Improving 10-deacetylbaccatin III-10-β-O-acetyltransferase catalytic fitness for Taxol production

The natural concentration of the anticancer drug Taxol is about 0.02% in yew trees, whereas that of its analogue 7-β-xylosyl-10-deacetyltaxol is up to 0.5%. While this compound is not an intermediate in Taxol biosynthetic route, it can be converted into Taxol by de-glycosylation and acetylation. Here, we improve the catalytic efficiency of 10-deacetylbaccatin III-10-O-acetyltransferase (DBAT) of Taxus towards 10-deacetyltaxol, a de-glycosylated derivative of 7-β-xylosyl-10-deacetyltaxol to generate Taxol using mutagenesis. We generate a three-dimensional structure of DBAT and identify its active site using alanine scanning and design a double DBAT mutant (DBATG38R/F301V) with a catalytic efficiency approximately six times higher than that of the wild-type. We combine this mutant with a β-xylosidase to obtain an in vitro one-pot conversion of 7-β-xylosyl-10-deacetyltaxol to Taxol yielding 0.64 mg ml−1 Taxol in 50 ml at 15 h. This approach represents a promising environmentally friendly alternative for Taxol production from an abundant analogue.

That said, this reviewer recommends that the authors rework the current article to de-emphasize the notion that their work describes novel activity of DBAT and xylosidases with taxanes. DBAT has been characterized and its substrate scope, including 10-deacetyltaxols with regards to the current work, has been done. Several early attempts to deglycosylate taxols were cited by the authors and thus reduces the novelty of the current work, IF emphasis is placed on enzyme discovery.
Emphasis should be placed mostly on directed evolution methods (there are several examples in the literature with other enzymes) to increase enzyme fitness by the popular method of alanine scanning mutagensis. Consequently, this will guide the authors to cite references that show successes in these approaches.
There are several instances where the phrasing of sentences needs work; otherwise, the meaning is lost or the flow of the reading is disrupted.
Attachments are included showing annotations within the text that need to be address before this can be a competitive manuscript.
Reviewer #2 (Remarks to the Author) Li et al. showed an alternative reaction for Taxol production. The alternative reaction does not consume natural sources as the reactants do not solely depend on the Taxus species and is greener in the sense that it produces less waste compared with the conventional production methods. Also, through use of molecular modelling and mutagenesis experiments, authors have showed a double mutant of DBAT enzyme with enhanced catalytic efficiency. Lastly, they used the modeling studies to propose a molecular mechanism for the DBAT enzyme and also its mutant. To my view, this study holds an advancement to the current literature highlighting the need for efficient production of drugs from the natural sources. Hence, it is suitable for publication. However, I would like list one major and a few minor points that should be answered/revised before publication. The major shortcoming of the paper is the reliability of the homology model. Authors have used their structural model to interpret the function/selectivity. To be sure about the structure and function relationship, the structural information has to be originated from a physical source like Xray, or at least the modeling should be defended well enough to attribute functional roles to it. As the authors have stated that the sequence homology between the template and target is only 30%. Percent homology lower than 30% is very close to the twilight zone of alignments (Fischer, Barret et al. 1999). Swiss-model server has straightforward (automated) online modeling applications that works best if the alignment is not in the twilight zone. In this case, although having pivotal importance for the modeling, the pairwise sequence alignment between the template and target is missing. It might be necessary to see how the overall sequences diverge/converge in the alignment. Hence, I would recommend authors to show pairwşsw sequence alignment used for the modeling step. Also justify why this modeling can be trusted to deduce conclusions on the functionality. Minor points Figure 8 is not clear. The ligand, in particular, cannot be distinguished. For all of the parts of Fig. 8 I recommend a revision to clarify the residues and the ligand. Moreover, as suggested by the authors the enhanced catalytic activity obtained for the mutant G38R is likely to be due to two factors. First one is a single hydrogen bond addition to the overall structure, formed by N35 and R38. The second one is the steric clash generated in case of G38 compared with R38. Although single hydrogen bond is very weak (also we cannot assess the distance from the figure) to alter catalytic efficiency greatly, it may still be reasonable to think that the formation of this H-bond made G38R is more active than the wild-type. However, arginine is much bulkier than glycine. Hence, it is highly unlikely to accept that G38 presents a steric clash, but R38 does not. In other words, the mutagenesis of G38R does not provide any evidence regarding a broader substrate entrance for the mutant. On the contrary, F301V as authors have claimed presents a bulkier space than the wild type. Indeed, F and V have comparable volumes. Overall I would recommend authors to quantitatively investigate the changes in the geometry of the catalytic cleft and then make inferences. Many computational methods, and also web servers, are available and might present a fast and simple alternative for volume calculations. In the manuscript entitled 'Semi-rational site-directed mutagenesis of 10-deacetylbaccatin III-10β-O-acetyltransferase (DBAT) and biosynthesis of taxol from 7-β-xylosyl-10-deacetyltaxol by "onepot reaction"', the authors describe the possibility of transforming 7-β-xylosyl-10-deacetyltaxol into taxol by means of the action of the enzyme 10-deacetylbaccatin III-10-β-O-acetyltransferase. This enzyme controls the formation of baccatin III from 10-deacetylbaccatin III, apart from the preliminary removal of the xylosyl group. This interesting study explores the possible use of heterologous systems for obtaining taxol from the related compound 7-β-xylosyl-10-deacetyltaxol. The manuscript is clearly written and the results are well discussed. The authors have assayed different DBATs obtained from several Taxus species and have used the L-alanine scanning mutagenesis methodology to increase the activity of the enzymes studied. For all these reasons, I think this manuscript deserves publication in Nature Communications, after the following minor points are amended. 1) Why do the authors say that DT is a "by-product"? Is the biosynthesis of this compound known? Do the authors think that DT is formed from taxol? This point should be clarified and expanded on in the manuscript. 2) On page 14/23, line 289: Why do the authors assume that the gene expression levels were not different among different species? In which conditions? This assumption is erroneous.
3) The strategy used for changing amino acids in order to increase enzyme activity should be supported with more up-to-date bibliography. 4) There are a number of minor typing and language mistakes that should be corrected. For example, in the Introduction, page 3/23, line 52: "it" should be "which" ("…which can also be obtained"); in Results, page 5/23, line 127: delete "in" before "consistent"; page 17/23, line 363: canadiensis should not begin with a capital letter. A careful reading of the text is necessary to eliminate other errors.

Reviewer #1 (Remarks to the Author):
Q: That said, this reviewer recommends that the authors rework the current article to de-emphasize the notion that their work describes novel activity of DBAT and xylosidases with taxanes. DBAT has been characterized and its substrate scope, including 10-deacetyltaxols with regards to the current work, has been done. Several early attempts to deglycosylate taxols were cited by the authors and thus reduces the novelty of the current work, IF emphasis is placed on enzyme discovery.
A: Thank you for your comments. We have reorganized this part to de-emphasize the notion on the novel activity of DBAT and the β-xylosidase. In the revised manuscript, we just introduced the general research background of the two kind enzymes to readers for their easier understanding of this work. A: Thank you very much for your help in revision of the text and annotations, we have followed the annotations in the text to re-write or address them (see below and the revised manuscript).

Main questions or annotations in the text:
Q1: In what source (Taxol is about 0.02%, XDT is up to 0.5%)?
A: Line 16 (in the revised manuscript), "in yew trees" was added; line 17, "in the plants" was added.

Taxol by de-glycosylation and acetylation)?
A: Line 18-19, "by chemical or biological methods" was added.

Q3: This is NOT a new discovery (see ref 29).
A: The elucidation work was de-emphasized in the revised manuscript.

Q4: What is a "sent coat"?
A: Line 39, according to the description in the ref., the sentence was changed into "for which Taxol is used for the coronary stent or balloon catheter coating", and more ref. were added.

Q5: In what other genus species is it produced?
A

Q6: To what are the authors referring, T and pH or just pH?
A: Line 130, the sentence has been revised to define the optimum pH.

Q7: There are no points between 0 and 3h. How could a proper inflection point outside the true linear range be determined. More points are needed between 0-3h.
A: Thank you for your advice, actually, more points had been detected around 3h, and these points have been added in Figure 6a.

Q8: Why did the authors assume the enzyme activity was degrading as the rate approached what could be equilibrium? These acyltransferase reactions are reversible!
A: In the pre-test, we found that the enzyme was not much stable, especially at a relatively higher temperature, and the linear portion of the time course plot was within 3 h, so, we chose the time point of 3 h to initiate the enzyme supplementing to maximize the Taxol yield.

Q9: Why was the concentration chosen? Why not 1.5 mg/mL?
A: Line 278-279, the sentence has been reorganized: "All of the yields met the requirement for the next reaction from DT to Taxol and the relatively lower concentration of 0.5 mg/mL LXYL-P1-2 was chosen in the following "one-pot reaction" system". It is also to limit the enzyme consumption.
Q10: The entire 50 mL lot was purified on the HPLC? 3 / 5 A: Yes, at first, the product from the entire 50 mL was extracted through ethyl acetate, then, the product was concentrated into a small volume, and finally, the product was purified and prepared by HPLC.

Q11: How many (steps are there in Taxol biosynthesis)?
A: Line 297-302, the sentences have been reorganized. From cited ref., there are 20 enzymes involved in 19 steps of the biosynthetic pathway from GGPP to Taxol.
Other changes have also been made mainly based on the instructions of the reviewer (see the revised manuscript).

Reviewer #2 (Remarks to the Author):
Q: The major shortcoming of the paper is the reliability of the homology model. Authors have used their structural model to interpret the function/selectivity. To be sure about the structure and function relationship, the structural information has to be originated from a physical source like X-ray, or at least the modeling should be defended well enough to attribute functional roles to it. As the authors have stated that the sequence homology between the template and target is only 30%. Percent homology lower than 30% is very close to the twilight zone of alignments (Fischer, Barret et al. 1999). Swiss-model server has straightforward (automated) online modeling applications that works best if the alignment is not in the twilight zone. In this case, although having pivotal importance for the modeling, the pairwise sequence alignment between the template and target is missing. It might be necessary to see how the overall sequences diverge/converge in the alignment. Hence, I would recommend authors to show pairwise sequence alignment used for the modeling step.

Also justify why this modeling can be trusted to deduce conclusions on the functionality.
A: Thank you for your comments and suggestions. Since the crystal structure of DBAT is not available, the homology modeling has to be used in this work. Generally, the term "homology" may not be measured by percentage, you probably mean the "similarity". It's our fault that we did not involve these very important data in the original manuscript. These data have been added in the revised manuscript. The sequence similarity between DBAT and the template was 45%, which should guarantee the accuracy of the predicted structure. Thank you for your suggestion, the pairwise sequence alignment between the template and DBAT has been added to Fig. 3 and in the supplementary materials, which made the article more readable. And the corresponding paragraphs have been reorganized with more ref. The predicted model was assessed to be qualified by a Ramachandran plot (Supplementary Fig. 3). DBAT also showed similar hydrophobicity and distribution of electronegativity residues with the template ( Supplementary  Fig. 4). Therefore, we think that the homology model is reasonable for the protein engineering research of DBAT, although it may have some shortcomings. Figure 8 is not clear. The ligand, in particular, cannot be distinguished. For all of the parts of Fig. 8 I recommend a revision to clarify the residues and the ligand. A: Fig. 8 has been revised to highlight the ligand and the two mutation sites.

Minor points Q:
Moreover, as suggested by the authors the enhanced catalytic activity obtained for the mutant G38R is likely to be due to two factors. First one is a single hydrogen bond addition to the overall structure, formed by N35 and R38. The second one is the steric clash generated in case of G38 compared with R38. Although single hydrogen bond is very weak (also we cannot assess the distance from the figure) to alter catalytic efficiency greatly, it may still be reasonable to think that the formation of this H-bond made G38R is more active than the wild-type. However, arginine is much bulkier than glycine. Hence, it is highly unlikely to accept that G38 presents a steric clash, but R38 does not. In other words, the mutagenesis of G38R does not provide any evidence regarding a broader substrate entrance for the mutant.
A: Thank you for your reminding. In the revised manuscript, we did not emphasize the single hydrogen bond formation between N35 and R38, since further analysis showed that the H bond was also formed between N35 and G38 in the wild type; and you are right, arginine is much bulkier than glycine, there must be other reasons for the improved catalytic efficiency of the G38R mutation. We found that the amino acid residue of 38 site is near the surface, and the side-chain of Arg is more hydrophilic than that of Gly, so the increased hydrophilic property of G38R mutation is probably one of the reasons for the elevated catalytic efficiency. The related paragraphs have been reorganized in the revised manuscript. A: Thank you for your comments. During the extraction process, XDT is often obtained as a companion of Taxol, so it is defined as a by-product of Taxol. How XDT is formed is unknown, and we have discussed this question in the revised manuscript (Line 306-310). Again, thank you for your kind letter and the constructive comments.
Reviewer #1 (Remarks to the Author) Lines 22-23: Reword: "The research not only exhibited potential substrate binding or catalytic sites of DBAT..." This is a vague remark. Rewrite: "In this study, the structure of an enzyme homolog was used to model DBAT, identify its active site, and to design a double mutant DBAT(G38R/F301V) that had a kcat/Km of XXXX." L25: Change to: "An in vitro..." L25: Omit quotation marks around "one-pot reaction"...unnecessary. (CHANGE ALL instances of this) L28: Change to: ...another choice for biocatalytic production of Taxol... L48: Change to: ..."percentage found in the bark.9" L64: Xylosyl (capital 'X' at start of sentence) L64: Explain here very briefly why XDT is considered a by-product and not a precursor of Taxol.