Cm28, a scorpion toxin having a unique primary structure, inhibits K_V1.2 and K_V1.3 with high affinity

Voltage-gated potassium (K_V) ion channels play a key role to maintain the proper physiological functions of both excitable and non-excitable cells. Pharmacological manipulation of these K_V channels has a significant therapeutic prospect in the management of autoimmune diseases, cancer, and neurological and cardiovascular disorders (Coetzee et al., 1999; Cahalan and Chandy, 2009; Panyi et al., 2014; Yang and Nerbonne, 2016; Hofschröer et al., 2021). K_V1.3 channels are expressed in peripheral immune cells and are upregulated in effector memory T (T_EM) cells in states of autoimmunity and inflammation. Their activity maintains the electrical driving force for Ca⁺ entry during T cell activation by the K⁺ efflux counterbalancing the persistent Ca⁺ influx required for proliferation and excessive release of cytokines (Cahalan and Chandy, 2009; Feske et al., 2012). Several studies have validated that specific and persistent blockade of K_V1.3 suppresses the T_EM cell activation and proliferation. This dependence of T_EM cells on K_V1.3 channels for proliferation brings K_V1.3 blockers into the spotlight as a potential therapeutic immunosuppressant to treat a range of autoimmune diseases such as multiple sclerosis, rheumatoid arthritis, type 1 diabetes mellitus, psoriasis, and others (Wulff et al., 2003; Beeton et al., 2006; Panyi et al., 2006; Toldi et al., 2010; Lam and Wulff, 2011; Serrano-Albarrás et al., 2019; Varga et al., 2021). Moreover, recent studies have demonstrated that K_V1.3 channels are also expressed in microglia, brain-resident macrophages, and are essential for their proliferation. Thus, K_V1.3 is also emerging as an attractive drug target in the treatment of neuroinflammatory disorders such as Parkinson’s disease (Sarkar et al., 2020; Tajti et al., 2020; Wang et al., 2020).

Several ion channel modulator peptide toxins have been identified from venomous animals such as charybdotoxin from scorpions, ShK from sea anemone, mambaquaretin-1 from snakes, and ProTx from spiders. Among the scorpion families, the Buthidae family has the most studied venoms due to its great toxicity, and several peptides reported to affect K⁺ channels were purified from them (Rokyta and Ward, 2017). Centruroides margaritatus belongs to this family and until now, only two peptides have been reported from its venom: (1) margatoxin, which blocks different K_V channels (Bartok et al., 2014), and (2) CmERG1, which completely blocks the K_V11.1 channel (Garcia-Calvo et al., 1993; Beltrán-Vidal et al., 2021). K⁺ channel inhibitor scorpion toxins (KTxs) have been classified into seven different families based on their structural and functional features: α-KTx, β-KTx, γ-KTx, δ-KTx, ε-KTx, κ-KTx, and λ-KTx (Tytgat et al., 1999; Rodríguez de la Vega and Possani, 2004; Tabakmakher et al., 2019). The α-KTx family contains peptides with 23–42 amino acids and share a common structural motif known as the cysteine-stabilized α/β scaffold, in which the α-helix and β-sheets are held together by 3–4 disulfide bridges. Based on the sequence similarity, 31 subfamilies of α-KTx were described previously (https://kaliumdb.org). A typical “functional dyad” consists of a critically positioned lysine residue and an aromatic residue nine positions downstream (∼6.6 Å α-carbon-benzene ring center distance in the 3-D structure), which is also considered a common characteristic of these peptides (Dauplais et al., 1997; Rodríguez de la Vega et al., 2003; Panyi et al., 2006). The critical lysine that protrudes into the selectivity filter of the channel is essential for the high-affinity block, and the aromatic residue seems responsible to determine selectivity among K_V1.x channel subtypes (Goldstein and Miller, 1993; Corzo et al., 2008; Papp et al., 2009; Bartok et al., 2015). Interestingly, there are scorpion toxins that inhibit K_V channels despite lacking the functional dyad (Batista et al., 2002; Abdel-Mottaleb et al., 2006). Moreover, for stable toxin–channel interaction, other influential residues of toxins interact pairwise with the channel residues contributing to their selectivity profile among different K_V1.x channel (Aiyar et al., 1995; Mouhat et al., 2004; Varga et al., 2021).

The ongoing discovery of K⁺ channel blocker peptides suggests that scorpion venoms are remarkably rich sources of these peptides. The diverse nature of their primary sequence and valuable therapeutic potential encourage the exploration of novel peptides in different scorpion venoms (Ortiz et al., 2015; Gubič et al., 2021; Varga et al., 2021). A comprehensive characterization of a Colombian scorpion C. margaritatus venom was reported previously by our group in order to investigate its effect on various voltage-gated K⁺ and Na⁺ channels, and a new γ-KTx (CmERG1, γ-KTx 10.1) from C. margaritatus, which fully blocks the human ether-à-gogo-related gene (hERG1) potassium channel (K_V11.1) with high affinity, was also described (Beltrán-Vidal et al., 2021). In this work, during further electrophysiological characterization of C. margaritatus venom components we discovered another exciting peptide, named Cm28. This peptide obeys a unique and unusual primary structure and was shown to be a potent and selective pore blocker of human K_V1.2 and K_V1.3 channels. In addition, Cm28 did not inhibit a panel of ion channels including K_V, voltage-gated sodium (Na_V), and proton (H_V) channels. Cm28 also suppressed human CD4⁺ effector memory T cells activation in vitro by downregulating the IL2R and CD40 ligand expression. Phylogenetic analysis conducted with the amino acid sequence of Cm28 compared with the other known K⁺-channel blocking peptides of scorpions, strongly support the conclusion that Cm28 is the first example of new subfamily of α-KTx blocking peptides. The uniqueness of its primary structure would provide a novel drug template for designing a highly selective K_V1.3 inhibitors.

A comprehensive description of venom preparation and purification approach of several peptide toxins including a short peptide Cm28 with 2,820 D mol wt from the venom of C. margaritatus was reported previously (Beltrán-Vidal et al., 2021). Briefly, venom was milked from scorpions by electric stimulation, dissolved in sterile water, and centrifuged at 15,000 rpm and 4°C for 15 min. The supernatant was collected, lyophilized, and stored at −20°C. To achieve high yield and purity, a three-step purification scheme was exploited. The soluble venom was first subjected to gel filtration using Sephadex G-50 column in 20 mM ammonium acetate buffer (pH 4.7) at 2 ml/min flow rate, and three fractions were collected. Fraction FII, which typically contains toxic peptides, was purified through ion-exchange chromatography (IEC) as a second step using carboxy-methylcellulose column. Peptides were eluted at a flow rate of 2 ml/min with a linear gradient 0–100% of 500 mM ammonium acetate buffer over 200 min. Fractions from IEC were further purified by reverse-phase high-performance liquid chromatography (RP-HPLC) using an analytical grade C₁₈ reverse-phase column (Vydac). A linear gradient from 100% of solution A (0.12% trifluoroacetic acid [TFA] in water) to 60% of solution B (0.1% TFA in acetonitrile) over 60 min was run at 1 ml/min flow rate to elute pure peptides from the column. Absorbance was monitored at 230 nm. Fractions were collected manually and stored at −20°C until further use after vacuum drying. A sample from single peaks of RP-HPLC was analyzed in LCQ Fleet mass spectrometer coupled with an electrospray ionization (Thermo Fisher Scientific, Inc.).

The primary structure of pure peptide was determined by automated Edman degradation using Biotech PPSQ-31A Protein Sequencer equipment (Shimadzu Scientific Instruments, Inc.) following the same procedure as described for another component from the same venom (Beltrán-Vidal et al., 2021). First, a pure native peptide was applied directly for sequencing, and then a reduced and alkylated sample of the same peptide was sequenced to identify cysteine residues.

The search for potential homologs of Cm28 was performed by BLAST using the NCBI-Non-redundant protein sequences (nr) and Uniprot Swiss-Prot databases. An additional search was performed with the blastp option of Diamond v2.0.13.151 (Buchfink et al., 2021) against the 195 scorpion KTx sequences available in Kaliumdb (potassium channel polypeptide ligand database; Tabakmakher et al., 2019) using an e value = 1 × 10⁻⁵ as the significance cutoff. This database has >300 sequences, but here only the ones specific for K⁺ channels isolated from scorpion venom were used. Identification of conserved domains was performed using Pfam (Mistry et al., 2021) and InterPro (Blum et al., 2021).

All amino acid sequence alignments were performed with mafft v7.475 (Katoh and Standley, 2013). The phylogenetic analysis by maximum likelihood was performed with iqtree v2.1.3 (Minh et al., 2020). Iterative maximum likelihood analyses were performed using all 146 α-KTx sequences of Kaliumdb to determine the group of α-KTx closest to Cm28. The sequences of the remaining families were included as outgroups. The best substitution model was determined with the modelfinder (Kalyaanamoorthy et al., 2017). Phylogenetic analysis of Cm28 was determined using the WAG + R3 model with 10,000 ultrafast boostraps (Hoang et al., 2018). Tree was edited using FigTree1.4.4.

The 3-D modeling of Cm28 was performed with the help of the AlphaFold2 “colab notebook” (Mirdita et al., 2019; Jumper et al., 2021) enabling refinement with Amber-Relax (Eastman et al., 2017).

Chinese hamster ovary (CHO) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, cat# 11965084; Gibco) containing 10% FBS (Sigma-Aldrich), 2 mM L-glutamine, 100 µg/ml streptomycin, and 100 U/ml penicillin-g (Sigma-Aldrich) in a humidified incubator at 37°C and 5% CO₂. Cells were passaged twice per week following a 5-min incubation in PBS containing 0.2 g EDTA/L (Invitrogen).

Human peripheral blood monocytes (PBMCs) were isolated from the venous blood of anonymous healthy donors through Histopaque1077 (Sigma-Aldrich) separation technique. PBMCs were grown (density 5 × 10⁵ cells/ml) in RPMI 1640 medium (cat# 11875085; Gibco) supplemented with 10% FBS, 2 mM L-glutamine, 100 µg/ml streptomycin, and 100 U/ml penicillin-g in a humidified incubator at 37°C and 5% CO₂ for 3–6 d. Phytohemagglutinin A (PHA; Sigma-Aldrich) was also added at a concentration of 2, 5, and 10 µg/ml to activate the PBMCs and amplify the K_V1.3 expression. CHO cells and PBMCs were washed gently twice with 2 ml of extracellular (bath) solution (see Electrophysiology) for the patch-clamp experiments.

CHO cells were transiently transfected using Lipofectamine 2000 kit (Invitrogen), as per manufacturer’s protocol with the following ion channel coding vectors: hK_V1.1 (hKCNA1 gene) and hK_V1.2 (hKCNA2 gene) in pCMV6-AC-GFP plasmid (cat# RG211000 and RC222200; OriGene Technologies), hK_V1.5 in pEYFP plasmid (a kind gift from A. Felipe, University of Barcelona, Barcelona, Spain), hK_Ca3.1 (hKCNN4 gene) in pEGFP-C1 vector (a kind gift from H. Wulff, University of California, Davis, Davis, CA), hNa_V1.5 (hSCN5A, a kind gift from H. Abriel, University of Bern, Bern, Switzerland), and hHv1 (hVCN1, GenBank accession no. BC007277.1, a kind gift from Kenton Swartz, National Institutes of Health, Bethesda, MD). At 24 h after transfection, GFP-expressing transfectants were identified with Nikon TE 2000U fluorescence microscope using bandpass filters of 455–495 and 515–555 nm for excitation and emission, respectively, and used for current recordings (∼60–70% success rate for co-transfection). In general, currents were recorded 24–36 h after transfection.

Human embryonic kidney 293 cells stably expressing hK_V11.1 (hERG1 and hKCNH2 genes, a kind gift from H. Wulff), mK_Ca1.1 (BK_Ca, mKcnma1, a kind gift from C. Beeton, Baylor College of Medicine, Houston, TX), and hNa_V1.4 (hSCN4A gene, a kind gift from P. Lukács, Eötvös Loránd University, Budapest, Hungary) were used.

Whole-cell currents were measured using patch-clamp technique in voltage-clamp mode following standard protocols (Hamill et al., 1981). All recordings were performed using Multiclamp 700B amplifier connected to a personnel computer with Axon Digidata1440 digitizer and Clampex 10.7 software was used for data acquisition (Molecular Devices). In general, current traces were lowpass filtered by using the built-in analog four-pole Bessel filters of the amplifiers and sampled (4–50 kHz) at least twice at the filter cutoff frequency. Micropipettes were pulled from GC150F-7.5 borosilicate capillaries (Harvard Apparatus) with tip resistance typically ranging 3–6 MΩ in the bath solution. Only those records were used for data analysis when the leak current at holding potential was <10% of peak current at the test potential. Recordings were carried out at room temperature (20–25°C). Control and test solutions were perfused into the cell through a gravity flow perfusion system. The excess bath solution was removed constantly with vacuum suction.

For the measurement of K_V1.1–K_V1.3, K_V1.5, mK_Ca1.1, and Na_V1.4–Na_V1.5 currents, the normal bath solution consisted of (in mM) 145 NaCl, 5 KCl, 2.5 CaCl₂, 1 MgCl₂, 5.5 glucose, and 10 HEPES, pH 7.35. To record the tail current of K_V1.2, the bath solution (HK-20) contained 130 mM NaCl and 20 mM KCl, and the other components remained unchanged. In the HK-150 bath, all Na⁺ was substituted by K⁺ to yield 150 mM K⁺ concentration. In the Na⁺-free extracellular solution, all Na⁺ was substituted by choline-Cl, and other components remained unchanged. Equimolar substitution of Na⁺ for TEA-Cl was used in the various TEA⁺-containing solutions (Fig. 6). The bath solution for K_V11.1 consists of (in mM) 140 choline-Cl, 5 KCl, 2 MgCl₂, 2 CaCl₂, 10 HEPES, 20 glucose, 0.1 CdCl₂, pH 7.35; for K_Ca3.1 (in mM), 145 L-aspartic Na⁺ salt, 5 KCl, 10 HEPES, 5.5 glucose, 2.5 CaCl₂, and 1 MgCl₂, pH 7.4; and for hH_V1 (in mM), 60 L-aspartic acid Na⁺ salt, 80 MES, 5.5 glucose, 6 MgCl₂, pH 7.4. The osmolarity of the extracellular solutions was between 302 and 308 mOsm/liter. All the bath solutions were supplemented with 0.1 mg/ml BSA (Sigma-Aldrich) to prevent toxin adsorption to the plastic surfaces of the perfusion system. The composition of the internal solution was (in mM) 140 KF, 2 MgCl₂, 1 CaCl₂, 10 HEPES, and 11 EGTA, pH 7.22 for recording K_V1.1–K_V1.3, K_V1.5, and mK_Ca1.1 currents. For Na_V1.4–Na_V1.5 currents, the internal solution consisted of (in mM) 10 NaCl, 105 CsF, 10 HEPES, and 10 EGTA, pH 7.2; and for K_V11.1 (in mM), 140 KCl, 2 MgCl₂, 10 HEPES, and 10 EGTA, pH 7.3. The internal solution for K_Ca3.1 recording contained (in mM) 150 K-Asp, 5 HEPES, 8.5 CaCl₂, and 1.0 MgCl₂, pH 7.22; and for hH_V1 (in mM), 90 L-Aspartic acid with Na, 80 MES, 6 MgCl₂, and 3.3 glucose, pH 6.17. The measured osmolarity of internal solutions was ∼295 mOsm/liter.

All salts and positive control test chemicals for ion channel assay (ClGBI-5-chloro-2-guanidinobenzimidazole, TRAM-34, and TEA-Cl) were purchased from Sigma-Aldrich. rMgTx was in-house produced as described elsewhere (Naseem et al., 2021).

In general, for all the measurements the holding potential (V_h) was kept at −120 mV and the pulses were delivered every 15 s except when indicated. For recording the currents of K_V1.1–K_V1.3 and K_V1.5 ion channels, 15–300 ms long voltage pulses to +50 mV were applied. To record the K_V1.3 currents for conductance–voltage (G-V) relationship, activated T cells were depolarized to voltages ranging from −70 mV to +50 mV in steps of 10 mV every 15 s. For instantaneous current–voltage (I-V) relationships of K_V1.2 and K_V1.3, currents were evoked with 200-ms-long voltage ramps to +50 mV. For recording K_V11.1 current, voltage step to +20 mV for 1.25 s from a V_h of −80 mV followed by a step to −40 mV for 2 s was applied every 30 s, and the peak (tail) currents were recorded during the latter step. mK_Ca1.1 currents were evoked by depolarizing the cells to +100 mV for 600 ms from a V_h of −100 mV. For K_Ca3.1 currents, 150-ms-long voltage ramp to +50 mV from −120 mV was applied every 10 s. Current through the human proton channel (hH_V1) was elicited by applying a 1.0-s-long voltage ramp to +100 mV from a V_h of −60 mV every 15 s. For sodium currents through Na_V1.4 and Na_V1.5, 15-ms-long voltage steps to 0 mV were applied every 10 s.

PBMCs were isolated from a healthy donor’s blood and cultured as explained earlier. Dead Cell Removal Microbead Kit (Miltenyi Biotec B.V & CO. KG) was used to eliminate the dead cells, and CD4⁺ T_EM lymphocytes were isolated through magnetic cell sorting (negative selection) with the CD4⁺ Effector Memory T Cell Isolation Kit (Miltenyi Biotec B.V & CO. KG). Briefly, all types of cells except CD4⁺ T_EM lymphocytes were labeled with a cocktail of monoclonal antibodies (biotin-conjugated anti-CD8, -CD14, -CD15, -CD16, -CD19, -CD34, -CD36, -CD45RA, -CD56, -CD123, -CD235a, -TCR γ/δ, and APC-conjugated anti-CCR7). Next, cells were incubated with anti-APC and anti-biotin secondary antibodies, both coupled with magnetic microbeads. The cell preparation was passed through LD Column (Miltenyi Biotec B.V & CO. KG) mounted on MidiMACS Separator (Miltenyi Biotec B.V & CO. KG), and untouched CD4⁺ T_EM lymphocytes were collected as flow through.

CD4⁺ T_EM lymphocytes were divided into five different treatment groups: (1) unstimulated and non-treated, (2) unstimulated and treated with Cm28 (1.5 µM), (3) stimulated only, (4) stimulated and treated with Cm28 (1.5 µM), and (5) stimulated and treated with recombinant margatoxin (rMgTx, 5 nM, produced in-house as previously described; Naseem et al., 2021). The high concentrations of peptide toxins were used to ensure the complete blockade of K_V1.3 channels throughout the entire treatment duration and to counterbalance peptide adsorption to plastic surfaces and biological degradation (Beeton et al., 2011; Veytia-Bucheli et al., 2018; Naseem et al., 2021). To stimulate lymphocytes through the T cell receptor (TCR), anti-human CD3 monoclonal antibody (clone OKT3; BioLegend) was bound to the surface of a 96-well cell culture plate (cat# 3599; Corning) at 1 µg/well in PBS at 4°C overnight. Wells were washed twice with PBS to get rid of the unbound antibody. CD4⁺ T_EM cells were suspended at a density of 1 × 10⁶ cells/ml in RPMI 1640 medium (cat# 11875085; Gibco) supplemented with 10% FBS, 2 mM L-glutamine, 100 µg/ml streptomycin, and 100 U/ml penicillin-g. To ensure the complete blockade of K_V1.3 prior to activation, cells were incubated with the Cm28 (at 1.5 µM) or rMgTx (at 5 nM) for 30 min. Subsequently, cells were loaded in the wells (0.2 ml of cell suspension per well) and the plate was incubated in a humidified incubator at 37°C in 5% CO₂ for 24 h. Each experiment included two technical duplicates and was performed on three different donors with the same conditions.

For assessing the cell viability, Zombie NIR fixable viability Dye (cat# 423105; BioLegend) was used. Cells were washed with PBS and incubated with Zombie NIR dye (at 1:500 dilution in 100 μl PBS) at room temperature for 20 min in dark. For staining the cells with fluorescent antibodies, cells were washed with PBS supplemented with 1% FBS and stained with PerCP/Cyanine5.5 conjugated anti-human CD25 (IL2R) antibody (clone BC96; BioLegend) and fluorescein isothiocyanate (FITC) conjugated anti-human CD154 (CD40L) antibody (clone 24–31; BioLegend) in 100 μl PBS + 1% FBS at 4°C for 20 min in dark. Cells were then washed with PBS + 1% FBS buffer and finally resuspended in 150 μl PBS + 1% FBS for flow cytometer analysis. Samples were measured using a NovoCyte 3000 RYB flow Cytometer (ACEA Bioscience, Inc.) and analyzed with FCS Express 6.0 (De Novo Software). Briefly, cells were gated in an FSC-H versus SSC-H density plot. Histograms corresponding to IL2R (CD25) and CD40L were generated as peak-normalized overlays. Unstained cell controls (negative) were always used for comparison, and mean fluorescent intensities were normalized to that of their stimulated but not treated control. To determine cell viability, positive staining with the Zombie NIR dye and changes in FSC were considered as indicators of dead cells. Cells treated with 30% of DMSO were used as a positive control for the viability dye.

Cells treated with sterile water and lysis buffer for 45 min at 37°C were used as spontaneous and maximum LDH activity controls, respectively. For experimental positive control, cells were treated with 50 mM sodium azide (NaN₃). The experiment was repeated for three different donors.

Statistical analyses and graph plotting were executed in GraphPad Prism software (version 8.0.1). Data were presented as mean ± SEM. For pairwise comparison, student’s t test with Mann-Whitney rank sum test and for multiple comparisons, one-way ANOVA with post-hoc Tukey’s test was performed. Statistical significance is indicated in terms of P values.

Fig. S1 shows the comparison of representative 3-D structures of each KTx family with the modeled 3-D structure of Cm28.

A detailed description about the purification of peptides from the soluble venom of C. margaritatus and their proteomic analysis was reported in our previous publication (Beltrán-Vidal et al., 2021). A three-step purification approach at the preparative level was followed to achieve a generous quantity of peptides for proteomic and functional characterization. First, soluble crude venom was fractionated into three fractions (FI, FII, and FIII) by gel filtration chromatography using Sephadex G-50 column (not shown). Then, peptide-based toxic components were separated from the main fraction FII through IEC using carboxy-methylcellulose column. Finally, the 10 sub-fractions (FII.1–10, not sown) from IEC were individually subjected to reverse-phase HPLC. A peptide with mol wt 2,820.5 D was discovered in HPLC fraction of FII.6 (Fig. 1 A) in addition to another novel peptide blocker (CmERG1, γ-KTx 10.1) of K_V11.1 (hERG1) channel (Beltrán-Vidal et al., 2021). The 2,820.5 D peptide was called “Cm28,” corresponding to the scorpion’s name C. margaritatus and its mol wt. Cm28 peptide was eluted at 24.5 min retention time from C₁₈ HPLC column as indicated in Fig. 1 A. The full-length amino acid sequence of Cm28 was determined by direct automatic Edman degradation of native peptide and, alkylated and reduced forms of peptide. This novel peptide contains 27 amino acids with 6 cysteines and 3 potential disulfide bridges (Fig. 1 B). The primary structure of Cm28 is unique and unusual because it has fewer residues and is a completely different sequence than other known scorpion toxins blocking K⁺ channels.

No homologous amino acid sequences of Cm28 were identified by BLAST in the NCBI and UniProt databases. However, diamond blastp analysis against kaliumdb sequences revealed α-KTx13.4 (Tityus stigmurus, UniProt accession no. P0C8L2), ε-KTx1.2 (Titus serrulatus, UniProt accession no. P0C175) and ε-KTx1.1 (T. serrulatus, UniProt accession no. P0C174) as potential homologs of Cm28. Alignment of the representative sequences shows that their percent identity ranges from 23 to 42% (Fig. 2), with the C-terminal region of the peptides being the most conserved. No Pfam domain was identified in Cm28 as reported by http://pfam.xfam.org/ (Mistry et al., 2021). Phylogenetic analysis was performed by comparing the amino acid sequence of Cm28 with 75 other reported scorpion toxins (Kaliumdb/UnitProt; Fig. 3). Cm28, ε-KTx1.1, and ε-KTx1.2 toxins clustered within the α-KTx family suggest that these three toxins belong to this KTx family. Moreover, the 3-D structure of Cm28 (Fig. S1) shows one short α-helix connected to an antiparallel β-sheet stabilized by three disulfide bonds (CSα/β), a similar motif found in the structure of the α-KTx toxins.

The primary structural features make Cm28 an exceptional scorpion toxin; therefore, we assessed if such an unusual peptide has any pharmacological activity against K⁺ channels. First, we aimed at testing the effects of Cm28 on two human K⁺ channels, K_V1.2 and K_V1.3. The macroscopic K_V1.2 currents were measured in transiently transfected CHO cells (see Materials and methods for detail). Channels were activated by a series of depolarization pulses to +50 mV from −120 mV. Due to the highly variable activation kinetics of K_V1.2 (Rezazadeh et al., 2007), 15–500-ms long pulses every 15 s were applied to maximize the open probability of the channel. The slower inactivation rate of K_V1.2 prevented inactivation even at 500-ms-long depolarization pulses. For K_V1.3 current measurements, human peripheral T lymphocytes were activated by Phytohemagglutinin A (PHA) to boost up the expression of K_V1.3 channels, and the pipette-filling solution was Ca²⁺ free to avoid K_Ca3.1 channel activation. Thus, the whole-cell currents were recorded exclusively through K_V1.3 channels, as shown previously (Gurrola et al., 2012; Varga et al., 2012). The K_V1.3 currents were evoked by 15-ms-long depolarization pulses to +50 mV. The use of short pulses every 15 s ensured that there is no cumulative inactivation of K_V1.3 channel. Cm28 dissolved freshly in the extracellular solution was applied through a custom-built micro perfusion system at the rate of 200 μl/min. The complete exchange of solution in the bath chamber, i.e., the proper operation of the perfusion apparatus, was confirmed frequently using fully reversible inhibitors as positive controls at a concentration equivalent to their K_d values (i.e., 14 nM charybdotoxin [ChTx] for K_V1.2 [Fig. 4 A] and 10 mM TEA⁺ for K_V1.3 [Fig. 4 B]), and the ∼50% inhibition in peak current was an indicator of both the ion channel and the proper operation of the perfusion system.

Fig. 4 A represents the whole-cell currents through K_V1.2 recorded sequentially in the same cell, before (control trace, black) and after perfusing the cell with 2 nM Cm28 till the equilibrium block (purple trace). At equilibrium block, Cm28 showed ∼70% reduction in current amplitude. The block was almost fully reversible upon washing the perfusion chamber with toxin-free solution (wash-out trace, green in Fig. 4 A). The onset and recovery from the block of K_V1.2 currents at 2 nM Cm28 are shown in Fig. 4 C. Normalized peak currents were plotted as a function of time. Both the association and dissociation processes of Cm28 were very slow and, accordingly, the development of equilibrium block and recovery up to ∼85% of control current took several minutes. Similar sets of experiments were carried out for K_V1.3: 2 nM Cm28 inhibited ∼58% of the K_V1.3 current upon reaching the block equilibrium. Fig. 4 B displays the current traces recorded sequentially in the same T lymphocyte, in the presence (red trace) and absence (black trace) of Cm28 peptide. Like K_V1.2, the block of K_V1.3 was also reversible (90% recovery took 10 min) upon perfusing the cell with toxin-free solution (wash-out trace, green in Fig. 4 B). The onset of steady-state block and relief from the block took comparatively less time than for K_V1.2 as shown in Fig. 4 D, indicating that the association and dissociation steps are faster for K_V1.3 than K_V1.2.

We performed a concentration–response experiment to determine the concentration-dependent block of K_V1.2 and K_V1.3 channels by Cm28. Different concentrations of Cm28 were applied to the cell for an adequate duration to reach the complete equilibrium block, considering the slow blocking kinetics, especially at low toxin concentrations. The remaining current fractions were calculated as (I/I₀, where I₀ is the peak current in the absence of the toxin and I is the peak current at equilibrium block in the presence of Cm28 at a given concentration). Data points were fitted with Hill equation (see Materials and methods for details) to obtain dose–response curves. The resulting dissociation constant (K_d) values and Hill coefficients (H) were K_d = 0.96 nM, H = 1.04 for K_V1.2 (Fig. 4 E) and K_d = 1.3 nM, H = 0.93 for K_V1.3 (Fig. 4 F). Cm28 showed a similar affinity for both channels.

Traditionally, most of the known toxins inhibit the K_V channels by following two modes of action. First, simple pore blocking in which toxins physically occlude the pore region preventing the permeation of K⁺ ions. In the second mechanism, toxins bind to the voltage-sensing domain of the K_V channels and modulate its gating by causing a prominent shift in the voltage dependence of steady-state activation toward more depolarized potentials and consequently, reduce the K⁺ current (Swartz and MacKinnon, 1997; Moreels et al., 2017). The blocking mechanism of Cm28 was assessed by determining the voltage dependence of steady-state-activation and the threshold voltage of activation for both K_V1.2 and K_V1.3 ion channels. Instantaneous I-V relationship was recorded using CHO cells for K_V1.2 and activated T cells for K_V1.3 (Fig. 5, A and B). Currents were evoked by applying 200-ms-long voltage ramps from −120 mV to +50 mV every 15 s. Cm28 did not shift the threshold voltage of activation of either current as shown in Fig. 5, A and B. The current traces in the control solution and at the equilibrium block with 2 nM Cm28 showed a similar threshold voltage of activation; approximately −23 mV for K_V1.2 (Fig. 5 A) and approximately −40 mV for K_V1.3 (Fig. 5 B).

For the construction of G-V relationship for K_V1.2, isochronal tail peak currents were recorded in CHO cells at −120 mV followed by 300-ms long depolarizations ranging from −70 to +80 in 10 mV steps from V_h −120 mV in HK-bath containing 20 mM K⁺ to increase the tail currents. Due to highly variable activation properties (Rezazadeh et al., 2007), only those records were considered for analysis that had a similar gating mode. Normalized tail peak currents were plotted as a function of membrane potential (E_m) in Fig. 5 C, the solid lines represent the best-fit Boltzmann sigmoidal function. Cm28 did not introduce any substantial shift in the G-V curve of K_V1.2 (Fig. 5 C). The midpoint voltage (V₅₀) of the G-V relationship for K_V1.2 was 21 ± 3 mV in the control solution (n = 5) and 15 ± 4 mV at equilibrium block with 2 nM Cm28 (n = 5). Fig. 5 E indicates that the difference between V₅₀ for K_V1.2 in the presence or absence of Cm28 was statistically nonsignificant. For G-V relationship of K_V1.3, whole-cell currents in the activated human T lymphocytes were measured in response to voltage pulses ranging from −70 to +50 mV in 10 mV steps from V_h of −120 mV, and the conductance values were calculated for each test potential and normalized for the maximal conductance. The best fit of the Boltzmann sigmoidal function to the averaged data points yielded the superimposed solid lines as shown in Fig. 5 D, indicating that there is no change in the voltage dependence of steady-state activation of K_V1.3 in the presence of Cm28 at 2 nM, similar to K_V1.2. The V₅₀ values for K_V1.3 were similar in the control solution (V₅₀ = −20 ± 3, n = 4) and at an equilibrium block with 2 nM Cm28 (V₅₀ = −19 ± 2, n = 4) as shown in Fig. 5 E. As the voltage dependence of steady-state activation and the threshold voltage of activation were not affected by the Cm28 for both K_V1.2 and K_V1.3 ion channels, it suggests that Cm28 is not a gating modifier, rather that it interacts with the pore region of ion channels.

Similarly, the association (τ_on) and dissociation (τ_off) time constants of K_V1.2 blockade at 2 nM Cm28 were determined by fitting the single exponential function to data points during the wash-in procedure and wash-out procedure (Fig. 4 C). Like K_V1.3, assuming the bimolecular interaction between the toxin and the channel k_on and k_off rate constant were calculated using the above-mentioned equations and time constants (Table 1).

The dissociation constants (K_d = k_off/k_on) calculated from the block kinetics yielded 1.18 nM for K_V1.2 and 2.15 nM for K_V1.3, as given in Table 1, that are in good agreement with the K_d values obtained from equilibrium block (Fig. 4, E and F).

To reveal the selectivity profile of Cm28, we assayed the effect of Cm28 on two members of voltage-gated Shaker family channels, hK_V1.1 (Fig. 6 A) and hK_V1.5 (Fig. 6 B), that are closely related to the channels inhibited by Cm28. In addition, we also tested the effect of Cm28 on hK_V11.1 (hERG1, Fig. 6 C), a voltage-gated cardiac K⁺ channel; hK_Ca3.1 (IKCa1, SK4, Fig. 6 D), the Ca²⁺ activated K⁺ channel expressed in T lymphocytes; mK_Ca1.1 (BK, Slo1, MaxiK, Fig. 6 E), the large conductance voltage- and Ca²⁺-activated channel; two voltage-gated sodium channels, hNa_V1.4 (Fig. 6 F) and hNa_V1.5 (Fig. 6 G), expressed in skeletal and cardiac muscles, respectively; and hH_V1 (Fig. 6 H), a voltage-gated proton channel. We found that, except K_V1.1, none of the ion channels tested (Fig. 6, A–I) were inhibited by Cm28 at 150 nM concentration, which is >150-fold concentration than the K_d for K_V1.2 and >100-fold than the K_d for K_V1.3. The application of 150 nM Cm28 reduced ∼27% of K_V1.1 current and the RCF value was 0.73 ± 0.03 (n = 3). The estimated K_d value for K_V1.1 from a single concentration, based on a bimolecular interaction, yielded ∼0.4 μM. The amount of native peptide was not sufficient to construct a complete dose-response curve at this very high peptide concentration range.

After characterizing the pharmacological properties of the unique Cm28 peptide with electrophysiology, we investigated whether these features reflect in biological functional assays. As the main aim of this assay was to evaluate the effect of Cm28 on the TCR-mediated activation of CD4⁺ T_EM cells, first we had to determine whether the peptide compromises cell viability. Following a 24-h culture period in the presence of 1.5 μM Cm28 or 50 nM rMgTx, the viability of CD4⁺ T_EM cells was not impaired either in quiescent or TCR-activated cells as analyzed by two different assays (Fig. 7). Staining the cells with a fixable viability dye Zombie NIR followed by flow cytometry identified that ≥90% of cells were viable in the presence of either toxin. 30% DMSO was added to the cells for 30 min as a positive control for dead cells (Fig. 7 A). In parallel, the LDH activity assay revealed that cytotoxicity of Cm28 was <1% and that of rMgTx was <2% for either quiescent or TCR-activated T_EM cells after 24 h culture period. 50 mM NaN₃, as the positive control, showed 15% cytotoxicity (Fig. 7 B).

In human T lymphocytes, the expression of the Ca²⁺-dependent early activation markers in the cell membrane, such as IL2R and CD40 ligand, is upregulated upon TCR-mediated activation. These activation markers have been used as a readout to assess K_V1.3-dependence of T cell activation (Balajthy et al., 2016; Veytia-Bucheli et al., 2018; Naseem et al., 2021). CD4⁺ T_EM cells were pre-incubated for 30 min with either 1.5 μM Cm28 (>1,000× concentration of its K_d for K_V1.3) or 5 nM rMgTx (100× concentration of its K_d for K_V1.3, as positive control), and the cells were then activated for 24 h with plate-bound anti-human CD3 antibody in the continuous presence of toxins. The flow cytometric overlayed histograms in Fig. 8, A and C show that Cm28 (red traces) significantly reduced the fraction of CD40L (Fig. 8 A) and IL2R (Fig. 8 C) expressing T_EM cells, similar to rMgTx (blue trace), as compared with the control cells stimulated identically in the absence of toxin (green trace). The expression of CD40L (Fig. 8 B) and IL2R (Fig. 8 D) in Cm28-treated T cells, normalized to that of stimulated but not treated cells, is reduced to ∼47 and ∼55%, respectively. Similarly, positive control rMgTx decreased the CD40L and IL2R expression levels to ∼51 and ∼48%, respectively. There was no change in the expression of CD40L or IL2R in CD4⁺ in T_EM cells not exposed to anti-CD3 antibody regardless of the presence (US + Cm28) or absence of Cm28 (US; Fig. 8).

In this article, we characterized the in vitro pharmacological and immunological activities of Cm28, a novel peptide isolated from C. margaritatus belonging to the Buthidae family of scorpions. Cm28 consists of only 27 amino acids with 6 cysteine residues. It is a high-affinity blocker of human K_V1.2 and K_V1.3 channels with K_d values of 0.96 and 1.3 nM, respectively. It also inhibited K_V1.1 channel with low affinity. The application of high concentration (∼100× of K_d for K_V1.3) of Cm28 did not inhibit the several other ion channels tested in this study including four other subtypes of K⁺ channels (K_V1.1, K_V1.5, K_V11.1, K_Ca1.1, and K_Ca3.1), two subtypes of Na⁺ channels (Na_V1.5 and Na_V1.4), and the voltage-gated H⁺ channel hH_V1. In biological functional studies, Cm28 (at 1.5 µM concentration, ∼1,000× of K_d for K_V1.3) substantially inhibited the expression of IL2R and CD40L in activated human CD4⁺ T_EM lymphocytes in vitro without compromising cell viability.

Cm28 has a unique primary structure, and it is quite different (Fig. 1 B) from all the other 195 peptide toxins described thus far from scorpion venoms. The closest similarities were found with the ε-KTx family (39–40% identity) and to a less extent with the peptides of α-KTx subfamily 13 (29–33% identity; Fig. 2). In phylogenetic tree analysis, Cm28 and members of ε-KTx family were in the same clade which belongs to α-KTx family (Fig. 3). So far, only two peptide toxins belonging to the ε-KTx family have been described. They lack a classical secondary structure and exhibit an inhibitor-cystine knot (ICK) type scaffold (Cremonez et al., 2016). In our phylogenetic tree, the ε-KTx family was clade with the α-KTx family; however, its phylogenetic position needs to be confirmed by evaluating a larger number of orthologs to substantiate, at least by phylogenetic analysis, whether the ε-KTx could be a subfamily of the α-KTx with an ICK motif. The branch support values (92 and 80) suggest that Cm28 is completely separated from ε-KTx1.1 and ε-KTx1.2. In addition, the modeled structure of Cm28 shows more similarity to the structure of α-KTx toxins that lack the ICK scaffold. For these reasons, we suggest that Cm28 is the first example of a new subfamily of α-KTxs for which the proposed systematic number is α-KTx 32.1 and its primary sequence has been deposited in Zenodo (Naseem et al., 2022), and will be available in the Uniprot Knowledgebase under accession no. C0HM22. However, a structural study is needed to determine whether Cm28 has an ICK scaffold like the ε-KTxs or whether Cm28 has the characteristic scaffold of the α-KTxs.

The mechanism through which scorpion toxins block K⁺ channels involve (1) plugging the pore of the channel through binding to the extracellular vestibule (Goldstein and Miller, 1993), and (2) modulating the gating of the channel through binding to the voltage-sensor domain (Swartz and MacKinnon, 1997; Moreels et al., 2017). We argue that Cm28 is not a gating modifier because it did not change the voltage dependence of steady-state activation and the threshold voltage of activation of either K_V1.2 or K_V1.3 ion channel (Fig. 5, A–E). On the contrary, we propose that Cm28 could be a pore blocker. The blocking kinetics follows a simple bimolecular interaction between toxin and channel described previously for classical pore blockers such as ChTx (Goldstein and Miller, 1993). The apparent first-order association rate depended on the concentration of Cm28, being faster at higher toxin concentration and the dissociation rate remained constant regardless of change in Cm28 concentration (Fig. 5 F). In addition, the dissociation constants calculated by the k_off/k_on ratio (K_d = 1.15 nM for K_V1.2 and K_d = 2.15 nM for K_V1.3) are very close to the one determined by fitting the Hill equation to the concentration dependence of current inhibition (K_d = 0.96 nM for K_V1.2 and K_d = 1.3 nM for K_V1.3). Numerous studies have suggested other mechanisms of block in which a diverse family of toxins target the turret region of K_V channels. The precise molecular processes behind the turret block mechanism remain obscure, however a general concept (turret-block) whereby the toxin acts as a lid above the pore entry was proposed (Saikia et al., 2021). We have recently shown that a turret-modulating toxin may exert its blocking effect by modifying the rates of structural water exchange at the inactivation cavities that are involved in controlling inactivation (Szanto et al., 2021) rather than directly blocking the permeation pathway (Karbat et al., 2019). The turret block mechanism was not addressed in this study as the primary sequence of Cm28 is totally different from Cs1. Cs1 is a low affinity (in the order of micromolar concentrations) and partial blocker of the mammalian K_V channels (Karbat et al., 2019) as opposed to Cm28.

The characteristics of Cm28 block of K_V1.3 are consistent with pore blocker toxins; however, in general, pore blocker peptides contain the typical functional dyad or at least the critical lysine residue which protrudes into the selectivity filter of the channel (Goldstein and Miller, 1993). In Cm28, the typical lysine together with an aromatic residue that interacts with the selectivity filter is not found. However, as demonstrated, Cm28 inhibits K_V1.2 and K_V1.3 with high affinity. There are other small α-KTx peptides (<30 residues long) that show high affinity for K_V1.3 channel in the absence of the functional dyad. BmP02 (α-KTx 9.1) and BmP03 (α-KTx 9.2), for example, are toxins that differ by only one amino acid. A Lys at position 16 in BmP02 is replaced by Asn in BmP03. Both toxins inhibit the K_V1.3 channel with IC₅₀ values of 7 and 85.4 nM, respectively. This 12-fold affinity variation suggests that the Lys16 is a part of the functional surface, even though the typical functional dyad is not found in these toxins (Zhu et al., 2012). Kbot1 toxin (α-KTx 9.5) is 93% identical to BmP02 differing only by two amino acids (N14H and K16V). Nevertheless, Kbot1 is also a potent inhibitor of K_V1.3 with an IC₅₀ value of 15 nM. It is worth noting that Kbot1, like the BmP03 toxin, does not contain Lys16 but has an IC₅₀ closer to that of BmP02, which can be explained by compensating for the loss of cationic charge through the addition of a histidine residue at position 14. Moreover, Kbot1 inhibits the ChTx binding in the rat brain synaptosomes with an IC₅₀ of 10 nM (Mahjoubi-Boubaker et al., 2004; Zhu et al., 2012). BmP02, BmP03, and Kbot1 all three belong to the same α-KTx subfamily 9, so it could be that this is a characteristic of this family. However, a short peptide Tt28, from α-KTx 20.1 subfamily lacking the dyad motif, blocks K_V1.3 channel with an IC₅₀ value of 7.9 nM, although it shares only 25% identity with BmP02 (Abdel-Mottaleb et al., 2006), meaning these short peptides without the functional dyad can be found in other α-KTx subfamilies.

The lack of a functional dyad was also observed for larger peptides (>30 residues in length). By using these toxins as models and applying different approaches, the mechanism underlying the interaction between these peptides and K_V1 channels is revealed. An example is the toxin A24-A33-Pi1, an analog of the toxin Pi1 (α-KTx 6.1). In contrast to the native version, A24-A33-Pi1 lacks the functional dyad K24-Y33. Nevertheless, the toxin was able to bind the channel (K_d = 22 µM), indicating that the functional dyad does not appear to be a requirement for recognition and binding to the channel (Mouhat et al., 2004). On the other hand, it has been reported that in the interaction between Tc32 (α-KTX 18.1) and the K_V1.3 channel, the differences in the electrostatic properties of the toxin and the channel, the contact surfaces, and the total dipole moment orientations, lead to a lysine residue, even if it located at a different position from the functional dyad, physically blocking the pore of the channel (Stehling et al., 2012). According to the computer simulation, this effect of rearrangement was also observed during the interaction of BmP02 and the K_V1.3 channel. After the electrostatic interaction, the side chain of Lys11 was oriented to enter the pore directly, although it has a different structure from the classical dyad motif (Wu et al., 2016). Thus, it could be speculated that the basic residues in Cm28 are involved in the recognition of the channel and the electrostatic forces may rearrange the toxin in such a way that either Lys1 or Lys22 side chain protrude into the pore. Moreover, considering the possibility that another Lys in the toxin takes the place of the canonical lysine of the dyad, it could also be that the dyad is oriented in the opposite way. This phenomenon was previously reported for the toxin κ-KTx1.1, which interacts with the ion channel through a reversed dyad motif, consisting of an aromatic residue Tyr5 and the Lys19 (Srinivasan et al., 2002). In Cm28, the reversed dyad might consist of Tyr13 and Lys22, which could have a similar interaction as the reversed dyad in κ-KTx1.1 toxin. Clearly, extensive structural analysis and molecular docking should be performed to determine whether any of the above mechanisms may be involved in the interaction of Cm28 with K_V1.1-K_V1.3 channels or whether Cm28 hides an undescribed interaction mechanism.

Several scorpion toxins inhibit K_V1.3 with great affinity; however, they also show off-target effects by inhibiting other K⁺ channels, thereby compromising their therapeutic potential. For example, HsTX1 (α-KTx-6.3) inhibits K_V1.1 in addition to K_V1.3 (Lebrun et al., 1997; Regaya et al., 2004) and MgTx (α-KTx-2.2) and ChTx (α-KTx-1.1) blocks more than one K_V1.x channel subtypes with high affinity (Bartok et al., 2014). Vm24 (α-KTx-23.1, previously reported by our group) is the only natural peptide toxin that showed 1,500-fold selectivity for K_V1.3 over 10 other ion channels tested (Gurrola et al., 2012; Varga et al., 2012). The selectivity of these attractive peptides could be improved by peptide engineering for therapeutic development. For example, HsTX1[R14A] mutant retained high affinity for K_V1.3 and showed 2,000-fold selectivity over K_V1.1 (Rashid et al., 2014). An engineered analog of Anuroctoxin (α-KTx-6.12, AnTx) with double substitution (N17A/F32T) was developed previously by our group which preserved its natural potency for K_V1.3, while gaining 16,000-fold selectivity over K_V1.2 (Bartok et al., 2015). ShK-186 (originally isolated from sea anemone) is the best example of an engineered analog with high affinity and selectivity for K_V1.3 and is under clinical trial for autoimmune diseases called Dalazatide (Pennington et al., 2015; Tarcha et al., 2017). As demonstrated, Cm28 blocks the K_V1.2 and K_V1.3 with a similar potency (Fig. 4) and shows ∼400-fold less affinity for K_V1.1 (Fig. 6 A). The order of the blocking potency of Cm28 for various ion channels was hK_V1.2 ≈ hK_V1.3 ≫ hK_V1.1 > hK_V1.5 ≈ hK11.1 ≈ hKCa3.1 ≈ mKCa1.1 ≈ hNav1.4 ≈ hNav1.5 ≈ hHv1 (Fig. 6, B–H). After identifying the key residues in a unique primary structure of Cm28 responsible for interaction with K_V1 channel subtypes, Cm28 peptide can be engineered to improve the selectivity for K_V1.3 over K_V1.1 and K_V1.2.

The potency of Cm28 as a potential immunosuppressor agent was shown in a biological functional assay where the expression of IL2R and CD40L in human CD4⁺ T_EM lymphocytes was determined following TCR-mediated activation. IL2R and CD40L are early, Ca²⁺-, and NFAT-dependent activation markers of the T_EM cells (Schuh et al., 1998; Schonbeck et al., 2000). Since K_V1.3 regulates Ca²⁺ signaling, hence the expression level of the IL2R and CD40L, inhibition of K_V1.3 leads to a reduced expression level of these early activation markers (Chimote et al., 2017; Veytia-Bucheli et al., 2018; Naseem et al., 2021). Cm28 significantly downregulated the expression of IL2R and CD40L upon TCR-mediated activation of CD4⁺ T_EM cells, similar to the positive control MgTx (Fig. 8), which is consistent with the literature and validates the role of the K_V1.3 ion channels in T cell activation through maintaining the Ca²⁺ influx (Panyi et al., 2006; Feske et al., 2015; Veytia-Bucheli et al., 2018). The high concentration of Cm28 (∼1,000-fold the K_d for K_V1.3) was used in this assay to ensure practically a complete blockade of the K_V1.3 channels. Moreover, it was shown that the cell viability after 24 h was not compromised at a high Cm28 concentration (Fig. 7). The use of high toxin concentration is in accordance with the previous reports where significantly higher concentrations of Vm24 and Shk were used in biological assays than the K_d of the toxin for K_V1.3 (Beeton et al., 2011; Veytia-Bucheli et al., 2018).

It is worth noting that the path from the discovery of such potential peptides to its therapeutical application requires a variety of steps, most notably the generation of modifications that improve its affinity and selectivity or proteolytic stability, increasing its serum half-life. As mentioned above, there is a wide variety of toxins whose pharmacological targets are potassium channels. However, the size of Cm28 may provide an advantage over other larger peptides that also inhibit K_V1.2 or K_V1.3 channels. Shorter analogs of ShK toxin have been shown to have lower susceptibility to proteolysis. Reducing the peptide length made the structure of the analogs more constrained, and also reduced the number of positively charged (Lys and Arg) residues and aromatic residues, making the peptides more resistant to degradation by trypsin and chymotrypsin (Krishnarjuna et al., 2018). In addition, other techniques such as peptide cyclization have been used to improve not only proteolytic resistance but also serum stability of short peptides (González-Castro et al., 2021). Although Cm28 is a promising new peptide, the approaches for increasing the selectivity of Cm28 for K_V1.3 discussed above must be utilized and the advantages of the shorter Cm28 over other peptides must be experimentally confirmed to exploit the benefits of Cm28 in the treatment of autoimmune disorders.

Christopher J. Lingle served as editor.

The authors thank Cecilia Nagy and Adrienn Bagosi for expert technical assistance. Also, the support of Dr. Jimena I. Cid. Uribe during the phylogenetic analysis is recognized.

This work was supported by the following: research grants from the Hungarian National Research, Development, and Innovation Office (K143071 to G. Panyi, K142612 to T.G. Szanto, and K132906 to J. Borrego) and grant CONACYT 303045 from the National Council of Science and Technology of Mexico (to L.D. Possani). This work was supported by the Stipendium Hungaricum Scholarship by the Tempus Public Foundation (to M.U. Naseem).

The authors declare no competing financial interests.

Author contributions: M.U. Naseem contributed to conceptualization, investigation, formal analysis, and writing of the original draft. E. Carcamo-Noriega, J. Beltrán-Vidal and G. Delgado-Prudencio contributed to investigation and formal analysis. J. Borrego contributed formal analysis and writing of the original draft. T.G. Szanto contributed to methodology, manuscript review and editing. F.G. Zamudio contributed to investigation. L.D. Possani contributed to conceptualization, manuscript review and editing, funding acquisition, and methodology. G. Panyi contributed to conceptualization, writing of the original draft, manuscript review and editing, funding acquisition, and methodology.