Discovery and structural characterization of newcastle disease virus mimetic peptides identified through phage display Authors: D. Caixeta Moreira, E. Resende Vaz, H. Oliveira Almeida Souza, S. Lorenti, T. Velikkakam, B. Beatriz Fonseca, V. Alonso-Goulart Published: 2026 ABSTRACT Highlights•Three NDV-mimetic peptides (NDV-1, NDV-2, NDV-3) are selected by Phage Display.•NDV-1 exhibits strong residue reactivity and high structural stability.•NDV-1 may serve as a promising candidate for immunogenic applications.•NDV-2 displayed lower free energy values and affinity scores.•NDV-3 shows the highest free energy and potential for oncological applications. Introduction Newcastle Disease Virus (NDV), also classified as Avian Orthoavulavirus 1 (formerly AMPV-1), is an enveloped, single-stranded, negative-sense RNA virus belonging to the Paramyxoviridae family. Since its identification in the 1950s, NDV has been recognized as a highly contagious pathogen affecting over 240 bird (Flanagan et al., 1955). Its non-segmented genome encodes six structural proteins, which orchestrate its complex replication cycle (Cornax et al., 2013; Sinkovics and Horvath, 2000). Interestingly, nearly a decade after its discovery, NDV began to be explored as a potential oncolytic agent (Cassel and Garrett, 1965). Since then, extensive research has focused on elucidating the mechanisms by which NDV selectively targets and eliminates neoplastic cells. Scientific interest in NDV arises from its intrinsic capacity to selectively infect human tumor cells while largely sparing healthy tissues (Schirrmacher; Van Gool; Stuecker 2019). By infecting tumor cells, the virus triggers both innate and adaptive immune responses, increasing the expression of immunomodulatory molecules and promoting the activation of various immune effector cells (Ravindra et al., 2009). Beyond its direct oncolytic effects mediated by necrosis, apoptosis, or autophagy, NDV acts as a potent immunomodulator (Matveeva and Chumakov, 2018). Despite promising preclinical and clinical evidence regarding the use of NDV in oncology, several challenges remain. Optimizing delivery methods and formulations is crucial to maximizing clinical efficacy. Furthermore, as research into viral engineering advances, NDV-derived strategies have the potential to become valuable tools in the therapeutic arsenal against cancer (Burman et al., 2020). Given that cancer remains a leading cause of global mortality characterized by high pathophysiological diversity, the search for novel biological agents such as targeted therapies and immunotherapies is a continuous priority in medical research (Diori Karidio and Sanlier, 2021; Panteleakou et al., 2009). In this context, the identification of bioactive molecules that mimic viral functions or structural components offers a strategic alternative to using the whole virus. Phage display is a powerful molecular biology technique used to identify peptides or proteins that interact with specific targets (Kim and Samal, 2016; Paschke, 2006) . By utilizing vast libraries of bacteriophages expressing random peptides, this technology allows for the selection of high-affinity ligands through a process called biopanning. The identified sequences can then be determined via DNA sequencing, facilitating the development of new diagnostic platforms, vaccines, and targeted drug candidates (Sidhu and Koide, 2007; Jaroszewicz et al., 2022). Therefore, the present study aimed to identify and characterize novel peptides mimetic to NDV using phage display, evaluating their structural properties and prospects as molecular tools for biotechnological applications. Results and discussion NDV antibody purification It was observed that injecting 0.5 mL of undiluted serum into a 1 mL loop under a pressure of 0.4 MPa and a flow rate of 10 mL/min optimized the performance of the HisTrap column (Redwan et al., 2021). All purified samples were quantified via BCA assay, yielding concentrations of 1.6 mg/mL for immunized IgY and 3.4 mg/mL for the control IgY. The flow-through concentrations were 4.5 mg/mL and 6.2 mg/mL for the immunized and control groups, respectively. Samples were standardized to 1 mg/mL for SDS-PAGE analysis (Fig. 1).Fig. 1Electrophoretic Profile of Purified IgY and Control Proteins. Coomassie Blue-stained SDS-PAGE highlighting the purity and structural integrity of the samples. Lane (A): Molecular weight protein marker (250 kDa); Lane (B): Bovine Serum Albumin (BSA) used as a molecular weight reference (∼66 kDa); Lane (C): Purified IgY from immunized serum under non-reducing conditions, showing the characteristic high-molecular-weight band (∼180 kDa); Lane (D): Purified IgY from control (non-immunized) serum under non-reducing conditions.Fig 1: dummy alt text Electrophoretic Profile of Purified IgY and Control Proteins. Coomassie Blue-stained SDS-PAGE highlighting the purity and structural integrity of the samples. Lane (A): Molecular weight protein marker (250 kDa); Lane (B): Bovine Serum Albumin (BSA) used as a molecular weight reference (∼66 kDa); Lane (C): Purified IgY from immunized serum under non-reducing conditions, showing the characteristic high-molecular-weight band (∼180 kDa); Lane (D): Purified IgY from control (non-immunized) serum under non-reducing conditions. As expected, the column was able to perform, even with a sample of a different nature than that intended by the manufacturer. Considering that the molecular weight of IgY is slightly variable and studies suggest finding IgYs around 180 kDa, we can observe IgY antibodies between the markers 250 kDa and 130 kDa (Yaghubi et al., 2020; Schirrmacher; Van Gool; Stuecker 2021). We noted a slight decrease in purification efficiency compared to egg-white-derived samples previously purified with the same column, likely due to the higher complexity and protein content of the serum matrix. The presence of a single, faint band at a lower molecular weight, alongside the prominent intact immunoglobulin band, confirms the overall stability and high purity of the purified IgY fractions. NDV mimetic peptide Isolated antibodies were used to select peptides using phage display. Non immunized antibodies, control, were used to eliminate phages binding to basal antibodies of immunized species. Antibodies immunized with NDV were used to obtain peptides mimetic to the virus. A total of 48 clones were isolated after three rounds of selection. Reactivity was confirmed by phage-ELISA (Fig. 2A). The 10 clones with the highest reactivity (ratio > 1.0) were selected for sequencing, resulting in three peptides, named: NDV-1, NDV-2, and NDV-3, all showing statistical difference with negative control (Fig. 2B).Fig. 2Binding Specificity of Selected Phage Clones to NDV-IgY. (A) Indirect Phage-ELISA screening of 10 highest individual clones after three rounds of biopanning. The reactivity of the clones was evaluated against NDV-specific IgY (target) and control IgY. (B) Comparative binding affinity of the most reactive mimotopes (NDV-1, NDV-2, and NDV-3) identified during the selection process. Data are expressed as the mean ± standard deviation (SD) of three independent experiments performed in triplicate.Fig 2: dummy alt text Binding Specificity of Selected Phage Clones to NDV-IgY. (A) Indirect Phage-ELISA screening of 10 highest individual clones after three rounds of biopanning. The reactivity of the clones was evaluated against NDV-specific IgY (target) and control IgY. (B) Comparative binding affinity of the most reactive mimotopes (NDV-1, NDV-2, and NDV-3) identified during the selection process. Data are expressed as the mean ± standard deviation (SD) of three independent experiments performed in triplicate. In silico analysis of NDV-1, NDV-2, and NDV-3 Peptides deduced sequences were made by Expasy, generation the following sequences: KNIGAPW, LNTTRWF, and IHSPTAL (NDV-1, NDV-2, and NDV-3, respectively). BLAST analysis showed that NDV-2 shares 60% identity with a fragment of the NDV large polymerase protein, suggesting it may act as a linear mimotope. In contrast, NDV-1 and NDV-3 showed no significant linear alignment with viral proteins. These likely represent conformational (topological) mimotopes, where the spatial arrangement of residues replicates the viral epitope's binding characteristics independently of primary sequence homology (Andreatta et al., 2017). The fact that NDV-1 and NDV-3 successfully bind to NDV-specific IgY in ELISA assays, despite the lack of BLAST alignment, strongly suggests they act as structural mimics of topographical epitopes on the virus surface, even in the absence of linear sequence homology. The value of phage-derived peptides lies in their functional mimicry, which is independent of direct genomic homology, a common finding in the discovery of novel diagnostic and therapeutic ligands. Using Clustal Omega server, we aligned our peptides sequences. It was observed that although there are no longer conserved regions among the peptides, the is a moderate conservation on their sequences, observed by the green and blue colors (Fig. 3A). Considering that peptides sequence are only 7 aa, such alignment regions indicate that peptides may bind to the target using such conserved sequence, with may represents their sensitivity and affinity do anti-NDV (Fig. 3A) (Crooks et al., 2004; Frank and Sippl, 2008). The infographic shows the peptide aa distribution after the alignment. Proline (P) and Tryptophan (W) were frequent at positions 6 and 7, while Glycine (G) and Serine (S) appeared consistently in NDV-1 and NDV-3 (Fig. 3B). This visualization highlights the consensus motifs potentially responsible for the binding interaction with the target IgY.Fig. 3Conservation and Frequency Analysis of Selected Peptides. A) Multiple Sequence Alignment (MSA): Primary amino acid sequences of the selected mimotopes aligned using the Clustal Omega algorithm to identify conserved residues and motifs. B) Sequence Logo Representation: Quantitative infographic demonstrating the positional frequency of amino acids within the aligned peptides. The height of each letter on the y-axis is proportional to the residue frequency at that specific position, while the x-axis represents the sequential position (1 to 9, including flanking residues).Fig 3: dummy alt text Conservation and Frequency Analysis of Selected Peptides. A) Multiple Sequence Alignment (MSA): Primary amino acid sequences of the selected mimotopes aligned using the Clustal Omega algorithm to identify conserved residues and motifs. B) Sequence Logo Representation: Quantitative infographic demonstrating the positional frequency of amino acids within the aligned peptides. The height of each letter on the y-axis is proportional to the residue frequency at that specific position, while the x-axis represents the sequential position (1 to 9, including flanking residues). Structural modeling and comparison We compared structural predictions from AlphaFold2 and PEP-FOLD3 to evaluate peptide stability and conformation (Fig. 4A). AlphaFold2 predictions yielded pLDDT scores ranging from 59.1 to 60.2, indicating medium confidence. This level of confidence is expected for ultra-short 7-mer peptides, which inherently possess high conformational flexibility and structural instability in solution (Jumper et al., 2021; Mariani et al., 2013; Varadi et al., 2022). Modeling such short linear sequences presents significant biophysical challenges due to their extreme flexibility. Consequently, the reported conformations, such as the helical motif in NDV-3 or the extended backbone in NDV-2, are predicted predominant states within a computational environment rather than rigid, static structures.Fig. 4Structural Modeling and Conformational Analysis of Identified Mimotopes. A) AlphaFold2 Predictive Modeling: Structural representations highlighting the Predicted Local Distance Difference Test (pLDDT) scores. The color-coded scale indicates the per-residue confidence levels for each 7-mer peptide. B) PEP-FOLD3 De Novo Modeling: Best-ranked 3D tertiary structures predicted for NDV-1, NDV-2, and NDV-3. C) Structural Superposition: Overlay of the 3D models generated by PEP-FOLD3 (blue) and AlphaFold2 (variable colors). Alignments were anchored by the N-terminal residue to evaluate conformational convergence between the two algorithms. D) Local Secondary Structure Propensity: Probability profiles for each amino acid residue across the peptide sequences. The structural states are defined by the following color code: red (helical), green (extended/beta-strand), and blue (random coil).Fig 4: dummy alt text Structural Modeling and Conformational Analysis of Identified Mimotopes. A) AlphaFold2 Predictive Modeling: Structural representations highlighting the Predicted Local Distance Difference Test (pLDDT) scores. The color-coded scale indicates the per-residue confidence levels for each 7-mer peptide. B) PEP-FOLD3 De Novo Modeling: Best-ranked 3D tertiary structures predicted for NDV-1, NDV-2, and NDV-3. C) Structural Superposition: Overlay of the 3D models generated by PEP-FOLD3 (blue) and AlphaFold2 (variable colors). Alignments were anchored by the N-terminal residue to evaluate conformational convergence between the two algorithms. D) Local Secondary Structure Propensity: Probability profiles for each amino acid residue across the peptide sequences. The structural states are defined by the following color code: red (helical), green (extended/beta-strand), and blue (random coil). While these in silico models provide valuable insights into potential bioactive motifs, the inherent flexibility of 7-mers remains a critical factor. Performing an extension or grafting these 7-mer sequences into more stable structural scaffolds could achieve better stability to the bioactive conformation. Therefore, these predictions serve as a structural basis for future molecular optimization, acknowledging that structural instability is a known hurdle in the translation of short mimotopes into functional therapeutic or vaccine candidates. Furthermore, for future therapeutic or diagnostic applications, these peptides would likely require multimerization (e.g., tandem repeats or MAPs) to increase avidity, or conjugation to larger carrier proteins or nanoparticles to enhance stability and reduce off-target risks. Additionally, PEP-FOLD3 uses a physical simulation algorithm searching for the structure with the lowest predicted free energy scores, which may not necessarily correspond to the native structure (Fig. 4B) (Thevenet et al., 2012). PEP-FOLD3 provided ranked models based on predictive free energy scores: −11.82 (NDV-1), −10.37 (NDV-2), and −8.18 (NDV-3). It is important to clarify that these are computational modeling scores and not experimental thermodynamic binding energies (KD). Interestingly, NDV-3 exhibited the highest predictive energy (least stable in isolation), which correlated with its superior reactivity in ELISA (Fig. 2B). We hypothesize that this relative instability might facilitate a more flexible induced fit during antibody-target interaction (Lamiable et al., 2016; Shen et al., 2014; Wang and Miao, 2020). Lastly, we compared the 3D structures generated by both programs (Fig. 4C). NDV-3 displayed a spiral-like tendency in both models, while NDV-1 and NDV-2 favored helical and extended conformations, respectively. However, given the short length of these peptides, these secondary structures should be interpreted as predicted conformational tendencies rather than rigid, definitive states. The modeling (PEP-FOLD3 and AlphaFold2) included the specific flanking residues used in the phage display construct (AC- at the N-terminus and -CGGS at the C-terminus). This approach ensures that the in silico predictions account for the structural constraints imposed by the neighboring amino acids in the pIII fusion protein. It is worth mentioning that in addition to the 3D structure of peptides (Fig. 4B), the local structural propensity for each aa residue was calculated using the PEP-FOLD3 server (Fig. 4D). A comparative analysis of the structural models and the probability profiles reveals that the NDV-3 peptide has a greater tendency to a spiral/helical structure. Given that peptide binding efficacy depends not only on primary sequence but also on the adoption of a bioactive native conformation (Li et al., 2021), this well-defined helical structure in NDV-3 likely contributes to its superior binding affinity toward anti-NDV IgY observed in our assays. In contrast, NDV-2 presented a more extended, flexible backbone, while NDV-1 exhibited a less stable helical arrangement. These structural characteristics, coupled with their intermediate affinities, suggest that these conformations may be less energetically favorable for paratope interaction compared to the robust helical motif identified in NDV-3 (Levin et al., 2020). These findings reinforce the importance of secondary structure stability in the functional mimicry of viral epitopes. Immunogenicity prospecting Using BepiPred-3.0, the peptides were screened for B-cell interaction potential (Fig. 5). The dashed horizontal line indicates the default threshold (0.15); residues exceeding this baseline are considered to possess significant potential immunogenicity. These data highlight the structural segments within the 7-mer sequences that are most likely to mimic native viral antigenic determinants. All three peptides scored above the default threshold (0.1512), with an average score of approximately 26.6, suggesting a high sensitivity for epitope recognition. NDV-1 showed the highest reactivity score (yellow regions), which, combined with its lower predicted free energy, suggests it may be a candidate for future immunogenicity studies. These results are preliminary and hypothesize that the tertiary structure tendencies observed in Fig. 4 influence the accessibility of these predicted epitopes (Hoo and Siak, 2020; Muhammad et al., 2023).Fig. 5Mimotopes predicted by BepPred-3.0. Scores are shown by color gradient, the brighter meaning higher possibility to interact with B-cell’s receptors. The propensity scores for each amino acid residue are represented by a color gradient, where increased intensity (brightness) correlates with a higher probability of serving as a functional B-cell epitope.Fig 5: dummy alt text Mimotopes predicted by BepPred-3.0. Scores are shown by color gradient, the brighter meaning higher possibility to interact with B-cell’s receptors. The propensity scores for each amino acid residue are represented by a color gradient, where increased intensity (brightness) correlates with a higher probability of serving as a functional B-cell epitope. Conclusion In conclusion, three mimetic peptides (NDV-1, NDV-2, and NDV-3) were successfully identified by phage display as specifically reactive to Newcastle Disease Virus (NDV) antibodies. Although the use of a single biological replicate for immunization enabled effective mimotope selection, further studies involving larger cohorts are required to assess inter-individual variability and the reproducibility of these findings across distinct immune backgrounds. Among the identified candidates, NDV-3 exhibited the highest binding sensitivity, which may be associated with its higher predicted free energy and “spiral-like” conformational profile, potentially favoring a more flexible induced-fit interaction. These features support its potential application in targeted strategies, including peptide-based diagnostics or as a scaffold for investigating virotherapy-related mechanisms, pending additional validation. In contrast, NDV-1 and NDV-2 displayed lower predicted free energy and more stable helical or extended conformations, which were associated with reduced binding sensitivity in the applied models. Nevertheless, the elevated epitope scores and conformational stability of NDV-1 indicate its potential relevance for immunogenicity studies and vaccine development. It is important to emphasize that these findings are based on computational predictions and preliminary binding assays. Therefore, comprehensive in vitro and in vivo investigations including cytotoxicity, neutralization assays, and stability assessments in biological fluids are necessary to fully characterize the biological properties of these mimotopes and to establish their translational potential in immunobiotechnology and oncology. Methods Ethics statement and immunization All animal procedures were conducted in accordance with institutional ethical guidelines and approved by the INATA under protocol number 003/01.0658.2021. To obtain Newcastle Disease Virus (NDV) antibodies, White Leghorn chickens were immunized with either an attenuated form of the virus or phosphate-buffered saline (PBS) as a negative control. Immunizations and serum collections were performed by the research department at Inata Biológicos, and the resulting serum samples were donated for this study. IgY Purification via HPLC Antibodies were purified using a HisTrap IgY Purification HP column (GE Healthcare). Although primarily designed for egg white samples, the protocol was adapted for serum samples. The column was equilibrated with a cleaning buffer (20 mM sodium phosphate, 30% isopropanol, pH 7.5), followed by a pre-run with binding buffer (20 mM sodium phosphate, 0.5 M potassium sulfate, pH 7.5). Undiluted serum (0.5 mL) was injected into a 1 mL loop at a flow rate of 10 mL/min under 0.4 MPa of pressure. Elution was performed using 20 mM sodium phosphate buffer, and purified polyclonal IgY raised against the whole attenuated virus fractions were stored at −20 °C. Protein quantification and SDS-PAGE Protein concentrations were determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific) according to the manufacturer's instructions. Absorbance was measured at 562 nm using a VICTOR Nivo™ microplate reader (PerkinElmer). For purity assessment, samples were standardized to 1 mg/mL, denatured at 95 °C for 5 min in reducing sample buffer, and resolved on a 16% SDS-PAGE gel with a 3% stacking gel. Electrophoresis was performed at 120 V for 60 min. Gels were stained with Coomassie Blue for 1 hour and destained overnight in a solution of 50% methanol and 10% acetic acid. Biopanning and Phage selection To identify NDV-mimetic peptides, we performed three rounds of affinity selection (biopanning) using the Ph.D.−7™ Phage Display Peptide Library (New England Biolabs), which consists of randomized 7-mer linear peptides fused to the pIII coat protein of M13 bacteriophages, with an initial complexity of approximately 109 unique sequences. First Round and Subtractive Selection: High-affinity microplates (Greiner Bio-One) were coated overnight at 4 °C with 100 μg/mL of either NDV-specific IgY (target) or control IgY (negative control) in bicarbonate buffer (0.1 M NaHCO3, pH 8.6). Each well received 150 μL of the protein solution. On the following day, the wells were blocked with 5% Bovine Serum Albumin (BSA) in PBS for 1 hour at 37 °C to prevent non-specific binding. To ensure the selection of NDV-specific ligands, a subtractive biopanning strategy was employed. The phage library (input titer of 1011 pfu in 100 μL) was first incubated in a negative control well (PBS-immunized IgY) for 30 min at 37 °C. The supernatant containing unbound phages was sequentially transferred to two additional negative control wells. This depleted library was then transferred to the first positive well (NDV-immunized IgY) and incubated for 60 min at 37 °C. Washing, Elution, and Amplification: Non-binding phages were removed by washing the positive wells 10 times with PBS containing 0.1% Tween-20 (PBST). Bound phages were eluted using 100 μL of 0.2 M Glycine-HCl (pH 2.2) for 10 min and immediately neutralized with 15 μL of 1 M Tris-HCl (pH 9.1). The eluate was then used to infect E. coli ER2738 cultures in the early log phase (OD600 ≈ 0.3) for amplification. Subsequent Rounds and Titration: The amplified phages were purified by double precipitation with 20% Polyethylene Glycol (PEG 8000)/2.5 M NaCl. Two additional rounds of biopanning were conducted following the same protocol, but with increased selection stringency by raising the PBST concentration to 0.5% in the washing steps. After the third round, the final eluate was titrated on LB/IPTG/X-gal plates to isolate individual blue plaques for DNA extraction, sequencing, and Phage-ELISA validation. Phage-ELISA High-affinity microplates (Greiner) were filled with a solution of 100 μg/mL of IgY in bicarbonate buffer (0.1 M NaHCO3 – pH 8.6) overnight at 4 °C. With 3 wells for the negative IgY control and 3 positive IgY wells, for each isolated bacteriophage. On following day, the plate was washed with PBS-T 3 times and blocked with PBS-BSA for 1 hour at 37 °C. Subsequently, the plate was washed with PBS-T 3 times and incubated for 1 hour at 37 °C. After incubation, it was again washed again 5 times with PBS-T and then secondary antibodies against bacteriophages conjugated with peroxidase (SIGMA) diluted in PBS-BSA were added, again being incubated for 1 hour at 37 °C. HRP-conjugated anti-M13 secondary antibody (Sigma-Aldrich) was added, followed by detection with an OPD solution (H2O2 and o-phenylenediamine) in citrate buffer. The reaction was stopped with 2 M H2SO4, and absorbance was measured at 492 nm. DNA extraction and sanger sequencing DNA was extracted from selected high-reactivity clones using an iodide-based lysis buffer followed by ethanol precipitation. The gIII gene region was sequenced using the BigDye™ Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems). Capillary electrophoresis was performed on an ABI 3730xl Genetic Analyzer. Sequence chromatograms were analyzed using Chromas Lite and Geneious 4.8.3 software. In silico analysis DNA sequences were translated into amino acid sequences using ExPASy. 3D structural modeling was performed using PEP-FOLD3 and ColabFold v1.5.5 (AlphaFold2). Sequence alignments were conducted via BLAST (against NDV proteins) and Clustal Omega. Sequence patterns were visualized using SeqLogos, and potential B-cell mimotopes were predicted using the BepiPred-3.0 server. Statistical analysis Statistical analysis was performed using GraphPad Prism (8.0.1). Data from Phage-ELISA assays were expressed as the mean ± standard deviation (SD) of three independent experiments (n = 3). Differences between groups were evaluated using one-way ANOVA followed by Tukey's post-hoc test. A p-value < 0.05 was considered statistically significant. Associated content This study does not have any supplementary materials. Funding This work was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), the Fundação de Amparo à Pesquisa de Minas Gerais (FAPEMIG), the Rede Mineira de Pesquisa Translacional em Imunobiológicos e Biofármacos no Câncer (REMITRIBIC, RED-00031–21), and SisNANO-MCTI. CRediT authorship contribution statement D. Caixeta Moreira: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Formal analysis. E. Resende Vaz: Methodology, Data curation. H. Oliveira Almeida Souza: Methodology, Data curation. S. Lorenti: Methodology, Data curation. T. Velikkakam: Methodology, Data curation. B. Beatriz Fonseca: Resources, Investigation, Conceptualization. V. Alonso-Goulart: Writing – review & editing, Writing – original draft, Supervision, Project administration, Funding acquisition. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.