Comparative in silico analysis of transporters coded within biosynthetic genes clusters for ramoplanin and related antibiotics

Размещено на http://www.allbest.ru/

Размещено на http://www.allbest.ru/

Comparative in silico analysis of transporters coded within biosynthetic genes clusters for ramoplanin and related antibiotics

K. Zhukrovska, V. Fedorenko

Ivan Franko National University of Lviv

Glycopeptide antibiotics (GPAs), like teicoplanin and vancomycin, have been the first-line treatment for infections caused by Gram-positive multidrug-resistant pathogens. GPAs appear to be related to ramoplanin-like lipodepsipeptides (LDPs), yet another significant class of lipid II binders. Major compounds among LDPs are ramoplanin (the key representative), enduracidin, and chersinamycin; each with known biosynthetic gene clusters (BGCs). Five additional BGCs for the putative LDPs were recently described. LDP BGCs are poorly investigated; one particular aspect that deserves further investigation is transporters coded within BGCs. These proteins most likely take part in the export of antibiotics out of the cell, as well as in the producer's resistance to its own secondary metabolite. In this work, we performed in silico analysis of genes encoding transporters from ramoplanin and other LDP BGCs. We investigated the domain architecture of these transporters, discovered their homologues in BGCs from MIBiG and beyond, generated models of secondary and tertiary structures, and compared the overall LDP BGCs transport genes blueprint. We were able to identify previously uncharacterized gene encoding ABC transporter within ramoplanin BGC - ramo3. Ramol and Ramo3 in ramoplanin BGC appear to be paralogues coding for a permease subunit of the ABC transporter. In every other LDP BGCs, except for chersinamycin BGC, we found only one corresponding homologue encoding this type of protein. Similarly, we found that Ramo2 and Ramo23 are also homologous proteins, which appear to be ATP-binding subunits of the ABC transporter; Ramo2 and Ramo23 have only one homologue in each other LDP BGCs. Next, we were able to describe Ramo8 as ATP-binding ABC transporter, containing both ATPase and transmembrane parts, similar to those encoded in GPA BGCs. For Ramo8, we modelled 3D structure as well as quaternary structure for homodimer of this protein. Finally, our in silico analysis revealed Ramo31 to be a proton membrane antiporter, having distant homologue only in chersinamycin BGC; most likely Ramo31 is not connected to ramoplanin biosynthesis.

Keywords: biosynthetic gene clusters, ramoplanin, membrane transport proteins, secondary metabolites, soil microorganisms

ПОРІВНЯЛЬНИЙ IN SILICO АНАЛІЗ БІЛКІВ-ТРАНСПОРТЕРІВ, ЗАКОДОВАНИХ У КЛАСТЕРАХ БІОСИНТЕТИЧНИХ ГЕНІВ РАМОПЛАНІНУ ТА СПОРІДНЕНИХ АНТИБІОТИКІВ

К. Жукровська, В. Федоренко

Львівський національний університет імені Івана Франка

Глікопептидні антибіотики (ГПА), такі як тейкопланін і ванкоміцин, є одними з препаратів першої лінії для лікування інфекцій, спричинених грампозитивними мікроорганізмами, стійкими до різних лікарських засобів. ГПА пов'язані з ліподеп- сипептидами (ЛДП), ще одним важливим класом антибіотиків, здатних зв'язувати ліпід II. Основними сполуками, що належать до групи ЛДП, є рамопланін (ключовий представник), ендурацидин і черсинаміцин. Для цих антибіотиків відомі кластери біо- синтетичних генів (КБГ), що кодують їхній біосинтез. Нещодавно було описано п'ять додаткових КБГ, які кодують імовірні ЛДП. КБГ ЛДП недостатньо досліджені; одним із аспектів, які потребують подальшого вивчення, є білки-транспортери, закодовані в межах цих КБГ. Ці білки, скоріш за все, беруть участь в експорті антибіотиків із клітини, а також у забезпеченні стійкості продуцента до власного вторинного метаболіту. У цій роботі ми провели in silico аналіз генів, які кодують транспортери, в межах КБГ рамопланіну й інших ЛДП. Ми дослідили доменну архітектуру цих транспортних білків, виявили їхні гомологи в КБГ, депоновані у репозиторії MIBiG та за його межами, створили моделі вторинних і третинних структур, і порівняли розташування транспортних генів у КБГ ЛДП. Нам вдалося ідентифікувати раніше неохарактеризований ген, що кодує ABC-транспортер у КБГ рамопланіну - ramo3. Ramo1 і Ramo3 у КБГ рамопланіну є паралогами, що кодують пермеазну субодиницю ABC-транспортера. У всіх інших КБГ ЛДП, окрім КБГ черсинаміцину, ми знайшли тільки один відповідний гомолог, який кодує цей тип білка. Подібним чином ми виявили, що Ramo2 і Ramo23 також є гомологічними білками, які, найімовірніше, є АТФ-зв'язуючими субодини- цями ABC-транспортера; Ramo2 і Ramo23 мають лише по одному гомологу в інших КБГ ЛДП. Ми описали Ramo8 як АТФ-зв'язуючий ABC-транспортер, що містить як АТФазну, так і трансмембранну частини, і виявляє схожість до транспортерів, що кодуються в КБГ ГПА. Для Ramo8 ми змоделювали третинну структуру мономера, а також четвертинну структуру гомодимера цього білка. Також аналіз in silico виявив, що Ramo31 є протонним мембранним антипортером, віддалений гомолог якого закодований лише в КБГ черсинаміцину та, скоріш за все, цей білок не пов'язаний із біосинтезом рамопланіну.

Ключові слова: кластери біосинтетичних генів, рамопланін, мембранні транспортні білки, вторинні метаболіти, ґрунтові мікроорганізми

The fast rise of bacteria resistant to existing antibiotics outpaces the present pipeline of new drugs, creating a severe danger to our ability to treat infections successfully. In fact, the antibiotic pipeline is unequipped to deal with the growing bacterial resistance of current antimicrobials. For a long time, glycopeptide antibiotics (GPA), teicoplanin and vancomycin particularly, have represented the frontline treatment of infections caused by Gram-positive multidrugresistant (MDR) pathogens. The potent action of GPA antibiotics against Gram-positive bacteria depends on their remarkable ability to disrupt cell wall biosynthesis by specifically binding to the D-Ala-D-Ala motif of lipid II [34]. However, teicoplanin and vancomycin-resistant strains of staphylococci and enterococci have continued to emerge in the last decades [30].

Another promising class of antimicrobials are lipodepsipeptides (LDPs), exemplified with ramoplanin. The latter is produced by a soil-dwelling «rare» actinobacterium Actinoplanes ramoplaninifer ATCC 33076. Mode of action of this antibiotic is considered promising for the treatment of infections caused by Gram-positive MDR pathogens. Ramoplanin inhibits cell wall biosynthesis by binding to the lipid II and consequential inhibition of transglycosylation reactions [31]. The clinical development of ramoplanin was initially hampered due to its low local tolerability when injected intravenously. More recently, this LDP has been evaluated to treat Clostridioides difficile infections since ramoplanin is not-absorbable and achieves high colonic concentrations [25].

Today, except for ramoplanin, chemical structures for only two ramoplanin-like LDPs are known: enduracidin, which is produced by Streptomyces fungicidicus ATCC 21013 [8], and the recently discovered chersinamycin, produced by Micromonospora chersina DSM 44151 [23]. In addition, five more biosynthetic gene clusters (BGCs) of potential LDPs have recently been described in silico [23]. Despite antimicrobial potential of LDPs, corresponding BGCs remain poorly explored. To date, few aspects of ramoplanin and enduracidin biosynthesis, such as non-ri- bosomal synthesis of aglycone [10], mannosylation [4], halogenation [15], pathway-specific regulation [5] (only for enduracidin), and some others, have already been studied experimentally. However, the mechanisms that can potentially ensure the producer's resistance to its own secondary metabolite have escaped comprehensive elucidation.

Although extracellular lipid II is the most probable target of LDPs, evidence exists that LDPs can also bind lipid I, inhibiting intracellular lipid II biosynthesis [31]. Hence, intracellular accumulation of LDPs might be an issue for the producing culture. Transmembrane transporters encoded within LDP BGCs might contribute to the self-resistance of LDP producers. Some transport proteins have been associated with bacterial self-resistance to synthesized product, such as DrrA and DrrB in daunorubicin and doxorubicin producer Streptomyces peucetius ATCC 27952 [13]. DrrA and DrrB interact to generate an ATP-dependent efflux pump that transports daunorubicin and doxorubicin out of the cell, thereby conferring resistance [13]. With overexpression of arrA and arrB, doxorubicin synthesis was increased by a factor of 2,2 [18].

Recently, the features of the distribution, structure and phylogeny of ABC transporters in GPA and related peptide antibiotics BGCs were investigated [35]. Notably, these transporters share similarities in their amino acid composition, and categorization as MdlB(MsbA)-like, characterized by a six-helix N-terminal transmembrane domain [35]. Despite their widespread presence, experimental investigations into these transporters have primarily focused on tba gene from the BGC of balhimycin in Amycolatopsis balhimycina DSM 5908. This protein likely functions as a homodimer, and the knockout of the corresponding gene leads to an increased intracellular and decreased extracellular concentration of balhimycin [21].

The ramoplanin BGC was first described in 2005 [24], and after studies of mannosylation and halogenation of ramoplanin aglycone in 2016 [16, 17], no in silico or in vitro studies of the functions of the genes responsible for the synthesis of this antibiotic has been performed. Modern bioinformatics analysis benefits from a broad and constantly growing toolbox of data analysis resources and algorithmic approaches. Therefore, the purpose of this work was to investigate the properties of genes and encoded transport proteins in ramoplanin and related LDP BGCs using contemporary in silico methods. Such an evaluation can reveal new peculiarities of transporters encoded in ramoplanin and related LDP BGCs. Also, in this work, we offer the detailed description of all genes that code for transport proteins in the ramoplanin BGC, a comparison of the domain structures of their products, and distribution among other LDP BGCs.

Methods

BGC analysis. For nucleotide and amino acid sequence analysis, the nucleotide sequence of ramoplanin BGC (DD002243) [24] was used. All other LDP BGCs were obtained from genome sequences deposited in GenBank or the MIBiG repository under accession numbers: VFOE01000001 (Streptomyces sp. SLBN-134 Ga0314649_11), BGC0000341 (S. fungicidicus ATCC 21013), NZ_KB913037 (Amycolatopsis balhimycina FH 1894), CP016174 (Amycolatopsis orientalis B-37), LT629775 (Streptomyces sp. TLI_053), NZ_FMIB01000002 (M. chersina DSM 44151). Analysis of BGCs was carried out using the AntiSMASH 6.0.1 [1] and Geneious 4.8.5 [14] software.

Search for homologous proteins. A search for homologues of all genes encoding transport proteins of the ramoplanin BGC was conducted using the MIBiG repository [20] and the Protein BLAST [27]. The search for homologous proteins encoded in known LDP BGCs was performed using the built-in algorithm in the Geneious 4.8.5 software.

Predicting the domain organization of exporter proteins and modeling their location in the cell membrane. The presence of conservative domains was determined using CD-search algorithm from NCBI [17]. The transmembrane a-helices within the amino acid sequences of ABC transporters were identified using the TMHMM 2.0 software [16]. Subsequently, the two-dimensional topology of these transporters relative to the cytoplasmic membrane was reconstructed using TMRPres2D [29].

Modeling of the tertiary and quaternary structure of the ABC transporter Ramo8. The tertiary structure of Ramo8 was modeled based on of the experimentally determined crystal structure of the ABC transporter 9ACTN from Streptomyces sp SLBN-134 using the Swiss-mode server [32]. To model the quaternary structure - the Ramo8 homodimer - AlphaFold2-based prediction and visualization of secondary and 3D structures of proteins were used [26-28]. The best model according to AlphaFold2, was visualized using Mol* Viewer [28].

Results and Discussion

glycopeptide antibiotic bio-synthetic gene

We began by in silico searching for genes encoding transport proteins in ramoplanin BGC. In addition to the previously identified in silico putative transport protein genes - ramo1, ramo2, ramo8, ramo23, and ramo31 [6, 7, 17] - we discovered one more gene with a similar function in ramoplanin BGC, namely, ramo3. The literature lacks detailed functional characterization of these genes; however, given the domain structure of their products and similarity to ABC transporters, it is likely that they play a role in ramoplanin export. With these prerequisites, we will further describe the above-mentioned probable transporter genes, the domain structures of their encoded proteins, the homologous proteins found in MIBiG and using BLAST search, and, in particular, homologues found in other LDP BGCs using Geneiuos 4.8.5. This information, along with amino acid sequence identity (a.s.i,) to transporter proteins from ramoplanin BGC, is summarized in Table 1 and Table 2.

The ramo1 gene is 1002 bp long (the product is 333 aa). The protein encoded by this gene contains the ABC-2 transporter permease domain (75-198 aa, Fig. 1). The most similar described protein in the MIBiG database is the transmembrane transport protein from the BGC of dyne- micin in M. chersina (64 % a.s.i., 331 aa), and the closest homologue we found using the BLAST algorithm is the permease subunit of the ABC transporter in S. vitiensis (WP_018215178, 81 % a.s.i., 336 aa). According to the AntiSMASH analysis of S. vitiensis genome sequence (GenBank NZ_KB900388), no secondary metabolite BGCs are detected in this region (Table 2).

We found six similar gene products in the BGCs of LDP: TQL19422 (50,3 % a.s.i.) from Streptomyces sp. SLBN-134 Ga0314649_11, ABD65951 (50,0 % a.s.i.) from S. fungicidicus ATCC 21013, ctg1_5219 (48,8 % a.s.i.) from Am. balhimycina FH 1894, WP_037306103 (48,2 % a.s.i.) from Am. orientalis DSM 40040/KCTC 9412, ANN17109 (47,6 % a.s.i.) from Am. orientalis B-37, and SDT44233 (43,5 % a.s.i.) from Streptomyces sp. TLI_053. It is worth noting that no such protein was discovered in chersinamycin BGC (Table 1).

Table 1 Homologues of ramoplanin BGC transporter proteins encoded in other LDP BGCs

Table 2. Homologues of ramoplanin BGC transporter proteins found in MIBiG repository and using the BLAST search

Proteins of this type are typically one of the components of transport complexes of the ABC-2 type, which facilitate ATP-mediated transport of one or more diverse substrates. Well- known examples of such proteins include CcmB, responsible for transporting haeme in Escherichia coli and Mycobacterium tuberculosis, or DrrB in the doxorubicin producer S. peucetius [3].

Based on previous reports, ramo2 is predicted to code for an ATP-binding subunit of the ABC transporter complex [10]. The 915 bp coding sequence of this gene translates into a 304 aa polypeptide. In the MIBiG database, the most similar protein is also from the BGC of dynemicin, namely ACB47083 (77 % a.s.i., 305 aa). According to BLAST results, a similar protein is present in the genome of S. vitiensis (WP_018215179, 89 % a.s.i., 304 aa). The gene encoding this protein in S. vitiensis genome is located alongside the gene encoding Ramo1 homologue and is also not a part of any secondary metabolite BGC. Homologues are present in all LDP BGCs (Table 1).

The ATP-binding subunit of the ABC transporter protein has no transmembrane regions and operates by a mechanism that enables movement across membranes of almost any type of molecule, from large polypeptides to small ions. It can use a large number of other proteins as mediators [9]. Thus, a particular ABC-ATPase evolved specifically to function in complex with its cognate membrane protein. Together, they form a transport pathway necessary for the transport of a specific type of molecule, or in the case of some ABC transporters, multiple types of molecules. Specific transport molecules possess recognition motifs that allow them to bind selectively to their cognate transporters. Binding triggers conformational changes in the transporter, activating the ATPase activity and initiating the appropriate transport pathway for the bound substrate [11]. Ramo2 is most likely an ABC-ATPase capable of establishing one transport mechanism with a membrane transport protein; however, the precise function of such a protein in ramoplanin biosynthesis remains unknown.

We discovered a previously unidentified gene encoding an ABC transporter within ramo- planin BGC - ramo3. The nucleotide sequence of ramo3 (1011 bp) codes for a protein with 54 % a.s.i. to Ramo1. Accordingly, the products of these genes both have six transmembrane a-helices and the ABC-2 transporter permease domain (Fig. 1, the ABC-2 domain is marked in green).

Fig. 1. Secondary structure of ABC transporters Ramol (336 aa) and Ramo3 (321 aa). Prediction of transmembrane a-helices and placement of proteins in the cell membrane was carried out as described in Methods section. The ABC-2 domain of Ramo1 (75-198 aa) and Ramo3 (70-201 aa) is marked in green

As summarized in Table 2, the closest hit in the MIBiG database is the protein from the dynemicin BGC ACB47082 (71 % a.s.i., 339 aa). The closest homologue found using the BLAST algorithm is the protein WP_223874070 in S. vitiensis (83 % a.s.i., 336 aa), located not in secondary metabolite BGC. Analyzing Table 1, we can conclude that except for chersinamycin, all the other LDP BGCs contain only one homologue to both Ramo1 and Ramo3. Six similar gene products in the BGCs of LDP with a.s.i. percentage are listed: WP_037306103 (58,3 % a.s.i.) from Am. orientalis DSM 40040/KCTC 9412, ctg1_5219 (58,2 % a.s.i.) from Am. balhimycina FH 1894, ANN17109 (57,7 % a.s.i.) from Am. orientalis B-37, TQL19422 (56,3 % a.s.i.) from Streptomyces sp. SLBN-134 Ga0314649_11, ABD65951 (56,0 % a.s.i.) from 5. fungicidicus ATCC 21013, and SDT44233 (44,0 % a.s.i.) from Streptomyces sp. TLI_053.

The ramo8 gene (1923 bp long) codes for the most similar protein to the typical ABC transporters present in GPA BGCs. This gene product contains the MdlB (MsbA) superfamily domain, characterizing this protein as an ATP-binding ABC transporter containing both ATPase and transmembrane parts. The presence of this domain characterizes this protein as a probable component of the antibiotic transport system [35]. Genes encoding exporter proteins with the described structure are present in all sequenced glycopeptide BGC [6].

As seen in Table 2, in the MIBiG database, the greatest similarity of Ramo8 was found with the ABC transporter present in the BGC of enduracidin (ABD65953, 651 aa, 72,5 % a.s.i.). The most similar ABC transporter found using BLAST belongs to Micromonospora sp. MH33 (WP_107078706, 642 aa, 78 % a.s.i.). In this region of the Micromonospora genome, AntiSMASH annotates possible BGC for type III polyketide synthase metabolite. Homologous proteins are also present in all other LDP BGCs (Table 1): WP_091321314 (77,3 % a.s.i.) from M. chersina DSM 44151, WP_037306101 (73,7 % a.s.i.) from Am. orientalis DSM 40040/KCTC 9412, ANN21821 (73,7 % a.s.i.) from Am. orientalis B-37, ctg1_5221 (74,4 % a.s.i.) from Am. balhimycina FH 1894, TQL19420 (72,5 % a.s.i.) from Streptomyces sp. SLBN-134 Ga0314649_11, and SDT44201 (63,3 % ide a.s.i.) from Streptomyces sp. TLI_053. For comparison, we also chose previously described [35] ABC transporter encoded in teicoplanin BGC Tei4*, which has 54,3 % a.s.i. to Ramo8.

The secondary structure of Ramo8, ABD65953, WP_091321314, PSK62646, and Tei4* ABC transporters with color-coded functional motifs is shown in Fig. 2. The C-terminal ATPase domains have a complete set of motifs [33] necessary for their functioning, these include the Walker motif A (P-loop), the Q-loop, the Walker motif B, the D-loop, the H-loop, and the signature motif of the ABC transporter [35].

Fig. 2. Secondary structure of ABC transporters Ramo8, ABD65953, WP_091321314, PSK62646, and Tei4*. Prediction of transmembrane a-helices and placement of proteins in the cell membrane was carried out as described in Methods section. Colors indicate C-terminal functional motifs (see the legend to the figure)

It is interesting to note that almost all ABC transporters encoded in the glycopeptide antibiotics BGCs have an N-terminal transmembrane domain with six a-helices, in contrast to the ABC transporter encoded in the ramoplanin BGC, which, according to the prediction of the secondary structure, has a transmembrane domain consisting of five a-helices (the region corresponding to the fourth transmembrane a-helix is missing). Corresponding ABC transporters encoded in chersinamycin and enduracidin BGCs have six a-helices. This arrangement of a-helices in Ramo8 can be explained by the inaccuracy of the construction of the secondary structure model. When constructing the 3D model of Ramo8, the third and fourth a-helices are formed but do not cross the membrane completely (Fig. 3). They can form a kind of «anchor», which may be important for stabilizing the protein structure in the membrane.

Fig. 3. Model of the tertiary structure of the ABC transporter Ramo8. The tertiary structure of Ramo8 was modeled based on the experimentally established crystal structure of the ABC transporter 9ACTN from Streptomyces sp SLBN-134 using the Swiss-mode server [32]. Elements of secondary structures are highlighted in colors

Based on the observed homodimer behavior of many ABC transporters [21], we constructed a homodimer model for Ramo8. AlphaFold2 analysis reveals a highly plausible homodimeric architecture for Ramo8, characterized by precise positioning of all transmembrane a-helices and critical components within the C-terminal ATPase domain active site (Fig. 4). The quaternary structure of Ramo8, as illustrated in Fig. 4, conforms to the characteristic features of ABC exporters. They are homodimers, each transmembrane domain contains six transmembrane a-helices [36].

The ramo23 (930 bp) gene codes for the ATP-binding subunit of the ABC transporter, as discovered previously in silico [23]. The product of the ramo23 gene contains the CcmA domain, which is also observed in Ramo2. As summarized in Table 2, the most similar described protein in the MIBiG database is the ATP-binding subunit of the ABC transporter encoded in the BGC of dynemicin in M. chersina (59 % a.s.i.). The closest homologue found by the BLAST algorithm is the ATP-binding subunit of the ABC transporter in Micromonospora sp. A3M-1- 15 (WP_254341048, 60 % a.s.i.). The gene for this transporter is not located in any secondary metabolite BGC. The exact functions of these transporter proteins are not known, but it is most likely that they participate in the transport of substances through the membrane in a complex with a membrane protein.

Fig. 4. Model of the quaternary structure of the ABC transporter Ramo8. The model was built using the AlphaFold2 service [12]. Structural elements are highlighted in colors (see the legend to the figure)

The Ramo23 shares 59,9 % a.s.i. to Ramo2. Correspondingly, in each LDP BGCs, there is one gene encoding homologue of both Ramo23 and Ramo2.

The last gene encoding a transport protein in the ramoplanin BGC is ramo31 (1290 bp). The in silico predicted product of this gene is a proton membrane antiporter. The most similar described protein in the MIBiG database is a membrane antiporter protein encoded in tiacumicin B BGC in D. aurantiacum subsp. hamdenensis (82 % a.s.i.), and the closest homologue found by the BLAST algorithm is a proton antiporter in A. deccanensis (81 % a.s.i.), which gene is located in tiacumicin-like BGC (see Table 2).

Ramo31 contains the domain of the PLN03159 superfamily, characterized by the CD- search algorithm as a cation-proton antiporter that performs the functions of maintaining cation homeostasis and cell pH. Among all LDP BGCs, only chersinamycin BGC encodes a homologous protein WP_091305532 with 34,8 % a.s.i. to Ramo31 (Table 1). The secondary structures of mentioned proteins are shown in Fig. 5.

Fig. 5. Secondary structure of antiporter proteins Ramo31 (429 aa), WP_091305532 (M. chersina, 453 aa), ADU86006. (D. aurantiacum subsp. hamdenensis, 428 aa). The prediction of transmembrane a-helices and the localization of proteins in the cell membrane was performed as described in Materials section

The structure of the above-mentioned proteins is partially different. Ramo31 has 13 transmembrane a-helices, WP_091305532 has 12 transmembrane a-helices, and ADU86006 has 11 transmembrane a-helices. In addition to the PLN03159 superfamily domain, Ramo31 also contains KefB domains (typical for the potassium ion transport system described in E. coli, involved in the transport and metabolism of inorganic ions) and the Na+-H+ exchanger domain cl01133. The canonical E. coli KefB protein has a structure similar to Ramo31, with 13 transmembrane a-helices, but is characterized by the presence of the PRK03659 superfamily domain instead of PLN03159. The PRK03659 domain is typical for proteins responsible for potassium exchange and is regulated by glutathione adducts, leading to transient acidification of the cytoplasm [2].

The ramo31 gene product appears to be involved in cation transport and the regulation of homeostasis and pH. However, whether this protein is directly related to ramoplanin biosynthesis and whether its inclusion in the ramoplanin BGC is warranted remains to be determined.

Analysis of LDP BGCs reveals a single gene encoding a protein homologous to ramol and ramo3 products in all cases, except for chersinamycin BGC, which does encode a similar protein at all (Fig. 6).

From Fig. 6 we can also conclude that there is only one corresponding protein homologous to the ramo2 and ramo23 gene products in all LDP BGCs. All BGCs, except for chersinamycin, also lack a homologue of the proton membrane antiporter encoded by the ramo31 gene.

It is intriguing to consider that some gene products in the ramoplanin biosynthetic pathway might not be directly involved in the final product export, but their contributions to other vital processes are likely significant.

In all LDP BGCs, except for ramoplanin BGC, only one gene encoding ATP-binding subunit and one gene for permease subunit of the ABC transporter are present. In these BGCs, the corresponding genes are located side by side, and corresponding proteins likely form one transport system, although the substrate of this system in unclear. This is why it is difficult to explain

the presence of the two genes for ATP-binding subunit and two for permease subunit in ramo- planin in BGC. It is possible that one functional pair is not essential for antibiotic export.

Fig. 6. Schematic arrangement of genes encoding transport proteins in LDP BGCs. Different types of transport proteins are indicated by colors (see the legend to the figure)

The precise explanation for this phenomenon is yet to be determined, but the chersinamycin BGC lacks the gene encoding the permease subunit, although it possesses a gene for the ATP- binding subunit of the ABC transporter. We did not find a corresponding gene for the permease subunit of the ABC transporter in the genomic region near chersinamycin BGC either.

The Ramo31 homologue is present only in chersinamycin BGC with low percentage of a.s.i. (34,8 %). Also, considering the probable function and location of the genes encoding these transport proteins on the edges of both BGCs, we can assume that they do not participate in the export of the corresponding antibiotics.

On the contrary, the observed features of Ramo8, including its potential for homodimerization, suggest it plays a crucial role in ramoplanin transport, likely acting as a key component of the antibiotic's export system across the cell membrane.

References

1. Blin K., Shaw S., Kloosterman A. M. et al. AntiSMASH 6.0: improving cluster detection and comparison capabilities // Nucleic Acids Res. 2021. Vol. 49. P. 29-35. doi: 10.1093/nar/ gkab335

2. Booth I. R. The regulation of intracellular pH in bacteria // Novartis Found. Symp. 1999. Vol. 221. P. 19-28. doi: 10.1002/9780470515631.ch3.

3. Chang G., Chen L., Dassa E. et al. Structural and functional diversity calls for a new classification of ABC transporters // FEBS Letters. 2021. Vol. 594. P. 3767-3775. doi: 10.1002/18733468.13935.

4. Chen J. S., Wang Y.-X., Shao L. et al. Functional identification of the gene encoding the enzyme involved in mannosylation in ramoplanin biosynthesis in Actinoplanes sp. // Biotech- nol. Lett. 2013. Vol. 35. P 1501-1508. doi: 10.1007/s10529-013-1233-3.

5. Chen Y. W., Liu X. C., Lv F X., Li P. Characterization of three regulatory genes involved in enduracidin biosynthesis and improvement of enduracidin production in Streptomyces fungi- cidicus // J. Appl. Microbiol. 2019. Vol. 127. P 1698-1705. doi: 10.1111/jam.14417.

6. Donadio S., Sosio M., Stegmann E. et al. Comparative analysis and insights into the evolution of gene clusters for glycopeptide antibiotic biosynthesis // Mol. Genet. Genomics. 2005. Vol. 274. P 40-50. doi: 10.1007/s00438-005-1156-3.

7. Han J., Chen J., Shao L. et al. Production of the ramoplanin activity analogue by double gene inactivation // PLoS One. 2016. Vol. 11. P 1-11. doi: 10.1371/journal.pone.0154121.

8. Higashide E., Hatano K., Shibata M., Nakazawa K. Enduracidin, a new antibiotic. Streptomyces fungicidicus no. b 5477, an enduracidin producing organism // J. Antibiot. 1968. Vol.

21. P 126-137. doi: 10.7164/antibiotics.21.126.

9. Higgins C. F ABC transporters: from microorganisms to man // Annu. Rev. Cell Biol. 1992. Vol. 8. P 67-113. doi: 10.1146/annurev.cb.08.110192.000435.

10. HoertzA. J., Hamburger J. B., Gooden D. M. et al. Studies on the biosynthesis of the lipodep- sipeptide antibiotic ramoplanin A2 // Bioorganic Med. Chem. 2012. Vol. 20. P 859-865. doi: 10.1016/j.bmc.2011.11.062.

11. Holland I. B., Blight M. A. ABC-ATPases, adaptable energy generators fuelling transmembrane movement of a variety of molecules in organisms from bacteria to humans // J. Mol. Biol. 1999. Vol. 293. P 381-399. doi: 10.1006/jmbi.1999.2993.

12. Jumper J., Evans R., Pritzel A. et al. Highly accurate protein structure prediction with AlphaFold // Nature. 2021. Vol. 596. P 583-589. doi: 10.1038/s41586-021-03819-2.

13. Kaur P. Expression and characterization of DrrA and DrrB proteins of Streptomyces peuce- tius in Escherichia coli: DrrA is an ATP binding protein // J. Bacteriol. 1997. Vol. 179. P 569-575. doi: 10.1128/jb.179.3.569-575.1997.

14. Kearse M., Moir R., Wilson A. et al. Geneious basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data // Bioinformatics. 2012. Vol. 28. P 1647-1649. doi: 10.1093/bioinformatics/bts199.

15. Kittila T., Kittel C., Tailhades J. et al. Halogenation of glycopeptide antibiotics occurs at the amino acid level during non-ribosomal peptide synthesis // Chem. Sci. 2017. Vol. 8. P 5992-6004. doi: 10.1039/c7sc00460e.

16. Krogh A., Larsson B., Von Heijne G., Sonnhammer E. L. L. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes // J. Mol. Biol. 2001. Vol. 305. P 567-580. doi: 10.1006/jmbi.2000.4315.

17. Lu S., Wang J., Chitsaz F. et al. CDD/SPARCLE: the conserved domain database in 2020 // Nucleic Acids Res. 2020. Vol. 48. P. 265-268. doi: 10.1093/nar/gkz991.

18. Malla S., Niraula N. P., Liou K., Sohng J. K. Self-resistance mechanism in Streptomyces peucetius: overexpression of drrA, drrB and drrCfor doxorubicin enhancement // Microbiol. Res. 2010. Vol. 165. P. 259-267. doi: 10.1016/j.micres.2009.04.002.

19. McCafferty D. G., Cudic P., Frankel B. A. et al. Chemistry and biology of the ramopla- nin family of peptide antibiotics // Biopolym. - Pept. Sci. Sect. 2002. Vol. 66. P. 261-284. doi: 10.1002/bip.10296.

20. MedemaM. H., Kottmann R., Yilmaz P. et al. Information about a biosynthetic gene cluster // Nat. Chem. Biol. 2017. Vol. 11. P. 625-631. doi: 10.1038/nchembio.1890.Minimum.

21. Menges R., Muth G., Wohlleben W., Stegmann E. The ABC transporter Tba of Amycolatopsis balhimycina is required for efficient export of the glycopeptide antibiotic balhimycin // Appl. Microbiol. Biotechnol. 2007. Vol. 77. P. 125-134. doi: 10.1007/s00253-007-1139-x.

22. Mirdita M., Schutze K., Moriwaki Y. et al. ColabFold: making protein folding accessible to all // Nat. Methods. 2022. Vol. 19. P. 679-682. doi: 10.1038/s41592-022-01488-1.

23. Morgan K. T., Zheng J., McCafferty D. G. Discovery of six ramoplanin family gene clusters and the lipoglycodepsipeptide chersinamycin // ChemBioChem. 2021. Vol. 22. P. 176-185. doi: 10.1002/cbic.202000555.

24. Pat. EP 1 326 983 B1 Europian Union. Gene cluster for ramoplanin biosynthesis / Farnet C. M., zazopulous E., Staffa A. Stat. 15.10.2001. Pub. 18.04.2002. P.1-222. WO 2002/031155.

25. Petrosillo N., Granata G., CataldoM. A. Novel antimicrobials for the treatment of Clostridium difficile infection // Front. Med. 2018. Vol. 5. doi: 10.3389/fmed.2018.00096.

26. Pettersen E. F, Goddard T D., Huang C. C. et al. UCSF ChimeraX: structure visualization for researchers, educators, and developers // Protein Sci. 2021. Vol. 30. P. 70-82. doi: 10.1002/ pro.3943.

27. Sayers E. W., Bolton E. E., Brister R. J. et al. Database resources of the national center for biotechnology information // Nucleic Acids Res. 2022. Vol. 50. P. 20-26. doi: 10.1093/nar/ gkab1112.

28. Sehnal D., Bittrich S., Deshpande M. et al. Mol. Viewer: modern web app for 3D visualization and analysis of large biomolecular structures // Nucleic Acids Res. 2021. Vol. 49. P. W431-W437. doi: 10.1093/nar/gkab314.

29. SpyropoulosI. C, Liakopoulos T. D., BagosP. G., Hamodrakas S. J. TMRPres2D: High quality visual representation of transmembrane protein models // Bioinformatics. 2004. Vol. 20. P. 3258-3260. doi: 10.1093/bioinformatics/bth358.

30. Unni S., Siddiqui T J., Bidaisee S. et al. Reduced susceptibility and resistance to vancomycin of Staphylococcus aureus: a review of global incidence patterns and related genetic mechanisms // Cureus. 2021. Vol. 13. doi: 10.7759/cureus.18925.

31. Walker S., Chen L., Hu Y. et al. Chemistry and biology of ramoplanin: a lipoglycodepsipep- tide with potent antibiotic activity // Chem. Rev. 2005. Vol. 105. P. 449-475. doi: 10.1021/ cr030106n.

32. Waterhouse A., Bertoni M., Bienert S. et al. Swiss-model: homology modelling of protein structures and complexes // Nucleic Asid Res. 2018. Vol. 46. P. 296-303. doi: 10.1093/nar/ gky427.

33. Westfahl K. M., Merten J. A., Buchaklian A. H., Klug C. S. Functionally important ATP- binding and hydrolysis sites in Escherichia coli // Biochem. 2008. Vol. 47. P. 13878-13886. doi: 10.1021/bi801745u

34. Yushchuk O., Ostash B., Truman A. W et al. Teicoplanin biosynthesis: unraveling the interplay of structural, regulatory, and resistance genes // Appl. Microbiol. Biotechnol. 2020. Vol. 104. P. 3279-3291. doi: 10.1007/S00253-020-10436-Y.

35. Yushcuk O., Zhukrovska K., Fedorenko V. Insights into the phylogeny of transporters coded within biosynthetic gene clusters for glycopeptides and related antibiotics // Visnyk Lviv Univ. Biol. Ser. 2022. Vol. 86. P 33-46. doi.org/10.30970/vlubs.2022.86.

36. Zolnerciks J. K., Andress E. J., Nicolaou M., Linton J. K. Structure of ABC transporters // Essays Biochem. 2011. Vol. 50. P 43-61. doi: 10.1042/BSE0500043.

Размещено на Allbest.ru