Download PDF
Synlett 2023; 34(07): 793-806
DOI: 10.1055/s-0042-1751373
account
Chemical Synthesis and Catalysis in India

Recent Advancements in Bottromycin Biosynthesis

Krushnamurthy PH
,
Subramanya KS
,
Simita Das
,
Dhananjaya G.
,
Nilkamal Mahanta∗

We thank Science and Engineering Research Board (SERB), Department of Science and Technology (DST), Ministry of Science and Technology, India for providing financial assistance for this work (Grant No: SERB-SRG/2019/000948 to Dr. Nilkamal Mahanta).
› Further Information
 
 PDF Download Permissions and Reprints All articles of this category


Abstract

Bottromycin is a structurally complex cyclic peptidic compound isolated from Streptomyces bottropensis and related organisms and belongs to the RiPP family of natural products (ribosomally synthesized and post-translationally modified peptides). It exhibits potent antibacterial properties against gram-positive pathogens (including drug resistant strains such as MRSA, MIC 1 μg/mL and VRE, MIC 0.5 μg/mL) and mycoplasma. Bottromycin blocks the binding of the aminoacyl-tRNA to the A-site on the 50S ribosome and hence inhibits protein synthesis. Bottromycins contain structurally diverse post-translational modifications (PTMs) on a small peptide (GPVVVFDC) including a unique macrocyclic amidine, rare β-methylation, terminal thiazole heterocycle, oxidative decarboxylation, and Asp epimerization, among others. It exhibits a precursor peptide organization with a C-terminal follower peptide and a N-terminal core peptide. There are several new studies reported recently which gave detailed insights into the bottromycin biosynthesis pathway. This Account highlights the current advancements in understanding the biosynthetic pathway of bottromycin focusing mainly on the biochemically and structurally characterized enzymes and intricate details of the peptide–protein biophysical interactions. These studies have provided a strong foundation for conducting combinatorial biosynthesis and synthetic biological studies to create novel bottromycin variants for therapeutic applications.

1 Introduction

2 Biosynthetic Pathway for Bottromycin

3 Enzymology of Bottromycin Biosynthesis

3.1 Cleavage of Methionine (BotP)

3.2 Radical SAM Methyltransferases (BotRMT1, BotRMT2, BotRMT3)

3.3 ATP-Dependent YcaO Enzymes

3.3.1 Thiazoline Formation by BotC

3.3.2 Macrolactamidine Formation by BotCD

3.4 Follower Peptide Hydrolysis (BotAH)

3.5 Aspartate Epimerization (BotH)

3.6 Oxidative Decarboxylation (BotCYP)

3.7 O-Methyltransferase (BotOMT)

4 Heterologous Bottromycin Production and Analogue Preparation

5 Summary and Outlook


#

Key words

RiPP antibiotic - bottromycin - radical SAM - cytochrome P450 - YcaO

Biographical Sketches

Zoom Image

Krushnamurthy P. H. obtained his MSc in pharmaceutical chemistry from Kuvempu University, Karnataka, India in 2016. Currently he is pursuing PhD in chemistry at the Indian Institute of Technology ­Dharwad, India on the study of organic chemistry of various enzyme-catalyzed reactions in the biosynthesis of ribosomally derived and nonribosomally derived peptide natural products under the supervision of Dr. Nilkamal Mahanta.

Zoom Image

Subramanya K. S. obtained his MSc in organic chemistry from Karnataka University, Dharwad in 2016. He has worked as a project assistant in the Department of Chemistry at the Indian Institute of Technology Dharwad on a ribosomally synthesized and post-translationally modified peptide natural product under the supervision of Dr. Nilkamal Mahanta. Previously, he has worked as scientific analyst at Molecular Connections Pvt Ltd, Bengaluru from 2017–2018 and then worked as junior research fellow at Poornaprajna Institute of Scientific Research, Bangalore from 2018–2019.

Zoom Image

Simita Das completed her MSc in applied chemistry from the National Institute of Technology Silchar, Assam in the year 2019. Currently, she is pursuing PhD from the Department of Chemistry at the Indian Institute of Technology Dharwad. She is currently working on the ribosomally synthesized and post-translationally modified peptide natural products and their biosynthesis under the supervision of Dr. Nilkamal Mahanta. Formerly, she has worked as chemistry subject study matter expert at Studymode Technologies Pvt Ltd Bangalore for a year (2019–2020).

Zoom Image

Dhananjaya G, completed his MSc in pharmaceutical chemistry from Kuvempu University in the year 2020. Currently, he is pursuing PhD from the Department of Chemistry at the Indian Institute of Technology ­Dharwad, Karnataka under the supervision of Dr. Nilkamal ­Mahanta since December 2021 and is working on the biosynthesis of ribosomally synthesized and post-translationally modified peptide natural products. He has received DST-­Inspire Fellowship from the ­Government of India for pursuing his PhD studies.

Zoom Image

Nilkamal Mahanta obtained his PhD in chemistry (vitamin biosynthesis) from Texas A&M University, College Station, ­Texas, USA under the supervision of Prof. Tadhg Begley. After postdoctoral studies at the University of Illinois, Urbana-­Champaign, Illinois, USA in RiPP natural products with Prof. Douglas Mitchell, he joined ­Indian Institute of Technology Dharwad as an assistant professor of chemistry and presently serving as the head of the department. Currently, his laboratory is investigating various intriguing enzyme-catalyzed reactions involved in the biosynthesis of antibiotic and anticancer natural products.

1

Introduction

Ribosomally synthesized and post-translationally modified peptides (RiPPs) comprise an emerging class of peptide-derived natural products.[1] [2] [3] [4] [5] RiPPs have a diverse set of biological activities, ranging from antibacterial, antiviral, antitumor, antidiabetics, to antinociception agents.[2,3] RiPP biosynthesis has been extensively studied due to its vast repertoire of post-translational modifications (PTMs) and their therapeutic potential.[2,4,6,7] Moreover, since the biosynthesis commences on a ribosomally synthesized peptide, it offers excellent opportunity for bioengineering of RiPPs for better activities as compared to the less flexible nonribosomal peptide synthetase (NRPS) and polyketide synthase (PKS) derived natural products.[8] [9] [10] RiPP biosynthesis begins with chemical modifications (PTMs) on a precursor peptide produced by the ribosome, generally consisting of N-terminal leader region and C-terminal core region by several downstream enzymes. These clustered group of genes which encode for the enzymes that modify the core peptide to generate the mature RiPP is known as the biosynthetic gene cluster (BGC).[1] [2] Major biosynthetic enzymes bind to the recognition sequences of the leader peptide using a conserved region of the enzyme (ca. 90–100 amino acid sequence either at the N-terminal or C-terminal regions) known as the RiPP precursor peptide recognition element (RRE).[11] Sometimes, the RRE could be a standalone protein in the BGC as well. This is followed by several interesting chemical reactions on the core peptide by other enzymes of the BGC. The leader is subsequently cleaved from the modified core and the latter may undergo further modifications giving rise to the bioactive structure of the natural product.

Bottromycin was originally obtained from Streptomyces bottropensis DSM 40262 in 1957,[12] however, the correct structure was identified in 2009.[13] Later, other Streptomyces species were also found to be the producers of various bottromycin congeners including S. sp. WMMB272,[14] S. sp. BC16019,[15] and S. scabies.[16] These compounds (Figure [1]) contain several unique structural features including a class-defining macrocyclic amidine, β-methylated amino acids, decarboxylated C-terminal thiazole, and epimerized Asp (d-epimer), among others.[15]

The bottromycins show impressive antibiotic activities against gram-positive bacteria and mycoplasma. These include notorious human pathogens such as methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE) with minimal inhibitory concentrations (MICs) of 1.0 μg/mL and 0.5 μg/mL, respectively.[13] Bottromycin A2 functions as protein biosynthesis inhibitor by interfering with the A site of 50S ribosome and hinders the binding of incoming aminoacyl-tRNA to this site.[17] It was hypothesized that bottromycin binds near the A-site of ribosome, lowering its affinity towards aminoacyl- and peptidyl-tRNA. Additional studies also indicated that it inhibits both the peptidyl-tRNA translocation and mRNA transport on the ribosomes, however, absence of a ribosome–bottromycin structure eludes this confirmation.[5] [18] As no cross-resistance has been observed thus far, bottromycins are considered to be excellent candidates for new antibiotics. Unfortunately, bottromycins have not yet entered clinical use, despite its promising in vitro efficacy. One of the reasons could be its poor in vivo stability.[19] As per reports, in blood plasma, the methyl ester of bottromycin is hydrolyzed very rapidly,[20] which also increased the MIC to 64 μg/mL from 1.0 μg/mL against MRSA. However, the propyl ketone derivative (replacement of the –OMe group of the methyl ester by a propyl group) of bottromycin A2 exhibited excellent hydrolytic stability.[20] Total syntheses of bottromycin and its analogues have been reported by various groups[13] [21] [22] using elegant synthetic strategies, which have been covered in other reviews.[18]

Recently, there have been significant advancements in the characterization of the biosynthetic pathway of this promising antibiotic and preparation of derivatives using synthetic biology approaches. Derived from a core peptide (GPVVVFDC), bottromycins exhibit complex PTMs in their structure. However, unlike other RiPPs with N-terminal leader peptides studied so far, it has a 35-residue long C-terminal follower peptide which is regarded to be crucial for stability and recognition. Bottromycin BGC was discovered in 2012 establishing that it is a ribosomally synthesized peptide,[14] [15] [16] , [23] however, in vitro biochemical studies on various enzymes on the pathway have not been reported until recently.[24] [25] [26] [27] [28] [29] The BGC from S. sp. BC16019 is shown in Figure [2] with annotated functions of the enzymes. The core peptide undergoes many PTMs including β-C-methylation of the Pro, Val, and Phe residues, O-methylation of the Asp residue, thiazoline, macrolactamidine, and oxidative decarboxylation to form a thiazole (Scheme [1]). In this Account, we aim to highlight the core enzymatic steps and mechanisms involved in constructing the complicated skeleton of bottromycin followed by several tailoring modifications to complete its bioactive structure.

Zoom Image
Figure 2 Biosynthetic gene cluster (BGC) of bottromycin
Zoom Image
Figure 1 Structure of bottromycin and its variants. Red and black elliptical regions indicate demethylated and methylated parts, respectively. ­Bottromycin D has an Ala in place of Val3.[18]
Zoom Image
Scheme 1 Proposed biosynthetic pathway for bottromycin

# 2

Biosynthetic Pathway for Bottromycin

Based on the in vivo knockout studies and in vitro assays on some of the enzymes, a biosynthetic pathway has been proposed for bottromycin, and the order of reactions is largely established (Scheme [1]). It is proposed that hydrolysis of the methionine on BotA precursor peptide 1 by the hydrolase (BotP) results in 2 which will be followed by the methylation of sp 3 carbon centers (β-carbons on Phe, Val, and Pro side chains) by radical S-adenosylmethionine (SAM) dependent methyltransferases (BotRMT1, BotRMT2, and BotRMT3) to form 3. Thereafter, adenosine triphosphate (ATP) dependent YcaO enzyme (BotC) will carry out thiazoline formation to generate intermediate 4. Macrolactamidine formation, catalyzed by another YcaO enzyme (BotCD) will follow, leading to the generation of the bottromycin core scaffold 5. The follower peptide (35 residues long) will be cleaved off by the amidohydrolase BotAH, resulting in 6 and further tailoring will be performed by α/β-hydrolase (BotH) which epimerizes α-carbon of Asp7 (reversible), generating intermediate 7. Thereafter, cytochrome P450 enzyme BotCYP will catalyze oxidative decarboxylation, converting the thiazoline moiety into thiazole 8. Final O-methylation by SAM dependent methyltransferase (BotOMT) will complete the biosynthesis of bottromycin A2.[14] [15] [16] , [23] [30] Since the genes in different BGCs[14] [15] [16] , [23] from various organisms have been named differently (corresponding proteins perform the same functions), we have decided to use the bot (from S. sp. BC16019) nomenclature for discussion here. However, the actual annotations used by the original authors for experimental investigations will be presented in parentheses, such as bst, bmb, and btm from S. sp. WMMB272, S. bottropensis, and S. scabies, respectively.[18]


# 3

Enzymology of Bottromycin Biosynthesis

3.1

Cleavage of Methionine (BotP)

N-Terminal methionine elimination from the precursor peptide (BotA) is essential, as it makes Gly1 available for macrolactamidine formation with an internal amide bond (Val4), the class-defining PTM in bottromycins. BotP, the enzyme responsible for this modification, shows sequence homology with M17 leucine aminopeptidases (LAPs) derived from several organisms.[15] According to metabolomics data, BotP was believed to be the first enzyme on the pathway.[30] In vitro analysis showed the catalytic activity of BotP with BotA with metal ion as a cofactor with the latter probably activating water for cleaving the peptide bond.[25] BotP belongs to the Zn2+-dependent amino peptidase family.[31] Recently, the crystal structure of BotP was determined to 1.76 Å resolution, and the substrate scope was assessed using a truncated version of BotA.[25] BotP monomers have an unique ‘comma’ structure,[25] with a big C-terminal region which cover residues 164–499 and a smaller N-terminal region covering residues 35–163. It was found to be structurally similar to that of other M17 LAPs, but it lacks the divalent metal ions found in LAPs. BotP was pre-incubated with MnCl2 followed by crystallization to obtain BotP–Mn2+ crystals. The presence of Mn2+ ions in BotP was further confirmed by a X-ray fluorescence scan[25] of the BotP–Mn2+ crystals. This cocrystal structure was later solved to 2.32 Å resolution which revealed that both the putative metal binding sites in BotP were occupied with Mn2+.

In vitro studies with recombinant BotP were conducted in the presence of divalent metal ions including Zn2+, Mn2+, Co2+, and Mg2+ ions. Co2+ was found to be the most suitable cofactor with maximal activity, whereas lesser activity was observed with Mn2+. On the other hand, Zn2+ and Mg2+ ions showed almost no effect.[25] Competition assay of BotA and shortened BotA core (MGPVV) showed that BotA gets processed ca. 4-fold faster than the shortened form. First three residues of the core (MGP) provides the substrate specificity with the long and narrow S1 pocket of BotP accommodating the Met residue. It was observed that Thr375 and Thr373 from BotP form hydrophobic interaction and hydrogen bond with Met of BotA, respectively.[25] Several mutations in the active site pocket for BotP were accepted, highlighting its promiscuity.


# 3.2

Radical SAM Methyltransferases (BotRMT1-3)

SAM-dependent methyltransferases (MTs) catalyze the transfer of methyl groups from SAM to nucleophilic substrates in a SN2 reaction with an inversion of configuration, generating S-adenosyl-homocysteine (SAH) as the byproduct. These include all DNA MTs and some protein MTs.[32] On the other hand, radical SAM methyltransferases (rSAM MTs) methylate inert carbon centers (sp 2- or sp 3-hybridized) using diverse mechanisms.[33] These enzymes employ a [4Fe–4S] cluster as a redox cofactor and are subdivided into four classes presently. Class A includes rSAM MTs that modify RNA bases and utilize noncluster binding Cys residues for methylation. A C-terminal SAM-binding domain and a N-terminal B12 (cobalamin) binding domain are the hallmarks of class B enzymes that methylate carbon centers and P-atoms of phosphinates. On the other hand, B12-independent class C enzymes are close homologues of the rSAM enzyme coproporphyrinogen III oxidase HemN, which use two molecules of SAM for methylation. Finally, class D rSAM MTs use methylene tetrahydrofolate as the methyl donor for this purpose.[33]

Bottromycin BGC encodes three class B rSAM MTs (BotRMT1, BotRMT2, BotRMT3), which transfer methyl groups to β-carbon centers of Phe6, Val4/Val5, and Pro2 residues of the BotA core, respectively, as per in vivo gene deletion experiments in S. sp. BC16019[15] and S. scabies. [16] An unique CX7CX2C [4Fe–4S] cluster binding motif is present in these rSAM MTs instead of the conventional CX3CX2C motif found in rSAM enzymes.[34] Genetic knockout studies on btmC (botRMT1), btmG (botRMT2), and btmK (botRMT3) were performed using pYH7-based double crossovers[16] in S. scabies. Bottromycin biosynthesis was stopped once btmC was inactivated indicating partial methylation of Phe6 is essential for proper maturation. Bottromycin B2 (Figure [1]) was the product of the btmK deletion mutant indicating that it methylates Pro2, and deletion of btmG resulted in a bottromycin variant which lacks two methyl groups suggesting that it methylates Val (Val4/Val5). The Müller group[15] also performed gene deletions (for botRMT1, botRMT2, and botRMT3) on the expression construct DG2-kan-efflux-ermE using the Red/ET recombinant technology. Thereafter, S. coelicolor A3(2) was transformed with these knockout constructs and heterologous production of bottromycin variants were studied. Extensive LC–HRMS characterization of the S. coelicolor A3(2) culture extracts elucidated the methylation patterns of the new bottromycin derivatives. Based on the MS and MS/MS studies on the new bottromycin analogues, it was proposed that BotRMT1, BotRMT2, and BotRMT3 catalyze β-C-methylation of Phe6, Val4/Val5, and Pro2, respectively.[15]

Other than bottromycin, thiostrepton[35] [36] and polytheonamides[37] are other RiPPs that involve class B rSAM MTs in their biosynthetic pathways in addition to many antibiotics such as gentamicin,[38] clorobiocin,[39] chondrochloren,[40] moenomycin,[41] and l-phosphinothricin.[42] However, their enzymatic mechanisms are still poorly understood. Thus far, studies have been reported for PhpK,[42] TsrM,[36] Fom3,[43] PoyC,[37] and CysS[44] enzymes from phosalacin, thiostrepton, fosfomycin, polytheonamide, and cystobactamid biosynthesis pathways, respectively. According to these studies, two molecules of SAM are required for each round of catalysis. The catalytic 5′-deoxyadenosyl (5′-dA) radical is formed from one molecule of SAM by homolytic cleavage of the C–S bond, while the other one is involved in the delivery of the methyl group. Interestingly, B12-rSAM MTs in both polytheonamide A[5] and bottromycin A2[3] possess unusual CX7CX2C motif instead of canonical rSAM motif, CX3CX2C. Corresponding triple variant AX7AX2A in PoyC abolished its activity suggesting this motif coordinates the [4Fe–4S] cluster essential for the reductive cleavage of SAM.[5] Based on this observation, such Cys-rich motifs in BotRMT1, BotRMT2, and BotRMT3 are likely to coordinate the [4Fe–4S] cluster.

Moreover, these rSAM enzymes also harbor the RRE motifs (predicted residues based on HHPred homology analysis with PqqD are A492 –K593 for BmbB and L526 –R670 for BmbF)[24] which are predicted to engage the follower peptide to initiate the biosynthetic pathway.[11] Recent fluorescence polarization (FP) based binding studies with precursor BotA (BmbC) and truncated versions of RRE from BotRMT1 (BmbB) and BotRMT2 (BmbF) illustrated moderate binding affinities (K D: 876 ± 67 nM for BmbB and K D: 597 ± 25 nM for BmbF), indicating that these rSAM MTs with proposed RREs for substrate recognition are likely to carry out reactions early on the pathway.

So far, none of the rSAM MTs in bottromycin biosynthesis are reconstituted and characterized in vitro. Based on the previous reports on similar enzymes, a proposed mechanism[45] for these enzymes in bottromycin pathway is shown in Scheme [2]. At first, co(I)balamin (9) will react with SAM-1 (10) via SN2 reaction to give methyl co(III)balamin (11) and SAH (12). In the next step, SAM-2, which is presumably coordinated to the essential [4Fe–4S] cluster, will undergo reductive cleavage generating the byproduct methionine and the highly reactive 5′-deoxyadenosyl radical (5′-dA•) 13. This 5′-dA• radical, in turn, will abstract a hydrogen atom from the β-carbon (Phe6 in case of BotRMT1) from BotA (1), yielding the substrate radical 14 and the byproduct 5′-dAH. The substrate radical will finally abstract the methyl group from 11 (presumably generated by homolytic cleavage of the Co–C bond), yielding the methylated substrate 15 and Co(II)balamin. This, upon reduction possibly by the [4Fe–4S] cluster, will form 9, ready for another catalytic cycle. This mechanistic proposal, however, awaits biochemical and experimental validation.

Zoom Image
Scheme 2 Proposed mechanism for radical SAM methyltransferases (BotRMT1 is shown here as an example)

# 3.3

ATP-Dependent YcaO Enzymes (BotC and BotCD)

YcaO superfamily of proteins are known to catalyze the formation of azoline (oxazoline and thiazoline) heterocycles on peptidic backbones using nucleophilic side chains of Ser, Thr, and Cys with ATP as a cofactor[46] through a proposed hemiorthoamide intermediate. However, E1-ubiquitin-activating or ocin-ThiF-like proteins are needed as partner proteins to assist in substrate binding for some ­YcaOs[46] [47] through distinct RRE domains.

In-vitro reconstitution in 2012 demonstrated that YcaO and E1-like protein functioned together as a cyclodehydratase.[11]

They are usually fused to the N-terminal of the YcaO protein or as separate enzymes in the BGC.[46] [47] Currently, four functions of YcaO are known including thia(oxa)zoline formation, thioamide formation, macroamidine formation, and assisting in RimO-dependent methylthiolation.[47] Other than bottromycin, several other RiPPs such as cyanobactins, thiopeptides, and linear azol(in)e-containing peptides (LAPs) also employ YcaO enzymes for cyclodehydration reactions in their biosynthetic pathways.[47]

Two YcaO proteins have been reported in the bottromycin BGC, one of which catalyze the formation of thiazoline derived from Cys8 (BotC) while the other carry out the macrolactamidine formation, which involves Gly1 and Val4 (BotCD).[15] [16] [23] This was indicated from genetic knockout, comparative metabolomics, and mass-spectral networking studies [30]. The essentiality of these genes in the pathway was evident from complete abolishment of bottromycin production in the knockout mutants.[16] Interestingly, BGC does not contain any E1-like protein and neither of the ­YcaOs contain RREs demonstrating that these are standalone proteins, without requiring any partner proteins unlike other YcaOs.[47] The YcaO domain proteins work either by phosphorylating the amide backbone or by adenylating the carbonyl oxygen[46] [48] to perform these reactions. As the BGC lacks a flavin-dependent dehydrogenase, the conversion of thiazoline into thiazole was suggested to be catalyzed by BotCYP (BtmJ), a cytochrome P450 enzyme,[30] via oxidative decarboxylation.

3.3.1

Thiazoline Formation by BotC

In vitro reconstitution of thiazoline forming YcaO was achieved independently by two groups.[24] [29] As BotC could not be expressed in the soluble form, its homologue IpoC (from Streptomyces ipomoeae, 76% sequence identity to BotC) was used for in vitro studies by Koehnke group.[29] BotA was treated with IpoC in the presence of ATP and ­MgCl2 which resulted in a loss of 18Da mass, consistent with proposed cyclodehydration. Using tandem MS, it was confirmed that Cys8 in BotA precursor was converted into thiazoline. IpoC also lacked any predictive RRE domain. Hence, thermal shift assays (TSA) were performed to elucidate its interaction with the follower peptide.[24] [29] TSA is performed to measure the melting temperature (Tm) of peptide–enzyme interaction[49] and a rise in Tm is observed if a stabilizing interaction between a ligand/substrate and respective protein is formed. TSA assay showed that the region between residues 30 and 39 in the follower peptide is essential for the interaction of BotA with IpoC, indicating direct engagement between YcaO and the follower peptide.[29]

Around the same time, the Mitchell group independently showed that BotC (BmbD) catalyze ATP-dependent cyclodehydration of Cys8 using MALDI-TOF-MS, ESI-MS/MS analysis, and chemical-labelling studies.[24] Several core-peptide variants of BotA were tested to evaluate the substrate scope of BotC which indicated that Cys8 is indispensable for this reaction. Substitution of distal residues were not tolerated, illustrating that BotC has limited substrate promiscuity.

Proposed mechanism for BotC reaction with BotA (1) is shown in Scheme [3], however, it is yet to be experimentally validated fully. In the first step, the thiol of Cys8 in 1 will be deprotonated by a putative base in the active site of BotC. The resulting nucleophile 16 will attack the preceding Asp7 carbonyl group generating the oxyanion tetrahedral intermediate 17. Thereafter, 17 will be phosphorylated using ATP, forming the hemiorthoamide intermediate 18, and liberating the byproduct, ADP. This will be followed by subsequent elimination of the phosphate group from intermediate 18 to form the product 19 (Scheme [3]).

Zoom Image
Scheme 3 Proposed mechanism for BotC

# 3.3.2

Macrolactamidine Formation by BotCD

Macrocyclization is an important PTM on the pathway as it cyclizes the unstructured modified peptide to a rigid cyclic structure generating the core bioactive scaffold of bottromycin. In vivo, genetic deletion followed by comparative metabolomics and mass-spectral network analysis of individual BotCD and BotAH deletion mutants indicated that they collaborate to generate the macrolactamidine.[30]

As per the studies reported by Koehnke group, PurCD (from S. purpureus, 81% sequence identity to BotCD) was used to reconstitute the macrolactamidine formation with BotA and ATP/Mg2+ which was found to be reversible. This was confirmed by the formation of characteristic macroamidine fragments using MS/MS.[29] Using TSA, it was illustrated that the rate of macroamidine formation increases with increasing pH up to 9.5, after which its reactivity decreases probably due to enzymatic degradation. PurCD also lacks domains that are homologous to recognizable RREs. TSA showed that follower peptide region (residues 30–39) is essential for the interaction of BotA with PurCD.[29] Mechanistic studies on PurCD (BotCD) showed that it catalyzed reversible macrocyclization (both ring closure and opening).[29] Amidohydrolase (BotAH) was found to be essential for macroamidine formation in vivo [30] which was corroborated in vitro using PurAH, an amidohydrolase from S. purpureus, with 72% sequence identity to BotAH. PurAH was responsible for removing the follower peptide leading to irreversible formation of macroamidine, suggesting that PurCD collaborates with PurAH in generating the core scaffold of bottromycin.[27] However, the intricate details of this protein–­protein interaction remains to be investigated further.

The Mitchell group also reported independently that BotCD (BmbE) perform macrolactamidation on BotA (BmbC) in vitro which was found to be ATP-dependent.[24] This was confirmed by MALDI-TOF-MS, ESI-MS/MS analysis, and chemical-labelling studies. To assign the exact function of the YcaOs, labelling experiments were conducted involving alkylation of thiol with 2-bromoethylamine and reductive amination of some amino acids using formaldehyde and borane with BotA (BmbC) in the presence of BotC (BmbD) and BotCD (BmbE), respectively. Analysis of the reaction products revealed that dimethylation occurred at the Gly1 of the N-terminus in thiazoline product (BmbD reaction) and Cys8 side chain undergoes alkylation in the macrolactamidine product (BmbE reaction). In addition, BmbE also acted on the BmbC-C8A variant while BmbD did not. All these evidence suggested that thiazoline was formed by BmbD while macrocyclization was catalyzed by BmbE. Further, HR-ESI-MS/MS and NMR analysis unambiguously confirmed this finding.[24] BotA variants aimed to prepare different sized macrocycles were not accepted by BotCD (BmbE) suggesting that it has a limited substrate tolerance.[24]

A proposed mechanism for BotCD is shown in Scheme [4], which is yet to be tested experimentally. First, amino group of Gly1 in BotA (2) will be deprotonated by a putative base in the active site of BotCD. The resulting anion 20 will attack the Val4 carbonyl group generating the oxyanion tetrahedral intermediate 21. Thereafter, 21 will be phosphorylated using ATP, forming the hemiorthoamide intermediate 22 and releasing the byproduct, ADP. This will be followed by subsequent elimination of the phosphate group from 22 to yield the product 23 (Scheme [4]).

Zoom Image
Scheme 4 Proposed mechanism for BotCD

Simultaneously, the role of follower peptide in enzyme–substrate binding was also investigated by determining the activity of BotC (BmbD) and BotCD (BmbE) with BotA ­(BmbC) follower Ala mutants[24] followed by a qualitative MALDI-TOF-MS analysis. Since some portion of the BmbC follower region showed resemblance to lanthipeptide and cyanobactin recognition sequences, Ala variants of BmbC follower peptide were designed keeping this fact in mind (Figure [3]). Several amino acid residues in the follower peptide were found to be critical for processing by BmbD and BmbE, indicating that follower peptide region of BmbC is vital for both BmbD and BmbE functioning even though they lack any distinct RRE domains.[24] Critical binding residues in the follower peptide of BmbC from this assay are illustrated in Figure [3]. It was also evident that ATP-binding motifs conserved among the azoline-forming E1-ubiquitin/ocin-ThiF-dependent YcaOs contain essential Pro-rich C-termini, whereas BmbD and BmbE contain Glu-rich ATP binding motif which were experimentally found to be critical for their function.[24]

Zoom Image
Figure 3 Recognition sites in the BotA (BmbC) follower peptide assessed by reaction of BotC (BmbD) and BotCD (BmbE) with respective Ala variants (Panel A: for BmbD and Panel B: for BmbE). Residues colored in red indicate the sites of Ala substitution. The residues that are bold and underlined indicate the BmbC follower peptide variants that inhibited the activity of BmbD and BmbE during assay, respectively.[24]

#
# 3.4

Follower Peptide Hydrolysis (BotAH)

Metal-dependent amidohydrolase BotAH is the enzyme involved in the follower peptide cleavage after macrocyclization has been achieved. The Koehnke group determined the crystal structure of a related homologue PurAH (BotAH) at 1.73 Å resolution.[27] Two catalytic Zn2+ ions were observed in the active site. Coordinating residues for one Zn2+ ion included His210 and His229, while residues His94, His96, and Asp348 were bound to the other Zn2+ ion. In addition, it was also observed that carboxylated Lys183 coordinated both the metal ions.[27] A proposed mechanism for the follower peptide hydrolysis is given in Scheme [5]. Removal of a proton from water molecule by Asp348, coordinated by Zn2+ ions, results in activated hydroxyl which attacks the amide bond between the thiazoline and the follower peptide in 24, generating a tetrahedral intermediate 25. Upon resolution, this yields the modified core 26 and the follower peptide is liberated. From substrate scope analysis, Asp7 in precursor peptide BotA was found to be a critical residue, implicating its role in metal coordination during BotAH catalysis.[27]

Zoom Image
Scheme 5 Mechanistic proposal for amidohydrolase BotAH

# 3.5

Aspartate Epimerization (BotH)

The d-amino acids provide increased resistance to proteolytic degradation of peptides. To date, rSAM epimerases were reported to produce d-amino acids from l-amino acids through a radical-mediated mechanism[50] [51] and a two-step dehydration–hydrogenation sequence to convert l-serine into d-alanine.[52]

Radical SAM epimerases have been found in the biosynthetic pathways of several RiPPs including proteusins[50] and epipeptides.[53] However, in bottromycin BGC, none of the rSAM enzymes were found to be responsible for epimerization. It was elucidated that a α/β-hydrolase (ABH) enzyme BotH catalyzes Asp7 epimerization.[26] Crystal structure of BotH at a high resolution (1.18 Å) showed that four α-helices constitute a loop with a V-shape which lies above the putative active site revealing structural similarity to 3-oxoadipate-enol-lactonase.[54]

For in vitro reconstitution of BotH, macrolactamidine 2-thiazoline-4-carboxy intermediate 27 obtained by the reaction of N-terminal methionine cleaved core peptide 2 with YcaO enzymes BotC, BotCD, and amidohydrolase BotAH was used as a substrate (Scheme [6]).[8] [9] The epimerization from l-Asp to d-Asp was monitored by LC–MS and was found to be BotH concentration dependent. The reaction was found to be reversible with the d-Asp as the major epimer product. Both of these isomeric products displayed similar bioactivity.[26]

Zoom Image
Scheme 6 Proposed mechanism for aspartate epimerase BotH

The substrate mixture (l-Asp/d-Asp epimers) was cocrystallized with BotH, and the crystal structure was determined at 1.25 Å resolution.[26] As per the proposed mechanism (Scheme [6]), backbone NH group of Phe110 and Val41 from BotH are involved in hydrogen bonding with the thiazoline carboxyl anchoring the substrate in the active site. Generally, the mechanism for epimerization involves the removal of a proton from one side and addition of the same from the other side, however, there is no evident catalytic residue observed within 4 Å to Cα hydrogen of Asp7 for hydrogen abstraction. It was proposed that the abstraction of hydrogen in 27 could be substrate-assisted with the carboxyl group of Asp7 lying close (distance 2.2 Å) to its Cα hydrogen (Scheme [6]). Abstraction of proton from 27 by the carboxylate will result in enamine intermediate 28, which on protonation of the enamine double bond from the other side will generate the epimer 29. From the crystal structure, it was evident that four ordered water molecules are in close contact with the Asp7 carboxyl group, thus possibly allowing interchange of the abstracted proton with solvent during the reaction.[26] Reaction with various substrate variants D7A, D7N, and D7E showed that only D7E variant undergoes epimerization possibly because E also has a carboxyl side chain as that of D differing only in a methylene group which could be optimally located to participate in the reaction. This finding illustrated that the side-chain carboxyl group of Asp7 is most likely involved in anchimeric assistance during the epimerization reaction.[26]

Cocrystal structure of bottromycin–BotH showed that bottromycin bind in the same binding pocket as that of the BotH substrate (with 2-thiazoline-4-carboxyl moiety ), indicating that bottromycins could act as orthosteric inhibitors of BotH.[26] This was also supported by microscale thermophoresis (MST) experiments which showed that bottromycin A2 and its other variants bind BotH with high (nM) to low (μM) K D values. Using several variants of BotA, it was elucidated that BotH displays relaxed substrate specificity for epimerization at Asp7 when mutations were made at positions Pro2, Val3, and Val4. On the other hand, F6A variant was not epimerized, while epimerization was observed for F6Y and F6W substrates indicating the importance of the aromatic group in orienting Asp7. Since Val5 side chain interacts hydrophobically with the side chains of Phe6 and Val4, the V5T substrate did not undergo any epimerization. On the other hand, other variants such as V5L, V5E, and V5A were well tolerated.[26]

It was proposed that epimerization may take place in between thiazoline formation and its conversion to thiazole.[30] Imine–enamine tautomerization which make Asp7 Cα hydrogen acidic would be diminished once it is converted into the aromatic thiazole. The reason for rapid epimerization may be due to hydrogen bonding between the nitrogen of the thiazoline and the carboxylic acid which may facilitate imine–enamine tautomerization.[55] Such epimerization of amino acids close to carboxylated thiazolines have been reported earlier.[56]


# 3.6

Oxidative Decarboxylation (BotCYP)

Previous studies on RiPPs containing azol(in)e heterocycles such as cyanobactins or LAPs showed that a FMN-dependent dehydrogenase oxidizes azolines to the aromatic azole heterocycles.[57] However, a flavin-dependent dehydrogenase is not a part of bottromycin BGC for this proposed dehydrogenation step.[25] Instead, for this oxidative decarboxylation of thiazoline to thiazole, based on in vivo gene deletion study, a cytochrome P450 enzyme (BotCYP) was found to be responsible.[30] Recently, BotCYP was reconstituted using three enzymes: BotCYP, BmCPR, and Fdx2 (from Bacillus megaterium [5]) with substrate a/b (a = macroamidated, thiozoline product, b = epimerase, BotH-treated compound a) in the presence of NADPH resulting in a decarboxylated and oxidized reaction product with very poor yield (<25%).[28] To find a reason for low yield, equimolar mixture of l-Asp and d-Asp was used as a substrate and the product mixture was analyzed by LC–MS. Only d-Asp-containing substrate showed oxidative decarboxylation indicating that BotCYP preferentially accepts the d-isomer as a substrate. This confirmed that epimerization of l-Asp to d-Asp by epimerase BotH proceeds the BotCYP reaction. To know the configuration of Asp residue during oxidative decarboxylation, BotCYP assay was carried in water with a substrate analogue in which Cα-position of Asp was deuterated. This resulted in a peak with a similar LC retention time and MS/MS fragmentation patterns as the previous oxidative decarboxylated product. It indicates that during oxidative decarboxylation to the aromatic thiazole, the configuration of the Cα-position of Asp7 was retained.[28]

A mechanism for this unique cytochrome P450 enzyme mediated oxidative decarboxylation is proposed in Scheme [7]. Abstraction of β-hydrogen from the thiazoline intermediate 30 by the heme Fe (IV)-oxo cation radical intermediate 31 results in the substrate radical 32 and heme Fe (IV)-oxo intermediate 33. One-electron transfer from 32 to 33 will generate the cation intermediate 34 and Fe(III) intermediate 35. This will facilitate the elimination of CO2 from 34 to form the thiazole product 36.[58] However, experimental validation of this proposal awaits in-depth mechanistic studies.

Zoom Image
Scheme 7 Mechanistic proposal for BotCYP

# 3.7

O-Methyltransferase (BotOMT)

In bottromycin BGC, it was proposed that the protein-encoding botOMT gene carry out the final O-methylation. BotOMT is a member of the COG3315 family of SAM-dependent carboxyl MTs which perform methylation on the carboxylate groups.[32] Genetic knockout studies with ­botOMT confirmed its role in O-methylation of Asp7.[15] This metabolic study also suggested that O-methylation is possibly the last step in the biosynthesis pathway.[30] A probable mechanism (Scheme [8]) will involve SN2-dependent methylation of Asp7 carboxyl O-atom in intermediate 37 using SAM (10) as the methyl donor like other SAM-dependent MTs,[32] resulting in the final product 38 and SAH (12). However, in vitro reconstitution and characterization of this enzyme is yet to be achieved.

Zoom Image
Scheme 8 Proposed mechanism for BotOMT

#
# 4

Heterologous Bottromycin Production and Analogue Preparation

Heterologous production of bottromycin was performed by Müller,[15] Luzhetskyy,[59] and Truman[60] groups separately using engineered systems. For heterologous expression in a suitable host strain, entire bottromycin BGC from S. sp. BC16019 was subcloned into a cosmid (pOJ436) and directly transformed into S. coelicolor A3(2) and S. albus J1074. These expression hosts led to bottromycin production, however, with ca. 100 times lower productivity than the original producer. Deletion of ca. 16 kb insert from this construct and integration of kan resistant gene resulted in a new cosmid (DG2-kan) which showed improved production.[15]

To improve bottromycin production, a drug-resistance mutation in rpoB gene (RNA polymerase β subunit) was introduced. It was done by integration of rifampicin resistant gene in the rpoB gene region. These modified strains gave high productivity by 10-fold compared to the previous DG2-kan clones. BotT belongs to major facilitator superfamily (MFS) transporter which might affect bottromycin yield. Therefore, to increase botT gene expression, PermE* promoter was inserted in place of its native promoter using a spectinomycin antibiotic resistance marker. This resulted in a 2-fold increase in bottromycin production compared to the rif-resistant mutant.[15]

Luzhetskyy group[59] expressed the bot BGC containing DG2-kan in S. lividans TK24, which produced bottromycin A2 (concentration 0.23 mg/L). Randomly synthesized promotors based on distinct consensus sequences of ermEp1 promoter were incorporated in between botOMT and botRMT1 of bot BGC, and modified BGCs were introduced into S. lividans to increase the yield.[61] The cell extracts were analyzed from these mutant strains using LC–MS which showed that around 10 strains produced 5–50 times more bottromycin A2 than the native strain. Introduction of this modified BGC into the original bottromycin producer strain, S. sp. BC16019, resulted in a new strain containing both native as well as mutated BGCs. Analysis of the extracts by LC–MS showed 37-fold rise in the formation of bottromycin compared to the wild type S. sp. BC16019, suggesting that the promoter insertion was successful.[59]

In addition, the Wittmann group recently reported that the cultures of the engineered S. lividans TK24 DG2-Km-P41hyg+ strain[59] produced up to 109 mg/L of bottromycins (bottromycin A2; 60 mg/L and methyl bottromycin A2; 46 mg/L) on addition of inorganic talc microparticles (hydrous magnesium silicate, 3 MgO·4SiO2·H2O, 15 g/L) into the growth medium.[62] This is the highest titer of bottromycins production reported so far using a recombinant strain. Based on several experiments, it was proposed that talc microparticles affected the expression levels of individual genes of the bottromycin BGC. This resulted in a higher macrocyclization efficiency at the level of amidohydrolase (BotAH). It was observed that microparticles induced higher levels of advanced biosynthetic intermediates (such as nonmethylated pre-bottromycin and methylated pre-bottromycin) and reduced the levels of noncyclized shunt products and incomplete precursor peptides. This overall indicated a higher efficiency of the entire post-translational modification process resulting in higher yields of bottromycins.[62]

Zoom Image
Figure 4 New bottromycin variants observed by yeast-mediated pathway refactoring.[60] Red ellipses indicate the demethylated positions in the structure.

Recently, bottromycin BGC from S. Scabies DSM 41658 was cloned into an Saccharomyces cerevisiae cum Escherichia coli shuttle vector using another cloning technique named transformation-associated recombination (TAR) method.[60] Bottromycin BGC harboring vector (pCAPbtm) was then conjugated with heterologous host S. coelicolor M1146 for production of mature bottromycins. However, very low amounts of products were detected. To improve the productivity, heterologous promoters, genetic deletions and targeted mutations, repair of double-strand DNA breaks in yeast, etc. were introduced at several restriction sites in the BGC.[60] The efficient refactoring and remodeling of the BGC led to a pathway controlled by a riboswitch that exhibited about 120-fold enhancement in bottromycin production in S. coelicolor M1146 and resulted in new bottromycin variants with differing methylation patterns, which were not reported earlier (Figure [4]).[60]

Since bottromycin is a RiPP natural product, precursor peptide engineering allows for the introduction of new amino acids through mutations of codons in the precursor peptide encoding gene botA. Compared to the nonribosomal peptides, RiPPs are ideal for creating new derivatives through pathway engineering. However, bottromycins lack natural diversity within their core peptide.[24] The ­Luzhetskyy[59] and Truman[16] [63] groups generated several botA variants with altered core region in the precursor for in vivo production of bottromycin analogues. Unfortunately, in vivo expression of most of the mutants did not show bottromycin metabolites indicating that these unnatural substrates were not processed by the bottromycin biosynthetic machinery. Val3 is the highest tolerant position in the core peptide, while other positions amenable for alteration were Pro2, Val4, and Phe6. BotA-Val3Met and BotA-Val3Ile variants were successfully converted into mature bottromycin derivatives in vivo using engineered plasmids in S. sp. BC16019 ΔbotA (a strain that can not generate wild-type BotA) and were characterized by MS/MS. Based on preliminary MS studies, it was proposed that a new bottromycin derivative containing S-methylmethionine residue was also formed from the Val3Met variant with additional methylation probably occurring on sulfur. However, further experiments will be required to confirm its identity. On the other hand, β-C-methylation status of the Val4 variants and rate or extent of cyclization during biosynthesis in case of the Pro2Ala variant are not reported so far. Substitutions at Asp7 position were not tolerated.[59] These results are summarized in Figure [5].

Zoom Image
Figure 5 Tolerant sites in the BotA core peptide (Pro2Ala, Val3Met/Thr/Ile/Ser/Ala, Val4Ile/Ala, and Phe6Tyr) which gave rise to bottromycin-like unnatural variants upon in vivo expression in a heterologous host.[59] [63]

# 5

Summary and Outlook

Bottromycin is a promising RiPP antibiotic with impressive activities against several gram-positive pathogenic strains of bacteria and mycoplasma. It inhibits protein biosynthesis by exhibiting a unique mode of action. Lately, several studies have been reported on bottromycin biosynthesis which illuminated various new findings on the timing of the steps and mechanistic underpinnings of the pathway. Peptide–protein biophysical interactions and insights into the enzymatic mechanisms have been elucidated for several steps on the pathway including methionine cleavage, thiazoline and macrolactamidine formation, Asp epimerization, and oxidative decarboxylation, among others. These studies also have paved a way forward for combinatorial biosynthesis and synthetic biology attempts to generate bottromycin analogues for alternation and improvement of the biological activities. However, reconstitution and mechanistic studies of the radical SAM-dependent C-methylation reactions which presumably occur very early on the pathway are yet to be established. With several research groups investigating this promising antibacterial agent from mechanistic, therapeutic, and bioengineering perspectives, it is highly likely that complete in-depth biochemical, structural, and pharmacological features of bottromycin biosynthesis will be revealed in the near future. With RiPP compounds gaining immense attention from the natural products research community due to their diverse structures and wide range of biological activities, detailed understanding of this pathway will lay the foundation for elucidation of the biosynthetic pathways of various other RiPPs with novel therapeutic values.


#
#

Conflict of Interest

The authors declare no conflict of interest.


Corresponding Author

Nilkamal Mahanta
Department of Chemistry, Indian Institute of Technology Dharwad
WALMI campus, PB road, near high court, Karnataka 580011
India   

Publication History

Received: 21 June 2022

Accepted after revision: 13 September 2022

Article published online:
17 October 2022

© 2022. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


   Permissions and Reprints All articles of this category
Zoom Image
Zoom Image
Zoom Image
Zoom Image
Zoom Image
Zoom Image
Figure 2 Biosynthetic gene cluster (BGC) of bottromycin
Zoom Image
Figure 1 Structure of bottromycin and its variants. Red and black elliptical regions indicate demethylated and methylated parts, respectively. ­Bottromycin D has an Ala in place of Val3.[18]
Zoom Image
Scheme 1 Proposed biosynthetic pathway for bottromycin
Zoom Image
Scheme 2 Proposed mechanism for radical SAM methyltransferases (BotRMT1 is shown here as an example)
Zoom Image
Scheme 3 Proposed mechanism for BotC
Zoom Image
Scheme 4 Proposed mechanism for BotCD
Zoom Image
Figure 3 Recognition sites in the BotA (BmbC) follower peptide assessed by reaction of BotC (BmbD) and BotCD (BmbE) with respective Ala variants (Panel A: for BmbD and Panel B: for BmbE). Residues colored in red indicate the sites of Ala substitution. The residues that are bold and underlined indicate the BmbC follower peptide variants that inhibited the activity of BmbD and BmbE during assay, respectively.[24]
Zoom Image
Scheme 5 Mechanistic proposal for amidohydrolase BotAH
Zoom Image
Scheme 6 Proposed mechanism for aspartate epimerase BotH
Zoom Image
Scheme 7 Mechanistic proposal for BotCYP
Zoom Image
Scheme 8 Proposed mechanism for BotOMT
Zoom Image
Figure 4 New bottromycin variants observed by yeast-mediated pathway refactoring.[60] Red ellipses indicate the demethylated positions in the structure.
Zoom Image
Figure 5 Tolerant sites in the BotA core peptide (Pro2Ala, Val3Met/Thr/Ile/Ser/Ala, Val4Ile/Ala, and Phe6Tyr) which gave rise to bottromycin-like unnatural variants upon in vivo expression in a heterologous host.[59] [63]