Rational Design and Characterization of a Mutated Nanobody for Specific Targeting of Heparan Sulfate

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors applied sophisticated in silico tools to significantly improve the affinity of an existing nanobody targeting heparan sulfate. HS binding is the initial entry step of several pathogenic viruses and therefore, the newly developed nanobody can serve as a valuable tool for studying virus entry and also functionally interfere with it. A triple mutant that was identified by docking studies and MD simulations indeed revealed strongly enhanced affinity with a Kd in the single digit nanomolar range. These are interesting methodologies and data and the paper in principle should be considered for publication. The following points should be addressed beforehand:

• In view of the MD simulation results, it would be interesting to experimentally determine the Kd of the solitary Phe47Arg variant (also in presence of higher salt concentrations).
• Cell binding studies should be performed to confirm on-cell binding of Mut-Nb1  nanobody with a heparan sulfate positive cell line and one that is heparan sulfate negative such as e.g. BAF3. As an alternative  to a HS negative cell line, binding to a heparan sulfate positive cell line could be considered in vast excess of HS to estimate non-specific cell binding.
• Li 266: Wording: “The purified Mut-Nb1 protein was specifically recognized by the anti-His-tag monoclonal antibody, confirming its good immunoreactivity. This experiments serves to confirm identity of the purified protein, a good immunoreactivity of a His-antibody to a His-Tag can be expected. The statement confirming…. Should be deleted.
• Li 279 Please also provide the Ec50 value in nM
• Table 4 typo: MutationEnerg y

Author Response

List of Corrections – antibodies-4272529 Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate

 

Dear Editor and Reviewer,

Thank you for your comments and advice! We have revised the manuscript “Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate” carefully. The list of correction is as follows. Thank you very much!

 

 

Sincerely,

Dr. Junfang Hao,

College of Smart Animal Husbandry, College of Biology and Food, Shangqiu Normal University, China.

 

 

Response to Reviewer 1

1. In view of the MD simulation results, it would be interesting to experimentally determine the Kd of the solitary Phe47Arg variant (also in presence of higher salt concentrations).

Reply (Lines 390–393):

We acknowledge that experimental measurement of the Kd of the isolated Phe47Arg single mutant, particularly under higher ionic strength conditions, would provide valuable experimental validation and further enrich the interpretation of our MD simulation results.

Given the scope of the present study, which focuses on the combined mutational effects and overall binding mechanism, the Kd determination of the individual Phe47Arg variant under high salt conditions has not been included in the current experimental design. Nevertheless, we have now mentioned this meaningful research perspective in the Discussion section, and we will pursue it as an important direction in our future work to further verify the computational findings. Please refer to the fourth paragraph of the Discussion section in the revised version (lines 390–393).

2. Cell binding studies should be performed to confirm on-cell binding of Mut-Nb1 nanobody with a heparan sulfate positive cell line and one that is heparan sulfate negative such as e.g. BAF3. As an alternative to a HS negative cell line, binding to a heparan sulfate positive cell line could be considered in vast excess of HS to estimate non-specific cell binding.

Reply (Lines 395–401):

We fully agree that cell binding assays using heparan sulfate (HS)-positive and HS-negative cell lines (e.g., BAF3 cells) are necessary to further verify the specific on-cell binding ability of Mut-Nb1. Meanwhile, the alternative strategy of adding a large excess of free HS to HS-positive cells is also an effective approach to evaluate non-specific cell binding.

Nevertheless, this study primarily focuses on the molecular-level binding mechanism and affinity characterization of Mut-Nb1. Given the original research scope and limited revision time, it is not feasible to complete the cell culture and cell binding verification assays in the short revision period.

We have supplemented this important research perspective in the revised Discussion section. In our follow-up work, we will strictly follow the reviewer’s advice: perform cell binding tests with HS-positive and HS-negative cell lines, and adopt the excess free HS competition strategy to systematically evaluate the specific binding and non-specific background binding of Mut-Nb1 at the cellular level. Please refer to the fourth paragraph of the Discussion section in the revised version (lines 395–401).

3. Li 266: Wording: “The purified Mut-Nb1 protein was specifically recognized by the anti-His-tag monoclonal antibody, confirming its good immunoreactivity. This experiment serves to confirm identity of the purified protein, a good immunoreactivity of a His-antibody to a His-Tag can be expected. The statement confirming…. Should be deleted.

Reply (Lines 294–297): 

We fully agree with your opinion that this experiment was mainly designed to verify the identity and successful expression of the purified Mut-Nb1 protein. Specific recognition by anti-His-tag monoclonal antibody is an inherent and expected interaction between the Mut-Nb1 and His-tag. Therefore, we have removed the inappropriate statement “confirmed that it possessed good immunoreactivity” in Section 3.5 of the revised manuscript, and rephrased the relevant content to objectively present the identification result of the purified protein. The specific modifications are as follows:

Western blotting was performed using an anti-His-tag monoclonal antibody as the primary antibody. The purified MutNb1 protein exhibited a specific band at approximately 15 kDa (Fig 4B), verifying that the Histagged recombinant MutNb1 protein was successfully expressed and purified.

4. Li 279 Please also provide the EC₅₀value in nM.

Reply (Table 5): 

It has been revised and updated.

5. Table 4 typo: MutationEnerg y

Reply (Table 4):

Table 4 typo has been revised.

Once again, we sincerely thank you for your professional review and valuable suggestions.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In this paper the authors try to re-design an existing Nanobody so as to recognise heparin sulphate as this antigen is involved in entry of viruses in cells, and no good monoclonal antibodies are available. 

They started from Nanobody against chloramphenicol. (please mention this somewhere in the text, and why this Nanobody was chosen as a starting Nanobody. 

Then the chloramphenicol antigen was replaced in the complex by heparin sulphate by in silico calculations and a VIRTUAL Ala scanning was undertaken to improve the binding. Mutations at several spots were carried out (in silico) to search for higher affinity binding. 

Finally they selected the best possible mutant, synthesised and expressed the gene, purified the Nanobody and tested its binding to heparin sulphate in ELISA. 

Apart from the missing rational why they started with this Nanobody (pdb: 7TJC) (please make sure you also give the reference of this publication); there remain some flaws in the experimental data:

1. the redesigned Nanobody (with a His tag) was expressed and purified after lysis of the host bacteria by gel filtration. This is strain: Why not using IMAC and gel filtration. I never believe that gel filtration alone can purify the Nanobody to the extend she see on the Coomassie stained gel (and western blot). It is important that the material is highly purified to enable accurate affinity measurements.   

For the affinity measurement, the authors are referring to a paper of Beatty et al, (1987) with a protocol too measure affinity of BIVALENT monoclonal antibodies. Here we are talking about a MONOVALENT Nanobody. The Beatty formula for the bivalent IgG CANNOT be used for monovalent Nanobodies. Actually, the formula for monovalent affinity reagents becomes much easier. You have to use a fixed amount of antigen in consecutive wells; then add an increasing concentration of Nanobody (at around the same concentration as the concentration of the Ag used). Under the assumption that you reach a plateau signal (which is not the case here in Figure 4C) then the Nanobody concentration used to give 50% of plateau signal,  that corresponds roughly to the equilibrium dissociation constant, KD. 

Roughly as you have to assume that the binding equilibrium is not changed when binding to a, immobilised antigen, that you have on average 1 heparin sulphate happen on the carrier ovalbumin (to avoid avidity and rebinding effects), and that equilibrium will not change during washings. As it stands, the measured affinity (Ka= 5 x 108 M-1) corresponds to a KD of 2 nM, which is highly surprising. It is recommended to repeat the affinity measurement  by calorimetry (or SPR, although I doubt that SPR is sensitive enough to measure the affinity for a heparan sulphate hapten binder).  

Although the authors showed convincingly that their mutant Nanobody is now binding to heparan sulphate (and not to chondroitin sulphate or keratin sulphate). However, it is imperative to measure also the affinity of the original Nanobody (7TJC) on these antigens. This to avoid that the original 7TJC Nanobody binds chloramphenicol the way it is shown in the crystal structure of the complex (sandwiched between CDR3 and the FR2) and manages to bind to HS via its conventional paratope formed on top of the 3 CDRs. 

 

Minor comments:

Line 89: Ovalbumin instead of ovalbumen

Line 137: how much ovalbumin was used and how many HS haptens were conjugated per OVA molecule?

Line 199 and 200: I do understand the singel site saturation mutagenesis where 6 x 20 mutants are designed. But, what is C6E2 x 20 x 20 for the double site mutagenesis ? (what is C6E2?)

Line 206 : For triple mutants 1x7x3= 21 mutants were envisaged but in table 4 we have only 3 triple mutants

Line 297: what is "entry diverse pathogens'?  are those different pathogens making use of different types of cellular entry mechanisms? 

Author Response

List of Corrections – antibodies-4272529 Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate

 

Dear Editor and Reviewer,

Thank you for your comments and advice! We have revised the manuscript “Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate” carefully. The list of correction is as follows. Thank you very much!

 

 

Sincerely,

Dr. Junfang Hao,

College of Smart Animal Husbandry, College of Biology and Food, Shangqiu Normal University, China.

 

 

Response to Reviewer 2

1. They started from Nanobody against chloramphenicol. (please mention this somewhere in the text, and why this Nanobody was chosen as a starting Nanobody.

Apart from the missing rational why they started with this Nanobody (pdb: 7TJC) (please make sure you also give the reference of this publication).

Reply (Lines 328–331):

We fully agree that a clear rationale for selecting the initial nanobody scaffold is essential for evaluating the reliability of our findings. The original reference for this structure has been added, and the citation details can be found in reference [33] of the revised manuscript.

Rationale for selecting the initial nanobody scaffold. First, the crystal structure of wild-type Nb has been resolved (PDB: 7TJC), providing a highprecision structural model for subsequent molecular docking and sitedirected mutagenesis. Second, molecular docking showed that the wildtype Nb yielded a docking score of –7.30  kcal/mol with the target HS. According to empirical criteria, a docking score below –6.0  kcal/mol generally indicates good binding potential, suggesting that the wildtype Nb possesses a favorable recognition capability toward HS and is suitable as a starting scaffold for affinity maturation. The relevant content has been incorporated into the Discussion section of the revised manuscript.

In addition, preliminary ELISA experiments qualitatively verified that the wild-type Nb possesses a detectable binding affinity for HS (data not shown in the manuscript; see Figure 1 below), which provided a solid foundation for the subsequent rational engineering. In summary, this scaffold combines structural clarity, predicted binding activity, and experimental validation, thereby reducing uncertainty in the optimization process.

2. The redesigned Nanobody (with a His tag) was expressed and purified after lysis of the host bacteria by gel filt This is strain: Why not using IMAC and gel filtration. I never believe that gel filtration alone can purify the Nanobody to the extend she see on the Coomassie stained gel (and western blot). It is important that the materialis highly purified to enable accurate affinity measurements.

Reply (Lines 144–149):

The recombinant Mut-Nb1 protein in this study carried a His-tag. In fact, a two-step purification strategy combining Ni-NTA affinity chromatography and subsequent gel filtration chromatography (GFC) was adopted throughout the purification process.

In the original manuscript, Section 2.5 only described the final GFC purification step and omitted the essential Ni-NTA affinity enrichment procedure, resulting in an incomplete description of the experimental methods. Please accept my apologies for this oversight. Please see Section 2.5 of the revised document, where the newly added content has been marked in red font. The specific modifications are as follows:

Finally, the recombinant Mut-Nb1 was first enriched and purified via Ni-NTA affinity chromatography to remove most impurities. After concentration of the eluted fractions, further purification was performed by gel filtration chromatography (GFC). Additionally, the purified Mut-Nb1 protein was assessed by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis.

3. For the affinity measurement, the authors are referring to a paper of Beatty et al, (1987) with a protocol too measure affinity of BIVALENT monoclonal antibodies. Here we are talking about a MONOVALENT Nanobody. The Beatty formula for the bivalent IgG CANNOT be used for monovalent Nanobodies. Actually, the formula for monovalent affinity reagents becomes much easier. You have to use a fixed amount of antigen in consecutive wells; then add an increasing concentration of Nanobody (at around the same concentration as the concentration of the Ag used). Under the assumption that you reach a plateau signal (which is not the case here in Figure 4C) then the Nanobody concentration used to give 50% of plateau signal, that corresponds roughly to the equilibrium dissociation constant, KD.

Reply (Lines 164–171):

We sincerely thank the reviewer for the professional and insightful guidance. We have fully understood the comments and have revised our methodology as suggested.

You pointed out that “the calculation formula for monovalent affinity reagents is considerably simpler”. This perspective is correct. We have come to recognize that the method proposed by Beatty et al. (1987) was originally developed for bivalent IgG antibodies. As such, this complex formula is not applicable to the monovalent nanobody used in our study.

In our experimental setup, we have strictly followed the approach you recommended: antigen was coated at a fixed concentration of 2.0 μg/mL, and serially diluted nanobody concentrations ranging from 10 to 0.039 μg/mL were applied for the binding assay. Regarding the inappropriate use of the Beatty et al. (1987) complex formula in our original manuscript, we have made thorough corrections. Specifically, we have removed the citation of that reference and the associated bivalent antibody calculation formula. Instead, we re-analyzed the experimental data using a simple one-site specific binding model via nonlinear regression fitting, yielding the equilibrium dissociation constant (K_D) values that more reflect the binding characteristics of the monovalent nanobody. Correspondingly, Figure 4C has been updated, as presented in the revised manuscript.

As pointed out by the reviewer, K_D is routinely determined as the antibody concentration corresponding to 50% of the plateau signal in a saturation binding assay. Although the binding curve did not reach an obvious plateau at the highest tested concentration in Figure 4C, the binding signal still showed a significant concentration-dependent increasing trend.

In this study, the K_D was determined to be 65.87 nM by nonlinear regression fitting using the one-site specific binding model. This K_D corresponds to the half-maximal effective concentration (EC₅₀), representing the nanobody concentration required to achieve half of the theoretical maximum binding capacity (Bmax). Given the monovalent binding characteristic of nanobodies and the adoption of the 1:1 one-site fitting model in this work, this K_D can accurately reflect the intrinsic binding affinity of Mut-Nb1.

We have supplemented the revised manuscript with the above rationale in the Methods section (2.6).

4. Roughly as you have to assume that the binding equilibrium is not changed when binding to a, immobilised antigen, that you have on average 1 heparin sulphate happen on the carrier ovalbumin (to avoid avidity and rebinding effects), and that equilibrium will not change during washings. As it stands, the measured affinity (Ka= 5 x 108 M-1) corresponds to a K_Dof 2 nM, which is highly surprising. It is recommended to repeat the affinity measurement by calorimetry (or SPR, although I doubt that SPR is sensitive enough to measure the affinity for a heparan sulphate hapten binder).

Reply:

This is an excellent suggestion. The affinity constant (K_a) calculation method proposed by Beatty et al. (1987) was originally designed for bivalent IgG antibodies. IgG can form multivalent binding to immobilized antigens through bivalent cross-linking. Such multivalent avidity effect can substantially elevate the apparent affinity of antibodies, thereby leading to an underestimation of the K_D value calculated by this formula. We have now adopted the approach suggested by you: the experimental data were refitted by nonlinear regression using the simple one-site specific binding model. The recalculated K_D value, which is more consistent with the intrinsic binding characteristics of monovalent nanobodies, was determined to be 65.87 nM.

We fully agree with the potential avidity effect and equilibrium interference inherent to solid-phase assays. However, we sincerely regret that affinity re-measurement using sophisticated instruments such as ITC and SPR is currently not achievable in our laboratory due to the lack of relevant instrumental facilities. Nevertheless, we have strictly optimized the ELISA protocol and repeated the assay in multiple independent batches. All affinity data are presented as the statistical mean of three biological replicates combined with technical replicates, showing stable values and good reproducibility. Furthermore, we have tried our best to avoid non-specific binding and rebinding effects in the experimental design. A low-coupling ratio of HS-OVA conjugate system was applied with strictly controlled coupling conditions, ensuring that only a small amount of HS antigen was conjugated to each OVA carrier as far as possible. This strategy effectively reduced the interference caused by multivalent cross-linking and secondary rebinding.

In addition, the high-affinity characteristic of the nanobody is supported by structural evidence. Molecular dynamics simulations and residue energy decomposition confirmed that the triple-mutated nanobody obtained in this study forms extensive stable hydrogen bonds, electrostatic interactions, and hydrophobic stacking with HS. The overall binding free energy is lower, and the key core residues make prominent contributions, which structurally explains the ultrahigh affinity of the nanobody toward HS at the molecular level.

In our future work, we will further optimize the experimental system and adopt cross-validation with multiple affinity detection methods when experimental conditions permit.

5. Although the authors showed convincingly that their mutant Nanobody is now binding to heparan sulphate (and not to chondroitin sulphate or keratin sulphate). However, it is imperative to measure also the affinity of the original Nanobody (7TJC) on these antigens. This to avoid that the original 7TJC Nanobody binds chloramphenicol the way it is shown in the crystal structure of the complex (sandwiched between CDR3 and the FR2) and manages to bind to HS via itsconventional paratope formed on top of the 3 CDRs.

Reply:

Following your suggestion, we have provided the affinity measurement experiments of the original nanobody against the above three antigens using exactly the same ELISA protocol as that for the mutant nanobody. The ELISA results (Figure 1, data not shown in the manuscript) revealed that, compared with the PBS negative control group, the original nanobody (7TJC) exhibited differential binding affinities to HS, KS, and CS. Among them, it displayed a higher affinity toward HS, with an OD₄₅₀ value of 0.582 ± 0.062. This finding is consistent with our selection of the original nanobody as the framework antibody.

 

 Figure 1 Binding affinities of HS, KS, and CS to the wild-type Nb measured by ELISA.

6. Line 89: Ovalbumin instead of ovalbumen

Reply (Line 91): It has been revised

7. Line 137: how much ovalbumin was used and how many HS haptens were conjugated per OVA molecule?

Reply:

Amount of OVA used: In the conjugation reaction system of HS hapten and OVA in this experiment, the amount of OVA used was 2.0 mg, which was dissolved in 1 mL of 0.01 mol/L MES buffer (pH 5.2) to a final concentration of 2.0 mg/mL. This dosage is an optimized conventional dosage for the conjugation of small-molecule haptens with OVA, which can ensure the sufficiency of the conjugation reaction.

Conjugation ratio (number of HS haptens conjugated per OVA molecule): A single OVA molecule contains approximately 21 free amino groups available for conjugation. The HS-OVA conjugate prepared in this study was mainly applied to ELISA assays rather than animal immunization. Therefore, a low-conjugation strategy was adopted with a hapten-to-OVA molar feeding ratio of 20:1, and the reaction conditions were carefully optimized to avoid steric hindrance caused by excessive hapten modification. This design ensures the binding affinity and specificity between HS antigen and nanobody. According to the results calculated by ultraviolet spectrophotometry, an average of 1 to 2 HS hapten molecules were conjugated to each OVA molecule. Such a low conjugation ratio effectively reduces intermolecular steric hindrance, maximally preserves the native conformation of HS, and guarantees the specificity and result reliability of subsequent ELISA binding assays.

8. Line 199 and 200: I do understand the singel site saturation mutagenesis where 6 x 20 mutants are designed. But, what is C6E2 x 20 x 20 for the double site mutagenesis ? (what isC6E2?).

Reply (Line 224):

It has been revised as C₆²×20×20=6000, where C₆² represents the number of combinations for selecting 2 sites out of 6, i.e., C₆² = 15. Thus, for doublesite saturation mutagenesis, the total number of mutants is 15×20×20=6000.

9. Line 206 : For triple mutants 1x7x3= 21 mutants were envisaged but in table 4 we have only 3 triple mutants

Reply (Line 243):

Based on the results of single-site saturation mutagenesis, triple-site combinatorial mutagenesis was performed, yielding a total of 1 × 7 × 3 = 21 mutant variants (Table 4). Among them, only the top 3 triple mutants with the lowest mutation energy were selected and presented in Table 4.

We have now clarified this in the revised manuscript by adding: “The top three combinatorial mutants are listed in ascending order of mutation energy”. Additionally, we have revised and improved the annotations for Table 1, Table 2, and Table 3. 

10. Line 297: what is "entry diverse pathogens'? are those different pathogens making use of different types of cellular entry mechanisms?

Reply (Line 323):

It has been revised as “entry of diverse pathogens”, meaning that HS is involved in both the attachment and the entry processes of various pathogens.

Once again, we sincerely thank you for your professional review and valuable suggestions.

 

References

• Swofford, C.A.;Nordeen, S.A.; Chen, L.; Desai, M.M.; Chen, J.; Springs, S.L.; Schwartz, T.U.; Sinskey, A.J. Structure and specificity of an anti-chloramphenicol single domain antibody for detection of amphenicol residues. Protein Sci. 2022, 31, e4457. DOI:1002/pro.4457.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript titled "Rational Design and Characterization of a Mutated Nanobody for Specific Targeting of Heparan Sulfate " has provided a pipeline workflow for the generation of nanobodies against small molecules.

The study is particularly interesting, as most existing approaches rely on protein or peptide ligands, whereas the authors explore a non-proteinaceous ligand. Overall, the work provides useful insights and could be valuable to the scientific community.

Comments:

1. The authors should include a dedicated section discussing existing insilico tools and methodologies for nanobody generation, along with their limitations, to better position the novelty of the current work.
2. In Section 2.5, the authors state that size exclusion chromatography (SEC) was used after lysate clarification to purify the nanobody. Given that the construct contains a His-tag, it is important to clarify whether Ni-NTA affinity purification was performed prior to SEC.
3. The rationale for selecting the initial nanobody scaffold is not clearly explained. The authors should justify why this particular starting nanobody was chosen and discuss whether the workflow is generalizable to other nanobody scaffolds.
4. In Section 3.4, molecular dynamics (MD) simulations were performed on the nanobody–heparan sulfate complex to assess structural fluctuations. The authors should include a comparison with simulations of the nanobody alone to better evaluate the stabilizing effect of ligand binding.
5. The manuscript reports binding free energy calculations and identifies key contributing residues. The authors should clarify the methodology used to calculate residue-wise free energy contributions, including whether these were derived from single-point mutations or decomposition analysis. Additionally, a comparison between binding energies obtained from MD simulations and docking studies would strengthen the analysis.
6. The method used to determine EC50 values is not described and should be clearly outlined.
7. The authors should provide a detailed explanation of how binding specificity was assessed.

Author Response

List of Corrections – antibodies-4272529 Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate

 

Dear Editor and Reviewer,

Thank you for your comments and advice! We have revised the manuscript “Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate” carefully. The list of correction is as follows. Thank you very much!

 

 

Sincerely,

Dr. Junfang Hao,

College of Smart Animal Husbandry, College of Biology and Food, Shangqiu Normal University, China.

 

 

Response to Reviewer 3

1. The authors should include a dedicated section discussing existing insilico tools and methodologies for nanobody generation, along with their limitations, to better position the novelty of the current work.

Reply (Lines 331–349):

We have added a new paragraph in the Discussion section (second paragraph) of the manuscript, which systematically discusses the current computational tools and methods for nanobody generation, along with their major limitations. The specific supplementary content is as follows:

Currently, computeraided techniques are widely applied in the rational design and molecular engineering of Nbs. Common computational tools include homology modeling, molecular docking, MD simulation, virtual mutation, and binding energy analysis, represented by classic software such as SWISSMODEL, AutoDock, and GROMACS [34,35]. Moreover, machine learning has increasingly been applied to Nb sequence optimization [35]. Despite the capability of current computational methods to improve engineering efficiency and lower experimental costs, they present inherent limitations. First, these approaches rely heavily on highresolution protein structures and reliable homologous templates, resulting in considerably lowered prediction accuracy for targets with complex modifications and high conformational flexibility. Second, conventional docking is typically static and fails to fully reflect the impacts of solvent environment and dynamic conformational transitions on polysaccharide–Nb recognition. Third, traditional virtual mutation strategies mainly concentrate on individual residue functions, neglecting inter-residue synergistic effects and underlying mechanisms of interface remodeling [36]. In contrast, the computational strategy proposed herein integrates molecular docking, MD simulation, binding energy decomposition, and multisite cooperative mutation screening. This strategy enables systematic characterization of the binding interface between Mut-Nb1 and HS, allowing the identification of key residues and further optimization of intermolecular interactions. Notably, this approach is independent of any specific Nb framework. Therefore, the established workflow exhibits broad generalizability and can be readily extended to the rational molecular engineering of other target-specific Nbs.

2. In Section 2.5, the authors state that size exclusion chromatography (SEC) was used after lysate clarification to purify the nanobody. Given that the construct contains a His-tag, it is important to clarify whether Ni-NTA affinity purification was performed prior to SEC.

Reply ((Lines 144–149):

The recombinant Mut-Nb1 protein used in this study carried a His-tag, and a two-step purification strategy comprising Ni-NTA affinity chromatography followed by gel filtration chromatography (GFC) was employed.

In the original manuscript, Section 2.5 only described the final GFC purification step and omitted the essential Ni-NTA affinity enrichment procedure, resulting in an incomplete description of the experimental methods. Please accept my apologies for this oversight. Please see Section 2.5 of the revised document. The specific modifications are as follows:

Finally, the recombinant Mut-Nb1 was first enriched and purified via Ni-NTA affinity chromatography to remove most impurities. After concentration of the eluted fractions, further purification was performed by gel filtration chromatography (GFC). Additionally, the purified Mut-Nb1 protein was assessed by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis.

3. The rationale for selecting the initial nanobody scaffold is not clearly explained. The authors should justify why this particular starting nanobody was chosen and discuss whether the workflow is generalizable to other nanobody scaffolds.

Reply (Lines 327–330, 349–351): 

We fully agree that a clear rationale for selecting the initial nanobody scaffold is essential for evaluating the reliability of our findings. In addition, we also explore whether this workflow can be generalized to other nanobody scaffolds.

• Rationale for selecting the initial nanobody scaffold

First, the crystal structure of this Nb has been resolved (PDB: 7TJC), providing a highprecision structural model for subsequent molecular docking and sitedirected mutagenesis. Second, molecular docking showed that the wildtype Nb achieved a docking score of –7.30  kcal/mol with the target HS. According to empirical criteria, a docking score below –6.0  kcal/mol generally indicates good binding potential, suggesting that the wildtype Nb possesses a certain recognition capability toward HS and is suitable as a starting scaffold for affinity maturation. The relevant content has been incorporated into the Discussion section of the revised manuscript (lines 327–330).

In addition, preliminary ELISA experiments qualitatively verified that the wild-type Nb possesses detectable binding affinity to HS (data not shown in the manuscript; see Figure 1 below), which provided a solid foundation for the subsequent rational engineering.

 

Figure 1 Binding affinities of HS, KS, and CS to the wild-type Nb measured by ELISA.

The ELISA results (Figure 1) showed that, compared with the PBS negative control group, the wild-type Nb exhibited differential binding affinities to HS, KS, and CS. Among them, it displayed a higher affinity toward HS, with an OD₄₅₀ value of 0.582 ± 0.062. This finding is consistent with our selection of the starting nanobody as the framework antibody. In summary, this scaffold combines structural clarity, predicted binding activity, and experimental validation, thereby reducing uncertainty in the optimization process. 

• Generalizability of the workflow

The computational strategy employed in this study focuses on characterizing the binding interface between the nanobody and HS by integrating molecular docking, MD simulation, binding energy decomposition, and multi-site cooperative mutation screening, in order to identify key residues and optimize interactions. This approach does not depend on a specific nanobody framework. For other nanobody scaffolds, the workflow can be transferred by adjusting the simulation parameters (e.g., definition of the docking pocket and force field parameters for MD simulations) according to the threedimensional structure of the target nanobody. Thus, the proposed workflow is broadly generalizable and can be extended to the rational engineering of other nanobodies against different targets. The relevant content has been incorporated into the Discussion section of the revised manuscript (lines 349–351).

4. In Section 3.4, molecular dynamics (MD) simulations were performed on the nanobody–heparan sulfate complex to assess structural fluctuations. The authors should include a comparison with simulations of the nanobody alone to better evaluate the stabilizing effect of ligand binding.

Reply (Lines 250–254, 260–263):

We fully agree that comparing the molecular dynamics (MD) simulations of the HS-Mut-Nb1 complex with those of the Nb alone (free mutant nanobody, Free-Mut-Nb1) provides a more robust evaluation of the ligand-induced stabilizing effect.

In response to this suggestion, we have performed additional MD simulations of Free-Mut-Nb1 under identical conditions (simulation length, force field, solvent model, temperature, and pressure coupling parameters) as used for the complex system. Comparative analysis between the two simulation sets revealed the following key findings:

(1) Root Mean Square Deviation (RMSD) analysis

Free Mut-Nb1 (blue curve): It maintained an overall higher RMSD value, stabilizing at 0.16–0.18 nm, and exhibited markedly larger fluctuations than the RMSD of the HS-Mut-Nb1 complex. HS-Mut-Nb1 complex (black and red curves, representing two independent replicates): The RMSD converged rapidly within the first 10 ns and then remained stable at approximately 0.11 nm (black curve) and 0.15 nm (red curve), respectively, both with minor fluctuations. Upon HS binding, the HS fills the cavity of the antibody binding pocket and increases the interaction sites of the system. Such intermolecular interactions restrict the free movement of the Mut-Nb1 backbone, thereby reducing the overall RMSD value.

(2) Root Mean Square Fluctuation (RMSF) analysis

Free Mut-Nb1 (blue curve): The overall RMSF baseline was relatively low, suggesting a relatively compact conformation in the unbound state; nevertheless, a remarkably flexible region was observed around residue 74. In contrast, the RMSF of this region in the complex system decreased significantly, demonstrating that HS binding effectively stabilized the intrinsic flexible region of the Mut-Nb1, reflecting the structural stabilization effect induced by ligand (HS) binding. The RMSF curves (black and red curves) of the two replicated complex simulations overlapped well, indicating that the dynamic changes induced by small molecule binding exhibited good reproducibility.

(3) Radius of Gyration (Rg) analysis

Free Mut-Nb1 (blue curve): The values were relatively lower, ranging from 1.39 to 1.40 nm. HS-Mut-Nb1 complex (black/red curves): The Rg values concentrated around 1.41 to 1.43 nm. Free Mut-Nb1 adopted a flexible conformation in solution. After binding with the HS, the ligand acts as a molecular scaffold, opening the Mut-Nb1 binding pocket via induced fit and enlarging the local structure. Meanwhile, the Mut-Nb1 did not fully encapsulate the bound HS, and the incorporation of the ligand increases the overall size of the complex, leading to a slight Rg increase of approximately 1 Å.

(4) Solvent Accessible Surface Area (SASA) analysis

Free Mut-Nb1 (blue curve): The values were lower, fluctuating mainly within the range of 66–68 nm². HS-Mut-Nb1 (black/red curves): The values increased significantly and stabilized at 70–72 nm². The larger SASA of the complex is attributed to the partial exposure of the small molecule to the aqueous environment, and the exposed surface area of the ligand was incorporated into the total SASA of the system. The stable high plateau of the black and red curves indicates the formation of a stable interface between the small molecule and the antibody. The high stability of the complex SASA further verifies that no continuous collapse or dissociation occurs at the binding interface, confirming a stable and tight binding mode.

MD results of Free Mut-Nb1 were added to section 3.4 of the Results, and updated Figures 2A–D. 

 

Figure 2. MD simulation results of the HS-Mut-Nb1 complex. (A) Root-mean-square deviation (RMSD). (B) Root-mean-square fluctuation (RMSF). (C) Radius of gyration (Rg). (D) Solvent-accessible surface area (SASA). (E) Number of hydrogen bonds. (F–G) Hydrogen bond existence maps (F: 1st simulation; G: 2nd simulation). Simulations were performed using GROMACS 2022.4 with the CHARMM36 force field.

5. The manuscript reports binding free energy calculations and identifies key contributing residues. The authors should clarify the methodology used to calculate residue-wise free energy contributions, including whether these were derived from single-point mutations or decomposition analysis. Additionally, a comparison between binding energies obtained from MD simulations and docking studies would strengthen the analysis.

Reply (Lines 127–131, 188–189, and 371–390):

We have clarified our computational workflow and provided additional comparative analyses between docking and MD-based binding energy results, as detailed below.

First, we explicitly clarify that the per-residue binding energy contributions were quantitatively calculated via MMPBSA free energy decomposition analysis based on equilibrated molecular dynamics trajectories, rather than single-point mutation energy calculations. The detailed methodological description has been supplemented in Section 2.4 of the revised manuscript (lines 127–131).

Second, we have systematically compared the binding characteristics and energy parameters obtained from static molecular docking and dynamic MD simulation, and the relevant results were added to Section 3.1 (lines 188–189) and Discussion Section lines 371–390). The specific modifications for the discussion section are as follows:

In this study, static molecular docking analysis was first performed to characterize the initial recognition pattern between HS and Mut-Nb1. The docking score of −8.10 kcal/mol indicated the spontaneous binding propensity of the two molecules at the static structural level. The binding interface was collectively stabilized by polar interactions (hydrogen bonds and electrostatic interactions) mediated by Arg47 and Ser106, as well as a nonpolar π–sulfur interaction contributed by Phe37. Notably, distinct from the static docking scoring, subsequent MD simulations and MMPBSA decomposition calculations yielded a reliable overall binding free energy of −83.26 ± 3.06 kcal/mol. The numerical difference between the docking score and MMPBSA binding energy originates from their inherent calculation principles: docking relies on a single static conformation with simplified solvent conditions for rapid scoring, whereas MMPBSA averages thermodynamic parameters over dynamic trajectory frames under explicit aqueous environments, thus better reflects the real binding state in solution. This MD-derived result verifies the favorable thermodynamic stability of the HS-Mut-Nb1 complex in aqueous solution. Furthermore, per-residue energy decomposition analysis confirmed that Arg47 serves as the dominant residue contributing to binding energy, while Ser106 and Phe37 play auxiliary roles in consolidating the binding interface. Collectively, these findings demonstrate that Arg47-mediated polar interactions act as the core driving force for complex stability throughout the entire binding process, spanning from the initial static docking pose to the final dynamic equilibrium state.

6. The method used to determine EC₅₀values is not described and should be clearly outlined.

Reply (Lines 150–182):

In this study, indirect ELISA was performed to evaluate the binding activity of Mut-Nb1 toward HS. HS was coated at a concentration of 2.0 μg/mL. Mut-Nb1 was prepared by 2-fold serial dilution over a concentration range from 10 to 0.039 μg/mL. Subsequently, binding incubation, washing, color development and absorbance measurement were carried out under the same experimental conditions described in Section 2.6.

Using the logarithmic value of Mut-Nb1 concentration as the abscissa and the OD₄₅₀ value as the ordinate, nonlinear regression fitting was performed in GraphPad Prism using the one-site specific binding model to calculate the EC₅₀. Considering the monovalent property of nanobodies and the 1:1 binding model adopted in this study, the obtained EC₅₀ value can approximately reflect the intrinsic binding affinity of Mut-Nb1 for HS, and was therefore regarded as the equilibrium dissociation constant (K_D). The specific modifications are as follows:

2.6. Affinity and Specificity Assays of Mut-Nb1

The binding affinity of Mut-Nb1 to HS was determined by ELISA. Briefly, the pre-conjugated HS-OVA was diluted with carbonate buffer solution (CBS, pH 9.6) to 2.0 μg/mL, coated onto ELISA plates at 100 μL/well (in triplicate), and incubated overnight at 4 °C. Phosphate buffered saline (PBS) served as a negative control. After coating, the plates were blocked with 1% (w/v) BSA at 37 °C for 1 h. Following this, the plates were incubated with purified Mut-Nb1 (10–0.039 μg/mL) at 37 °C for 30 min. Subsequently, an anti-His-tag monoclonal antibody (diluted 1:1000) was added and the plates were incubated at 37 °C for 30 min. Then, goat anti-mouse IgG-HRP (diluted 1:1000) was added at 100 μL/well, and the plates were incubated at 37 °C for 30 min. Between each incubation step, the plates were washed five times with phosphate buffered saline with tween-20 (PBST) and blotted dry. For detection, 100 μL/well of 3,3',5,5'-tetramethylbenzidine (TMB) substrate solution was added for color development. The reaction was stopped by the addition of 50 μL/well of 2 M H₂SO₄. The OD₄₅₀ value of each well was immediately measured using a microplate reader. The equilibrium dissociation constant (K_D) was determined by fitting the saturation binding curve with the one-site specific binding model in GraphPad Prism, using the nanobody concentration producing 50% of the plateau absorbance as the K_D value. Under this model, based on the monovalent nature of the nanobody-antigen interaction, the fitted halfmaximal effective concentration (EC₅₀) corresponds to the antibody concentration required to achieve halfmaximal specific binding (Bmax/2) and is therefore reported as the intrinsic K_D.

The specificity of Mut-Nb1 was evaluated by measuring its binding to structural analogs of HS. All analogs were coated at a fixed concentration of 2.0 μg/mL, and Mut-Nb1 was tested at a concentration of 2.5 μg/mL using the same ELISA method described above. Specificity was assessed using the following two parameters:
(1) Positive/negative (P/N) ratio: P/N = OD₄₅₀ (positive control) / OD₄₅₀ (negative control).

A P/N ratio ≥ 2.1 was considered specific binding; P/N < 2.1 was regarded as nonspecific or negative.
(2) Crossreactivity (CR%): CR% = [EC₅₀ of HS] / [EC₅₀ of HS structural analogs] × 100%.

The EC₅₀ values of HS analogs were determined using the method described above. A lower CR% indicates weaker crossrecognition and thus higher specificity.

7. The authors should provide a detailed explanation of how binding specificity was assessed.

Reply (Lines172–182):

We have expanded Section 2.6 in the revised manuscript, as explained in detail below.

(1) Overall strategy
Specificity was evaluated by ELISA, measuring the binding of the Mut-Nb1 to HS and to structurally related analogs. Two metrics were used:

Positive/negative (P/N) ratio: to determine whether binding is statistically significant.

Crossreactivity (CR%): to quantitatively assess the relative recognition of HS analogs.

(2) Experimental design

Positive control: HSOVA (2.0 μg/mL in CBS buffer, pH 9.6) coated on ELISA plates.

Negative control: PBS (no antigen) coated under the same conditions.

Test groups: HS structural analogs, including CS and KS (each conjugated to OVA and coated at 2.0 μg/mL).

The MutNb1 was added at a fixed concentration (2.5 μg/mL) or as serial dilutions (10–0.039 μg/mL), followed by the standard ELISA procedure described in Section 2.6. OD₄₅₀ values were recorded.

(3) Specificity criteria

P/N ratio was calculated as OD₄₅₀ (positive control) / OD₄₅₀ (negative control). A P/N ratio ≥ 2.1 was considered specific binding; P/N < 2.1 was regarded as nonspecific or negative.

Crossreactivity (CR%) was calculated as: CR% = [EC₅₀ of HS)] / [EC₅₀ of HS structural analogs] × 100%. The EC₅₀ values of HS analogs were derived from dose–response curves fitted as described in Section 2.6. A lower CR% indicates weaker crossrecognition and thus higher specificity.

• Representative results

For MutNb1, the P/N ratio against HS was 3.84 (≥ 2.1). The CR% for HS was 100%, and for the other analogs all CR% values were below 6.60%, demonstrating that MutNb1 exhibits high specificity for HS. The detailed description above has been added to Section 2.6 (Methods) of the revised manuscript, rendering the specificity assessment procedure complete.

Once again, we sincerely thank you for your professional review and valuable suggestions.

 

References

[34] ValdésTresanco, M.S.; ValdésTresanco, M.E.; JiménezGutiérrez, D.E.; Moreno, E. Structural modeling of nanobodies: a benchmark of state-of-the-art artificial intelligence programs. Molecules. 2023, 28, 3991. DOI:10.3390/molecules28103991.

[35]Vecchietti, L.F.; Wijaya, B.N.; Armanuly, A.; Hangeldiyev, B.; Jung, H.K.; Lee, S.; Cha, M.; Kim, H.M. Artificial intelligence-driven computational methods for antibody design and optimization. mAbs. 2025, 17, 2528902. DOI:10.1080/19420862.2025.2528902.

[36]Salamouni, N.S.; Cater, J.H.; Spenkelink, L.M.; et al. Nanobody engineering: computational modelling and design for biomedical applications. FEBS Open Bio. 2025, 15, 236–253. DOI:10.1002/2211-5463.13850.

 

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors addressed my concerns in their revision but were not willing to confirm their simulation experimentally. They clearly indicate this and the corresponding limitations in the discussion section. Since this is a modelling and design paper, I can accept that.

Author Response

Response to Reviewer 1

1. The authors addressed my concerns in their revision but were not willing to confirm their simulation experimentally. They clearly indicate this and the corresponding limitations in the discussion section. Since this is a modelling and design paper, I can accept that.

Reply:

We sincerely thank Reviewer for the understanding and thoughtful consideration. We fully acknowledge the reviewer’s suggestion regarding experimental validation, and we have clearly indicated this limitation in the discussion section of the revised manuscript. We are truly grateful that the reviewer recognizes this work as a modeling and design study and accepts it as such. Thank you again for the constructive comments and for your final acceptance.

Reviewer 2 Report

Comments and Suggestions for Authors

The revised paper is already a net improvement to the first version. However, it still contains 2 major shortcomings. 

1. The rationale of the experiment should be: We want a binder (preferably a Nanobody) that will recognise Heparan sulphate (HS). So we look into the database of known Nanobody structures in complex with small organic molecules. The PDB contains several of such structures (1I3V, 1UOQ, 1QDO, 7TJC, etc..) The 7TJC woudl be a good starting point as it clamps its antigen, chloroamphenicol, between teh CDR3 and the Framework region 2, and the size and nature of chloramphenicol and HS bears some (weak) similarities. That should have been the reason to prefer and start with the 7TJC structure to try to fit and mutagenise the 7TJC Nanobody for associating with HS. As it stands now, we only find in the discussion; "We started from 7TJC", without any hint as to why this Nanobody was chosen. You could have chosen another Nanobody out of the >1000 Nanobody structures in the PDB. The introduction (it can be repeated in the discussion) should already contain this rationale.  
2. It is absolutely necessary to demonstrate that the original chloramphenicol-specific Nanobody is NOT recognising the HS as it is possible that this WT Nanobody cross reacts with HS. So, you should synthesize the gene for this WT Nanobody, expresse, purify and test it in ELISA in parallel with the mutated version for both chloramphenicol and HS. To proof that the specificity was changed completely. We do not believe in the accuracy of the in silico measurement of the complex stabilisation. It should be confirmed by an ELISA, SPR, BLI or  ITC measurement. ELISA will be the fastest. If you could have the crystal structure of the mutant Nanobody in complex with HS that would of course be preferred, but this is going to take too much time. 

Then you should also correct the data of Table 4. The triple mutations. The first and third are apparently identical F47R/D99Y/Y108/P, which is not the case as their mutation energy is different. 

Author Response

List of Corrections – antibodies-4272529 Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate

 

Dear Reviewer,

Thank you for your comments and advice! We have revised the manuscript “Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate” carefully. The list of correction is as follows. Thank you very much!

 

 

Sincerely,

Dr. Junfang Hao,

College of Smart Animal Husbandry, College of Biology and Food, Shangqiu Normal University, China.

 

 

Response to Reviewer 2

1. The rationale of the experiment should be: We want a binder (preferably a Nanobody) that will recognise Heparan sulphate (HS). So we look into the database of known Nanobody structures in complex with small organic molecules. The PDB contains several of such structures (1I3V, 1UOQ, 1QDO, 7TJC, etc..) The 7TJC woudl be a good starting point as it clamps its antigen, chloroamphenicol, between teh CDR3 and the Framework region 2, and the size and nature of chloramphenicol and HS bears some (weak) similarities. That should have been the reason to prefer and start with the 7TJC structure to try to fit and mutagenise the 7TJC Nanobody for associating with HS. As it stands now, we only find in the discussion; "We started from 7TJC", without any hint as to why this Nanobody was chosen. You could have chosen another Nanobody out of the >1000 Nanobody structures in the PDB. The introduction (it can be repeated in the discussion) should already contain this rationale.

Reply (Lines 328-337):

We greatly appreciate the reviewer’s valuable and insightful comments. We fully agree with your suggestion and have supplemented and clarified the detailed selection basis of wild-type Nb scaffold 7TJC in the revised manuscript.

Apart from the favorable molecular docking score between 7TJC and HS, another core reason for choosing 7TJC as the starting scaffold is its inherent structural binding characteristic: this native chloramphenicol-specific nanobody can clamp its ligand between the CDR3 loop and framework region 2. Importantly, the molecular size and physicochemical property of its original ligand chloramphenicol share certain structural similarities with HS, which makes 7TJC possess a naturally suitable binding pocket environment for further targeted modification and site-directed mutagenesis toward HS recognition.

Considering that more than 1000 nanobody structures are available in the PDB database, the above structural binding mode advantage combined with favorable docking affinity (–7.30  kcal/mol) jointly confirm the rationality of selecting 7TJC rather than other nanobodies as the modification template. The full selection basis has been added into the corresponding part of the manuscript as suggested.

 

2. It is absolutely necessary to demonstrate that the original chloramphenicol-specific Nanobody is NOT recognising the HS as it is possible that this WT Nanobody cross reacts with HS. So, you should synthesize the gene for this WT Nanobody, expresse, purify and test it in ELISA in parallel with the mutated version for both chloramphenicol and HS. To proof that the specificity was changed completely. We do not believe in the accuracy of the in silico measurement of the complex stabilisation. It should be confirmed by an ELISA, SPR, BLI or ITC measurement. ELISA will be the fastest. If you could have the crystal structure of the mutant Nanobody in complex with HS that would of course be preferred, but this is going to take too much time.

Reply (Lines 377-391)

We sincerely thank the reviewer for the rigorous and thoughtful guidance. We fully understand your request: to demonstrate a complete specificity switch by comparing the binding of the wild-type Nb and mutant nanobodies (Mut-Nb1) to both chloramphenicol and HS in parallel ELISA assays.

To be frank, although this study did not perform a parallel comparative validation of the binding activity and specificity of the wild-type Nb versus the Mut-Nb1 toward chloramphenicol, we have systematically characterized the binding properties of Mut-Nb1 to HS and its structural analogs in the main text. The K_D of Mut-Nb1 for HS was determined to be 65.87 nM, with a P/N ratio of 3.84 and a cross-reactivity rate of <6.60%. In addition, in response to Comment #5 from the previous round of review, we have already provided the binding data of the wild-type Nb to HS and its structural analogs. The results showed that the wild-type Nb exhibited only weak binding to HS, with an OD₄₅₀nm value of approximately 0.58 and a P/N ratio of 1.72 (relative to the PBS negative control). Its binding activities to KS and CS were even lower, with OD₄₅₀nm values of 0.20 and 0.29, respectively. This weak binding is likely attributable to partial similarities in functional groups and local spatial conformation among the ligands, rather than specific targeted recognition.

In addition to the binding free energy of the HS–Mut-Nb1 complex obtained from molecular dynamics simulations (ΔG = –83.26 ± 3.06 kcal/mol), the above series of in vitro experimental results consistently demonstrate that, following site-directed mutagenesis, both the binding affinity and recognition specificity of the Mut-Nb1 toward HS were significantly enhanced.

We acknowledge that the lack of parallel ELISA comparison of wild-type Nb and Mut-Nb1 against chloramphenicol may be a limitation of the current work. To address your request as much as possible, we have supplemented the Discussion section of the revised manuscript with comparative data on molecular docking scores and binding free energies, further supporting the binding characteristics of the Mut-Nb1 to HS.

We once again sincerely thank the reviewer for your understanding.

3. Then you should also correct the data of Table 4. The triple mutations. The first and third are apparently identical F47R/D99Y/Y108/P, which is not the case as their mutation energy is different.

Reply: (Table 4, third row)

We thank the reviewer for the careful reading. Upon verification, the triple mutation in the first row of Table 4 is F47R/D99Y/Y108P, while that in the third row is F47R/D99F /Y108P (not F47R/D99Y/Y108P). The two combinations differ at position 99 (Tyr vs Phe), and thus the different mutation energies are justified. To avoid any potential confusion, we have now colored the amino acid letters in red and adjusted the table formatting in the revised manuscript. We appreciate the reviewer’s attention to detail.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Comments and suggestions have been addressed in the updated manuscript. 

Author Response

Response to Reviewer 3

1. Comments and suggestions have been addressed in the updated manuscript.

Reply:

We greatly appreciate the reviewer’s careful review and positive feedback on our revised manuscript. Thank you for your valuable comments and suggestions, which have helped us significantly improve the quality of this work. Thank you very much!

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

Perhaps I was not clear enough. I had 2 remarks: First, one on the selection of the starting Nanobody and a second on generating the WT nanobody and the mutant Nanobody and testing their affinity (by ELISA, or other method) in parallel on chloramphenicol, and on Heparan sulphate.

The authors replied favourable on the first remark. The second remark was not done properly. It should not be too complicated or expensive. 

Design the codons of the WT Nanobody and add a His tag. Order the sequence and let it be cloned in an expression vector.(This should cost you 100-200 $, and within 2 weeks you might get the clone). Transform the clone and Express the sequence, purify the protein and in parallel express and purify again your mutated clone (or while waiting for the second clone to arrive). (should take less than 1 week)

Measure the OD280 of both preparation (IMAC and SEC purified), test on gel that the quality is comparable. And perform an ELISA on OVA-chloroform and on OVA-HS, use OVA as blank using seral dilutions of both Nanobodies, from 2 microM to 0.1 nM (for example) (should take a two days).

So, within about 3 weeks these experiments are done, and they will not be that expensive at all. 

As it stands, the authors try to convince the reader, that the in silico affinity measurements (or Delta G values) and the P/N values are sufficient to proof that there is no cross reactivity and that the WT doesn't bind to HS. It is impossible to accept these P/N values as trustable, however, they are done (as far as e can see) for one loading of antigen (we don't even know the yield of coupling of the hapten to carrier), one loading of Nanobody. So this could be in an overloaded or lag phase of the signal. 

Some time ago, I also tried to predict the affinity (enthalpy and entropic factors) of nanobody to antigen. Depending on the algorithm used and the settings within the algorithm one obtains a wide range of Delta G, and they not always correspond to the real affinity as measured by SPR. 

 

Author Response

List of Corrections – antibodies-4272529 Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate

 

Dear Editor and Reviewer,

Thank you for your valuable comments and constructive suggestions! We have thoroughly revised the manuscript “Rational design and characterization of a mutated nanobody for specific targeting of heparan sulfate” as advised. The list of correction is as follows. Thank you very much! 

We greatly appreciate your time and effort.

 

 

Sincerely,

Dr. Junfang Hao,

College of Smart Animal Husbandry, College of Biology and Food, Shangqiu Normal University, China.

 

Response to Reviewer 2

1. Perhaps I was not clear enough. I had 2 remarks: First, one on the selection of the starting Nanobody and a second on generating the WT nanobody and the mutant Nanobody and testing their affinity (by ELISA, or other method) in parallel on chloramphenicol, and on Heparan sulphate.

The authors replied favourable on the first remark. The second remark was not done properly. It should not be too complicated or expensive.

Design the codons of the WT Nanobody and add a His tag. Order the sequence and let it be cloned in an expression vector.(This should cost you 100-200 $, and within 2 weeks you might get the clone). Transform the clone and Express the sequence, purify the protein and in parallel express and purify again your mutated clone (or while waiting for the second clone to arrive). (should take less than 1 week)

Measure the OD₂₈₀ of both preparation (IMAC and SEC purified), test on gel that the quality is comparable. And perform an ELISA on OVA-chloroform and on OVA-HS, use OVA as blank using seral dilutions of both Nanobodies, from 2 microM to 0.1 nM (for example) (should take a two days).

So, within about 3 weeks these experiments are done, and they will not be that expensive at all.

As it stands, the authors try to convince the reader, that the in silico affinity measurements (or Delta G values) and the P/N values are sufficient to proof that there is no cross reactivity and that the WT doesn't bind to HS. It is impossible to accept these P/N values as trustable, however, they are done (as far as e can see) for one loading of antigen (we don't even know the yield of coupling of the hapten to carrier), one loading of Nanobody. So this could be in an overloaded or lag phase of the signal.

Some time ago, I also tried to predict the affinity (enthalpy and entropic factors) of nanobody to antigen. Depending on the algorithm used and the settings within the algorithm one obtains a wide range of Delta G, and they not always correspond to the real affinity as measured by SPR.

Rely (Lines 291–324, 402–405, Figure 4, Table 5)

We sincerely and greatly appreciate the reviewer’s rigorous comments and insightful suggestions, which have significantly improved the reliability and rigor of our study. We fully agree with the reviewer that the previous in silico predicted ΔG values and single-point P/N data were insufficient to solidly confirm the binding and cross-reactivity characteristics of the nanobodies, and parallel experimental validation of wild-type Nb and Mut-Nb1 was required.

To fully address the concerns raised by the reviewer, our team worked intensively and overtime within the limited 10-day revision period to complete required supplementary experiments strictly following the detailed experimental procedures recommended in the comments. First, we performed codon optimization of the wild-type Nb sequence, added a His-tag, and successfully constructed the wild-type Nb expression vector through gene synthesis and cloning. Subsequently, both wild-type Nb and Mut-Nb1 were expressed in parallel and purified via IMAC. The protein concentration of both preparations was determined by OD₂₈₀ measurement, and their consistent protein quality and purity were verified by SDS-PAGE. Following the reviewer’s advice, OVA was set as blank control, and OVA-conjugated heparan sulfate (HS) and OVA-conjugated Chloramphenicol (CAP) were separately coated onto ELISA plates.

The dose-dependent binding curves, K_D values, P/N ratios, and cross-reactivity (CR%) values of the two Nbs were quantitatively compared and comprehensively analyzed. All newly obtained data have been supplemented in Figure 4A–E and Table 5 in the revised manuscript. The obtained experimental data show that Mut-Nb1 exhibits a 6.78-fold higher binding affinity for HS and a 5.80-fold lower affinity for CAP compared with wild-type Nb. Meanwhile, this mutant displayed markedly decreased cross-reactivity against non-target compounds including CAP, CS and KS.

With these comprehensive parallel experimental results, our conclusions no longer rely merely on in silico thermodynamic prediction. All the experimental results have been incorporated into the results, figure 4, Table 5, and discussion of the revised manuscript. We believe that the newly added biochemical data greatly strengthen the validity and reliability of our findings.

Once again, we sincerely thank the reviewer for your patient guidance and valuable suggestions, which have substantially improved the quality of our manuscript. We hope that the revised manuscript will meet with your approval.

Author Response File: Author Response.pdf

Round 4

Reviewer 2 Report

Comments and Suggestions for Authors

The authors meet all comments and critics I provided earlier.