Structure-Guided Loop Grafting Improves Expression and Stability of Influenza Neuraminidase for Vaccine Development

Reviewer #1 (Public review):

Summary:

This manuscript described a structure-guided approach to graft important antigenic loops of the neuraminidase to a homotypic but heterologous NA. This approach allows the generation of well-expressed and thermostable recombinant proteins with antigenic epitopes of choice to some extent. The loop-grafted NA was designated hybrid.

Strengths:

The hybrid NA appeared to be more structurally stable than the loop-donor protein while acquiring its antigenicity. This approach is of value when developing a subunit NA vaccine which is difficult to express. So that antigenic loops could be potentially grafted to a stable NA scaffold to transfer strain-specific antigenicity.

Weaknesses:

However, major revisions to better organize the text, and figure and make clarifications on a number of points, are needed. There are a few cases in which a later figure was described first, data in the figures were not sufficiently described, or where there were mismatched references to figures.

More importantly, the hybrid proteins did not show any of the advantages over the loop-donor protein in the format of VLP vaccine in mouse studies, so it's not clear why such an approach is needed to begin with if the original protein is doing fine.

Reviewer #2 (Public review):

In their manuscript, Rijal and colleagues describe a 'loop grafting' strategy to enhance expression levels and stability of recombinant neuraminidase. The work is interesting and important, but there are several points that need the author's attention.

Major points

(1) The authors overstress the importance of the epitopes covered by the loops they use and play down the importance of antibodies binding to the side, the edges, or the underside of the NA. A number of papers describing those mAbs are also not included.

(2) The rationale regarding the PR8 hybrid is not well described and should be described better.

(3) Figure 3B and 6C: This should be given as numbers (quantified), not as '+'.

(4) Figure 5A and 7A: Negative controls are missing.

(5) The authors claim that they generate stable tetramers. Judging from SDS-PAGE provided in Supplementary Figure 3B (BS3-crosslined), many different species are present including monomers, dimers, tetramers, and degradation products of tetramers. In line 7 for example there are at least 5 bands.

Author response:

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript described a structure-guided approach to graft important antigenic loops of the neuraminidase to a homotypic but heterologous NA. This approach allows the generation of well-expressed and thermostable recombinant proteins with antigenic epitopes of choice to some extent. The loop-grafted NA was designated hybrid.

Strengths:

The hybrid NA appeared to be more structurally stable than the loop-donor protein while acquiring its antigenicity. This approach is of value when developing a subunit NA vaccine which is difficult to express. So that antigenic loops could be potentially grafted to a stable NA scaffold to transfer strain-specific antigenicity.

Weaknesses:

However, major revisions to better organize the text, and figure and make clarifications on a number of points, are needed. There are a few cases in which a later figure was described first, data in the figures were not sufficiently described, or where there were mismatched references to figures.

More importantly, the hybrid proteins did not show any of the advantages over the loop-donor protein in the format of VLP vaccine in mouse studies, so it's not clear why such an approach is needed to begin with if the original protein is doing fine.

We thank the reviewer for their helpful comments. We have incorporated feedback from the authors to improve the manuscript. Please see our point-by-point response.

The purpose of loop-grafting between H5N1/2021 (a high-expressor) and the PR8 virus was not to improve the expression of PR8, which is already a good expressing NA. Instead, the loop-grafting and the in vivo experiments were done to show the loop-specific protection following a lethal PR8 virus challenge.

Reviewer #2 (Public review):

In their manuscript, Rijal and colleagues describe a 'loop grafting' strategy to enhance expression levels and stability of recombinant neuraminidase. The work is interesting and important, but there are several points that need the author's attention.

Major points

(1) The authors overstress the importance of the epitopes covered by the loops they use and play down the importance of antibodies binding to the side, the edges, or the underside of the NA. A number of papers describing those mAbs are also not included.

We have discussed the distribution of epitopes on NA molecule in the Discussion section "The distribution of epitopes in neuraminidase" (new line number 350). In Supplementary Figures 1 and 2, we have compiled the epitopes reported by polyclonal sera and mAbs via escape virus selection or crystal structural studies. There are 45 residues examples of escape virus selection, and we found that approximately 90% of the epitopes are located within the top loops (Loops 01 and Loops 23, which include the lateral sides and edges of NA). We have also included the epitopes of underside mAbs NDS.1 and NDS.3 in Supplementary Figure 2. Some of the interactions formed by these mAbs are also within the L01 and L23 loops. All relevant references are cited in Supplementary Figures 1 and 2.

A new figure has been added [Figure 1b (ii)] to illustrate the surface mapping of epitopes on NA.

(2) The rationale regarding the PR8 hybrid is not well described and should be described better.

We described the rationale for the PR8 hybrid (new lines 247-250). For clarity, we have added the following sentence within the section "Loop transfer between two distant N1 NAs:...."

(new lines 255-258):

"mSN1 showed sufficient cross-reactivity to N1/09 to protect mice against virus challenge. Therefore, we performed loop transfer between mSN1 and PR8N1, which differ by 18 residues within the L01 and L23 loops and show no or minimal cross-reactivity, to assess the loop-specific protection."

(3) Figure 3B and 6C: This should be given as numbers (quantified), not as '+'.

We have included the numerical data in Supplementary Figure 6. The data is presented in semi-quantitative manner for simplification. To improve clarity, we have now added the following sentence to the Figure 3c legend: "Refer to Supplementary Figure 6 for binding titration data".

(4) Figure 5A and 7A: Negative controls are missing.

A pool of Empty VLP sera was included as a negative control, showing no inhibition at 1:40 dilution. In the figure legends, we have stated "Pooled sera to unconjugated mi3 VLP was negative control and showed no inhibition at 1:40 dilution (not included in the graphs)"

(5) The authors claim that they generate stable tetramers. Judging from SDS-PAGE provided in Supplementary Figure 3B (BS3-crosslinked), many different species are present including monomers, dimers, tetramers, and degradation products of tetramers. In line 7 for example there are at least 5 bands.

Tetrameric conformation of soluble proteins is evidenced by the size-exclusion chromatographs shown in Figures 3a and 6b. The BS3 crosslinked SDS-PAGE are only suggestive data, indicating that the protein is a tetramer if a band appears at ~250 kDa. However, depending on the reaction conditions, lower molecular weight bands may also be observed if crosslinking is incomplete.