**4. Vanadium-Dependent Haloperoxidases**

For several years after the discovery of heme-iron-dependent haloperoxidases, it was assumed that they are the only enzymes able to oxidize halides for the subsequent halogenation reaction. However, a new halogenating enzyme class was discovered in 1993 by van Schijndel et al. from *Curvularia inaequalis* using *ortho*-vanadate cofactor for the oxidation of halides [70,71]. Just two years later, a vanadate-dependent homolog from *Corallina o*ffi*cinalis* was crystallized [72]. These vanadium-dependent haloperoxidases became a popular research target as they were shown to exhibit high turnover numbers without suffering an oxidative inactivation and displaying a higher tolerance against hydrogen peroxide [73], In contrast to the heme-iron-dependent ones, however, they usually do not retain the formed hypohalous acid within the active site, leading to a freely diffusible strong oxidant. Resulting from this mechanistic aspect, random halogenations occur, even in the protein itself, leading to its destabilization and inactivation. Because of this free hypohalous species, the selectivity of the subsequent halogenation reaction is independent of the enzyme but from the electronic properties of the substrate. Most of the vanadium-dependent haloperoxidases originate from marine fungi and marine macroalgae (seaweeds) [74].

It is proposed that the catalytic cycle (Figure 9) forms a VV-peroxo-species as the key intermediate, where the halide is added and subsequently hydrolyzed to hypohalous acid. Identically to

heme-iron-dependent haloperoxidases, the presence of hydrogen peroxide may lead to the disproportion to singlet oxygen and the halide [55].

**Figure 9.** Proposed catalytic cycle of vanadium-dependent haloperoxidases. In its resting state (3 'o clock), vanadium contains four oxygen ligands, while the free coordination site is occupied by a catalytic histidine residue, resulting in a dative bond. In presence of hydrogen peroxide, a hydroxyl group is substituted by peroxide. Upon elimination of a hydroxide ion, a cycloperoxo-species is generated, which is stabilized by a catalytic lysine residue. This cyclic intermediate is opened by addition of a halide, in this case bromide, which can then be hydrolyzed by water, leading to the release of hypobromic acid, or in presence of another hydrogen peroxide molecule, be disproportioned to molecular oxygen and bromide. During catalysis, the vanadium does not alter its oxidation state (V) [55].

One of the best-investigated representatives of this class is the vanadium-dependent chloroperoxidase from the phytopathogenic fungus *Curvularia inaequalis* [70,71,75–77]. Even in absence of the vanadium-cofactor, the enzyme is stable in its *apo*-form and can easily be transformed to the holo-form by external addition of *ortho*-vanadate [70]. Although the gene can be heterologously expressed in *E. coli* and activated with vanadate, it was reported that the amount of enzyme obtained was very low. As an alternative, *Saccharomyces cerevisiae* was used as a host, yielding 100 mg/l *apo*-enzyme [75]. Kinetic experiments lead to a kcat/KM of 2.6 <sup>×</sup> <sup>10</sup><sup>6</sup> <sup>m</sup>−<sup>1</sup> <sup>s</sup>−<sup>1</sup> for hydrogen peroxide and 5.1 <sup>×</sup> <sup>10</sup><sup>7</sup> <sup>m</sup>−<sup>1</sup> <sup>s</sup>−<sup>1</sup> for bromide at pH 4.2, the optimal pH for bromoperoxidase activity [75].

It showed stability at high temperatures (TM of 90 ◦C) and tolerance against organic solvents like methanol, ethanol, and propan-2-ol (up to 40% *v*/*v*) [71]. *Ci*-VClPO was used as a hypohalogenite catalyst for the halogenation of phenols like thymol, while showing excellent stability towards hydrogen peroxide and organic solvents like methanol and ethyl acetate [76]. Furthermore, it was used for the mediation of *(Aza-)Achmatowicz* reactions in combination with cascades [78] and halofunctionalization reactions of aromatic and aliphatic alkenes like styrene and hexanol [77,79] (see Figure 10).

**Figure 10.** (**A**): *(Aza-)Achmatowicz* reaction transforming the furan **15** to the Michael-system [78] **14**. (**B**): Halofunctionalization of styrene (**16**) [77,79].

In contrast to the usually scarce selection of vanadium-dependent chloroperoxidases, many representatives of bromoperoxidases were researched in the past. One of the most prominent members of this group is the VBrPO from *Corralina o*ffi*cinalis*, a marine red algae. Similarly to the homolog from *C. inaequalis*, it excels with a high stability towards high temperatures up to 90 ◦C and in presence of organic solvents like ethanol, propanol, and acetone (up to 40% *v*/*v*) [72]. However, recombinant expression of the gene in *E. coli* BL21(DE3) proved difficult, as the amount of protein formed is high, but insoluble. Coupe et al. notably showed that by using a refolding procedure, 40 mg/L of active enzyme can be retrieved after expression and isolation [80].

The haloperoxidase was shown to accept a variety of substrates, like nitrogen-containing heterocycles, cyclic β-diketones, phenol, *o*-hydroxybenzyl alcohols, anisole (**19**), 1-methoxynaphthalene and thiophene in addition to alkene halogenations with styrene (**16**), cyclohexene (**22**) among others to yield various bromohydrins [81] (see Figure 11 and Table 2).

In most of the cases, no diastereoselectivity for the bromohydrin formation could be observed, except for the formation of bromohydrin from (*E*)-4-phenyl-buten-2-ol (**24**) [69]. Besides bromination reactions, haloperoxidases like the *Co*-VBrPO are able to catalyze sulfoxidations with 2,3-dihydrobenzothiopene (**26**), as well [82].


**Table 2.** Enzymological properties of vanadium-dependent haloperoxidases.

**Figure 11.** Selected reactions performed by the *Co*-VBrPO to illustrate the reaction spectrum [81].

#### **5. Metal-Free Haloperoxidases**/**Perhydrolases**

Although oxidative halogenation reactions are dominated by (transition) metal catalysis in nature, a group of enzymes was identified catalyzing halogenation without any metal cofactor. These metal-free haloperoxidases or perhydrolases were found to require hydrogen peroxide and halides as well, while forming percarboxylic acids from carboxylic acids using a catalytic triad of serine, histidine, and aspartate [84,85]. Their striking resemblance to lipases has initiated a general debate over the nature of these enzymes, as their characteristics resemble hydrolases with a halogenating sub-activity. This has led to controversies whether the metal-free haloperoxidases are not simply lipase-like enzymes moonshining as haloperoxidases. In fact, several lipases were tested positively for haloperoxidase activity despite low turnover numbers [81].

The key-step in catalysis is the formation of a peroxo-acid from a carboxylic acid by hydrogen peroxide, which subsequently forms an acylhypohalide acting as the halogenating agent (Figure 12) [86].

Many examples for metal-free HPOs in biotransformations are not known. The majority of investigations of this enzyme class were focused on determining and expanding the tolerance of these enzymes to organic solvents and temperatures. One recent example of a bioorganic application was the halogenation of nucleobases and analogues [87] (see Figure 13).

**Figure 12.** Proposed catalytic cycle of metal-free haloperoxidases/perhydrolases. This mechanism was compiled from several sources [81,85]. The catalytic cycle is adopted from the common hydrolase catalysis encountered in lipases and esterases, for instance. In presence of a carboxylic acid, in this case acetic acid, an ester is formed with the catalytic serin residue upon elimination of water (3 'o clock). In presence of hydrogen peroxide, the ester is cleaved, forming a percarboxylic acid. In the following step, a halide binds to the peroxoacid, which is hydrolyzed to the hypohalous acid, while the characteristic Ser-His-Asp triad is already regenerated.

**Figure 13.** Halogenation of indole (**28**) nucleobase analogs according to Lewkowicz and co-workers [87].

## *5.1. Flavin-Dependent Halogenases*

In addition to the long-known haloperoxidases, another class of enzymes has aroused much interest. It is suspected that flavin-dependent halogenases (FHals, Fl-Hals, or FDHs) evolved from monooxygenases that require flavin cofactors as well and, therefore, belong to the superfamily of flavin-dependent monooxygenases [88,89].

According to what is known so far, there are three natural target structures that can be addressed by FHals. The most studied and best understood group are the flavin-dependent tryptophan halogenases. In nature, there is the possibility to halogenate every position of the indole ring. Similar to this structure there is the group of flavin-dependent pyrrole halogenases and finally the flavin-dependent phenol halogenases (see Figure 14) [8]. The fact, that each and every position can be addressed by an individual enzyme demonstrates that FHals are selective halogenating catalysts in contrast to the majority of haloperoxidases. FHals must be differentiated according to the accessibility of their substrates. While a large number of these halogenases are involved in biosynthesis clusters of polyketides (PKS) and non-ribosomal protein synthesis (NRPS), some, such as tryptophan halogenases, can convert freely diffusible substrates and are not dependent on carrier proteins that activate or merely tether the substrate (Figure 14) [13].

For the application of this enzyme group, it is important to keep in mind that they need at least a two-component electron transport chain and therefore require a suitable flavin reductase [90–92]. In addition to the reductases that naturally belong to the biosynthesis clusters e.g., PrnF [93], applications with other reductases such as SsuE [91,94,95] or Fre [96,97] from *E. coli* have also been reported. To avoid the necessity of a second enzyme—the flavin reductase—or even a third enzyme for cofactor recycling, photochemical approaches are in the focus of current research in this field as well [98].

**Figure 14.** Regioselectivity of flavin-dependent halogenases and their dependency on carrier proteins. \* Natural products with halogenations are known, but so far, no enzyme is characterized. <sup>ˆ</sup> This tryptophan halogenase is one of the few examples that is carrier protein-dependent [99].

Figure 14 shows some representatives for the halogenation of the different positions of the different substrates (indoles [92,95,100–102], pyrroles [103] and phenols [104–107]), each with reference to the halogenating enzyme, the dependence on carrier proteins and the corresponding publication [108]. The halogenation of position four of indoles, as for example in 4-chloroindole-3-acetic acid, is known to date only from plants (*Pisum sativum*, *Lens culinaris*, *Vicia* sp., and in particular *Vicia faba*), as a growth hormone but no enzyme has yet been characterized responsible for its formation [7]. The publications

e.g., by Shepherd et al., the review of Latham et al. and other publications also show various mutants that led to changes in regioselectivity and substrate scope [8,96,100,109,110].

A lot of these enzymes that are dependent on carrier proteins produce well-known secondary metabolites like rebeccamycin (**4**) and vancomycin (**1**) but also a plethora of less investigated biosynthetic pathways [133]. The most important difference in the mechanism between flavin-dependent monooxygenases and halogenases is the conserved motif of two tryptophanes, one isoleucine and proline. This 10 Å long tunnel [89], first found in PrnA, serves to spatially separate the activated peroxy-flavin FAD(C4α)–OOH from the substrate binding site and thus prevents oxygenation [114,116,128]. After generating the hypohalides, a conserved lysine transfers the electrophilic chlorine as chloramine from the former peroxy-flavin to the substrate (Figure 15) [134].

**Figure 15.** Catalytic cycle of halogenation by flavin-dependent halogenases [96,115].

For the phenol halogenase the mechanism is proposed to be slightly altered. The phenolic hydroxyl group is deprotonated by an aspartate within the active site increasing the nucleophilicity of the enol α-carbon [130]. Based on these conserved motifs and the assumed reaction mechanism some putative halogenases have already been found and annotated. Recently even a viral halogenase VirX1 from cyanophages was discovered, which is the first FHal capable of in vitro iodination and stands out due to its broad substrate spectrum and preferred iodination [135].

Although the community has so far agreed that the preserved motif of separation tunnel and anchor lysine seems to be essential for the activity, current research shows a further class of flavin-dependent halogenases which lack these structural elements completely. One of these examples is the halogenase KerK that is under investigation by Piel and coworkers but has not yet been published except as a poster presentation on Biotrans 2019 in Groningen, the Netherlands [136].

In addition to the advantages of high regioselectivity and thus only few by-products, there are also some disadvantages in the use of this enzyme group. The low conversion rates speak against large-scale application and expression problems often occur. Many of the proteins produced in *E. coli* BL21(DE3) end up in the insoluble fraction as inclusion bodies. To deal with this issue, strains with co-expression of chaperones are used regularly (Table 3). The overall stability of these proteins also needs further optimization to be applicable in biocatalysis. As a promising result Kemker et al. the tryptophan halogenases were successfully scaled up in terms of a biocatalytic process employing immobilizing the enzymes by cross-linked enzyme aggregates (CLEAs). This yielded l-7-bromotryptophan on the gram scale [137].


**Table 3.** Examples of flavin-dependent tryptophan, pyrrole and phenol halogenases that can be carrier-dependent or independent.


**Table 3.** *Cont.*


**Table 3.** *Cont.*

<sup>1</sup> if not stated otherwise, the expression took place in the origin host.
