Discovery of fungal oligosaccharide-oxidising flavo-enzymes with previously unknown substrates, redox-activity profiles and interplay with LPMOs

Oxidative plant cell-wall processing enzymes are of great importance in biology and biotechnology. Yet, our insight into the functional interplay amongst such oxidative enzymes remains limited. Here, a phylogenetic analysis of the auxiliary activity 7 family (AA7), currently harbouring oligosaccharide flavo-oxidases, reveals a striking abundance of AA7-genes in phytopathogenic fungi and Oomycetes. Expression of five fungal enzymes, including three from unexplored clades, expands the AA7-substrate range and unveils a cellooligosaccharide dehydrogenase activity, previously unknown within AA7. Sequence and structural analyses identify unique signatures distinguishing the strict dehydrogenase clade from canonical AA7 oxidases. The discovered dehydrogenase directly is able to transfer electrons to an AA9 lytic polysaccharide monooxygenase (LPMO) and fuel cellulose degradation by LPMOs without exogenous reductants. The expansion of redox-profiles and substrate range highlights the functional diversity within AA7 and sets the stage for harnessing AA7 dehydrogenases to fine-tune LPMO activity in biotechnological conversion of plant feedstocks.

The manuscript by Momeni et al. is a well-executed investigation into a family of flavoenzymes (identified as AA7 in the CAZy database) that is particularly abundant in fungi, oomycetes and plants. The authors have integrated a comprehensive phylogenetic analysis with in vitro and structural characterisation of selected members, shedding new light on the structural basis underpinning different types of activity (oxidase vs dehydrogenase) seen within this enzyme family. The authors have characterised the activity of selected AA7 enzymes on a range of substrates, identified FgCelDH7C as an unusual dehydrogenase active on cello-oligosaccharides and shown its ability to deliver electrons to (previously characterised) AA9 LPMOs active on cellulose.
The range of experimental techniques and the quality of the data is commendable, and the authors make a number of interesting observations, however I feel that the story as a whole is not really ground breaking nor some of the findings entirely surprising. Several AA7 enzymes have been characterised both biochemically and structurally, and are known to catalyse the oxidation of several types of mono and oligosaccharides (including cello-oligosaccharides, xylo-oligosaccharides, chitooligosaccharides, oligo-galacturonides, lactose) into their lactones, although the authors here have dug deeper into the some specific aspects. The other reason of the limited novelty of current work is that several classes of redox proteins (notably FAD-dependent AA3 cellobiose dehydrogenases and AA12 pyrroloquinoline-quinone-dependent oxidoreductases) have been previously shown to be good electron donors for LPMOs, and adding AA7 to the list does not feel to me particularly revolutionary (nor unexpected). FgCelDH7C seems to be better than fungal CDH at driving one specific LPMO tested in this study, but does not necessarily mean it will do that with any LPMO.
Hence I feel that this (nice) piece of work would be better suited for a more specialised journal.
Suggestions and specific comments: -Line 56-58: not sure that this statement is correct. Unless I am mistaken, besides the four characterised fungal AA7s included in the CAZy database, several other non-fungal AA7s have been characterised before, although not included in CAZy (no idea why). -Line 219-220: typo ("is that is that") -Line 415: should be (Fig. 6, not Fig. 5).
- Supplementary Fig. 1: I have a few issues with this figure and the way the data are shown. The order in which the various substrates are listed in panel a vs c vs e (referring to the oxidases activities) is not consistent and should be corrected (for example, panel a shows cellobiose followed by maltose, while panel c inverts the order). Also, substrate GNB is in panel b and d but is missing from panel a and c. Also there are some substrates in panel e (e.g. ManNAc and galactosamine) that are not in panels a-d, and viceversa some are absent from panel e (for example LNB and GNB). And then panel f (showing dehydrogenase activity) has yet a different range of substrates compared to e (which should be the corresponding oxidase activity). It would also be nice to show in panel e the dehydrogenase activity on cellobiose, which in panel e seems like a decent substrate for the oxidase activity of PbChi7A. In summary, there should be better consistency between the various panels of this figure.
Reviewer #2 (Remarks to the Author): Referee report on "Discovery of fungal oligosaccharide-oxidising flavo-enzymes with previously unknown substrates, redox-activity profiles and interplay with LPMOs" Momeni et al. has undertaken a detailed and thorough study of several fungal oligosaccharideoxidizing flavo-enzymes capable of driving LPMO action. The study is well-planned and executed. There is no doubt that our understanding of LPMO action in nature and possibilities to fuel these enzymes in applied settings is greatly enhanced by this work. I strongly recommend publication in Nature Communications. Still, I would appreciate that two small issues were addressed as described below.
Minor comments: 1. Page 12.: Although the interplay between FgCelDH7C and PaLPMO9E and PaLPMO9H, respectively, is exquisitely demonstrated and described, I do struggle with the use of the term synergy. Especially how the results of the interplay or synergy is compared with the use of molecular oxygen in the presence of ascorbate for LPMO action. Synergy is an interaction or cooperation giving rise to a whole that is greater than the simple sum of its parts. Such synergy can be quantified using i.e. Abbot´s equation for synergy calculation. FgCelDH7C fuels LPMO action by producing H2O2, which is undoubtedly a better co-substrate than O2 with respect to enzyme speed (Bissaro et al. Nat. Chem. Biol 2017 and Kuusk et al. J. Biol. Chem. 2018). When O2 is used as a co-substrate, the rate of LPMO action is either dependent on an innate monooxygenase reaction i.e. building up the reactive oxygen species at the active site during turnover with substrate present or the LPMOs oxidase activity (reducing O2 to H2O2), which in turn is dependent on i.e. substrate concentration (equilibrium concentration of unbound LPMO), pH and so on. Without going into a further discussion or claiming which of these two scenarios describes the action of the AA9 in the presence of O2 and ascorbate, still, to a fairly large extent, the authors are comparing apples and oranges. I am fine with the authors claiming synergy, but the above issues should be mentioned with respect to the discussion with respect to observed differences in enzyme speed.
2. Page 14: "The fact that only partial reduction of the Cu(II) center was observed could be due to kinetic effects." This may be correct, but there is also a possibility that there is a thermodynamic reason. AA9 tend to have redox potentials of 0.19 to 0.23 V. What the potential of FgCelDH7C is I do not know, but the pre-incubation experiment with cellotriose and the pre-reduction with dithionite is in accordance with a close proximity of redox potentials. Vanillyl alcohol oxidase, as mentioned in the Introduction, has a reported redox potential of 0.055 V (van den Heuvel et al. J. Biol. Org, 2000), which is not too far from AA9 LPMOs.
Reviewer #3 (Remarks to the Author): Deeper comprehension of structure and activity of AA7 family members is important for understanding of the strategies employed by microorganisms for efficient lignocellulosic biomass enzymatic degradation and can have potential biotechnological applications.
The manuscript describes production, biophysical, structural and biochemical characterization of a set of recombinant AA7 enzymes from different clades within AA7 family. Tow crystal structures were solved and one enzyme was studied as an enzymatic tool to activate LPMOs. This is a nicely written and comprehensive manuscript, which can be further improved by following suggestions presented below: 1. Based on bioinformatics analysis, the authors subdivided AA7 family into 6 clades (I to VI) with some of them (clades II and V) being further subdivided into subclades A and B. They also state that all the characterized members of AA7 family map to the branch Va and lack of characterized members from clades II, III, IV, Vb and VI. Next the authors decided to focus on clades I, IIa, Va&b and VI. What is a rationale for such choice of the enzymes to study? If the idea was to make all the (uncharacterized) branches, some of them were not covered and some of the more characterized ones (Va) are present.
4. The authors show that both HPR and SOD, being added to LPMO+ FgCelDH7C+C3 reactions will cause dose-dependent decrease in LPMO activity (Fig. 5a). What is their interpretation of these results? The authors do not say if they added HRP substrate to the reaction (ABTS, AmplexRed, etc). How did the reaction proceed and what was the outcome? 5. SOD would act on dismutation of the superoxide radical into molecular oxygen and hydrogen peroxide. This should increase hydrogen peroxide concentration, which is a co-substrate of the LPMO. Nevertheless, the authors show that addition of SOD will lead to the decreased LPMO activity (Fig. 5b). One possible explanation of these results is that superoxide performs the LPMO priming reaction. However, the same suppression happens when DCIP is added to the reaction. The authors claim that the latter result indicates direct electron transfer from FgCelDH7C to LPMO and make a fair attempt to prove this using EPR. However, this apparently contradicts their own results with SOD. How the authors reconcile these observations?

Response to referees
We are very grateful to the constructive feedback and thorough work. Please find our detailed responses below: Reviewer #1: The manuscript by Momeni et al. is a well-executed investigation into a family of flavoenzymes (identified as AA7 in the CAZy database) that is particularly abundant in fungi, oomycetes and plants.
The authors have integrated a comprehensive phylogenetic analysis with in vitro and structural characterisation of selected members, shedding new light on the structural basis underpinning different types of activity (oxidase vs dehydrogenase) seen within this enzyme family. The authors have characterised the activity of selected AA7 enzymes on a range of substrates, identified FgCelDH7C as an unusual dehydrogenase active on cello-oligosaccharides and shown its ability to deliver electrons to (previously characterised) AA9 LPMOs active on cellulose. The range of experimental techniques and the quality of the data is commendable, and the authors make a number of interesting observations, however I feel that the story as a whole is not really ground breaking nor some of the findings entirely surprising. Several AA7 enzymes have been characterised both biochemically and structurally, and are known to catalyse the oxidation of several types of mono and oligosaccharides (including cello-oligosaccharides, xylo-oligosaccharides, chitooligosaccharides, oligo-galacturonides, lactose) into their lactones, although the authors here have dug deeper into the some specific aspects.
Response: We are very grateful for the thorough and constructive feedback. However, we respectfully do not share the same view regarding the novelty of the work based on: 1) All hitherto characterized AA7s are oxidases with a bi-covalently bound FAD, which has been the defining hallmark of this family. Our work changes the family definition, by revealing new clades and members with diverse modes of FAD-anchoring, substitutions in the FAD-domain "oxygen cage" and the correlation of these elements with a previously unknown oligosaccharide dehydrogenase activity. We are not sure if these findings can be considered as obvious without the comprehensive sequence, structural and biochemical work we present. 2) Our work identifies active site signatures that correlate with specificity towards carbohydrate/non-carbohydrate substrates and notably the identification of signatures for the different redox-profiles in this family, is another novel aspect of the work that considerably expands the inventory of AA7.
The other reason of the limited novelty of current work is that several classes of redox proteins (notably FAD-dependent AA3 cellobiose dehydrogenases and AA12 pyrroloquinoline-quinonedependent oxidoreductases) have been previously shown to be good electron donors for LPMOs, and adding AA7 to the list does not feel to me particularly revolutionary (nor unexpected).
Response: Thanks for this comment. However, we would like to kindly argue that we do not share the same view on the novelty of our findings, based on: 1) Importantly, both the PQQ-dependent enzyme (AA8-AA12) and the CDH (AA8-AA3) are bi-modular enzymes, where the electron transfer has been shown to be dependent on a cytochrome b haem N-terminal domain as opposed to direct electron from the single module AA7 flavo-dehydrogenase in the present work. 2) Indirect electron transfer via small organic molecules has been reported, but again this is different from the direct electron transfer, which, to the best of our knowledge, has not been explicitly shown before.

In summary, our work is the first to demonstrate direct electron transfer to the LPMO as opposed to haem or small molecule mediated transfer in the cases referred to by the reviewer. We have clarified these aspects in the revised manuscript to better highlight the novelty of the findings.
FgCelDH7C seems to be better than fungal CDH at driving one specific LPMO tested in this study, but does not necessarily mean it will do that with any LPMO. Hence I feel that this (nice) piece of work would be better suited for a more specialised journal.
Response: We would like to respectfully point out that the statement is not fully accurate. Our experiments have not been performed on a single LPMO, but on two different enzymes: the C1oxidising PaLPMO9E and the C4-oxidising PaLPMO9H (see L264-280 in the submitted manuscript and Supplementary Fig. 11) Supplementary Fig. 13). In conclusion, our previous and new experiments demonstrate that the new AA7 fuels the activity of three different AA9 LPMOs from two different fungal spp., suggesting the interplay of FgCelDH with LPMOs is not restricted to a specific LPMO.

Suggestions and specific comments:
-Line 56-58: not sure that this statement is correct. Unless I am mistaken, besides the four characterised fungal AA7s included in the CAZy database, several other non-fungal AA7s have been characterised before, although not included in CAZy (no idea why). Response: We thank the reviewer for the good comments. The statements in the introduction referred to the AA7 assigned enzymes. We are aware of and have cited the work on the other enzymes, where we argue that these cases fulfil the criteria that we have identified to be considered as AA7s. To clarify, the following statement has been added to the introduction: ". In addition, oligosaccharide oxidases from plants have been reported 13, 14 , but are currently not assigned into AA7." Here we cited Locci et al. (Plant J (2019) 98(3):540-554), but this enzyme is too divergent from canonical enzymes (<25% aa identity), so it did not fulfil our inclusion criterion in the tree. Please note that we have also communicated to the curator of the CAZy database, who indicated that the AA7 inventory will be expanded in as soon as their time permits to include additional sequences based on our work.
Response: We are grateful for the comment and the issues are corrected in the revised version.
- Supplementary Fig. 1: I have a few issues with this figure and the way the data are shown. The order in which the various substrates are listed in panel a vs c vs e (referring to the oxidases activities) is not consistent and should be corrected (for example, panel a shows cellobiose followed by maltose, while panel c inverts the order). Also, substrate GNB is in panel b and d but is missing from panel a and c. Also there are some substrates in panel e (e.g. ManNAc and galactosamine) that are not in panels ad, and vice versa some are absent from panel e (for example LNB and GNB). And then panel f (showing dehydrogenase activity) has yet a different range of substrates compared to e (which should be the corresponding oxidase activity). It would also be nice to show in panel e the dehydrogenase activity on cellobiose, which in panel e seems like a decent substrate for the oxidase activity of PbChi7A. In summary, there should be better consistency between the various panels of this figure.
Response: We thank the reviewer for the good observation. We have re-made the a-d panels to present the substrates consistently. The different substrates included in panels "e" and "f" compared to panels a-d, is due to the fact that the enzymes have different levels of activities for the different substrates, so for each enzymes, only the substrates that yield a signal higher than the noise are included. New experiments have been performed to add the missing dehydrogenase data in panel f. For GNB, the experiment has been done only in one assay based on availability of this pricy, but less preferred disaccharide, the relative activity level can be inferred from the dehydrogenase assay.

Reviewer
#2: Referee report on "Discovery of fungal oligosaccharide-oxidising flavo-enzymes with previously unknown substrates, redox-activity profiles and interplay with LPMOs" Momeni et al. has undertaken a detailed and thorough study of several fungal oligosaccharideoxidizing flavo-enzymes capable of driving LPMO action. The study is well-planned and executed.
There is no doubt that our understanding of LPMO action in nature and possibilities to fuel these enzymes in applied settings is greatly enhanced by this work. I strongly recommend publication in Nature Communications. Still, I would appreciate that two small issues were addressed as described below.
Minor comments: 1. Page 12.: Although the interplay between FgCelDH7C and PaLPMO9E and PaLPMO9H, respectively, is exquisitely demonstrated and described, I do struggle with the use of the term synergy. Especially how the results of the interplay or synergy is compared with the use of molecular oxygen in the presence of ascorbate for LPMO action. Synergy is an interaction or cooperation giving rise to a whole that is greater than the simple sum of its parts. Such synergy can be quantified using i.e. Abbot´s equation for synergy calculation. FgCelDH7C fuels LPMO action by producing H2O2, which is undoubtedly a better co-substrate than O2 with respect to enzyme speed (Bissaro et al. Nat. Chem. Biol 2017 and Kuusk et al. J. Biol. Chem. 2018). When O2 is used as a co-substrate, the rate of LPMO action is either dependent on an innate monooxygenase reaction i.e. building up the reactive oxygen species at the active site during turnover with substrate present or the LPMOs oxidase activity (reducing O2 to H2O2), which in turn is dependent on i.e. substrate concentration (equilibrium concentration of unbound LPMO), pH and so on. Without going into a further discussion or claiming which of these two scenarios describes the action of the AA9 in the presence of O2 and ascorbate, still, to a fairly large extent, the authors are comparing apples and oranges. I am fine with the authors claiming synergy, but the above issues should be mentioned with respect to the discussion with respect to observed differences in enzyme speed.
Response: We are very grateful for the positive and constructive feedback. We agree completely with the insightful reviewer and acknowledge that the ascorbate and AA7 reactions are not comparable and are not valid to claim synergy. We merely performed the ascorbate assay, due to the fact that ascorbate is the most common exogenous reductant in hitherto published LPMO assays. Our reasoning for using the term synergy for the AA7-reaction is that LPMO action (especially in a natural secretomes containing cellulases) generates cellooligosaccharide substrates for the AA7, and the action of the AA7 on cellooligosaccharides would generate the H2O2 co-substrate for the LPMOs. Indeed, we observe this type of "synergy" in the lack of added cellooligosaccharides, but the total level of the released cello-oligomers is enhanced markedly with addition of cellooligosaccharides (mimicking the presence of a cellulase). We have, however, revised the text and rephrased to omit the term "synergy" to avoid any misunderstandings based on the kind feedback.
2. Page 14: "The fact that only partial reduction of the Cu(II) center was observed could be due to kinetic effects." This may be correct, but there is also a possibility that there is a thermodynamic reason. AA9 tend to have redox potentials of 0.19 to 0.23 V. What the potential of FgCelDH7C is I do not know, but the pre-incubation experiment with cellotriose and the pre-reduction with dithionite is in accordance with a close proximity of redox potentials. Vanillyl alcohol oxidase, as mentioned in the Introduction, has a reported redox potential of 0.055 V (van den Heuvel et al. J. Biol. Org, 2000), which is not too far from AA9 LPMOs.
Response: We thank the reviewer and agree that the non-stoichiometric reduction could be theoretically due to either kinetic or thermodynamic barriers, so this is added in the revised manuscript. We previously determined the redox potentials of the PaAA9E and PaAA9H used in this study to +155 mV and +326 mV respectively ( [30990][30991][30992][30993][30994][30995][30996]. We expect that that FgCelDHA to have a redox potential in the +50 mV range, which is likely to be sufficient to allow the LPMO-Cu(II) to harvest the electron from the FAD-center (compare to redox potentials of the haem domain of CHDs which is in the +90 mM region (at pH 6). This argumentation is also added to the revised manuscript to aid the reader.
Reviewer #3 (Remarks to the Author): Deeper comprehension of structure and activity of AA7 family members is important for understanding of the strategies employed by microorganisms for efficient lignocellulosic biomass enzymatic degradation and can have potential biotechnological applications.
The manuscript describes production, biophysical, structural and biochemical characterization of a set of recombinant AA7 enzymes from different clades within AA7 family. Tow crystal structures were solved and one enzyme was studied as an enzymatic tool to activate LPMOs. This is a nicely written and comprehensive manuscript, which can be further improved by following suggestions presented below: