared to the monomer structure and no folding of the hinge helix. Notably, the CytR DBD behaves quite differently when bound to nonspecific DNA compared to LacR. A dearth of NOEs and complete lack of protection from hydrogen exchange are consistent with the protein populating a flexible, molten state when associated with DNA nonspecifically, similar to fuzzy complexes. The CytR DBD structure is significantly more stable when bound specifically to the udp half-site substrate. For CytR, the transition from nonspecific association to specific recognition results in substantial changes in protein mobility that are coupled to structural rearrangements. These effects are more pronounced in the CytR DBD compared to other LacR family members.

The nuclear receptor peroxisome proliferator-activated receptor gamma (PPAR{gamma}) regulates transcription via two activation function (AF) regulatory domains: a ligand-dependent AF-2 coregulator interaction surface within the C-terminal ligand-binding domain (LBD), and an N-terminal disordered AF-1 domain (NTD or A/B region) that functions through poorly understood structural mechanisms. Here, we show the PPAR{gamma} AF-1 contains an evolutionary conserved Trp-Pro motif that undergoes cis/trans isomerization, populating two long-lived conformations that participate in intradomain AF-1 and interdomain interactions including two surfaces in the C-terminal LBD ({beta}-sheet and the AF-2 surface), which are predicted in AlphaFold 3 models but not AlphaFold 2. NMR and chemical crosslinking mass spectrometry show that interdomain interactions occur for soluble isolated AF-1 and LBD proteins, as well as in full-length PPAR{gamma} in a phase separated state. Mutation of the region containing the Trp-Pro motif, which abrogates cis/trans isomerization of this region, impacts LBD interaction and reduces basal PPAR{gamma}-mediated transcription and agonist-dependent activation of PPAR{gamma}. Our findings provide structural insight into published in vitro and cellular studies that reported interdomain functional communication between the PPAR{gamma} AF-1 and LBD suggesting some of these effects may be mediated via AF-1/LBD interactions.

The intermediate filament (IF) protein vimentin is widely distributed in various cell types in the body and is vital for the proper maintenance of the cell cytoskeletal architecture, yet an extensive structural characterization of its head and tail domains remains elusive. Alterations in the assembly and organization of vimentin IFs, including filament network reorganization, have been associated with several diseases including cataracts, myopathies, and metastatic cancer. The C-terminal tail domain of vimentin is of increasing interest as it is essential for regulating the structure and mechanical properties of filament networks through interactions with divalent metal ions, but the molecular basis of these tail domain-metal interactions have not been characterized. In this work, we perform an in-depth analysis of the structural and metal binding properties of fragments of the vimentin tail domain. Mass spectrometry, and UV-Vis and circular dichroism (CD) spectroscopy reveal the direct binding of divalent copper (Cu(II)) to the last 11 residues of the tail domain. Solution nuclear magnetic resonance (NMR) and CD measurements show that in isolation, the complete vimentin tail domain is primarily disordered, and that Cu(II)-binding involves both the last 11 residues and another segment in the middle of the tail domain. Aside from these binding sites, Cu(II) does not induce any significant ordering of the tail domain. These findings further support the tail domain serving as a key metal binding region of vimentin and provide new insights into the important interplay between the tail domain and metals in vimentin IF physiology and pathology. O_FIG O_LINKSMALLFIG WIDTH=123 HEIGHT=200 SRC="FIGDIR/small/648257v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@fe39d0org.highwire.dtl.DTLVardef@85958eorg.highwire.dtl.DTLVardef@1dda6corg.highwire.dtl.DTLVardef@1ef0749_HPS_FORMAT_FIGEXP M_FIG C_FIG

Several proteins from the human proteome have been observed to adopt multiple distinct amyloid filaments, and specific protofilament folds are associated with different diseases. Thereby, it has become necessary to compare pairs of amyloid structures of a given protein. This paper describes the amyloid packing difference (APD), which quantifies the difference between such a pair as the percentage of residues that are involved in unique cross-{beta} packing interactions or that have different side chain orientations relative to the {beta}-strands. Clustering of -synuclein protofilament folds on pairwise APD values recapitulates previously reported clustering based on structural superpositions. Any pair of known protofilament folds of the prion protein, tau, -synuclein, TDP-43 or TAF15 from different diseases have APD values above 20%, whereas all pairs of structures that have been associated with the same disease have APD values below 40%. These observations provide context for the interpretation of APD values of new comparisons.

The omicron variant of SARS-CoV-2 is responsible for the COVID-19 pandemic, serving as a significant origin for the variants still being detected today. It affects the spike protein that most vaccines used to target when the Omicron strain was discovered. Here, we demonstrate that the receptor binding domain (RBD) of the Omicron variant of SARS-CoV-2 exhibits an increased affinity for human angiotensin-converting enzyme type 2 (hACE2) as a viral cell receptor compared to the prototype RBD. We also identified that {beta}- and {gamma}-actin are Omicron-specific binding partners of RBD. Protein complex predictions suggested that many of the Omicron-specific amino acid substitutions might be involved in the affinity of RBD and actin. Accordingly, we highlight the intriguing observation that proteins expected to localize to different cellular compartments exhibit strong binding.

SummarySite-saturation mutagenesis experiments have been transformative in our study of protein function. Despite the rich data generated from such experiments, current tools for processing, analyzing, and visualizing the data offer only a limited set of static visualization tools that are difficult to customize. Furthermore, usage of the tools requires extensive experience and programming knowledge, slowing the research process for those in the biological field who are unfamiliar with programming. Here, we introduce mutagenesis-visualization, a Python package for creating publication-quality figures for site-saturation mutagenesis datasets without the need for prior Python or statistics experience, where each of the graphs is generated with a one-line command. The plots can be rendered as native Matplotlib objects (easy to stylize) or Plotly objects (interactive graphs). Additionally, the software offers the possibility to visualize the datasets on Pymol. Availability and implementationThe software can be installed from PyPI or GitHub using the pip package manager and is compatible with Python [≥] 3.8. The documentation can be found at readthedocs and the source code on GitHub.

Hsp70 chaperones are crucial for maintaining protein homeostasis by regulating the stability and conformational states of client polypeptides through cycles of their binding and release. These cycles require conformational transitions of Hsp70 driven by ATP binding and hydrolysis. The ATPase activity of Hsp70 is controlled by J-domain protein (JDP) cochaperones, which allosterically stimulate ATP hydrolysis via interactions between their J-domains and Hsp70. The J-domain binds at the interface between the nucleotide (NBD) and substrate (SBD) binding domains of ATP bound Hsp70. Although, it was established that the JD interaction involves residues of helices II and III, and the interhelical loop critical for ATPase stimulation, the mechanism by which the allosteric signal induced by J-domain binding is transmitted to the distal nucleotide-binding pocket of Hsp70 remains unclear, as do the conformational changes leading to the ATP hydrolysis. Here, we addressed these questions by means of all-atom free energy simulations and dynamic network analysis, starting from the crystal structures of ATP-bound Hsp70 DnaK alone and in complex with the J-domain of DnaJ. We demonstrated that the presence of the J-domain results in the rearrangement of the nucleotide-binding pocket into a hydrolysis competent state, characterized by close contact between universally conserved T199 of NBD and {gamma}-phosphate of ATP. With network analysis we revealed that the allosteric signal for this rearrangement is transmitted along the {beta}-strand containing T199. Finally, we provide rationale for the signal transmission, where steric repulsion between the J-domains helix III and SBD induces a push of the T199 containing {beta}-strand. Overall, our study provides mechanistic insights into allosteric signal transmission within Hsp70, bridging the gap between J-domain binding and ATPase stimulation. TOC Graphic O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=109 SRC="FIGDIR/small/655504v2_ufig1.gif" ALT="Figure 1"> View larger version (36K): org.highwire.dtl.DTLVardef@1ff0f30org.highwire.dtl.DTLVardef@3e4157org.highwire.dtl.DTLVardef@133f159org.highwire.dtl.DTLVardef@12a35c2_HPS_FORMAT_FIGEXP M_FIG C_FIG

Metamorphic proteins, which can adopt multiple stable conformations, challenge the traditional understanding of protein structure and function. KaiB is a metamorphic protein that regulates the circadian clock, a central regulator governing gene expression in most light-perceiving organisms on Earth. An interesting aspect is that the circadian clock can be reconstituted in vitro by mixing Kai proteins (KaiA, KaiB, and KaiC) with ATP and Mg2+. The phosphorylation state of KaiC oscillates with a 24-hour period. The fold-switched form of KaiB binds to KaiC to activate the dephosphorylation of KaiC, while the other fold of KaiB dissociates from KaiC, allowing phosphorylation to be activated by the binding of KaiA to KaiC. To understand the metamorphic process of KaiB, we utilized AlphaFold2, a protein structure prediction program, and sequence alignments. We found that a proline residue determines the fold of KaiB. We also confirmed that mutating this proline to lysine changes KaiB to a fold-switched conformation. This validates that AlphaFold2 can be used for the study of metamorphic proteins.

Dysregulation of the biosynthesis of cholesterol and other lipids has been implicated in neurological diseases, including Parkinson's disease, where the misfolding of membraneassociated -Synuclein is a key molecular event. Recent research also suggests that -Synuclein aggregation is influenced by the lipid environment. The exact molecular mechanisms responsible for cholesterols effect on -Synuclein binding to lipids and how this binding may affect -Synuclein oligomerization and fibrillation remain elusive, as does the relative importance of cholesterol versus other lipid factors. We probed the interactions and fibrillation behaviour of -Synuclein using SMA nanodiscs, containing zwitterionic and anionic lipid model systems with and without cholesterol. SPR and ThT fluorescence assays were then employed to monitor -Synuclein binding, as well as fibrillation in the absence and presence of membrane models. 1H-15N correlated NMR was used to monitor the fold of -Synuclein in response to nanodisc binding, and we determined individual residue apparent affinities for the nanodisc-contained bilayers. Cholesterol inhibited -Synuclein interaction with lipid bilayers. We also find that cholesterol significantly promotes -Synuclein fibrillation, with a more than 20-fold reduction of lag-times before fibrillation onset. When -Synuclein-bilayer interactions were analysed for individual residues by solution-state NMR, we observed two different effects of cholesterol. In nanodiscs made of DOPC, cholesterol modulated the NAC part of -Synuclein, leading to stronger interaction of this region with the lipid bilayer. In contrast, in the nanodiscs comprising DOPC, DOPE and DOPG, the NAC part was mostly unaffected by cholesterol, while the binding of the N-terminal and the C-terminal were both inhibited.\n\n\n\nO_FIG O_LINKSMALLFIG WIDTH=194 HEIGHT=200 SRC=\"FIGDIR/small/725762v3_ufig1.gif\" ALT=\"Figure 1\">\nView larger version (54K):\norg.highwire.dtl.DTLVardef@1c9c126org.highwire.dtl.DTLVardef@a7e026org.highwire.dtl.DTLVardef@16d1a9forg.highwire.dtl.DTLVardef@1eeda2d_HPS_FORMAT_FIGEXP M_FIG C_FIG

Claspin, known to be highly disordered, plays important roles in replication fork progression, initiation and cellular responses to replication stress. However, regulation of its structure and molecular interactions is not completely understood. We show here, through Proximity-Ligation-Assays, the evidence for intramolecular interaction between the N- and C-terminal segments of Claspin, which depends on the Acidic-Patch [AP] segment near its C-terminus. Interaction of Claspin with DNA and replication factors is highly stimulated in {Delta}AP mutant and by prior dephosphorylation. The wild-type Claspin inhibits the helicase activity of MCM in an AP-dependent manner. {Delta}AP and dephosphorylated Claspin exhibit resistance to trypsin digestion compared to wild-type, suggesting the presence of structural domains in the formers. We propose that Claspin is converted from disordered (closed) to structured (open) conformation at initiation, which stimulates its DNA binding and interaction with replication factors and counteracts its helicase inhibitory activity to trigger initiation of DNA replication.

Many multi-domain proteins including the serpin family of serine protease inhibitors contain non-sequential domains composed of regions that are far apart in sequence. Because proteins are translated vectorially from N-to C-terminus, such domains pose a particular challenge: how to balance the conformational lability necessary to form productive interactions between early and late translated regions while avoiding aggregation. This balance is mediated by the protein sequence properties and the interactions of the folding protein with the cellular quality control machinery. For serpins, particularly 1-antitrypsin (AAT), mutations often lead to polymer accumulation in cells and consequent disease suggesting that the lability/aggregation balance is especially precarious. Therefore, we investigated the properties of progressively longer AAT N-terminal fragments in solution and in cells. The N-terminal subdomain, residues 1-190 (AAT190), is monomeric in solution and efficiently degraded in cells. More y-rich fragments, 1-290 and 1-323, form small oligomers in solution, but are still efficiently degraded, and even the polymerization promoting Siiyama (S53F) mutation did not significantly affect fragment degradation. In vitro, the AAT190 region is among the last regions incorporated into the final structure. Hydrogen-deuterium exchange mass spectrometry and enhanced sampling molecular dynamics simulations show that AAT190 has a broad, dynamic conformational ensemble that helps protect one particularly aggregation prone y-strand from solvent. These AAT190 dynamics result in transient exposure of sequences that are buried in folded, full-length AAT, which may provide important recognition sites for the cellular quality control machinery and facilitate degradation and, under favorable conditions, reduce the likelihood of polymerization.

Cytokines primarily interact with specific cytokine receptors on the cell surface as essential signal transduction pathways in many physiological and pathological processes. Therapeutic agents targeting cytokine-cytokine receptor (CK-CKR) interactions lead to the disruption in cellular signaling function and have been demonstrated effective in the treatment of many diseases including tumors. However, a lack of universal and quick access to annotated structural surface regions on CK/CKR has limited the progress of a structure-driven approach to the development of targeted macromolecular drugs and precision medicine therapeutics. Herein we develop CytoSIP (Single nucleotide polymorphisms (SNPs), Interface, and Phenotype), a rich internet application based on a database of atomic interactions around hotspots in experimentally determined CK/CKR structural complexes. The content of the CytoSIP database includes the following key features: (1) SNPs on CK/CKR; (2) interactions involving CK/CKR at the domain level, including CK/CKR interfaces, oligomeric interfaces, epitopes, or other drug targeting surfaces; and (3) diseases and phenotypes associated with CK/CKR or SNPs. The database introduces a unique tri-level SIP data model to link genetic variants (atomic level) to disease phenotypes (organism level) using protein structure (complexes) as an underlying framework (molecule level). Moreover, CytoSIP implements screening criteria and tools to allow customized selection of relevant subset of CK/CKR for the study of interest. This reduces the time and resources needed to interrogate large datasets and allows rapid screening of cytokines and cytokine receptor proteins interfaces for hotspots targeted drug design and any other specific cellular signaling/function mechanisms and their correlation to pathologies. The CytoSIP framework crafted herein bridges CK/CKR genotype with phenotype, facilitating not only the panoramic investigation of the context-dependent crosstalk between CK/CKR but also the development of targeted therapeutic agents. CytoSIP portal website is publicly accessible at https://CytoSIP.biocloud.top.

In vivo, the majority of nascent protein chains must begin folding during translation in order to obtain their native structure. While the importance of co-translational folding has become increasingly clear, the specific mechanisms underlying the coordination between the ribosome, nascent chain and molecular chaperones are still uncertain. Here, we have constructed a model of the co-translational folding pathway of the calcium responsive domain (CRD) of the human neuronal KV7.2 human neuronal ion channel, showing that calmodulin (CaM) is crucial. By combining Force Profile Analysis and single-molecule force spectroscopy techniques, we found that CaM, in a calcium-dependent manner, affects early folding events involving three key -helices in the CRD. In addition, this study suggests that CaM at early stages induces the formation of metastable hairpins, as a part of the co-translational folding pathway. These findings expand on the role of CaM as a key regulator of folding in eukaryotes: not only as an essential cellular signaling protein, but also as a bona fide co-translational molecular chaperone.

Ferroportin 1 (FPN1) is the only known mammalian iron efflux transporter. This multi-pass membrane protein, which adopts the Major Facilitator Superfamily fold, is tightly controlled by serum hepcidin to assure maintenance of adequate cellular and systemic iron levels. Earlier studies have shown that cholesterol-lowering drugs can reduce FPN1 expression in liquid-ordered plasma membrane microdomains and its sensitivity to hepcidin. However, the molecular mechanism by which cholesterol depletion regulates the localization of FPN1 at the cell surface remains unknown. In biochemical experiments, we show that cholesterol depletion reduces the iron export function of FPN1. Repletion with cholesterol restores FPN1 activity. This is not observed with the diastereoisomer epicholesterol, suggesting a direct interaction between cholesterol and FPN1. Consistent with this, we demonstrate that mutants affecting the key tyrosine residues of three cholesterol-recognition amino acid consensus (CRAC/CARC) motifs have a negative impact on FPN1 activity, in manner that also decreases its abundance in ordered plasma membrane microdomains. A complementary structural analysis allows us to focus on a conserved CARC motif (CARC-1) located in a deep hydrophobic groove between transmembrane helices 1 and 5. Molecular docking suggests that this groove is well suited to cholesterol binding. All these findings indicate that the interaction between FPN1 and cholesterol is of major importance for the localization of FPN1 in ordered microdomains of the plasma membrane, which is necessary for its optimal activity, and so its responsiveness to hepcidin.

Transcription by RNA polymerase II (Pol II) requires general transcription factors that bind with Pol II at the promoters of protein-coding genes to form preinitiation complexes (PICs). Among these is TFIIE, which recruits TFIIH to the PIC and stimulates the kinase and translocase activities of TFIIH, thereby regulating the fate of formed PICs. In this study, we used a purified reconstituted human Pol II transcription system and single molecule total internal reflection fluorescence (smTIRF) microscopy to monitor TFIIE binding dynamics in PICs under different conditions in real time. We observed highly dynamic interactions of the two subunits of TFIIE (TFIIE and TFIIE{beta}) with PICs. Measurement of rate constants for on/off binding of each subunit suggest they behave asynchronously. TFIIH exclusion increased the rates of association and dissociation for both subunits, with the strongest effect on TFIIE. Despite stabilization of TFIIE by TFIIH the TFIIE subunits remain dynamic in PICs. Additionally, two disease-related TFIIE{beta} point mutations destabilized TFIIE{beta} and altered its kinetic behaviors within PICs. Our results contribute to an emerging model that PICs are not static assemblies and highlight important connections between the structural arrangement and kinetic behaviors of GTFs in PICs.

Protein stability datasets contain neutral mutations that are highly concentrated in a much narrower {Delta}{Delta}G range than destabilizing and stabilizing mutations. Notwith-standing their high density, often studies analyzing stability datasets and/or predictors ignore the neutral mutations and use a binary classification scheme labeling only destabilizing and stabilizing mutations. Recognizing that highly concentrated neutral mutations would affect the quality of stability datasets, we have explored three protein stability datasets; S2648, PON-tstab and the symmetric Ssym that differ in size and quality. A characteristic leptokurtic shape in the {Delta}{Delta}G distributions of all three datasets including the curated and symmetric ones were reported due to concentrated neutral mutations. To further investigate the impact of neutral mutations on {Delta}{Delta}G predictions, we have comprehensively assessed the performance of eleven predictors on the PON-tstab dataset. Correlation and error analyses showed that all of the predictors performed the best on the neutral mutations while their performance became gradually worse as the {Delta}{Delta}G of the mutations departed further from the neutral zone regardless of the direction, implying a bias towards dense mutations. To this end, after unraveling the role of concentrated neutral mutations in biases of stability datasets, we described a systematic under-sampling approach to balance the {Delta}{Delta}G distributions. Before under-sampling, mutations were clustered based on their biochemical and/or structural features and then three mutations were systematically selected from every 2 kcal/mol of each cluster. Upon implementation of this approach by distinct clustering schemes, we generated five subsets varying in size and {Delta}{Delta}G distributions. All subsets notably showed amelioration of not only the shape of {Delta}{Delta}G distributions but also other pre-existing imbalances in the frequency distributions. We also reported differences in the performance of the predictors between the parent and under-sampled subsets due to the enrichment of previously under-represented mutations in the subsets. Altogether, this study not only elaborated the pivotal role of concentrated mutations in the dataset biases but also contemplated and realized a rational strategy to tackle this and other forms of biases. Under-sampling code is available on GitHub (https://github.com/narodkebabci/gRoR).