Determinants of trafficking, conduction, and disease within a K+ channel revealed through multiparametric deep mutational scanning

  1. Willow Coyote-Maestas
  2. David Nedrud
  3. Yungui He
  4. Daniel Schmidt

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This is a tour de force for mutagenesis and analysis of an ion channel protein, using a straightforward method the authors have developed for the comprehensive functional analysis of a deep mutational library. The approach introduced here will not only be of broad interest to the ion channel community, but it will also serve as a roadmap for performing similar studies on other proteins. The authors demonstrate the usefulness of this method by defining the functional domains of Kir2.1, thereby rediscovering known disease causing mutants, and highlighting a number of mutations with similar phenotypes that may also result in disease phenotypes.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

A long-standing goal in protein science and clinical genetics is to develop quantitative models of sequence, structure, and function relationships to understand how mutations cause disease. Deep mutational scanning (DMS) is a promising strategy to map how amino acids contribute to protein structure and function and to advance clinical variant interpretation. Here, we introduce 7429 single-residue missense mutations into the inward rectifier K + channel Kir2.1 and determine how this affects folding, assembly, and trafficking, as well as regulation by allosteric ligands and ion conduction. Our data provide high-resolution information on a cotranslationally folded biogenic unit, trafficking and quality control signals, and segregated roles of different structural elements in fold stability and function. We show that Kir2.1 surface trafficking mutants are underrepresented in variant effect databases, which has implications for clinical practice. By comparing fitness scores with expert-reviewed variant effects, we can predict the pathogenicity of ‘variants of unknown significance’ and disease mechanisms of known pathogenic mutations. Our study in Kir2.1 provides a blueprint for how multiparametric DMS can help us understand the mechanistic basis of genetic disorders and the structure–function relationships of proteins.

Article activity feed

  1. Version published to 10.7554/elife.76903 on eLife
  2. eLife

    Evaluation Summary:

    This is a tour de force for mutagenesis and analysis of an ion channel protein, using a straightforward method the authors have developed for the comprehensive functional analysis of a deep mutational library. The approach introduced here will not only be of broad interest to the ion channel community, but it will also serve as a roadmap for performing similar studies on other proteins. The authors demonstrate the usefulness of this method by defining the functional domains of Kir2.1, thereby rediscovering known disease causing mutants, and highlighting a number of mutations with similar phenotypes that may also result in disease phenotypes.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

  3. eLife

    Reviewer #1 (Public Review):

    Inward-rectifier Kir2.1 K+-permeable channels contribute to establishing the resting membrane potential in multiple cell types and are associated with several human diseases. In these and other proteins, most amino acid determinants for folding, stability, trafficking and function are not well understood. In the present manuscript, Willow Coyote-Maestas and collaborators apply deep mutational scanning and FACS-seq approaches to systematically assess the impact of all possible single-residue substitutions in the Kir2.1 channel on its folding/trafficking to the plasma membrane and its ability to conduct K+ ions. The value of integrating data from both surface-trafficking and function of an ion channel sheds light into clinically-relevant data and structure-function relationships. By using a fluorescence-based high-throughput trafficking assay, the authors identify multiple regions in the Kir2.1 channel that are highly sensitive to perturbation and that are organized into structural domains mainly involving intra-subunit interactions. Using a membrane-intercalating, charged dye, the authors also assay the ability of the channel variants to conduct K+. The results from the trafficking and functional assays are generally consistent with known structure-function relations for these proteins and raise multiple intriguing hypothesis related to different aspects of channel function. By comparing the results from both assays, the authors identify that regions associated with channel gating and allosteric regulation are spatially segregated from those associated with folding, stability and trafficking, and have distinct sensitivities to mutational perturbation. The mapping of these regions constitutes a relevant starting point to understand how the requirements for conformational dynamics and structural stability are balanced in these proteins to enable their biological functions. Finally, the authors identify that channel variants with a large negative impact on their surface trafficking ability are underrepresented in human gene variant databases, pointing to the essential function of these proteins and suggesting a strong negative selection for those variants in the population. The authors relate their measured fitness scores with expert-validated benign human variants to establish thresholds to predict the likelihood of pathogenicity for variants of unknown significance present in the databases.

    The data presented is of high quality, there is high reproducibility between replicates, and good coverage depth for the variants. This constitutes the first study of its kind for its size and for quantitating both surface trafficking and ion channel function, yielding reams of new and interesting information. However, for this same reason, a more critical assessment and validation is required to solidly establish this new approach, the significance of the datasets that are reported, and the many specific observations pointed out throughout the work. In general, sources of error are not adequately discussed throughout the work, some examples being sequencing errors, nonspecific antibody binding, and PCR errors. More specific improvements to the presentation, analysis, and discussion of the data are warranted to fully realize the potential of this approach and are detailed below.

    Firstly, the signal-to-noise ratio in these approaches is limited compared to methods with lower throughput. Even in best-case scenarios like this manuscript, there are much fewer individual observations per variant relative to e.g. a FLIPR multi-well assay averaging signals from millions of cells per variant. The FACS experiment for functionality (Suppl. Fig. 3F) in particular has a very limited dynamic range, and no additional data is shown to provide quantitative support for the choice of gates. Without additional data for cells expressing individual channel variants with defective permeation, it is hard to evaluate how accurate the selection assay is in discriminating between channels with loss of function or gain of function phenotypes. The results of the surface-trafficking experiment did yield distinct populations providing more confidence in the ability of the assay to discriminate between phenotypes, which is an excellent result. We unfortunately do not see similarly distinct populations in the function-based selection, suggesting that it has an overall decreased sensitivity.

    A concern is that no effective discussion or quantitation is provided to assess to extent to which there are variants with solidly negative surface expression scores and neutral or positive function scores. Mutations that severely disrupt folding or trafficking generally also lead to a complete lack of functional channels in the plasma membrane, as pointed out by the authors on lines 390-393, such that function fitness scores are expected to correlate strongly with variants with very negative surface expression scores. Analysis of this aspect of the data could provide an excellent means to critically assess the accuracy of the functional fitness scores.

    The use of the literature as means to test the accuracy of observations is not very effective and it is unclear how systematic the comparisons are between the experimental results presented here and phenotypes described in the literature using other approaches. Only observations from the literature that agree with the authors' observations are discussed, without addressing whether there are also discrepancies. Further, discussion of published experimental observations are hard to trace back to the data as presented in the manuscript, as the figures are very complex. Further, the text does not discuss relevant per-residue scores that seem inconsistent with the authors' interpretation. For example, the functional assay yields permissive results for certain substitutions at the GYG selectivity filter sequence Glycine residues, contrary to the idea that these are immutable. Similarly, substitution of the glycine hinge at position 176 with asparagine or lysine yields positive fitness scores in the permeation assay, counter to the expected flexibility requirements for that position. Analogously, certain substitutions at the PIP2 binding residues appear to be inconsistent with the expected receptor-ligand interactions (e.g., negative residues yield positive function fitness scores at R80 and W81). In general, it is unclear how the robustness of the per-residue results compares with the pre-position averaged data. Establishing this is important to adequately interpret the results and extract the most useful information. Many potentially interesting observations thus remain speculative and would benefit from additional experimental validation. Some of these examples include the increased surface expression of N386E, N386D, T150, Y153, the positive effect in function of N224K and of the residues surrounding the hinge at G176, and importantly, at least some of the predicted phenotypes for the variants of unknown significance.

    The existence of a co-translationally folded biogenic domain is overstated, as the nature of the assay precludes distinctions between different steps along the folding, tetramerization and trafficking of channels, and are based on findings on another protein. It is also unclear if any quantitative parameters were used by the authors to identify the specific regions of the C-terminus required for Golgi export, as some of the secondary structure domains listed on line 157 appear to be more permissive to perturbation than others.

    The distinction between intra- and inter-subunit interactions being primarily associated with surface expression or function could be analyzed in a more systematic manner, and otherwise remains anecdotal to those regions specifically mentioned in the text, rather than the entire protein.

    Finally, limited information on the sequencing method is provided; specifically, the length of the amplicon is not described, nor how stringent the conditions were to discriminate between SPINE-library mutations and those caused by errors at different stages of the study, and the strategy used to assess the presence of additional mutations outside the SPINE-derived variant tiles d. It is also unclear whether replicates reflect different libraries, different recombined cell populations, or different aliquots from one cell library after recombination.

  4. eLife

    Reviewer #2 (Public Review):

    The authors use deep mutational scanning to introduce almost every possible amino acid substitution at every residue in Kir2.1, and then assess the consequences for assembly, trafficking and channel activity, using single cell fluorescent activated cell sorting and deep sequencing approaches. Experimental determination of mutational consequences is a very meaningful advance over purely computational predictions, and the present analysis shows that trafficking mutations are under-represented in such computational variant effect databases. The results provide a deep catalog of the consequences of each mutation that will be of great potential use in assessment of clinical relevance of identified natural variants.

  5. eLife

    Reviewer #3 (Public Review):

    Coyote-Maestas et al. have elegantly and comprehensively addressed pressing questions in the fields of both protein biochemistry and ion channel physiology in this paper, using Kir2.1 as a model system for an application that could be widely adopted to many ion channels. There are, to my mind, two keys elements that any reader should take away from this manuscript. First, they outline a clear model for how to leverage deep mutational scanning (DMS) for the functional analysis of ion channels. DMS analysis has been used historically to map permissive and non-permissive sites for mutational substitutions. These are frequently paired with biophysical or biochemical analysis, and scans are usually limited to sub-domains or small proteins. Rarely do such studies probe the full capacity of a large, multi-subunit protein as done in this paper. Second, they adapt two flow cytometry based functional assays of ion channel activity (surface expression and channel conductance) to fully map the contributions of each mutation in their library to the common physiologically relevant deficits of these proteins driving human disease.

    They have maintained a clear focus on using this screen to fully enumerate and define the domains of Kir2.1 involved in its full maturation and activity. Impressively, their data analysis and highlighted mutational deficiencies map exceptionally well to the already known disease causing variants of Kir2.1. These observations build confidence that this method of analysis can reveal potential disease-causing mutants, including identified variants in this study that exist in the human population but have not yet been characterized as disease relevant (likely due to low population prevalence). In addition to clinical relevance, the authors also enumerate the boundaries of domains that had been previously hypothesized as key elements in the trafficking, export, assembly, and function of Kir2.1 and significantly improve the resolution of which residues and substitutions are deleterious to function and trafficking of this important protein.

    The data within this paper will serve as a blueprint for further dissection, analysis, and engineering of Kir2.1 and other ion channels for many years into the future.

    It should be noted that the mutational coverage in this study is not complete. "No data" is observed for roughly 7% of possible amino acid substitutions in their functional analysis studies. This is likely due to insufficient library coverage of these positions in their originating stable cell lines. This is not unusual for libraries of this size; however, the absence of data is not necessarily random. Specific amino acids and domains have less coverage than others (example, AA350-354, AA251, AA150-190, AA53, etc). Readers should be aware of this as a limitation in the analysis of these domains. While complete, and absolute data of every possible positional substitution would be valuable, the lack of data at these positions does not diminish the overall interpretations, functional observations, and conclusions of the study. Rather, this gaps should simply be carefully considered if using this dataset for comprehensive modeling or interpretation of Kir2.1 or other ion channels in computational or functional characterization efforts.

    Coyote-Maestas et al. achieve a critical milestone in this manuscript, creating a foundational dataset for the biochemical understanding of all residues of an ion channel in the two functions most highly correlated with disease, trafficking and conductance. This paper will serve as a road map to future studies aimed at developing similar datasets in the ion channel field and beyond.

  6. Version published to 10.1101/2022.01.06.475280v1 on bioRxiv