Alterations in secondary structure and binding affinity from mutations in the interleukin-1 receptor antagonist

2.1. Protein Sequence Acquisition

All cytogenetic abnormalities of known DIRA patients have been displayed in Table 1. As of August 8, 2024, there have been 34 DIRA patients and 16 variants that have been reported within literature from 22 studies [8–29]. Of all genetic variants, we identified 6 unique nonsense, 3 frameshift, 3 large deletions, 1 splice-site, 1 in-frame deletion, and 2 compound mutations displayed in Table 1. We precisely obtained these counts from a manual search on the PubMed database with the key-words on August 8, 2024: “Deficiency of Interleukin-1 Receptor Antagonist” OR “DIRA”. After review of literature, we found that most protein sequences were not directly listed from the publications or their supplementary materials. However, all publications included information on alterations from the wild type (WT) reference mRNA sequence provided by the National Center for Biotechnology Information (NCBI). All variants were directly referred to from transcript variant 1 (NM_173842.3) (n = 15) and variant 2 (NM_173841.2) (n = 1) of the IL1RN mRNA. After manually creating the mutations in each WT sequence, we verified most sequences with the Infevers database (http://fmf.igh.cnrs.fr/infevers/) [30]. As a final verification of the sequences, we translated the mutant mRNA sequences using an automated translation tool (https://web.expasy.org/translate/) to ensure they matched the reported amino acids and positions from individual publications. Overall, this study did not require consent or approval from an Institutional Review Board as all mRNA and protein sequences were obtained or derived from publicly available sources: NCBI, publications, and the Infevers database. Our analysis focused solely on the mutations reported and their impacts on the structure and binding affinity of the resulting IL-1Ra protein.

However, one limitation that occurred during sequence acquisition was that the large deletion variant of a 175-kb deletion was unable to be formulated. This was due to the complexity of the mutation, where it involved 6 different IL-1 related genes, as well as the lack of data availability and validation of the precise mRNA sequences from publications or Infevers database. Despite this limitation, this study was able to estimate the protein sequences of 15 variants.

2.2. 3D Structure Acquisition

Following acquisition of protein sequences, we followed the validated methodologies reported by Jesus et al. and Altiok et al. to predict alterations in secondary structure of the WT IL-1Ra protein in silico [13–14]. Both studies used the resolved protein structure of IL-1Ra in its binding mode to IL-1R1 from Schreuder et al. found with X-Ray crystallography [31]. Given the complexity of the mutations reported for DIRA as well as the fact that mutagenesis can only be evaluated for missense variants with current technological innovations, these studies naturally only highlighted the impacted regions of the protein, describing impacted β-strands, α-helices, and amino acids necessary for binding. In fact, these two studies did not delve further into detailed alterations in protein stability (i.e. free energy changes) or structural damages (e.g. bond breakages, etc). This study used this same protocol for 15 variants, 2 of which were completed by these studies, and will help validate the workflow of this study. It is important to note that the experimental model is missing the signal peptide cleavage as well as 5 amino acids following this sequence, but does not impact the structural analyses of this study or of the prior.

2.3. Predicting Structural Alterations

Within the WT IL-1Ra structure, the known structure is composed of 12 β-strands and 3-10 α-helices [32]. β1, β4, β5, β8, β9 and β12 are particularly important for the formation of the protein as they are critical to forming its six-stranded β-barrel [32]. From the PDB, the experimental model by Schreuder et al. reported 4 α-helices [31]. In order to streamline the process of identifying the secondary structures and impacted regions from individual sequences, we aligned the sequences and annotated the partial amino acid sequences to their corresponding secondary structure. In Supplementary Material 1, we have provided these annotated structures derived from the Schreuder et al. model.

After annotating all 15 sequences, we recorded the specific secondary structures that were retained from the experimental model to the mutant. We only considered the individual structures to be retained if all WT amino acids were present. Structures that contained partially correct sequences and additional novel amino acids were not considered a part of the count of original structures. Overall, from the 15 genetic variants, we have counted the β-strands and α-helices retained from the original model.

2.4. Predicting Alterations in Binding Sites

From a prior study by Evans et al., authors identified through site-directed mutagenesis that W16, Q20, Y34, Q36, and Y147 on IL-1Ra are essential for its binding affinity to IL-1R1, and later validated by the Schreuder et al. model [31, 33]. Evans et al. reported that upon binding of IL-1Ra, the free energy change of the entire structure was 52 kJ/mol, and found that 4 of the 5 sites (W16, Q20, Y34, and Y147) accounted for at least 18.3 kJ/mol (35.2%) [33].

At the binding interface of IL-1Ra, W16 forms hydrophobic interactions at domain 2 of the receptor, important for orientation and structural integrity of the cytokine. Q20 forms two hydrogen bonds with the main chain of the receptor on domain 2, critical for stabilization of the binding conformation. At the receptor side chain, Y34 similarly forms a hydrogen bond. Similarly, on domains 1 and 2 of IL-1R1, Q36 forms 3 hydrogen bonds likely for stabilization.

Lastly, Y147 does not directly interact with domain 2 of the receptor but uniquely with 2 water molecules to form a hydrogen bond with IL-1R1.

Overall, this study annotated the binding sites present in the WT IL-1Ra and its variants from the aligned sequences. This information has also been included in Supplementary Material 1. We have comprehensively counted the binding sites present in the sequences among the variants.

2.5. Validation of Structural Analyses

To further validate our findings of impacts on secondary structure and binding affinity, evidence from in vitro studies are necessary. However, considering the challenges of obtaining the individual patient data or performing a broad experimental protocol, it is not feasible to conduct these experiments and beyond the scope of this study. Yet, prior studies have already performed experiments that can help further validate the conclusions presented by this study. Within this work, we delve into the individual variants in the context of the IL1RN mRNA and subsequent protein expression from quantitative polymerase chain reaction (qPCR) and Western blotting experiments, respectively. If neither experiments were reported within studies, or any applicable experiments, we have noted this as well.

3.1. Nonsense Variants

Of the DIRA patients in literature, 6 unique nonsense variants have been reported up to date. The first variants were reported by Aksentijevich et al. including Q54X and E77X in 2009 [8]. From our prediction model, the Q54X variant would retain β1- and β2-strands and no α-helices, while E77X would have β1-, β2-, β3-strands and 1 α-helix. In terms of binding affinity, most sites were missing on Q54X, present with only W16 and Q20, while E77X was only missing Y147. Given that both variants were missing most critical secondary structures, and a significant number of binding sites, it would be expected that there would be no expression of the resulting proteins or at most of a non-functional truncated protein. From the clinical trial and the individual case report, unfortunately there were no experiments performed to evaluate the expression of the mRNA or protein expression of the Q54X variant [8, 15]. However, E77X was comprehensively evaluated in the clinical trial alone [8]. From qPCR, the heterozygote E77X variants showed diminished expression of IL1RN mRNA (n = 2), and even less in the homozygote (n = 1). In western blotting, a band corresponding to IL-1Ra was present but diminished in the heterozygotes, and completely missing in the homozygote.

Next, Ulusoy et al. reported a R26X variant in 2015 [16]. From the in silico model, it was predicted that no β-strands or α-helices would be retained from the WT, and therefore no binding sites would be present. Similarly, Kuemmerle-Deschner et al. reported a S21X variant in 2020, and had the same outcome of prediction [11]. The reasoning for these results is that the nonsense mutations occur in the signal peptide cleavage not present in the Schreuder et al. experimental model. Although the R26X had no in vitro experiments, S21X did have supporting data. From the S21X variant, there was induction of IL1RN mRNA expression, but no expression from the resulting IL-1Ra protein.

Lastly, the remaining two variants reported were by Abdwani et al. in 2021 and Altiok et al. in 2018 for a Q122X and Q119X variant, respectively [14, 21]. The prediction model displayed that the Q122X variant would retain β1-, β2-, β3-, β4-, β5-, β6-, and β7-strands with the first two α-helices, and therefore only missing the Y147 binding site. Unfortunately, this variant did not have any further experimental data presented. Conversely, the Q119X was predicted to retain the same secondary structures and binding sites as the Q122X variant. The potential explanation for these sequences being similar is that Q122X variant refers to different protein isoform that differs only in the segment prior to the signal peptide cleavage but retains the same amino acid sequences from the Schreuder et al. experimental model, and was the primary reason both structures were able to be predicted. Furthermore, our findings for the Q119X variant were validated by Altiok et al. that stated that the “β8-, β9- and β12-strands are missing” from the β-barrel of IL-1Ra [14].

3.2. Large Deletion and Splice-Site Variants

From the known DIRA patients, we predicted the alterations in structure and binding affinity of 2 unique large deletions and 1 splice-site variant. The first large deletion variant was first reported by Mendonca et al. in 2017 for a 22,216 bp deletion [9]. The authors reported that the mutation spanned the promoter region and the first 2 coding exons of the IL1RN gene, with no in vitro protein expression testing performed in the study. From our prediction of this variant, there would be no original β-strands, α-helices, or binding sites retained given the promoter sequence has been removed.

The other splice-site and large deletion variants, specifically E69GfsX2 and I40QfsX6, have been recently reported by Urbanaja et al. in 2024 [22]. For the E69GfsX2, the variant was predicted to retain the β1-, β2-, β3-strands and the first 1 α-helix, only missing the Y147 binding site total. Likewise, for the I40QfsX6 variant, no β-strands, α-helices, or binding sites are retained. From the study, the authors indirectly tested for protein expression, measuring IL1RN mRNA levels and gene expression from RNA-seq for both variant types. Both variants were found to have decreased levels of mRNA expression compared to the WT.

3.3 Frameshift Variants

There have been a total of 3 frameshift variants that have been reported for DIRA to date. The first variant was N52KfsX25, reported by Aksentijevich et al. in 2009 [8]. From the prediction model, the variant was expected to retain β1- and β2-strand but no α-helices, and W16 and Q20. Similarly to the 2 other variants reported from this clinical trial, analysis of IL1RN mRNA expression revealed diminished expression in the heterozygotes and even less in the homozygote. Likewise, western blotting revealed that the band of the heterozygotes were present but still diminished, and the homozygote being completely absent.

The other two variants reported, T133PfsX118 and N52KfsX4, were identified by Sözeri et al. and Ziaee et al. in 2018 and 2020, respectively [12, 20]. Unfortunately, no experimental studies were conducted in either of these case reports. From the T133PfsX118 variant, it is predicted that the structure would be missing β8-, β9-, β10-, β11-, and β12-strands and the last α-helix, ultimately retaining all major binding sites except Y147. For the N52KfsX4 variant, the structure is expected to only retain β1- and β2-strands, containing W16 and Q20.

3.4. In-frame Deletions & Compound Variants

Within current literature, there has been 1 variant with an in-frame deletion and 2 others reported for compound mutations. Jesus et al. first reported a D72_I76del variant in 2011, comprehensively reporting the expression of the resulting IL1RN mRNA and subsequent IL-1Ra [13]. From the prediction model, the deletion would impact most of the β4-strand in the WT structure. Like the data presented for other variants, mRNA expression was diminished in heterozygous parents, while the homozygotes themselves were even less. Interestingly, however, the variant was expressed in lower levels than the heterozygous parents and anakinra itself (positive control) during western blotting. However, from further tests, no binding was found between the mutant IL-1Ra and its IL-1R1. Additionally, our prediction analysis was also validated by the findings of this study, where the authors stated the “missing amino acids are in strand β4”.

Furthermore, the two compound variants were reported by Stenerson et al. in 2011 and Mendonca et al. in 2020 [18–19]. Stenerson et al. evaluated a E77X and unique T47TfsX4 variant, where only the β1-strand was retained and no α-helices, along with 2 binding sites (W16 and Q20). The case report for this patient did predict that the protein would be truncated to be ⅓ the size of the WT (177 amino acids), which was close to those predicted from this study of 49 amino acids. However, no other in vitro studies experiments were performed for this variant. Lastly, Mendonca et al. reported a variant containing a Q45X and I71_P75del mutation. From the structural analysis of this study, it retained the same number of original β-strands and α-helices to the other compound variant, but had only kept the W16 binding site.