Model

Introduction

We initially aimed to express the antibody sequence of Ustekinumab in Lactobacillus, but encountered frequent plasmid mutations during replication in the bacterial host, which compromised expression stability. To address this, we designed three feasible IL-23 scFv prototype constructs and utilized tools such as AlphaFold and ClusPro for structural modeling and binding prediction. These prototypes were subsequently used for expression and functional analysis in HEK293T cells.

Anti-IL23 scFv Design

  • Methodology

    We follow a no structure optimization workflow (no Rosetta/ABlooper at this stage). To build a human-compatible scaffold, we selected human germline frameworks IGHV3-23*01 (VH) and IGKV1-39*01 (VL) and grafted our designed CDRs onto these backbones. These frameworks are widely used in mAbs/scFvs and are well represented in public repositories, which supports annotation consistency and downstream developability.

    The target antigen IL-23 p19 features positively charged surface patches with shallow grooves and relatively flat regions. Consequently, our sequence strategy emphasizes multi-point interactions using polar residues (to form H-bond/ionic contacts on basic patches) and aromatic residues (for π–π/π–cation and shape complementarity). Candidate CDR motifs were first aggregated via LLM-assisted motif mining (fragment-level pattern extraction/rationalization), then manually curated and cross-checked against IMGT, SAbDab, OAS, and abYsis for observability and biochemical plausibility.

    1. VH: Framework similar to IGHV3-2301*
    2. VL: Framework similar to IGKV1-3901*
      • Source:Selected from the IMGT structural database.
      • IGKV1-3901* is a frequent human kappa light chain germline known for its structural stability and high expression yield.
      • A real Fab antibody structure based on this germline was chosen as the backbone to facilitate accurate modeling and loop grafting.
  • CDR Design (Hypothesis-Driven, Sequence-Only)

    Approach. Introduce controlled, rational mutations within CDRs and graft onto IGHV3-2301 / IGKV1-3901 to test hypotheses tailored to p19’s chemistry:

    1. HCDR3 as the affinity core: combine polar kernels (for H-bond/ionic networks) with compact aromatic tails to “pack” shallow grooves.
    2. LCDR3 as a positioning/pocket anchor: small aromatic + polar mixes (e.g., NWP-centered variants) to aid docking on shallow surfaces.
    3. HCDR2 as a polar density stripe: Tyr/Asn/Asp/Ser clusters to engage basic patches (Arg/Lys) on p19.
    4. HCDR1/LCDR1/LCDR2 tuned for flexibility, hydration, and minimal structural burden.

Reference

  1. Lloyd, S. B., Niven, K. P., Kiefel, B. R., Montefiori, D. C., Reynaldi, A., Davenport, M. P., Kent, S. J., & Winnall, W. R. (2017). Exploration of broadly neutralizing antibody fragments produced in bacteria for the control of HIV. Human vaccines & immunotherapeutics, 13(11), 2726–2737.

  • Features and Advantages:
    • This design emphasizes a short and compact HCDR3 loop featuring the ARDY motif, which effectively fits the small binding interface of the IL-23 p19 subunit, enhancing binding specificity and structural stability.
    • The HCDR2 contains multiple polar residues (Y, D, N) that provide a rich potential for hydrogen bonding and electrostatic surface interactions, thereby strengthening antigen contact.
    • The LCDR1 adopts an extended design that combines flexibility and reach, allowing it to increase the binding surface area and compensate for minor structural variations on the p19 surface.

      LCDR2 and LCDR3 remain minimalistic yet structurally stable, supporting the overall loop architecture and reducing the risk of misfolding or distortion.

  • Full Sequence:

    anti-IL-23_scfv1 (Signal peptide – VH – Linker – VL – HisTag):

    METDTLLLWVLLLWVPGSTGEVQLLESGGGLVQPGGSLRLSCAASGFTFGYTFTNYAMMSWVRQAPGKGLEWVSAIYSGSNDYYGYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARDYWGSSYFDYFGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITCRASRASKSVSTSGSLNWYQQKPGKAPKLLIYDTWTSLQSGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQYNNWPPTHHHHHH

  • CDR Sequence
    Chain Region Sequence
    VHHCDR1GYTFTNYAM
    VHHCDR2IYSGSNDYYG
    VHHCDR3ARDYWGSSYFDYF
    VLLCDR1RASKSVSTSGS
    VLLCDR2DTWT
    VLLCDR3QQYNNWPPT

  • Features and Advantages:

    The H3 loop is relatively short and compact, making it well-suited for binding to small surface areas.

  • Full Sequence:

    anti-IL-23_scfv2 (Signal peptide – VH – Linker – VL – HisTag):

    METDTLLLWVLLLWVPGSTGEVQLLESGGGLVQPGGSLRLSCAASGFTFGYTFTNYMSWVRQAPGKGLEWVSASGSNDYYYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCRDYWGSSYFDYGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITCRASRASQGSSWLNWYQQKPGKAPKLLIYDTYYSLQSGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQYSNWPLTHHHHHH

  • CDR Sequence
    Chain Region Sequence
    VHHCDR1GYTFTNY
    VHHCDR2SGSNDYY
    VHHCDR3RDYWGSSYFDY
    VLLCDR1RASRASKSV
    VLLCDR2DTYY
    VLLCDR3QQYSNWPLT

  • Features and Advantages:
    • The H3 loop is relatively long and contains aromatic and positively charged amino acids, enabling π–π stacking and salt bridge interactions with the acidic regions on the IL-23 p19 surface.
    • The design was guided by docking pose contact points, including residues such as Glu34, Asp70, and Tyr88 on IL-23 p19.
  • Full Sequence:

    anti-IL-23_scfv3 (Signal peptide – VH – Linker – VL – HisTag):

    METDTLLLWVLLLWVPGSTGEVQLLESGGGLVQPGGSLRLSCAASGFTFGFTFSNYMSWVRQAPGKGLEWVSASGSNSYYYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCGRGFDHRGYWYYGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITCRASRASQGTSWLNWYQQKPGKAPKLLIYDSSSSLQSGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQYSNWPPLHHHHHH

  • CDR Sequence
    Chain Region Sequence
    VHHCDR1GFTFSNY
    VHHCDR2SGSNSYY
    VHHCDR3GRGFDHRGYWYY
    VLLCDR1RASQGTSW
    VLLCDR2DSSS
    VLLCDR3QQYSNWPPL

  • Features and Advantages:
    • This construct directly uses the variable heavy (VH) and variable light (VL) regions of risankizumab as its framework.
    • We converted variable regions (VH/VL) of risankizumab into single-chain variable fragments (scFv) and assessed whether binding/function can be retained without the IgG architecture.
    • Directly use the publicly available VH/VL sequences of risankizumab for academic research. Build both orientations: VH–(G₄S)₃–VL
  • Figure 1. The schematic diagram of scFv p19.
  • Full Sequence:

    anti-IL-23_scfv_p19 (Signal peptide – VH – Linker – VL – HisTag):

    METDTLLLWVLLLWVPGSTGGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITCKASRDVAIAVAWYQQKPGKVPKLLIYWASTRHTGVPSRFSGSGSRTDFTLTISSLQPEDVADYFCHQYSSYPFTFGSGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGECHHHHHH

  • Features and Advantages:
    • This construct directly uses the variable heavy (VH) and variable light (VL) regions of ustekinumab as its framework.
    • We converted variable regions (VH/VL) of ustekinumab into single-chain variable fragments (scFv) and assessed whether binding/function can be retained without the IgG architecture
    • Directly use the publicly available VH/VL sequences of ustekinumab for academic research.Build both orientations: VH–(G₄S)₃–VL
  • Figure 2. The schematic diagram of scFv p40.
  • Full Sequence:

    anti-IL-23_scfv_p40 (Signal peptide – VH – Linker – VL – HisTag):

    METDTLLLWVLLLWVPGSTGEVQLVQSGAEVKKPGESLKISCKGSGYSFTTYWLGWVRQMPGKGLDWIGIMSPVDSDIRYSPSFQGQVTMSVDKSITTAYLQWNSLKASDTAMYYCARRRPGQGYFDFWGQGTLVTVSSSSTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSNTKVDKRVEPKSCDKTHGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITCRASQGISSWLAWYQQKPEKAPKSLIYAASSLQSGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQYNIYPYTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGECHHHHHH

Protein Structure Modeling

AlphaFold2 was used to model the protein structure and verify whether the grafting strategy results in a stable conformation.

Figure 3. Sequence Coverage and Identity Plot
Figure 4. AlphaFold2 Predicted scFv #1 Structure
Figure 5. PyMOL visualization of predicted scFv #1 structure

Figure 6. Sequence Coverage and Identity Plot.
Figure 7. AlphaFold2 Predicted scFv #2 Structure.
Figure 8. PyMOL visualization of predicted scFv #2 structure.

Figure 9. Sequence Coverage and Identity Plot.
Figure 10. AlphaFold2 Predicted scFv #3 Structure.
Figure 11. PyMOL visualization of predicted scFv #3 structure.

Figure 12. Sequence Coverage and Identity Plot.
Figure 13. AlphaFold2 Predicted scFv p19 Structure.
Figure 14. PyMOL visualization of predicted scFv p19 structure.

Figure 15. Sequence Coverage and Identity Plot.
Figure 16. AlphaFold2 Predicted scFv p40 Structure.
Figure 17. PyMOL visualization of predicted scFv p40 structure.

Antibody–Antigen Docking Analysis

To evaluate whether the designed antibody—generated through CDR grafting—has the potential to bind IL-23, we performed antibody–antigen docking using the ClusPro server.

The input structures included a modeled 3D structure of the antibody Fab region and the IL-23 antigen structure retrieved from the Protein Data Bank (PDB). Docking was carried out using ClusPro’s Antibody Mode, which is specifically optimized for antibody–antigen interactions.

Binding stability was assessed based on ClusPro’s default weighted energy scoring function:

Weighted Score = 0.4 × Repulsive + (−0.4 × Attractive) + 600 × Electrostatic + 1 × Desolvation.

This approach allowed us to preliminarily assess the feasibility of antigen binding for the grafted antibody design.

  1. ClusPro Docking Results

    A total of 30 clusters were generated from the ClusPro docking analysis. The top three clusters, based on their weighted scores for both center and lowest-energy poses, are summarized below:

    • Cluster 4:

      • Center pose: −327.5
      • Lowest energy pose: −390.0
      • Number of members: 43
      • Interpretation: This cluster exhibits the most favorable lowest-energy pose among all clusters (−390.0), suggesting a highly stable and energetically preferred binding conformation. Additionally, it has a large number of members (43), indicating that this binding mode is frequently sampled and potentially biologically relevant.
    • Cluster 13:

      • Center pose: −313.5
      • Lowest energy pose: −376.5
      • Number of members: 23
      • Interpretation: Although the number of members is slightly lower, the cluster demonstrates a well-converged and energetically favorable conformation, as reflected by the close agreement between center and lowest-energy poses. This consistency implies structural robustness and reliable docking behavior.
    • Cluster 8:

      • Center pose:−369.5
      • Lowest energy pose: −374.6
      • Number of members: 34
      • Interpretation: While its lowest-energy score is marginally higher than the others, Cluster 8 presents a tight energy distribution between its center and best pose. With 34 members, this cluster represents a recurrent and stable binding configuration, potentially corresponding to a native-like interaction.

    Overall Conclusion

    Taken together, Cluster 4 appears to be the most promising candidate due to its most favorable lowest-energy score (−390.0) and high member count, indicating both thermodynamic stability and frequent sampling. Cluster 13 displays excellent convergence and energetic stability, while Cluster 8 shows a consistent and recurrent binding pose with moderate energetic favorability. These three clusters represent strong candidates for further structural and functional analysis.

  2. Contact Residues (Deduplicated)
    • Antibody side (Ab)

      TYR31, PHE33, TYR36, MET38, ALA55, ILE56, TYR57, GLY59, SER60, ASN61, ASP62, TYR63, TYR64, TYR66, ASP69, LYS72, TRP108, GLY109, SER111, SER162, VAL163, SER164, SER168, TYR228, TRP231

    • Antigen side (Ag, IL-23)

      GLY21, GLY70, GLY72, ARG79, ASP80, ASN81, SER82, GLN83, PHE84, GLN87, ARG88, GLN91, GLU134, GLY135, HIS136, HIS137, TRP138, GLU139, THR140, GLN141, GLN142

      Figure 18. anti-IL-23 scFv 1 Docking with IL23 p19 subunit amination.
  3. Epitope Distribution and Structural Characteristics

    The IL-23 epitope presents a two-patch architecture:

    • Patch A centers around Arg79 and nearby polar/charged residues, forming a localized interaction hub.
    • Patch B spans residues 134–142, a region rich in His, Trp, Gln, and polar/charged side chains, typically forming a short loop or surface-exposed segment on IL-23.
  4. Interaction Chemistry and Interpretation
    1. Hydrophobic and Aromatic Patches
      • Antibody side features Phe33,Tyr31/36/57/63/64/66/228, Met38, Trp108/231, Ile56, forming a well-defined hydrophobic and aromatic patch.
      • Antigen side includes Leu-like (Phe84), His-rich residues (His136/137), and Trp138, suggesting possible π–π stacking, cation–π interactions, and hydrophobic packing at the Patch B interface.
      • This interaction contributes significantly to the low docking energy and structural stability observed in ClusPro.
    2. Electrostatic and Salt-Bridge Interactions
      • Asp62, Asp69, Glu139 (Ab) and Arg79, Arg88 (Ag) may form salt bridges or ion-pair interactions, anchoring the N-terminal patch.
      • Lys72 (Ab) and Glu134/139 (Ag) further complement the interface electrostatics, potentially stabilizing conformational docking.
    3. Hydrogen Bond Networks
      • Ser60, Ser111, Ser162/164/168, Gln141/142, Thr140, Asn61, Tyr223 (Ab) and Gln83/87/91, Ser82, His137, Thr140 (Ag) participate in multi-point hydrogen bonding.
      • These polar contacts help lock the geometry, reinforcing the core hydrophobic/π interactions with directional stability.
  5. Geometrical and Positional Features
    1. Two-Patch Binding
      • Patch A (N-terminal): Ab residues Tyr31–Asp62, especially charged/polar residues, interact with Ag Arg79, Asp80, forming a charge-dominated anchor zone.
      • Patch B (residues 134–142): Ab residues Trp108, Tyr228, Trp231, Ser/Gln-rich loops pack against His-Trp-rich Ag region, forming aromatic–hydrophobic + H-bond clusters.
    2. CDR Involvement (Estimated)
      • Contact residues span 31–38, 55–72, 108–111, 162–168, 228–231, suggesting contributions from HCDR1, HCDR3, LCDR1, LCDR3, aligning with the expected H3-dominant + L3/H1-assisted binding mode.
  6. Conclusion

    The scFv binds IL-23 through a dual-patch interface involving:

    • A charged/polar anchor near Arg79 (Patch A),
    • A hydrophobic–aromatic core around residues 134–142 (Patch B), with multi-point hydrogen bonding, π–π stacking, and salt-bridge formation.

      These complementary interactions collectively explain the favorable docking energy and suggest a biochemically robust and geometrically coherent binding interface.

  1. ClusPro Docking Results Overview

    A total of 30 clusters were generated from the ClusPro docking analysis. The top three clusters, based on their weighted scores for both center and lowest-energy poses, are summarized below:

    • Cluster 6:

      • Center pose: −320.0
      • Lowest energy pose: −390.7
      • Number of members: 32
      • Interpretation: This cluster contains the most favorable lowest-energy pose (−390.7), indicating a highly stable and energetically favorable binding conformation. Despite its moderate member count, the extreme energy minimum suggests a strong and specific interaction in at least one sampled orientation.
    • Cluster 2:

      • Center pose: −379.0
      • Lowest energy pose: −379.0
      • Number of members: 54
      • Interpretation: The identical values for center and lowest-energy poses indicate excellent convergence, suggesting that the sampled binding conformations are highly consistent and structurally reliable. With 54 members, this cluster represents a frequently sampled and potentially biologically relevant binding mode.
    • Cluster 0:

      • Center pose:−366.5
      • Lowest energy pose: −372.8
      • Number of members: 78
      • Interpretation: Although the energy scores are slightly less favorable compared to Clusters 6 and 2, this cluster has the largest number of members (78), implying a dominant and recurrent binding pattern during the docking simulation. The narrow energy gap between center and best poses further supports its structural coherence.

    Overall Conclusion

    Taken together, Cluster 6 exhibits the most favorable lowest-energy docking conformation (−390.7), suggesting a particularly stable interaction. Cluster 2, with perfect convergence and a substantial number of members, presents a compelling and biologically plausible binding mode. Meanwhile, Cluster 0, characterized by its large membership (78), may reflect the most frequently sampled interaction pose, reinforcing its relevance for further investigation.

  2. Contact Residues (Deduplicated)
    • Antibody side (Ab)

      THR32, PHE33, THR34, GLY103, SER104, SER105, TYR106, PHE107, ASP108, TYR109, GLY112, GLY113, GLY155, SER156, TRP158, THR177, TYR178, TYR218

    • Antigen side (Ag, IL-23)

      HIS49, MET50, ASP51, LEU52, ARG53, GLU54, GLU55, GLY56, ASP57, GLU58, GLU59, THR61, ASN62, HIS66, GLN68, CYS69, LYS160, ARG163

      Figure 19. anti-IL-23 scFv 2 Docking with IL23 p19 subunit amination.
  3. Epitope Distribution and Structural Characteristics

    The IL-23 epitope is distributed across two structurally distinct patches:

    • Patch A: residues 49–55, a segment enriched with polar and charged side chains including His49, Arg53, Glu54/55, likely forming a surface loop.
    • Patch B: residues 160–163, a short region near Lys160 and Arg163, potentially located near the C-terminal side of a domain or flexible linker.
  4. Interaction Chemistry and Interpretation
    1. Hydrophobic and Aromatic Patches
      • Antibody sideshows a robust cluster of Phe33, Tyr106/109/178/218, Trp158, forming a dense aromatic patch.
      • Antigen side includes Leu52, Met50, and His49/66, suggesting potential π–π stacking, hydrophobic interactions, and cation–π pairing (e.g., His49↔Phe33/Tyr106 or Arg53↔Trp158).
      • This hydrophobic-aromatic interface is likely a major contributor to favorable docking energies.
    2. Electrostatic and Salt-Bridge Interactions
      • Ab residues Asp108, Glu-rich loop (if present), and Thr/Ser side chains may form salt bridges or ion-dipole contacts with Arg53, Lys160, Arg163 on IL-23.
      • Patch B with Arg163–Lys160 is highly cationic, complementing acidic or polar patches on the antibody surface.
    3. Hydrogen Bond Networks
      • Ser104/105/156, Thr32/34/177, Asn62 (Ag), Thr61, Gln68, Glu54–59 form a complex hydrogen-bonding web, stabilizing the interaction.
      • This polar network is interwoven with the aromatic and electrostatic interactions, enhancing overall binding affinity and specificity.
  5. Geometrical and Positional Features
    1. Two-Patch Binding
      • Patch A (Ag residues 49–55): Interacts with Ab residues 32–34, 103–109, mostly involving polar, aromatic, and acidic residues, forming a polar–aromatic clamp.
      • Patch B (Ag residues 160–163): Engages Ab residues 155–158, 177–178, featuring Trp158–Arg163 interaction and polar reinforcement.
    2. CDR Involvement (Estimated)
      • Contacts span 32–34, 103–109, 155–158, 177–178, 218, suggesting engagement from LCDR1, HCDR3, and possibly LCDR3/HCDR1, indicating a balanced interface where both chains contribute meaningfully — especially through HCDR3’s aromatic-rich region.
  6. Conclusion

    The scFv engages IL-23 through a dual-region interface:

    • Patch A leverages aromatic–polar clamping of IL-23’s N-terminal loop (residues 49–55),
    • Patch B anchors on the positively charged Lys160/Arg163 patch, stabilized by Trp158/Tyr178-rich segments.

      These interactions involve a synergy of π-stacking, hydrogen bonds, and salt bridges, resulting in energetically favorable and geometrically coherent docking. The balanced involvement of multiple CDRs supports a stable and specific antigen recognition profile.

  1. ClusPro Docking Results Overview

    Note on Docking Parameters
    Due to technical issues encountered when attempting to use ClusPro's antibody mode, the docking analysis for scFv #3 was instead performed using the normal mode. While antibody mode is typically optimized for antibody-antigen interactions, the normal mode still provides meaningful structural insights. Therefore, the clustering and energy evaluations discussed in this report were all generated under the standard ClusPro docking protocol.

    A total of 30 clusters were generated from the ClusPro docking analysis. The top three clusters, based on their weighted scores for both center and lowest-energy poses, are summarized below:

    • Cluster 0:

      • Center pose: −951.9
      • Lowest energy pose: −1283.3
      • Number of members: 184
      • Interpretation:This cluster demonstrates the most favorable lowest-energy pose (−1283.3) among all clusters, suggesting an exceptionally stable and tight binding interaction. With 184 members, it is also the most populated cluster, strongly indicating a dominant and highly sampled binding mode with significant biological relevance.
    • Cluster 1:

      • Center pose:−1102.3
      • Lowest energy pose: −1146.7
      • Number of members:113
      • Interpretation: This cluster shows a very favorable energy landscape, with both center and lowest-energy poses exhibiting strong binding affinities. The high member count (113) reinforces the likelihood that this pose is robust and consistently sampled, reflecting a reliable docking outcome.
    • Cluster 3:

      • Center pose:−1017.5
      • Lowest energy pose: −1118.1
      • Number of members: 54
      • Interpretation: Though it has fewer members, Cluster 3 still presents a remarkably favorable energy profile. The tight energy distribution between center and best poses suggests structural coherence, making this a noteworthy candidate despite its lower sampling frequency.

    Overall Conclusion

    Cluster 0 clearly emerges as the most promising candidate due to its extremely favorable lowest-energy pose (−1283.3) and the largest number of members (184), suggesting a dominant and highly stable binding mode. Cluster 1 also demonstrates strong energetic favorability with consistent sampling. Cluster 3, while less populated, shows good energy convergence and remains a viable alternative for further structural or functional analysis.

  2. Contact Residues (Deduplicated)
    • Antibody side (Ab)

      TRP41, PRO45, LEU46, VAL47, GLY48, HIS49, MET50, ASP51, LEU52, ARG53, GLU54, GLU55, GLY56, ASP57, GLU58, THR60, THR61, VAL64, LEU99, TRP152, GLN153, LEU155, LEU156, ARG158, PHE159, LYS160, ILE161, ARG163, SER164, GLN166

    • Antigen side (Ag, IL-23)

      PHE31, PHE33, ASN57, TYR59, TYR61, ASP103, ARG105, TYR107, TRP108, TYR109, TYR110, GLY111, GLY112, GLY113, GLY114, SER115, GLY116, GLY117, GLY118, GLY119, SER120, GLY121, GLY122, GLY123, GLY124, GLN155, GLY156, THR157, TRP159, SER220, ASN221, TRP222

      Figure 20. anti-IL-23 scFv 3 Docking with IL23 p19 subunit amination.
  3. Epitope Distribution and Structural Characteristics

    The IL‑23 epitope presents a multi‑patch surface:

    • Patch A (aromatic N‑terminal loop): Phe31/Phe33 with nearby Asn57–Tyr61, forming an aromatic‑polar loop likely exposed at the domain surface.
    • Patch B (central aromatic–glycine belt): Asp103–Tyr110 followed by a Gly/Ser‑rich segment (111–124), consistent with a flexible loop/turn that can adapt to antibody packing.
    • Patch C (mid–C‑terminal ridge): Gln155–Trp159 plus a distal Ser220–Trp222 tip; the presence of Trp159/Trp222 suggests an additional aromatic hotspot.
  4. Interaction Chemistry and Interpretation
    1. Hydrophobic & Aromatic Packing
      • Antibody: dense aromatic/hydrophobic cluster (Trp41, Leu46, Val47, Met50, Leu52, Trp152, Phe159, Ile161) with auxiliary apolar residues (Leu99, Leu155/156, Val64).
      • Antigen: multiple aromatics (Phe31, Phe33, Tyr59, Tyr61, Tyr107, Tyr109, Tyr110, Trp108, Trp159, Trp222).
      • Likely π–π (e.g., Trp41/Trp152/Phe159 ↔ Tyr/Phe/Trp on IL‑23) and cation–π pairings (see below), forming a hydrophobic–aromatic core across Patches A–C that strongly stabilizes the interface.
    2. Electrostatics & Salt Bridges
      • Acidic Ab residues (Asp51, Asp57, Glu54/55/58) are well‑positioned to engage Arg105(Ag) and possibly modulate the local pKa of nearby tyrosines/tryptophans.
      • Basic Ab residues (Arg53, Arg158, Lys160, Arg163) can pair with Asp103(Ag) and acidic microenvironments in Patch B, creating salt bridges / ion–dipole clamps that anchor the central loop.
    3. Hydrogen Bond Networks
      • Polar Ab residues (Thr60/61, Gln153, Ser164, Gln166, His49, Glu54/55) align with Asn57, Thr157, Gln155, Ser115/120, backbone carbonyls in 111–124 on IL‑23 to build a multi‑point H‑bond web.
      • This network complements the aromatic core, locking the binding geometry while allowing minor loop adjustments in the Gly/Ser belt.
    4. Cation–π Highlights
      • Arg105 (Ag) near Trp152 / Phe159 / Trp41 (Ab) is a strong candidate for Arg–π stabilization.
      • Additional Arg163/Lys160 (Ab) may participate in long‑range electrostatic steering toward Patch B/C.
  5. Geometrical and Positional Features
    1. Multi‑patch engagement
      • Patch A (Ag 31–33, 57–61) engages Ab 41–61: an aromatic–polar clamp (Trp41 + acidic loop) around Tyr/Phe ring systems.
      • Patch B (Ag 103–110 + 111–124) docks against Ab 99–166: Trp152/Phe159 pack onto Tyr/Trp‑rich Ag residues while Gly/Ser belt forms adaptable H‑bonds with Thr/Gln/Ser on the Ab.
      • Patch C (Ag 155–159, 220–222) makes additional aromatic contacts (Ab Trp152/Phe159/Leu155–156) and polar tethers (Ab Gln153/Gln166).
    2. CDR Involvement (estimated)
      • Contacts spanning 41–61 suggest strong CDR1/CDR2‑like involvement on the heavy/light chain depending on numbering.
      • The 99–166 region (including Trp152, Gln153, Phe159, Lys160, Arg163, Gln166) indicates HCDR3‑proximal / loop‑loop participation, consistent with H3‑assisted packing against the central patch.
  6. Conclusion

    The scFv recognizes IL‑23 via a three‑patch interface:

    1. an aromatic N‑terminal loop (Phe31/Phe33, Tyr59/61),
    2. a central aromatic–glycine belt (Asp103–Tyr110 + Gly/Ser 111–124), and
    3. a mid–C‑terminal ridge (Gln155–Trp159 and Ser220–Trp222).
    4. The binding is driven by π–π/cation–π stacking (notably around Arg105 and multiple Tyr/Trp residues), reinforced by salt bridges (e.g., Ab Asp/Glu ↔ Ag Arg105 / Ab Arg ↔ Ag Asp103) and a distributed H‑bond network across flexible loops. Together, these features rationalize the favorable docking energies and support a geometrically coherent, chemically complementary interface suitable for stable neutralization of IL‑23.

  1. ClusPro Docking Results Overview

    A total of 30 clusters were generated from the ClusPro docking analysis. The top three clusters, based on their weighted scores for both center and lowest-energy poses, are summarized below:

    • Cluster 6:

      • Center pose: −258.4
      • Lowest energy pose: −355.1
      • Number of members: 41
      • Interpretation: This cluster provides the most favorable lowest-energy pose (−355.1) among all clusters in this run, indicating an energetically preferred binding conformation. The moderate member count suggests it is well sampled and structurally plausible.
    • Cluster 20:

      • Center pose:−252.3
      • Lowest energy pose: −353.7
      • Number of members:14
      • Interpretation: Despite a smaller population, the best pose energy is highly competitive (−353.7). The gap between center and best poses implies some conformational diversity, yet the deep minimum supports a potentially specific interaction.
    • Cluster 0:

      • Center pose: −259.3
      • Lowest energy pose: −351.4
      • Number of members: 138
      • Interpretation: This is the most populated cluster, indicating a frequently sampled and recurrent binding mode. While its lowest-energy score is slightly less favorable than Clusters 6 and 20, the large membership supports robust convergence and possible biological relevance.

    Overall Conclusion

    Taken together, Cluster 6 is the strongest candidate by energy minimum (−355.1), Cluster 20 shows a comparably deep minimum with fewer but focused samples, and Cluster 0 represents the dominant, highly recurrent pose family. Prioritizing experimental follow-up could start with Cluster 6 (energy lead), then validate Cluster 0 (sampling dominance), with Cluster 20 as a complementary conformation worth testing.

  2. Contact Residues (Deduplicated)
    • Antibody side (Ab)

      THR28, ASP31, TYR50, TYR52, ARG54, ASP55, SER57, LYS59, LYS65, ASP100, ARG101, SER102, TYR104, TRP106, TYR109, ILE268, TYR286, TRP287, THR290, SER329, SER330, TYR331

    • Antigen side (Ag, IL-23)

      TRP41, VAL47, GLY48, HIS49, ARG53, GLU55, GLY56, ASP57, GLU58, ASN62, GLN68, CYS69, GLY70, GLY72, ASP74, PRO75, GLN76, GLY77, ARG163, GLN166, ALA167, ARG174

      Figure 21. anti-IL-23 scFv p19 Docking with IL23 p19 subunit amination.
  3. Epitope Distribution and Structural Characteristics

    The IL-23 epitope adopts a multi-patch layout:

    • Patch A (N-terminal loop, 41–58): Aromatic/charged segment including Trp41, Arg53, Glu55, Asp57, Glu58, adjacent to Val47/His49—a classic solvent-exposed loop suitable for π interactions and electrostatics.
    • Patch B (flexible loop, 70–77): Gly/Ser/Pro/Gln-rich (Gly70, Gly72, Asp74, Pro75, Gln76, Gly77) consistent with a conformationally adaptable belt.
    • Patch C (cationic ridge, 163–174): Arg163/Gln166/Ala167/Arg174 forming a short basic ridge that often mediates long-range steering and local anchoring.
  4. Interaction Chemistry and Interpretation
    1. Hydrophobic & Aromatic Packing
      • Antibody: clustered aromatics/hydrophobes (Tyr50, Tyr52, Tyr104, Tyr109, Tyr286, Trp106, Trp287, Ile268) supported by Val/Leu neighbors.
      • Antigen: prominent aromatics Trp41 (Patch A) plus apolar Val47/Pro75.
      • Likely π–π / cation–π contacts (e.g., Ab Trp106/Tyr50/Tyr52 ↔ Ag Trp41; Ab aromatics ↔ Ag Arg53/Arg174), creating a hydrophobic–aromatic core across Patches A and C that lowers the docking energy.
    2. Electrostatics & Salt Bridges
      • Ab basic residues (Arg54, Lys59, Lys65, Arg101) can pair with Ag acidic residues (Glu55, Asp57, Glu58) in Patch A, forming salt bridges / ion–dipole clamps.
      • Ab acidic residues (Asp31, Asp55, Asp100) can complement Ag cations (Arg53, Arg174) at Patches A/C, providing bidirectional charge compensation that stabilizes the pose.
    3. Hydrogen Bond Networks
      • Ab polar donors/acceptors (Thr28, Ser57, Ser102, Tyr104, Tyr109, Thr290, Ser329/Ser330, Tyr331, Gln side chains if nearby) align with Ag Asn62, Gln68/Gln76/Gln166, Cys69 (thiol H-bonding context-dependent), and backbone carbonyls of Gly70–Gly77, forming a multi-point H-bond web.
      • The Gly/Pro/Ser belt (70–77) likely conforms to the antibody surface, adding adaptable polar contacts that reinforce the aromatic core.
  5. Geometrical and Positional Features
    1. Multi‑patch engagement
      • Patch A (Ag 41–58) ↔ Ab 28–65: Aromatic–electrostatic clamp—Ab Tyr50/Tyr52/Trp106 stack with Trp41, while Arg54/Lys59/Lys65 salt-bridge with Glu55/Asp57/Glu58.
      • Patch B (Ag 70–77) ↔ Ab 100–109: The flexible Gly/Pro/Gln loop adapts to Ab Tyr104/Trp106/Tyr109 with H-bonds + π proximity, smoothing the interface between Patches A and C.
      • Patch C (Ag 163–174) ↔ Ab 268–331: Arg163/Arg174 approach a distal aromatic/polar platform (Ile268, Tyr286, Trp287, Thr290, Ser329/Ser330, Tyr331), enabling cation–π and polar tethering that helps lock orientation.
    2. CDR Involvement (estimated)
      • Ab 28–65 suggests strong CDR1/CDR2-like participation (numbering-scheme dependent).
      • Ab 100–109 indicates HCDR3-proximal contributions.
      • Ab 268–331 likely reflects framework/extended loop contacts that broaden the footprint and stabilize Patch C engagement.
  6. Conclusion

    This scFv engages IL-23 via three cooperative patches:

    • Patch A (41–58): an aromatic/charged anchor centered on Trp41 / Arg53 / acidic triad, clamped by Ab aromatics + basic residues;
    • Patch B (70–77): a flexible Gly/Ser/Pro loop that adapts through H-bonding to Ab Tyr/Trp;
    • Patch C (163–174): a cationic ridge (Arg163/Arg174) stabilized by Ab distal aromatics/polar tethers.

  1. ClusPro Docking Results Overview

    A total of 30 clusters were generated from the ClusPro docking analysis. The top three clusters, based on their weighted scores for both center and lowest-energy poses, are summarized below:

    • Cluster 0:

      • Center pose: −333.6
      • Lowest energy pose: −333.6
      • Number of members: 82
      • Interpretation: Center and best poses are identical, indicating excellent convergence. Combined with the largest member count (82), this suggests a robust, frequently sampled binding mode with consistent energetics.
    • Cluster 2:

      • Center pose:−327.0
      • Lowest energy pose: −327.0
      • Number of members:44
      • Interpretation: Also perfectly converged (center = best). The solid member count supports a reliable and reproducible binding conformation, albeit with slightly less favorable energy than Cluster 0.
    • Cluster 10:

      • Center pose: −326.3
      • Lowest energy pose: −326.3
      • Number of members: 19
      • Interpretation: Despite a smaller population, the identical center/best scores denote tight convergence and a clean energy landscape, making it a credible alternative pose family.

    Overall Conclusion

    All three clusters show perfect convergence (center equals lowest energy), which is uncommon and strengthens confidence in the docking outcomes. Cluster 0 is the lead candidate due to its most favorable score (−333.6) and highest membership. Clusters 2 and 10 provide converged, slightly higher-energy alternatives worth retaining for cross-validation or ensemble modeling.

  2. Contact Residues (deduplicated)
    • Antibody side (Ab)

      SER28, THR31, TRP33, SER52, VAL54, ASP55, ARG59, ARG99, ARG100, PRO101, GLY102, GLN103, SER271, TRP273, SER308, GLY309, THR310, ASP311, TYR332, ASN333, ILE334, TYR335, TYR337

    • Antigen side (Ag, IL-23)

      ILE1, TRP2, VAL11, GLU12, LEU13, ASP14, TYR16, PRO17, PRO20, GLY21, GLU22, MET23, GLN56, LYS85, ASP87, GLY88, TRP90, SER140, SER141, GLU187, MET189, GLN200

      Figure 22. anti-IL-23 scFv p40 Docking with IL23 p40 subunit amination.
  3. Epitope Distribution and Structural Characteristics

    The IL-23 epitope displays a multi-patch topology:

    • Patch A (N-terminal, 1–23): Aromatic/hydrophobic core (Trp2, Tyr16, Ile1, Val11, Leu13, Met23) interleaved with acidic residues (Glu12, Glu22, Asp14)—a solvent-exposed loop/ridge ideal for π interactions and electrostatics.
    • Patch B (middle loop, 56–90): Gln56–Lys85–Asp87–Gly88–Trp90—a flexible loop with a strong cationic (Lys85) and aromatic (Trp90) character.
    • Patch C (distal tip, 140–200): Ser140/141 and Glu187–Met189–Gln200, a polar/acidic segment that commonly forms stabilizing H-bonds and salt bridges with antibody framework or distal CDR regions.
  4. Interaction Chemistry and Interpretation
    1. Hydrophobic & Aromatic Packing
      • Antibody: multiple aromatics/hydrophobes (Trp33, Trp273, Tyr332, Tyr335, Tyr337, Ile334, Val54) cluster into a hydrophobic–aromatic patch.
      • Antigen: Trp2, Tyr16, Trp90 (plus Met23/Met189, aliphatic Ile/Val/Leu) provide rich partners for π–π stacking and van der Waals packing.
      • Expect Trp33/Tyr332/Tyr335Trp2/Tyr16/Trp90 as key π–π hotspots across Patches A–C.
    2. Electrostatics & Salt Bridges
      • Basic Ab residues (Arg59, Arg99, Arg100) complement acidic Ag residues (Glu12, Glu22, Asp14, Glu187), forming salt bridges / ion–dipole clamps that anchor Patch A and Patch C.
      • Acidic Ab residues (Asp55, Asp311) can pair with Lys85 (Ag) in Patch B, reinforcing the central interface and aiding pose orientation.
    3. Hydrogen-Bond Networks
      • Polar Ab residues (Ser28, Thr31, Ser52, Gln103, Ser271, Ser308, Thr310, Asn333) align with Gln56, Ser140/141, Gln200 and backbone atoms in Gly21/Gly88, forming a multi-point H-bond web.
      • This polar network interlocks the aromatic core while permitting modest loop adaptability—consistent with stable yet low-energy docking.
  5. Geometrical and Positional Features
    1. Multi‑patch engagement
      • Patch A (Ag 1–23)Ab 28–59: An aromatic–electrostatic clamp, where Trp33/Tyr stack on Trp2/Tyr16, and Arg59/Arg99/Arg100 salt-bridge with Glu12/Glu22/Asp14.
      • Patch B (Ag 56–90)Ab 99–106: Lys85/Trp90 engage Ab Arg99/Arg100 + Trp/ Tyr (e.g., Trp33/Trp273/Tyr104-region if proximal), combining cation–π with local H-bonds.
      • Patch C (Ag 140–200)Ab 271–337: Glu187–Met189–Gln200 contact a distal Ab platform (Trp273, Tyr332, Tyr335, Tyr337, Ser308/Thr310/Ser330-like) that provides π contacts + polar tethers to lock orientation.
    2. CDR Involvement (estimated)
      • Ab 28–33 suggests CDR-L1/H1 participation; 52–59 is consistent with CDR-L2/H2 support.
      • 99–103 (Arg99/Arg100/Gln103) implies H3-proximal contacts steering Patch B.
      • 271–337 indicates framework/extended loop assistance stabilizing Patch C.
  6. Conclusion

    The scFv recognizes IL-23 through three cooperating patches:

    • a N-terminal aromatic/acidic hub (1–23) clamped by Ab aromatics and arginines,
    • a central flexible loop (56–90) centered on Lys85/Trp90 with cation–π and H-bonding, and
    • a distal polar/acidic segment (140–200) anchored by Ab distal aromatics and polar tethers.