**3. Mechanism of Action of Benzothiazinones**

An investigation of mechanism of action of BTZ was carried out within the European Union (EU)-funded consortium New Medicines for Tuberculosis (NM4TB), which was formed in 2006 under the leadership of Prof. Stewart Cole. The compound BTZ038 was initially chosen as a lead molecule. Since it contains a chiral center, pure enantiomers BTZ043 (S) and BTZ044 (R) were prepared and subjected to MIC determination. Both forms showed comparable activities in vitro in *M. tuberculosis*. Accordingly, only one of them, BTZ043, was subjected to mechanistic studies [9].

Genetic experiments aimed at identification of the BTZ target were carried out by Prof. Giovanna Riccardi and her team at University of Pavia. The two approaches which were chosen for this purpose pointed out to the role of the gene *rv3790* in conferring resistance against the drug. In the first approach, *M. smegmatis* was transformed with a cosmid library containing DNA fragments from the same organism. An ortholog of the *rv3790* gene from *M. smegmatis* mc2155, *MSMEG\_6382*, was identified on a cosmid, which resulted in an increase in MICs towards BTZ043. In the second approach, BTZ-resistant strains of *M. smegmatis*, *M. bovis* BCG and *M. tuberculosis* were isolated and their resistance was traced to mutations in the same codon in *rv3790* and its orthologs. In all mutant strains codons corresponding to Cys387 in *rv3790* gene product were replaced by Ser or Gly, which increased the MICs of the resistant strains up to 10,000 times. Accordingly, the mycobacterial species *M. avium* and *M. aurum*, in which this position was occupied by Ala or Ser, respectively, were naturally resistant against BTZ, which supported identification of *rv3790* as a target [9]. Coincidentally, the function of the protein encoded by this gene was described only a short while before the data on the BTZ target became available. In the course of studies focused on the metabolism of arabinose in mycobacteria carried out by Prof. Michael McNeil with collaborators at Colorado State University in Fort Collins, the *rv3790* and *rv3791* gene products were proposed to be involved in a unique epimerisation of decaprenylphosphoryl ribose (DPR) to decaprenylphosphoryl arabinose (DPA) [14]. The latter molecule serves as a sole donor of arabinose residues for synthesis of the essential arabinan polymers of the mycobacterial cell wall [15]. It was shown that the conversion takes place via oxidation of the DPR substrate, followed by reduction of the resulting keto-intermediate and thus the two oxidoreductases, *rv3790* and *rv3791*, from the so-called arabinogalactan biosynthesis cluster [16] were ideal candidates to catalyse these reactions. However, at the time it was not possible to conclude which of the two is responsible for oxidation and which for the reduction step. Examination of the effects of BTZ043 on enzymatic conversion of DPR to DPA proved that the compound inhibited the oxidation step. Given the confirmed roles of Rv3790 and Rv3791 in epimerization of DPR, we named them DprE1 and DprE2, respectively (Figure 2) [9].

An important step towards understanding the mechanism of action of BTZs was isolation of a covalent adduct of BTZ043 and DprE1 by Trefzer et al. in 2010 [17]. The authors concluded that reduction of the nitro-group to nitroso derivative is critical for formation of the semimercaptal adduct of BTZ043 and DprE1 through the Cys387, which was verified by mass spectrometry. The following study proved that DprE1 itself is responsible for the reductive activation of DprE1. The electrons needed for the reaction are provided by the enzyme's prosthetic group, FAD, which gets reduced during the catalysis of DPR oxidation (Figure 2) [18]. Accordingly, BTZ represents a typical example of a suicide inhibitor. The proposed mechanism of action was confirmed by crystallographic studies in which structures of DprE1 from *M. smegmatis* were obtained in native and BTZ043-bound forms. As expected, the semimercaptal adduct was formed only in the presence of the substrate analog, farnesylphosphoryl ribose, which provided electrons for reduction of the FAD cofactor necessary for BTZ activation [19].

**Figure 2.** Mechanism of action of macozinone. *Upper line:* Reactions catalyzed by DprE1 and DprE2. *Bottom line:* Activation of macozinone and its covalent attachment to DprE1 from *M. tuberculosis* through Cys387 of the enzyme's active site. The scheme is based on mechanistic studies performed initially with benzothiazinone BTZ043 [17–19] and subsequently confirmed for piperazine-benzothiazinone PBTZ169 (macozinone) [20].

#### **4. From BTZ043 to PBTZ169**

At the time of its discovery, BTZ043 was the most efficient inhibitor of in vitro mycobacterial growth ever described. While MICs for first line drugs isoniazid or ethambutol for *M. tuberculosis* were 0.02–0.2 μg/ml or 1–5 μg/ml, respectively, the value for BTZ043 was just 1 ng/ml. The compound proved to be active also in intracellular *M. tuberculosis* infection in Raw 264.7 macrophages and in a mouse model of chronic tuberculosis and it did not show any signs of toxicity or mutagenicity in concentrations well above the amounts used for the experiments. Nevertheless, the extremely low MIC did not translate to comparably high efficiencies in animal studies [9].

In the effort to improve pharmacological properties of benzothiazinone series, structure-activity relationship (SAR) analysis was performed, with main focus on position 2 of BTZ scaffold [20]. It was previously shown that the rest of the positions harbor substituents critical for the antimycobacterial activity, e.g., replacement of 8-nitro-group with any other group, including hydroxylamino-, aminoand nitryl-, led to at least a 500-fold increase in the MIC against *M. tuberculosis* [9]. The new series contained a piperazine at the position 2 of the BTZ structure, which allowed extensive modifications of the N-4 position of this substituent. Evaluation of the 60 prepared piperazine-benzothiazinones (PBTZ) for in vitro activity on *M. tuberculosis* pointed to a strong correlation between the MICs and lipophilicity. The alkyl-PBTZ derivatives were most active and some of them reached the MICs between 0.19–0.75 ng/ml for *M. tuberculosis*. Five of these derivatives were tested in the chronic TB mouse model, which allowed for selection of PBTZ169 as the most promising candidate [20].

Comparison of BTZ043 with PBTZ169 confirmed superior properties of the latter in all tested parameters. It was three to seven times more active in vitro against *M. tuberculosis*, *M. bovis* BCG, *M. marinum*, *M. smegmatis*, *Corynebacterium diphtheriae* and *C. glutamicum* as established by resazurin microtiter assay (REMA). The molecule was also highly active against a panel of 9 MDR- and XDR-clinical isolates of *M. tuberculosis* [20]. Cross-resistance between BTZ043 and PBTZ169 was confirmed for BTZ-resistant strains of *M. tuberculosis*, *M. bovis* BCG and *M. smegmatis*, indicating the common mechanism of action. Interestingly, PBTZ169 was less susceptible to inactivation by nitro-reduction by the nitroreductase NfnB from *M. smegmatis* compared to BTZ043, as confirmed by the liquid chromatography-mass spectrometry analyses. This property could be critical for in vivo activity, due to inevitable actions of host nitroreductases or microbial nitroreductases in the digestive tract of the patients, which would potentially reduce the amount of the available active drug. Enzymologic studies confirmed that PBTZ169 is more efficient inhibitor of its DprE1 target than BTZ043; the enzyme was completely inactivated after 5 min of incubation with 5 μM PBTZ169, while four times higher concentration was needed to achieve the same result with BTZ043. The crystal structure of

*M. tuberculosis* DprE1 with covalently bound PBTZ169 showed similarities with the BTZ043-bound structure of the *M. smegmatis* enzyme. However, higher flexibility of the methyl-cyclohexyl group on piperazine, could account for better affinity of this drug for its target and explain its higher capacity to inactivate the enzyme [20].

Efficacies of PBTZ169 and BTZ043 were also evaluated against *M. marinum* in a zebrafish embryo model. Although after 5 days of treatment both drugs efficiently decreased the bacterial burden at concentrations of 25 nM and 50 nM, examination of BTZ043-treated embryos pointed to developmental defects. By contrast, application of PBTZ169 up to 10 μM did not cause any pathological changes. In vivo studies in the mouse chronic model of TB proved the superiority of PBTZ169 over BTZ043 in lowering colony forming unit counts in the spleen and lungs, which, however, could not be attributed to better pharmacokinetics, since the two compounds had similar properties except for faster uptake of PBTZ169 [20]. Aiming at the use of the novel PBTZ derivative in designing a new regimen for TB, combination studies with selection of approved and experimental drugs were performed. The combination of PBTZ169 and bedaquiline, which proved to be synergistic in the in vitro REMA assay, was tested in the mouse chronic model of TB. Simultaneous administration of PBTZ, bedaquiline and pyrazinamide was more efficient in reducing the mycobacterial counts both in the lungs and spleen compared to the standard mixture of isoniazid, rifampicin and pyrazinamide [20].

Among favorable properties of PBTZ169 over BTZ043 is the lack of the chiral center, which simplified the production of the drug and significantly decreased its price. This is critical for developing treatment for a disease greatly affecting low-income and middle-income countries in particular. Taken together, PBTZ169 was identified as a promising candidate for development as a novel TB medicine [20].
