Comprehensive insights into the production of long chain aliphatic aldehydes using a copper-radical alcohol oxidase as biocatalyst

1 Comprehensive insights into the production of long chain aliphatic aldehydes using a copper-radical alcohol oxidase as biocatalyst David Ribeaucourt1,2,3, Bastien Bissaro1, Victor Guallar4,5, Mehdi Yemloul2, Mireille Haon1, Sacha Grisel1, Véronique Alphand2, Harry Brumer6, Fanny Lambert3, Jean-Guy Berrin1* and Mickael Lafond2* 1INRAE, Aix Marseille Univ, UMR1163 Biodiversité et Biotechnologie Fongiques, 13009, Marseille, France 2 Aix Marseille Univ, CNRS, Centrale Marseille, iSm2, Marseille, France 3 V. Mane Fils, 620 route de Grasse, 06620 Le Bar sur Loup, France 4Joint BSC-CRG-IRB Research Program in Computational Biology, Barcelona Supercomputing Center, Jordi Girona 29, E-08034 Barcelona, Spain 5Institució Catalana de Recerca i Estudis Avançats (ICREA), 08010 Barcelona, Spain 6Michael Smith Laboratories and Department of Chemistry, University of British Columbia, 2185 East Mall, Vancouver, BC, V6T 1Z4, Canada * corresponding authors: Jean-Guy Berrin ([email protected]) Mickael Lafond (mi[email protected]) “This document is the unedited Author’s version of a Submitted Work that was subsequently accepted for publication in ACS Sustainable Chemistry & Engineering , copyright © American Chemical Society after peer review. To access the final edited and published work see https://pubs.acs.org/doi/10.1021/acssuschemeng.0c07406 2 ABSTRACT The oxidation of alcohols is a cornerstone reaction in chemistry, notably in the flavors and 1 fragrances industry where long chain aliphatic aldehydes are major odorant compounds. In a 2 context where greener alternatives are sought after, biocatalysis holds many promises. Here, we 3 investigated the ability of the alcohol oxidase from Colletotrichum graminicola (CgrAlcOx) – an 4 organic cofactor-free enzyme belonging to the copper-radical oxidases (CROs) class – to convert 5 industrially-relevant long chain aliphatic alcohols. CgrAlcOx is a competent catalyst for the 6 conversion of octan-1-ol, when supported by the accessory enzymes peroxidase and catalase. 7 Detailed examination of the products revealed the occurrence of an overoxidation step leading to 8 the production of carboxylic acid for some aliphatic aldehydes and benzaldehyde derivatives. The 9 partition between aldehyde and acid products varied upon substrate properties (chain length and 10 propensity to form geminal-diols), enzyme specificity, and could be tuned by controlling the 11 reaction conditions. In silico analyses suggested an inhibitory binding mode of long chain aliphatic 12 geminal-diols and a substrate-induced fit mechanism for a benzyl alcohol-derivative. By 13 demonstrating their natural ability to perform long chain aliphatic alcohol oxidation, the present 14 study establishes the potential of fungal CRO-AlcOx as promising candidates for the green 15 production of flavors and fragrances compounds. 16 17 18 KEYWORDS: Biocatalysts, Copper-Radical Oxidases, Alcohol-Oxidases, Long Chain Aliphatic 19 Alcohols, Fragrant Aldehydes. 20 21 22 23 3 INTRODUCTION 24 25 The oxidation of alcohols to aldehydes is a major reaction in the fine chemical industry1–3. 26 Aldehydes are key intermediates for organic synthesis in applications such as pharmaceuticals or 27 alkene synthesis4, but also valuable final products such as flavors and fragrances ingredients5–7. 28 Traditional chemical processes for the oxidation of alcohols usually entail the use of toxic catalysts 29 such as chromium VI8, hence calling for the development of eco-friendly alternatives. Yet, many 30 challenges obstruct the different biocatalytic paths that can be envisioned, especially for the water 31 insoluble and poorly reactive unactivated long chain aliphatic alcohols9,10. Amongst the broad class 32 of aldehydes, long chain aliphatic aldehydes from C6 to C1311–13 are of main importance for the 33 flavor and fragrance industry. Indeed, such aldehydes have been identified as one of the most 34 prominent classes of “key food odorants”11, and are also major fragrance ingredients, used in 35 quasi-all types of perfumes6. Typically, long chain aliphatic aldehydes provide green, fruity, fresh, 36 citrus-like, fatty, or the so-called aldehydic notes14,15. Long chain aliphatic aldehydes can be 37 enzymatically obtained via reduction of their acid counterpart using carboxylic acid reductases 38 (CARs - EC 1.2.1.30). However, CARs are intracellular FAD-dependent enzymes, requiring ATP 39 supply and NADP/NADPH recycling systems16, which renders their use hardly compatible with 40 industrial constraints. Alternatively, long chain aliphatic aldehydes can be obtained via the 41 oxidation of the corresponding alcohols but only a handful of long chain alcohol oxidoreductases 42 have hitherto been discovered, characterized and engineered for this purpose12,17,18. These 43 oxidoreductases include NAD(P)+-dependent alcohol dehydrogenases (ADHs - EC 1.1.1.1) and 44 flavin-dependent alcohol oxidases (FAD-AOXs - EC 1.1.3.13)19. ADHs are well-established 45 biocatalysts despite the reversible, unfavorable and nicotinamide-dependent nature of the alcohol 46 4 oxidation reaction they catalyze. On the other hand, FAD-AOXs offer irreversible oxidation of 47 alcohols with the aid of molecular O2 and a flavin cofactor tightly bound to the enzyme17,20. 48 49 Copper radical oxidases (CROs) belonging to the Auxiliary Activity Family 5 subfamily 2 50 (AA5_2) – according to the CAZy classification21,22 (www.cazy.org) – represent a promising 51 alternative to these two systems. They are organic cofactor free enzymes bearing two redox 52 centers: a copper ion and a 3′-(S-cysteinyl)-tyrosine (Cys-Tyr) free radical23. They catalyze the 53 oxidation of alcohols to aldehyde with the concomitant reduction of O2 to H2O2 (Scheme 1). They 54 are often use in conjugation with catalase – to remove deleterious H2O2 – and peroxidase (e.g. 55 horseradish peroxidase – HRP) for their activation24. For many years, the only characterized 56 member from this family was the canonical galactose 6-oxidase (EC 1.1.3.9) from Fusarium 57 graminearum (FgrGalOx)25. Recently, a new type of alcohol oxidases (CRO-AlcOx), was found 58 in this family. These enzymes, from Colletotrichum graminicola (CgrAlcOx) & C. gloeosporoides 59 (CglAlcOx), were described as competent aromaticand aliphaticprimary alcohol oxidases26. 60 Two homologues from C. higginsianum and Magnaporthe oryzae – anamorph Pyricularia oryzae 61 – have also been described, but only tested on short-chain aliphatic-alcohols27. A paralogous 62 enzyme from C. graminicola (CgrAAO) was recently reported to be highly active on aromatic 63 alcohols (EC 1.1.3.7) and 5-hydroxymethylfurfural (HMF, EC 1.1.3.47), but lacked activity on 64 long chain aliphatic alcohols28. Overall, the broad substrate scope covered by these recently 65 characterized fungal enzymes highlights the catalytic potential within the AA5_2 protein family. 66 A striking example of such catalytic promiscuity was recently unveiled in a variant of the 67 FgrGalOx, unlocking production of nitriles from alcohols in presence of ammonia29. Yet, despite 68 their intrinsic and unique biocatalytic abilities, and in contrast to ADH and AOX systems19 or the 69 5 archetypal FgrGalOx and its engineered variants that have been harnessed for multiple 70 applications24,30–37, the CRO-AlcOx have hitherto received little attention as biocatalysts for the 71 oxidation of industrially relevant alcohols. 72 73 Scheme 1: Reaction scheme of alcohol oxidation to aldehyde by CROs from the AA5_2 subfamily. 74 The main states of the two redox centers (copper ion and Cys-Tyr free radical) are depicted in blue. 75 Accessory enzymes commonly used to activate the CROs (i.e. peroxidase) and to remove deleterious H2O2 76 (i.e. catalase) are shown in orange and green boxes respectively. 77 78 Here, we present a new biocatalytic route for the production of odorant aldehydes, with a focus 79 on industrially-relevant aromatic and long chain aliphatic compounds. We report for the first-time 80 large-scale production of the CgrAlcOx and propose general guidelines for its use as green catalyst 81 for the oxidation of primary long chain aliphatic alcohols. Combining biochemical assays and in 82 silico modelling we provide unprecedented insights into the reaction determinants driving the 83 formation of aldehyde and controlling the subsequent, potential and multifactorial overoxidation 84 into carboxylic acid. 85 86 Catalase Cu 2+ Cys-Tyr • Cu + Cys-Tyr CRO AA5_2 CRO AA5_2 CRO AA5_2 Cu 2+ Cys-Tyr Peroxidase Semi-reduced INACTIVE Reduced ACTIVE Oxidized ACTIVE 6 RESULTS 87 88 Large-scale production of CgrAlcOx and analytical set up for bioconversion. 89 Biotechnological application of enzyme is frequently hampered by low recombinant production 90 yield. As a first step towards a scalable process, we developed larger-scale heterologous production 91 of CgrAlcOx in bioreactor using the yeast Pichia pastoris, which yielded up to 250 mg of purified 92 enzyme per liter of culture (Figure S1). The specific activities of both CgrAlcOx recombinant 93 enzymes produced in flask and in bioreactor were similar (Figure S1D). The recombinant 94 CgrAlcOx produced in bioreactor was further used in all the subsequent experiments described in 95 the manuscript. 96 The analysis of reaction mixtures involving poorly water-soluble long chain aliphatic 97 alcohols/aldehydes calls for the use of alternative methods than the indirect ABTS/HRP coupled 98 assay, routinely used for CROs. To this end, we implemented a gas chromatography (GC)-FID 99 analytical method. The first assays, run with the reference substrate benzyl alcohol (BnOH), 100 confirmed the requirement of the accessory enzymes catalase and/or HRP, to fulfill complete 101 conversion of BnOH38 (Figure S2). Nevertheless and interestingly, we here showed that as little 102 as 5 nM of catalase (Figure S2C) are enough to reach full conversion while the HRP must be added 103 in quasi-stoichiometric amounts (relative to the AlcOx, i.e. in the µM range; Figure S2D). 104 105 CgrAlcOx is a competent catalyst for the full conversion of octan-1-ol. 106 The next step was dedicated to the study of oxidation of octan-1-ol, often used as a model of 107 non-activated primary long chain aliphatic alcohols9. The corresponding aldehyde, i.e. octanal, is 108 a molecule with valuable aroma properties, naturally found in citrus essential oils6,14. As observed 109 for the conversion of BnOH (Figure S2), the conversion of octan-1-ol to octanal did not surpass 110 7 60% with CgrAlcOx only, while almost full consumption of the substrate was reached with the 111 addition of both accessory enzymes (12 µM HRP and 0.5 µM catalase) to the reaction (Figure 1A). 112 Catalase, when added alone, showed only minor enhancement of CgrAlcOx-mediated conversion 113 of octan-1-ol (Figure 1C), while HRP at high concentration (>3.5 µM) exhibited a much more 114 significant effect (Figure 1D). This result is the first report of complete turnover of an aliphatic 115 unactivated primary alcohol by a CRO-AlcOx (with a turnover number – TON – of 3000). Indeed, 116 previous attempts to convert the shorter alcohol butan-1-ol with the same enzyme (CgrAlcOx), 117 but under non-optimized reaction conditions, failed to surpass 30 % conversion (TON = 1413)26. 118 Importantly, the formation of a new product, namely octanoic acid, was monitored and identified 119 when HRP was added at high loading (> 3.5 µM) (Figure 1A, B & D & S3). The presence of acid 120 might indicate an overoxidation process, never highlighted before for any CRO-AlcOx. Of note, 121 this product was not observed for any of the tested catalase concentrations (Figure 1C). Such 122 overoxidation is undesired in the scope of flavors and fragrances inasmuch as the acid generates 123 off-flavors and additional purification steps for isolation of the aldehyde. Understanding and 124 control of this phenomenon is therefore required. 125 8 126 Figure 1. CgrAlcOx-mediated oxidation of octan-1-ol. (A) Oxidation of octan-1-ol by the 127 CgrAlcOx in the presence or absence of HRP (12 µM) and catalase (8 µM). (B) GC-FID 128 chromatograms of reactions catalyzed by CgrAlcOx in the presence or absence of HRP (12 µM) 129 or catalase (8 µM). (C & D) Oxidation of octan-1-ol by the CgrAlcOx with increasing 130 concentrations of catalase and HRP respectively. All reactions were incubated for 16 hours. 131 CgrAlcOx was used at 1 µM with 3 mM octan-1-ol. Error bars show s.d. (independent 132 experiments, n = 3). 133 134 Investigation of the overoxidation process 135 To investigate the mechanism underlying the overoxidation observed in our enzymatic reactions, 136 two hypotheses were probed: 137 H O OH O OH CgrAlcOx OCTANOIC ACID OCTAN-1-OL OCTANAL CgrAlcOx only + HRP & Catalase CgrAlcOx OCTANOIC ACID OCTANAL OCTAN-1-OL A B C D RATIO (%) RATIO (%) [CATALASE] (µM) 0 20 40 60 80 100 0 00.005 0.05 0.5 8 20 40 60 80 100 0.07 0.7 3.5 12 [HRP] (µM) RATIO (%) 0 0 20 40 60 80 100 +HRP +Cat (12 µM) (8 µM) 4 5 6 7 8 9 10 11 9 (1) The aldehyde is overoxidized to the acid by CgrAlcOx via a geminal-diol (gem-diol) 138 intermediate that would act as a secondary substrate (Scheme 2). This oxidation pathway is a 139 plausible route39,40, as notably reported for some ADHs41,42, FAD-AOXs17,43, AA5_2 GalOx44–48 140 and AA5_149 (glyoxal oxidases, GLOX, EC 1.2.3.15) . 141 (2) The aldehyde is overoxidized via a non-enzymatic mechanism or by the accessory enzymes 142 (i.e. HRP and/or catalase) added into the reaction. 143 R C H OH OH R C O H H2O R C O OH ADHs AOXs CROs (AA5_2) 144 Scheme 2. Aldehyde oxidation to carboxylic acid via non-enzymatic hydration followed by 145 enzymatic oxidation of the gem-diol intermediate. 146 In the gem-diol hypothesis, as suggested in a previous study38, the mechanism would be substrate 147 dependent. Indeed, some aldehydes (such as aliphatic aldehydes) are more prone than others to 148 undergo hydration50. For instance, hydration constant (KH) values of 0.75 M-1 vs 0.01 M-1 were 149 reported38 for hexanal and PhCHO, respectively51,52. Of note, KH values for octanal and hexanal 150 are expected to be similar as the increase in carbon chain-length has a minor effect51,53. To probe 151 this first hypothesis, starting with the benchmark substrate BnOH, and on the basis of former 152 studies43,47,54, we firstly chose two BnOH analogues bearing either an electron-withdrawing group 153 (EWG) or an electron-donating group (EDG) to affect the KH value. 1H-NMR analysis confirmed 154 that the gem-diol was formed (24 %) in aqueous conditions only for the aldehyde bearing an EWG 155 (4-nitro-benzaldehyde; 4-NO2-PhCHO) in contrast to 4-methyl-benzaldehyde (4-Me-PhCHO) 156 (Figure 2A). Subsequently, we carried out conversion experiments using these BnOH analogues 157 as substrates, in the presence of CgrAlcOx, HRP and catalase. Quasi-full conversion to the 158 corresponding carboxylic acid was obtained for 4-NO2-BnOH while in the same conditions 4-Me-159 16 whether the enzyme is substrate-free (Figure S11B) or in complex with octan-1-ol (Figure S11C) 254 or BnOH (Figure S11D) while a wider pocket is visible in the case of 4-NO2-BnOH (Figure S11E). 255 Such flexibility in the active site could play a significant role in the ability of CgrAlcOx to accept 256 a broad range of substrates. 257 During the course of PELE modelling experiments, we also probed the recognition by CgrAlcOx 258 of various aliphatic alcohols with chain lengths ranging from C2 to C10. A clear trend stood out: 259 the longer the chain length the lower the binding energy (see below). These simulations prompted 260 us to verify this trend by enzymatic assays. 261 The determination of kinetic parameters (Table 1 & Figure S12) revealed that while kcat values 262 for C7-C10 alcohols remained constant, the Km values showed a slight but constant decrease with 263 the increase in aliphatic chain length, a trend previously observed for shorter alcohols (C2-C7)26. 264 As a comparative control, we verified that the kinetic parameters for C7 alcohol determined here 265 and by Yin and colleagues26 were similar. 266 Table 1. Kinetic parameters of CgrAlcOx for several long chain aliphatic alcohols 267 determined by the spectrophotometric ABTS/HRP coupled assay using 1 nM CgrAlcOx. 268 Substrate k cat (s-1) K m (mM) k cat/ K m (s-1.M-1) Reference Heptan-1-ol 100 ± 1 0.19 ± 0.01 5.3 x 105 Yin et al.26 85.8 ± 1.016 0.19 ± 0.01 4.52 x 105 This work Octan-1-ol 84.4 ± 1.3 0.14 ± 0.01 6.03 x 105 This work Decan-1-ol 87.5 ± 2.2 0.12 ± 0.02 7.29 x 105 This work 269 Attempting to relate experimental kinetic parameters to docking results, a few facts need to be 270 laid out. The characteristic ‘ping-pong’ CRO’s catalytic cycle can be divided into two half-271 17 reactions, a reductive half-reaction that entails enzyme active site reduction upon substrate 272 oxidation, and a subsequent oxidative half-reaction where the enzyme is re-oxidized into its initial 273 state while O2 is reduced into H2O2. Studies performed on the FgrGalOx indicate that the oxidative 274 half-reaction is not rate-limiting for AA5_2s58,59, implying that kcat reflects directly the rate of the 275 reductive half-reaction. Furthermore, the facts that the kcat value is not impacted by the nature of 276 the oxidized substrate (Table 1), and is theoretically independent from the Kd, indicate that once 277 the substrate is bound, the chemical reaction (i.e. CH2-OH → CHO) occurs at the same rate-278 limiting speed. All these considerations allow us to predict that variations in Km and Kd values 279 should be directly correlated, meaning that the observed decrease in Km for longer chain lengths 280 should reflects an increase in affinity. The correlation observed when plotting PELE’s substrate 281 binding energy for C2 to C8, and C10 straight chain saturated primary alcohols against 282 experimental log (Km) values obtained here and previously26 provides support to our simulation 283 results and further suggests that the observed decrease in Km for longer chain lengths may reflects 284 an increase in affinity (Kd) (Figure 6). 285 18 286 Figure 6. Relationship between the average (top quartile) binding energies (C-Cu distances 287 < 4Å) determined by PELE and experimental log (Km) values for C2 to C8, and C10 straight 288 chain saturated primary alcohols with CgrAlcOx. 289 290 Preparative-scale experiment 291 292 To support the industrial applicability of CgrAlcOx catalyst, the knowledge acquired in the 293 present study was harnessed in a scaled-up experiment. To this end, we harvested the supernatant 294 of a CgrAlcOx-producing P. pastoris bioreactor and used it directly as crude enzyme solution. 295 One liter of crude enzyme was mixed to 2 g of octan-1-ol and allowed to react during 30 min to 296 limit overoxidation. After extraction with organic solvent, GC analysis revealed that 0.72 g of 297 octanal, 0.61 g of octan-1-ol and 0.03 g of octanoic acid were recovered. Interestingly, a parallel 298 reaction run with the same concentrations of substrate and accessory enzymes but at 10 mL volume 299 resulted in almost full conversion into aldehyde (92 % of octanal; 4.1 % of octan-1-ol and 3.6 % 300 of octanoic acid), suggesting that the limited rate of conversion observed for the gram-scale 301 experiment might be due to experimental constraints (e.g. stirring, vessel etc.) and should be 302 further optimized. 303 C10 C8 C2 C3 C4 C5C6 C7 19 DISCUSSION 304 305 In the current context, biotechnologies constitute the most promising alternative for the 306 production of long chain aliphatic aldehydes as odorants in a flavors and fragrances market 307 affected by an increasing demand for green and/or natural products that traditional production 308 means cannot provide60,61. We herein demonstrate that a fungal CRO-AlcOx can achieve full 309 conversion of some long chain aliphatic alcohols (i.e. hexan-1-ol and octan-1-ol) to yield the 310 corresponding industrially-relevant aldehydes, within a few minutes, under mild conditions: water 311 as solvent and O2 as co-substrate. To the best of our knowledge, very few studies have reported 312 the efficient conversion of unactivated long chain aliphatic alcohols. A recent contribution to this 313 field indicates that a mutant from a FAD-dependent choline oxidase from Arthrobacter 314 cholorphenolicus exhibited activity on some long chain aliphatic alcohols, with best performance 315 on C4-C7 substrates18. The authors reported for this mutant a kcat/Km value on hexan-1-ol 1000-316 fold lower than for the CgrAlcOx and described full conversion of C6-C9 long chain aliphatic 317 alcohols (10 mM) within 24 hours at 30°C, using 1 mg.mL-1 of enzyme, with an approximative 318 turnover frequency of 7 x 10-3 s-1. Comparatively, the CgrAlcOx was here used at 52 µg.mL-1 to 319 convert C6 and C8 alcohols (3 mM) in 15 minutes at 23°C exhibiting a turnover frequency of 3.3 320 s-1. These promising results prompted us to attempt a gram scale reaction, which lead to the 321 recovery of 0.72 g of octanal using a crude CgrAlcOx from bioreactor supernatant. This first proof 322 of concept highlights the potential of CRO-AlcOx as biocatalysts for the production of long chain 323 aliphatic aldehydes. 324 In this study, we also provide insights into the mechanism of the CgrAlcOx by comparing the 325 oxidation of benzyl alcohol and derivatives and long chain aliphatic alcohols (Figure 7). 326 20 327 328 Figure 7. Overview of the proposed oxidation mechanism by CgrAlcOx of benzyl alcohol (A) 329 and derivatives and long chain aliphatic alcohols (B). 330 It is now well established that recombinant CROs require accessory enzymes to harness their 331 full potential: catalase for protection against H2O2 and HRP for activation, the main resting-state 332 of the enzyme being inactivated62–64. Yet, as reported for GalOx, a minor fraction of the enzyme 333 pool is activated and likely explain the basal level of conversion observed without HRP or catalase, 334 before reaching an inactivated form due to the H2O2 effect or undergoing “off-cycle 335 inactivation”65. Here, results suggest different requirement in accessory enzymes according to the 336 substrate to be oxidized. Indeed, rapid conversions require both HRP and catalase, confirming that 337 the two accessory enzymes are not interchangeable but bring distinct improvement to the 338 reaction38. Long-chain unactivated aliphatic substrates can only be fully converted by providing 339 high amount of HRP coupled to catalytic amount of catalase. This observation is likely related to 340 the inhibition by long chain aliphatic gem-diols resulting from the hydration of aldehydes products 341 R O R H HO OH R OH O R OH +hemin Cgr AlcOx +HRP +Cat Cgr AlcOx +HRP +Cat Cgr AlcOx +HRP +Cat H n OH HO O n OH n6-8 O OH N Cl NN HO OHO O N Fe A B INHIBITION 21 (see discussion below). Although the molecular mechanism of activation of CROs needs to be 342 better understood, it is clear that HRP acts as an oxidizing agent to restore the active form of the 343 enzyme bearing the Cu2+ ion and the crosslinked Cys-Tyr free radical24,48. The stoichiometric 344 proportion of HRP needed suggests a protein-protein interaction rather than a catalytic reaction66. 345 It is worth noting that peroxidases are also routinely used to activate AA5_1 GLOXs67,68 pointing 346 out a common activation mechanism of CROs. 347 The present work sheds light on the overoxidation reaction occurring during AlcOx catalysis. 348 This reaction seems directly linked to the propensity of the produced aldehydes to form gem-diols 349 upon hydration, which can be used as substrates in a late subsequent reaction, yielding the 350 corresponding carboxylic acids. Gem-diols exist in equilibrium with their aldehyde counterparts 351 but are often unfavored unless the carbonyl function is destabilized. For instance, electron-352 attracting substituents, such as NO2, are known to disturb the dipole at the carbonyl group and 353 thence, foster the nucleophilic addition of H2O52. We show here that benzylic-aldehydes, when 354 prone to form gem-diols (i.e. 4-NO2-PhCHO), are oxidized into carboxylic acid by the AlcOx 355 whereas, when stabilized in their carbonyl form (i.e. 4-MePhCHO), are not further processed by 356 the enzyme. However, the latter condition is not necessarily sufficient per se, since the presence 357 of the substituting group may also affect the intrinsic reactivity of the carbon undergoing catalysis. 358 Long chain aliphatic aldehydes are also susceptible to form gem-diols. However, our results 359 suggest that these gem-diols are not oxidized by the AlcOx, and likely yield inhibitory binding at 360 the active site of the enzyme. Yet, partial oxidation to the carboxylic acid, promoted by 361 HRP/hemin, has been observed for >C6 aldehydes and seems therefore independent from 362 CgrAlcOx. This might account for a phenomenon related to the so-called “autoxidation”, a radical 363 chain process mediated by oxygen, and favored by a large number of catalysts, hemin (free or 364 22 bound to HRP) being likely one of them69,70. A catalytic action of HRP is unlikely in the 365 conversions we performed with CgrAlcOx and octan-1-ol, as H2O2 should be dismutated in-situ 366 by the catalase. It is however difficult to evaluate the amount of H2O2 that would potentially 367 accumulate in the reaction due to multiple crossed-production and consumption fluxes. 368 Other AA5_2 CROs have shown evidence of gem-diol-dependent overoxidation such as the 369 GalOx46,47 or the raffinose-specific galactose oxidases from C. graminicola71 and Penicillium 370 rubens72. Nevertheless, not all the CROs appear to exhibit this overoxidation phenomenon, such 371 as the CgrAAO that selectively oxidizes the primary alcohol function of HMF yielding as single 372 end product the bis-aldehyde compound 2,5-diformylfuran despite the presence of gem-diol28. The 373 formation of carboxylic acid can be either an unwanted product in the case of flavors and 374 fragrances, or a sought after compound for applications in synthetic organic chemistry73. Fine 375 tuning of reaction conditions (i.e. reaction length, substrate choice, quantity of accessory enzymes 376 added, pH) allows steering the reaction towards either aldehyde or carboxylic acid. In this regard, 377 the CgrAlcOx could be an efficient “all-in one” catalyst for direct oxidation of some benzylic-378 alcohols or other aromatic alcohols prone to form gem-diols to carboxylic acids. In contrast, 379 formation of long chain aliphatic gem-diols should be avoided to prevent inhibition. 380 Building on a previously exposed hypothesis38, we showed here that the conversion yield of 381 unactivated aliphatic alcohol is limited by an inhibition phenomenon that most likely involves the 382 corresponding aliphatic gem-diols present in the reaction. Computational studies further strengthen 383 this assumption, showing a possible inhibitory binding mode at the active site of the enzyme with 384 octanal gem-diol. Such hypothesis is consistent with the fact that (i) larger amounts of accessory 385 enzymes are required to convert efficiently octan-1-ol, since these enzymes activate the CgrAlcOx 386 and probably allow to decrease the amount of inhibitory aldehyde or gem-diol (via conversion into 387 23 acid) and that (ii) CgrAlcOx is unable to oxidize aliphatic gem-diols into acids. The inhibition 388 phenomenon seems common to all unactivated aliphatic alcohol as also reported previously for 389 butan-1-ol and glycerol26. 390 Docking studies also highlighted a possible substrate-induced fit for 4-NO2-BnOH into 391 CgrAlcOx active site as shown by a local twist of the side chain of key residues which could 392 explain the strong substrate tolerance observed for this enzyme. In addition, despite 60 years of 393 intensive research on this class of enzymes, almost exclusively focused on the FgrGalOx, such 394 flexibility at the active site of AA5_2s has not been reported and could contribute to the 395 unexplained inability to obtain crystal structures of CRO-substrate complexes28. 396 397 Conclusion 398 Notwithstanding their intrinsic potential, CRO-AlcOx have been underexploited for alcohol 399 oxidation, while efforts were focused on the carbohydrate-active FgrGalOx and mutants thereof19. 400 Our study strives to unveil the potential of these promising catalysts and to outline the framework 401 and boundaries of its scope of application. We herein provide guidelines for controlled oxidation 402 of some long chain aliphatic and aromatic alcohols in the context of the flavors and fragrances 403 market. The present work paves the way for scale-up experiments encouraged by high-yield 404 protein production in bioreactor and mild-condition catalysis achieved with the CgrAlcOx 405 biocatalyst. Engineering studies – sustained by a better understanding of the substrate-dependent 406 oxidation initiated here – and process development such as immobilization and biphasic systems 407 could foster the use of CRO-AlcOx for industrial-scale catalysis. 408 24 AUTHOR CONTRIBUTIONS 409 410 DR carried out most of the experimental work. BB provided guidance in the experimental work. 411 DR and BB interpreted the data and wrote the manuscript; VG performed the in silico experiments, 412 interpreted the corresponding results and was involved in the manuscript writing; MY drove the 413 NMR experiments and the corresponding result interpretations, and was involved in the manuscript 414 writing; MH and SG were involved in enzymes productions; VA was involved in the design of the 415 gas chromatography experiments; HB and FL were involved in the study design and manuscript 416 writing. JGB and ML conceptualized the study, supervised the work, and finalized the manuscript. 417 All authors approved the final version of the manuscript. 418 419 420 ACKNOWLEDGMENT 421 This study was supported by the French National Agency for Research (“Agence Nationale de 422 la Recherche”) through the “Projet de Recherche Collaboratif International” ANR-NSERC 423 (FUNTASTIC project, ANR-17-CE07-0047), by the Natural Sciences and Engineering Research 424 Council of Canada through the "Strategic Projects - Natural Resources and Energy - Project 425 (STPGP) entitled "FUNTASTIC - Fungal copper-radical oxidases as new biocatalysts for the 426 valorization of biomass carbohydrates and alcohols" and by the PID2019-106370RB-I00 grant 427 from the Spanish Ministry of Innovation and Sciences. We are grateful to MANE & Fils and the 428 “Association Nationale Recherche Technologie” (ANRT) for funding the PhD fellowship of D.R. 429 entitled “Discovery and structure-function study of new fungal copper radicals oxidases used as 430 biocatalysts for the valorisation of alcohols and plant biomass.” This Convention Industrielle de 431 Formation par la RecherchE (CIFRE) grant no. 2017/1169 runs from 1 April 2018 to 1 April 2021. 432 25 COMPETING INTEREST 433 The authors declare that they have no competing interests. 434 435 436 ASSOCIATED CONTENT Supporting information Experimental section; detailed results of scaled-up expression in bioreactor of CgrAlcOx; GC-FID results of benzyl alcohol, hexan-1-ol and octan-1-ol conversion by CgrAlcOx; GCFID chromatograms of chemical standards and control reactions; 1 H-NMR spectra of hexanal and octanal in D 2 O and quantification of gem-diols thereof; quantification of aldehydes oxidation to carboxylic acids in various control conditions; PELE’s energy plots for benzyl alcohol and benzaldehyde docking into CgrAlcOx active site; CgrAlcOx inhibition assay; computational models of CgrAlcOx in complex with octan-1-ol, 4-NO 2 -BnOH and BnOH; time-course oxidation of heptan-1-ol, octan-1-ol and decan-1-ol by CgrAlcOx used for determination of kinetic parameters; effect of acetone on CgrAlcOx activity; oven programs for GC-analyses; supplementary references. 32 FOR TABLE OF CONTENTS USE ONLY 671 672 673 674 675 676 SYNOPSIS 677 Establishing Copper-Radical Oxidases as promising biocatalysts for the sustainable production of 678 natural aliphatic aldehydes for flavors and fragrances industry. 679