Structural and functional analyses of the echinomycin resistance conferring protein Ecm16 from Streptomyces lasalocidi

Echinomycin is a natural product DNA bisintercalator antibiotic. The echinomycin biosynthetic gene cluster in Streptomyces lasalocidi includes a gene encoding the self-resistance protein Ecm16. Here, we present the 2.0 Å resolution crystal structure of Ecm16 bound to adenosine diphosphate. The structure of Ecm16 closely resembles that of UvrA, the DNA damage sensor component of the prokaryotic nucleotide excision repair system, but Ecm16 lacks the UvrB-binding domain and its associated zinc-binding module found in UvrA. Mutagenesis study revealed that the insertion domain of Ecm16 is required for DNA binding. Furthermore, the specific amino acid sequence of the insertion domain allows Ecm16 to distinguish echinomycin-bound DNA from normal DNA and link substrate binding to ATP hydrolysis activity. Expression of ecm16 in the heterologous host Brevibacillus choshinensis conferred resistance against echinomycin and other quinomycin antibiotics, including thiocoraline, quinaldopeptin, and sandramycin. Our study provides new insight into how the producers of DNA bisintercalator antibiotics fend off the toxic compounds that they produce.

www.nature.com/scientificreports/ Echinomycin is a prototypical DNA bisintercalator produced by multiple actinomycetes, including Streptomyces echinatus and Streptomyces lasalocidi (formerly known as S. lasaliensis) [7][8][9] . It is a cyclic depsipeptide that contains two quinoxaline groups and an unusual thioacetal bridge (Fig. 3) 10 . Echinomycin shows potent antimicrobial activity against methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococci 11,12 but it is not used clinically due to solubility and toxicity issues. The echinomycin biosynthetic gene cluster from S. lasalocidi contains genes that encode for enzymes that synthesize the quinoxaline group, enzymes that construct the peptide backbone, and genes that encode for proteins with unknown function 13 . One of the functionally uncharacterized proteins is Ecm16, which was postulated to provide self-protection against echinomycin based on the sequence identity (~ 30%) it shares with the prokaryotic UvrA proteins that function in the NER pathway 13 .
We have previously reported the in vivo and in vitro functional characterization of Ecm16 14 . The main findings of that study are (1) the echinomycin sensitive Escherichia coli K12 becomes echinomycin resistant upon transformation with the ecm16 encoding plasmid, (2) Ecm16 does not require participation of the NER proteins UvrA, UvrB, UvrC, or UvrD to provide echinomycin resistance, (3) Ecm16 does not complement UvrA function, (4) Ecm16's ATPase activity is essential for its anti-echinomycin activity, (5) Ecm16 binds to doublestranded DNA in a nucleotide sequence independent manner, and (6) Ecm16 binds to echinomycin-containing DNA ~ two-fold more strongly than to echinomycin-free DNA. In the current study, we have determined the crystal structure of Ecm16 to provide a structural context to its function. We have also performed mutational studies to dissect the role of Ecm16's insertion domain. In UvrA proteins, the insertion domain is involved in damage-specific DNA binding 15 . Lastly, we have probed the substrate specificity of Ecm16 by challenging ecm16expressing cells with a series of quinomycin and non-quinomycin DNA targeting antibiotics.

Results
X-ray crystal structure of Ecm16. We have determined the structure of Ecm16 bound to adenosine diphosphate (ADP) at 2.0 Å resolution (Fig. 1a). Our final model consists of the Ecm16 homodimer, four ADP, two Mg 2+ , four Zn 2+ , and 612 water molecules (Table 2). Some residues, including 183-293 of chain A and 185-295 of chain B that includes the entire insertion domain, were not modelled due to missing electron density (Supplementary Table 1). Each protomer of Ecm16 contains two ABC ATPase motifs, referred to as nucleotidebinding domain I and II (NBD-I and NBD-II). NBD-I consists of the ATP-binding I domain, signature I domain, and insertion domain. The insertion domain was not visible in the crystal structure, presumably because it is disordered. NBD-II consists of the ATP-binding II domain and the signature II domain. NBD-II is missing the helix-turn-strand that corresponds to residue 66-99 of NBD-I ( Supplementary Fig. 1). Other than these two differences, NBD-I and NBD-II have a relatively similar overall structure (RMSD = 1.5 Å for 208 C α atoms). The dimer interface of Ecm16 buries ~ 3900 Å 2 of surface area and is comprised of residues from the ATP-binding I, signature I, and signature II domains ( Supplementary Fig. 2). The ventral side of Ecm16 features an extended groove that is lined with numerous basic residues (K136, K143, K381, R384, R567, K568, R537, K549, K572, K577) ( Supplementary Fig. 3). This ~ 10 nm long and ~ 2 nm wide groove can potentially accommodate a ~ 32 bp B-form DNA and provides a structural basis for the previously reported DNA-binding activity of Ecm16 14 .
Ecm16 has a total of four nucleotide-binding sites, two proximal and two distal nucleotide-binding sites (Fig. 1a). The proximal nucleotide-binding site is located ~ 19 Å from the Ecm16 dimer interface and it is sandwiched between ATP-binding domain I and signature domain II. The distal nucleotide-binding site is located ~ 26 Å from the dimer interface and it is sandwiched between ATP-binding domain II and signature domain I. Each ATP-binding domain contains a Walker A motif, Walker B motif, and the α-helical ABC signature subdomain containing the LSGGQ sequence typically found in ABC transporters 16 and DNA repair proteins 17 . The ATP-binding and signature domains are connected by the Q-loop, which is the site of major conformational change and coupling of energy converting domains of NBD on other ATPases 18 . Electron density at both the proximal and distal nucleotide-binding sites showed the presence of ADP ( Supplementary Fig. 4). To further confirm the identity of the nucleotide bound to Ecm16, we performed liquid chromatography analysis of the protein extract. Only ADP was detected in this experiment ( Supplementary Fig. 5), indicating that the nucleotide observed in the Ecm16 crystal structure is ADP, and not ATP.
Mg 2+ ions are observed only at the two proximal nucleotide-binding sites, and not at the distal sites, even though Ecm16 was crystallized in the presence of 10 mM MgCl 2 (Fig. 1b, c). The reason for this is not clear but a Table 1. UvrA-like proteins expressed by producers of quinomycin antibiotics. a Resistance activity of Ecm16 and DrrC have been experimentally confirmed. Others are predictions based on sequence similarity and the natural product biosynthetic gene cluster that the encoding gene is found in. www.nature.com/scientificreports/  www.nature.com/scientificreports/ similar observation was made for the crystal structure of UvrA from Bacillus stearothermophilus 19 . The phosphate groups of ADP participate in an extensive hydrogen bond network involving the residues of the Walker A motif, while the ribose sugar and adenine base form relatively few interactions with Ecm16. The conserved histidine residue at position 501 stacks well against the adenine ring of ADP at the distal site. Each Ecm16 protomer contains two zinc-binding modules, which correspond to the UvrA zinc-binding module 2 and 3 observed in all UvrA crystal structures reported so far. In module 2, Zn 2+ is coordinated to C176, C179, C296, and C299, while in module 3, Zn 2+ is coordinated to C589, C592, C612, and C615 (Fig. 1d). These zinc-coordinating residues are conserved in UvrA and UvrA2 proteins 14,15 . The three-dimensional structure of Ecm16 resembles that of UvrA from B. stearothermophilus (RMSD = 2.6 Å for 1002 C α atoms) and UvrA2 from Deinococcus radiodurans (RMSD = 1.7 Å for 933 C α atoms) (Fig. 1e, Supplementary Fig. 6). The structural similarity of Ecm16, UvrA, and UvrA2 explains their common functionalities such as DNA binding and ATP hydrolysis 15,20 . However, Ecm16, like UvrA2, lacks the UvrB-binding domain and its associated zinc-binding module 1 which are found in all UvrA proteins, indicating that Ecm16 does not interact with UvrB from the NER pathway. This is consistent with our previous report that both the wild type  Table 2). Electrophoretic mobility shift assay (EMSA) showed that Ecm16 bound more tightly to echinomycin-bound DNA than normal DNA, whereas Ecm16-Δ ID did not bind to either type of DNA (Fig. 2a). Next, we measured the DNA binding affinity of Ecm16 and Ecm16-Δ ID using fluorescence polarization. The dissociation constant for Ecm16-DNA-echinomycin and Ecm16-DNA was 11.8 nM and 60.2 nM, respectively (Fig. 2b, Table 3). For Ecm16-Δ ID , no binding was observed for either DNA substrate. Therefore, EMSA and fluorescence polarization both showed that the insertion domain of Ecm16 is required for DNA binding. Next, we prepared Ecm16* in which the insertion domain of Ecm16 was exchanged with the insertion domain of DrrC, an Ecm16 homolog from Streptomyces peucetius (Supplementary Figs. 7,8). DrrC was reported to confer resistance against the DNA monointercalator antibiotic daunorubicin, although the molecular mechanism of DrrC is not known 22 . The insertion domain of Ecm16 and DrrC share 32% amino acid sequence identity. Ecm16* bound to echinomycin-containing DNA 2.4-fold more tightly than to normal DNA (K D = 37.7 nM vs. 90.8 nM) (Fig. 2b, Table 3). This result indicates that having a homologous insertion domain is sufficient for Ecm16 to distinguish echinomycin-bound DNA from normal DNA through differential binding. To investigate whether Ecm16* has anti-echinomycin activity, E. coli K12 cells expressing Ecm16* were challenged with 10 µM www.nature.com/scientificreports/ echinomycin. Ecm16* expressing cells were sensitive to echinomycin (Fig. 2d), indicating that the native insertion domain must be present to provide anti-echinomycin activity.
DNA-echinomycin stimulates the ATP hydrolysis activity of Ecm16. We reported previously that the ATP hydrolysis activity of Ecm16 is required to render echinomycin resistance in vivo 14 Fig. 10), which is consistent with the inability of Ecm16Δ ID to bind DNA (Fig. 2a, b). These results indicate that only the native DNA substrate stimulates the ATP hydrolysis activity of Ecm16 and that this property is lost when the insertion domain is deleted or when it is substituted with the insertion domain of a homologous protein. Therefore, the insertion domain plays an important role in determining Ecm16's substrate specificity and in supporting Ecm16's anti-echinomycin activity.
Ecm16 provides resistance against a variety of quinomycin antibiotics. We probed the substrate specificity of Ecm16 by testing whether Ecm16 can provide resistance against other DNA-binding drug molecules-doxorubicin, mitomycin C, daunorubicin, actinomycin D, cisplatin, thiocoraline, quinaldopeptin, and sandramycin. Because the permeability of these compounds is limited in Gram negative (diderm) bacteria, we used the Gram positive (monoderm) bacterium Brevibacillus choshinensis, instead of E. coli K12. Ecm16expressing B. choshinensis cells displayed resistance only against the DNA bisintercalator antibiotics echinomycin, thiocoraline, quinaldopeptin, and sandramycin (Fig. 3). The degree of resistance provided by Ecm16 was most pronounced at the highest antibiotic concentration tested. Cells containing the control vector grew very slowly in the presence of 0.1 µM echinomycin, 4 uM thiocoraline, 6 µM quinaldopeptin, or 2 µM sandramycin making it impossible to determine the doubling time, whereas cells expressing Ecm16 had doubling times which were more similar to cells which were not treated with the respective antibiotic (only 1.1-to 1.2-fold longer) (Fig. 3). Our result showed that Ecm16 is most effective against echinomycin, but it also provides some resistance against other structurally similar quinomycin antibiotics.

Discussion
Here we report the crystal structure of Ecm16 from the echinomycin producer S. lasalocidi. Ecm16 is a homolog of UvrA, the DNA damage sensor protein from the prokaryotic NER pathway. The main structural difference between Ecm16 and UvrA is that Ecm16 lacks the UvrB-binding domain and a zinc-binding module which are present in all UvrA structures reported to date (PDB ID: 2R6F, 3PIH, 3UWX, 3UX8, 3ZQJ, 6N9L). Another potential structural difference is the conformation of the ~ 100 residue insertion domain, although this remains to be verified since the insertion domain is not visible in the Ecm16 crystal structure, presumably because this domain is mobile in the absence of a bound DNA substrate. Overall, the three-dimensional structure of Ecm16 and UvrA are highly similar. They share the same protein fold and they both contain four ATP-binding sites and one continuous DNA-binding groove. Accordingly, Ecm16 and UvrA both display ATPase activity and bind double-stranded DNA 14 . However, Ecm16 lacks the UvrB-binding domain and its associated zinc-binding module, suggesting that Ecm16 and UvrA have distinct molecular mechanisms acquired potentially through divergent evolution. Ecm16Δ ID , which lacks the insertion domain, failed to bind DNA. Furthermore, expression of Ecm16Δ ID in E. coli K12 did not protect the cells from echinomycin. Ecm16*, which possesses the insertion domain from the daunorubicin resistance protein DrrC, showed 2.4-fold higher binding affinity to echinomycin-bound DNA than normal DNA. However, Ecm16*, in contrast to Ecm16, did not display the dramatic increase in ATP hydrolysis rate in the presence of echinomycin-containing DNA. Based on these results, we propose a two-step model for detection of echinomycin-bound DNA by Ecm16. In the first step, Ecm16 discriminates echinomycin-bound DNA from normal DNA by differential DNA binding affinity. This initial screening step requires the presence www.nature.com/scientificreports/ of the insertion domain independent of sequence. In the second step, Ecm16-bound DNA substrates are further discriminated by their ability to stimulate the ATPase activity of Ecm16. This second step appears to require an insertion domain that is specifically matched to echinomycin. Additional structural studies are needed to determine how these two steps are achieved at the molecular level. In particular, atomic structure of Ecm16 in complex with echinomycin-bound DNA and with echinomycin-free DNA will help decipher the molecular mechanism. Interestingly, Ecm16 also provided protection against the natural product DNA bisintercalators thiocoraline, quinaldopeptin, and sandramycin. This is reminiscent of UvrA protein's ability to detect a wide variety of DNA lesions. Understanding antibiotic resistance mechanisms is important because it enables the development of new therapeutic strategies. Our work has started to unravel a potentially novel antibiotic resistance mechanism, but further studies are needed to fully understand how Ecm16 confers echinomycin resistance. This includes determining the structure of Ecm16 bound to various DNA substrates and identifying, if any, Ecm16's partner proteins. Assuming Ecm16 requires partner proteins, they are likely to be proteins which are conserved throughout different phylogenetic lineages since Ecm16 confers echinomycin resistance when expressed in three distantly related organisms, S. lasalocidi, E. coli, and B. choshinensis.  Ecm16/Ecm16-Δ ID /Ecm16* cloning and protein expression. Codon optimized genes for ecm16 and ecm16-Δ ID were subcloned into the pUC19 vector using the NdeI and EcoRI sites (GenScript) for expression in E. coli. The genes were digested using enzymes NdeI and EcoRI and gel purified. pET28a (+) expression vector was cut with NdeI and EcoRI and gel purified. The ecm16 and ecm16-Δ ID inserts were ligated into the pET28a(+) vector at a 3:1 insert:vector ratio using the quick Ligation Kit (New England Biolabs Inc). Ligation products were transformed into chemically competent E. coli DH5α cells and grown overnight on LB-kan plates (50 µg ml −1 ) at 37 °C. E. coli BL21 (DE3) (Novagen) cells were transformed with the expression vector and then cultured to exponential phase at 37 °C in Luria Bertani (LB) medium containing 50 µg/ml kanamycin. Ecm16* was cloned into pET28a (+) vector using the same procedure using EcoR1 and HindII restriction digestion enzymes at 5′ and 3′ site respectively. Expression of ecm16 was induced using 0.25 mM isopropyl-d-1-thiogalactopyranoside (Thermo Scientific) at an optical density at OD 600 of 0.6-0.8. Cells were further grown for 16 h at 18 °C and then harvested by centrifugation at 6000 × g for 15 min at 4 °C. Cell pellet was resuspended in lysis buffer (50 mM HEPES pH 7.5, 200 mM NaCl, 10 mM imidazole, and 5% glycerol (v/v), 1 mM phenylmethylsulphonyl fluoride, 10 µg/ml DNase and 10 mM MgCl 2 ) and lysed by sonication. Cell lysate was centrifuged at 30,000 × g for 60 min at 4 °C and insoluble material was removed. Soluble fraction was loaded on a 5 ml Ni-NTA column (GE Healthcare) that was equilibrated with binding buffer (50 mM HEPES pH 7.5, 200 mM NaCl), washed with 250 ml washing buffer (50 mM HEPES pH 7.5, 200 mM NaCl, 30 mM imidazole, and 5% (v/v) glycerol), and bound protein was eluted with elution buffer (50 mM HEPES pH 7.5, 200 mM NaCl, 250 mM imidazole). Eluted protein solution was applied to a 5 mL HiTrap QHP HP column (GE Healthcare), and bound protein was eluted using a linear 50-500 mM NaCl gradient. Lastly, the Ecm16 samples were purified by size exclusion chromatography in 50 mM HEPES pH 7.5, 50 mM NaCl using a Superdex 200 10/300 GL (GE Healthcare). The purity of protein was assayed by SDS-PAGE. Protein samples were concentrated using Amicon 10-kDa MWCO centrifugal filter. All purification steps were performed at 4 °C. Ecm16-Δ ID and Ecm16* variants were prepared in the same manner as the wild type Ecm16 protein.
For in vivo studies using E. coli, codon-optmized ecm16 and ecm16-Δ ID from previously constructed pET28a vectors were amplified with SacI and EcoRI sites using Phusion polymerase (New England Biolabs). The genes were electrophoretically separated on a 0.7% agarose gel and gel purified. Following purification the genes were digested with SacI and EcoRI (New England Biolabs) and cloned into a similarly digested pBAD-Myc-HisA vector (Thermo Fisher). Following digestion, the vector was treated with shrimp alkaline phosphatase (New England Biolabs). Both the digested vector and digested fragments were electrophoretically separated on a 0.7% agarose gel and gel purified again.The fragments were ligated into the pBAD vector using T4 DNA ligase (New England Biolabs), then transformed into electrocompetent DH-5α cells and grown overnight on LB-amp plates (100 µg ml −1 ) at 37 °C. E. coli BW25113 (Coli Genetic Stock Center) cells were transformed with the expression vector and then cultured to exponential phase at 37 °C in LB medium containing 50 µg/ml ampilcillin. Construction of BW25113 cells with ecm16* was performed similarly to BW25113-ecm16 and BW25113-ecm16-Δ ID , with the exception that HindIII and EcoRI sites were used for restriction digestion.
UvrA expression in E. coli. UvrA gene from Thermotoga maritima was encoded into the pET28a (+) vector and transformed into E. coli Rosetta (DE3) pLysS cells. Protein expression was induced using 50 µM isopropyl-ß-d-galactoside (IPTG) and incubated for 5 h at 30 °C. Bacterial cells were resuspended in buffer containing 50 mM Tris pH 8.0, 150 mM NaCl, 2 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride (PMSF) and 5% (v/v) glycerol and sonicated. The lysate was centrifuged at 18,000 rpm for 45 min and the supernatant was applied on His-Trap crude 5 ml column (GE Healthcare). UvrA was eluted with a 200 ml imidazole gradient from 0 to 500 mM concentration. The purified protein was diluted with buffer containing 50 mM Tris pH 8.0, 2 mM DTT, and 5% (v/v) glycerol and applied to 5 mL HiTrap Q HP column (GE Healthcare). UvrA was eluted with 50 ml linear NaCl gradient from 50 mM to 1 M. The protein was further purified by size exclusion chromatography using Superdex 200 10/300 GL (GE Healthcare) equilibrated with buffer containing 50 mM Tris pH 8.0, 200 mM NaCl, 2 mM DTT and 5% (v/v) glycerol. The volume of the collected protein sample was reduced to a final concentration of 10 mg/ml and flash-frozen in liquid nitrogen.
Data collection and structure determination. Initial X-ray diffraction experiments were carried out at the Stanford Synchrotron Radiation Lightsource. The final X-ray diffraction data set was collected at beamline 17-ID-B of the Advanced Photon Source, Argonne National Laboratory and processed using autoPROC 24 . Molecular replacement was carried out using PHASER 25 and UvrA2 structure (PDB: 2VF7) 15 as the search model. Structure refinement was performed using PHENIX 26 and REFMAC 27 . Model building was done using COOT 28 with alternate sessions of refinement using PHENIX 26 .
Ecm16 expression in B. choshinensis. Codon optimized ecm16 gene for expression in B. choshinensis was synthesized (GenScript) and inserted into pUC19 vector using the BamHI and XbaI sites. pNI, a shuttle vector between B. choshinensis and E. coli, was purchased from Takara Bio. The pNI vector is under the P2 promoter, which is a portion of 5' sequence upstream of the cell wall protein, which is expressed strongly in B. choshinensis. ecm16 was inserted into the pNI vector following the ligation protocol described above. Ligation products were transformed into chemically competent E. coli DH5α cells and grown overnight on LB-amp (50 µg ml −1 ) plates at 37 °C. ecm16_pNI clone was verified by performing restriction enzyme digestion using BamHI and XbaI. Plasmids ecm16_pNI or pNI were transformed in B. choshinensis using the New Tris-PEG (NTP) method following the manufacturers guidelines (Takara Bio) 29  Electrophoretic mobility shift assay. PAGE-purified 32-bp DNA substrate (Integrated DNA Technologies) was dissolved in annealing buffer (30 mM HEPES, pH 7.5, 100 mM potassium acetate). This DNA contained the 5′ ACGT 3′ echinomycin binding site (Table S2). Echinomycin-DNA complex was formed by incubating echinomycin and DNA at molar ratio of 1.1: 1. Different concentrations of purified Ecm16, Ecm16-Δ ID , Ecm16* (0, 100, 200, and 300 nM) were incubated with 50 nM DNA in the presence or absence of echinomycin in binding buffer (50 mM HEPES, pH 7.5, 50 mM NaCl, 0.1 mg ml −1 bovine serum albumin) for 15 min at room temperature. The reaction mixture was separated in a 6% native polyacrylamide gel at 4 °C using 1× TBE (40 mM Tris acetate, 0.5 mM EDTA) as a running buffer for 30 min at 40 mA. The gels were stained using 1× SYBR gold nucleic acid stain in 1× TBE buffer and imaged using an ultraviolet transilluminator (Azure c200).
Fluorescence polarization assay. Fluorescein labeled 32-bp DNA oligonucleotide (Integrated DNA Technologies) was dissolved in binding buffer (50 mM HEPES, pH 7.5, 50 mM NaCl) and aliquoted to 50 nM final concentration to the reaction well. Purified Ecm16 or Ecm16-Δ ID or Ecm16* in presence or absence of echinomycin was serially diluted in binding buffer and added to each reaction well to final volume of 100 µl at a concentration ranging from 4 to 512 nM. To detect the change in the light polarization of the FAM-labeled DNA, fluorescent measurements were performed in a 384-well format on a black low-volume plate (Corning) using Synergy HT (BioTek) plate reader with excitation and emission wavelengths of 490 nm and 520 nm, respectively.
Reported polarization values are the average of three independent experiments. Data was analyzed in Graph Pad Prism 5 using the Hill equation. The calculated dissociation constants (K D ) are listed in Table 3.
ATPase activity assay. The ATPase activity of purified proteins was determined using the EnzChek™ phosphate assay kit (Thermo Fisher). In this assay, inorganic phosphate produced from ATP hydrolysis is utilized by purine nucleoside phosphorylase (PNP) to convert 2-amino-6-mercapto-7-methylpurine (MESG) to ribose 1-phosphate and 2-amino-6-mercapto-7-methylpurine. Product formation is followed by measureing absorbance at 360 nm. Prior to the ATPase assay, echinomycin was incubated with 32-bp DNA at 1. www.nature.com/scientificreports/ To convert optical density at 360 nm to the amount of degraded ATP, a calibration curve was constructed by plotting OD 360nm with phosphate standards.

Data availability
The Ecm16 coordinates and structure factors have been deposited in the Protein Data Bank with the accession code 7SH1.