Biochemical and Structural Analysis of an Eis Family Aminoglycoside Acetyltransferase from Bacillus anthracis

Abstract

Proteins from the enhanced intracellular survival (Eis) family are versatile acetyltransferases that acetylate amines at multiple positions of several aminoglycosides (AGs). Their upregulation confers drug resistance. Homologues of Eis are present in diverse bacteria, including many pathogens. Eis from Mycobacterium tuberculosis (Eis_Mtb) has been well characterized. In this study, we explored the AG specificity and catalytic efficiency of the Eis family protein from Bacillus anthracis (Eis_Ban). Kinetic analysis of specificity and catalytic efficiency of acetylation of six AGs indicates that Eis_Ban displays significant differences from Eis_Mtb in both substrate binding and catalytic efficiency. The number of acetylated amines was also different for several AGs, indicating a distinct regiospecificity of Eis_Ban. Furthermore, most recently identified inhibitors of Eis_Mtb did not inhibit Eis_Ban, underscoring the differences between these two enzymes. To explain these differences, we determined an Eis_Ban crystal structure. The comparison of the crystal structures of Eis_Ban and Eis_Mtb demonstrates that critical residues lining their respective substrate binding pockets differ substantially, explaining their distinct specificities. Our results suggest that acetyltransferases of the Eis family evolved divergently to garner distinct specificities while conserving catalytic efficiency, possibly to counter distinct chemical challenges. The unique specificity features of these enzymes can be utilized as tools for developing AGs with novel modifications and help guide specific AG treatments to avoid Eis-mediated resistance.

Graphical abstract

Bacillus anthracis (Ban), a Gram-positive bacterium, is the cause of the deadly infectious disease anthrax. The Bacillus genus includes other pathogens such as Bacillus cereus¹ and Bacillus thuringiensis, although these are less infectious. B. anthracis resides in soil and typically infects plant-eating mammals. Infection of carnivores and humans occurs usually through direct contact with highly resilient endospores. Upon infection, endospores germinate into active bacilli and multiply. The combined release of three proteins from these bacilli (lethal factor, edema factor, and protective antigen), which interact with their specific targets at the mammalian cell surface, leads to severe toxemia, known as anthrax disease (cutaneous and gastrointestinal forms). When acquired through inhalation of B. anthracis spores (pulmonary form), anthrax initially causes flu-like symptoms, but eventually leads to a fatal respiratory collapse.² This acute pulmonary infectious ability underlies potential use of B. anthracis as a bioweapon. Deliberate dissemination of an aerosolized form of virulent strains of B. anthracis (such as the Ames, Vollum, and other potential man-made derivatives) as a bioweapon is a real threat to both humans and livestock. Vaccines based on spores from the attenuated Sterne strain of B. anthracis are effective against anthrax, but vaccination of a majority of the human population is a difficult task,^3,4 and presently available vaccines are not entirely safe.^5,6 Therefore, antibiotics are needed for prophylactic treatment prior to potential exposure as well as postexposure emergency treatment of inhalation anthrax.⁷ Existing drugs (large doses of intravenous and oral antibiotics, e.g., ciprofloxacin, doxycycline, erythromycin, vancomycin, or penicillin) are only effective if started in the early stages of infection. In addition, some B. anthracis strains have already developed resistance to some of the aforementioned antibiotics.^8–11 For effective treatment of inhalation anthrax in humans and infected animals, new antibiotics are needed.

Biochemical and structural studies are underway to develop new drugs against B. anthracis and explore new drug targets in order to inhibit spore germination,¹² DNA replication, and the vegetative growth,^13–18 disable the released toxins and other virulence factors,^19–22 and utilize aminoglycosides (AGs) as toxin inhibitors and potential anti-anthrax drugs.^20,23–25 In this study, we investigated a highly potent AG acetylator encoded by gene bas2743 of B. anthracis (Eis_Ban). Eis proteins from Mycobacterium tuberculosis (Eis_Mtb), Mycobacterium smegmatis (Eis_Msm), and Anabaena variabilis (Eis_Ava) catalyze the transfer of an acetyl group to the amino group(s) of various AGs^26–28 and amino acid residues in lysine-containing drugs and proteins.^29,30 When upregulated, Eis_Mtb is known to cause resistance to the AG kanamycin A (KAN) in tuberculosis patients.³¹ Similarly, upregulation of Eis in B. anthracis may lead to resistance to AG antibiotics. To address the acetylation potential and possible differences in substrate specificity between Eis_Ban and Eis_Mtb, and Eis from other organisms, we carried out a combined structural and functional investigation of Eis_Ban. Our study revealed a highly divergent AG binding site endowing a remarkably distinct substrate specificity and highly potent AG acetylating activity.

MATERIALS AND METHODS

Bacterial Strains, Plasmids, Materials, and Instrumentation

The chemically competent Escherichia coli TOP10 and BL21 (DE3) strains were purchased from Invitrogen (Carlsbad, CA). All restriction enzymes, T4 DNA ligase, and Phusion DNA polymerase were purchased from NEB (Ipswich, MA). PCR primers were purchased from Integrated DNA Technologies (IDT; Coralville, IA). The pET15b vector was purchased from Novagen (Gibbstown, NJ). DNA sequencing was performed at the University of Michigan DNA Sequencing Core. All reagents were used as received without further purification. DTNB, AcCoA, AGs (apramycin (APR), amikacin (AMK), gentamicin (GEN), hygromycin (HYG), KAN, neomycin B (NEO), sisomicin (SIS), spectinomycin (SPT), streptomycin (STR), and ribostamycin (RIB)) (Figure S1, Supporting Information), ampicillin, chloramphenicol, ciprofloxacin, erythromycin, isoniazid, norfloxacin, and chlorhexidine (1) were purchased from Sigma-Aldrich (Milwaukee, WI). The AG geneticin (G418) was purchased from Research Products International (Mt Prospect, IL). The rest of the AGs (neamine (NEA), netilmicin (NET), paromomycin (PAR), and tobramycin (TOB)) (Figure S1) were purchased from AK Scientific (Mountain View, CA). The spectrophotometric assays were performed on a multimode SpectraMax M5 plate reader using 96-well plates (Fisher Scientific; Pittsburgh, PA). Silica gel 60 F₂₅₄ plates (Merck) were used for thin-layer chromatography (TLC) analysis. Liquid chromatography mass spectrometry (LCMS) was performed on a Shimadzu LCMS-2019EV equipped with a SPD-20AV UV–vis detector and a LC-20AD liquid chromatograph.

Cloning, Overproduction, and Purification of Seleno-methionine-Substituted Eis_Ban for Structural Studies

The Eis_Ban acetyltransferase (GI: 753454082) coding sequence cassette was amplified by PCR from chromosomal DNA of B. anthracis Sterne strain using primers compatible with the ligation-independent cloning vector pMCSG7 and cloned into pMCSG7 using the ligation-independent protocol.³² The recombinant Eis_Ban with an N-terminal His₆-tag and a TEV protease recognition site (ENLYFQ↓S) was expressed in an E. coli BL21 (DE3) strain harboring a pMAGIC plasmid encoding one rare E. coli tRNAs for Arg codons AGG/AGA. Expression of the His₆-tagged fusion protein in E. coli BL21 (DE3) strain carrying the pMAGIC vector was induced with isopropyl β-D-thiogalactoside (IPTG). A selenomethionine (SeMet) derivative of the expressed protein was prepared and purified using Ni-affinity chromatography as described previously.^33,34 Briefly, the harvested cells, containing SeMet-labeled protein, were resuspended in lysis buffer (500 mM NaCl, 5% [v/v] glycerol, 50 mM HEPES pH 8.0, 10 mM imidazole, and 10 mM β-mercaptoethanol), and the lysate was clarified by centrifugation, filtered through a 0.44 μm membrane, and applied to a 5 mL HiTrap Ni-NTA column (GE Health Systems) on an AKTAxpress system. The eluted His₆-tagged protein was further purified by size exclusion chromatography (Hiload 26/600 Superdex 200 pg GE) with crystallization buffer containing 100 mM NaCl, 20 mM HEPES pH 8.0, and 2 mM dithiothreitol (DTT) and then concentrated to 30 mg/mL for crystallization using an Amicon Ultra centrifugal filter device with a 10 000-MW cutoff (Millipore), flash-frozen, and stored in liquid nitrogen.

Overproduction and Purification of Eis Proteins for Biochemical Studies

The Eis_Mtb,²⁷ Eis_Msm,²⁸ and Eis_Ava²⁶ proteins (with a N-terminal His₆-tag) were prepared as previously reported. Eis_Ban was overexpressed and purified using the exact procedure used for Eis_Mtb²⁷ and stored at 4 °C in 50 mM Tris pH 8.0. After purification, 3.2 mg of the 48 213-Da Eis_Ban (NHis₆-tagged) protein was obtained per liter of culture (Figure S2, Supporting Information).

Determination of the AG Selectivity Profile of Eis_Ban by a Spectrophotometric Assay

The substrate specificity and acetlytransferase activity of Eis_Ban were determined using the Ellman method as previously reported.²⁷ The free thiol of CoA generated as a byproduct of the acetylation reaction reacts with the DTNB indicator releasing 2-nitro-5-thiobenzene, which absorbs at 412 nm (ɛ₄₁₂ = 14 150 M⁻¹ cm⁻¹). Reactions (200 μL) containing Tris (50 mM pH 8.0), AG (100 μM), AcCoA (500 μM), and DTNB (2 mM) were initiated by addition of Eis_Ban (0.5 μM final concentration) and were monitored every 30 s for 30 min at 25 °C.

Determination and Confirmation of Number of Acetylation Sites by the Spectrophotometric Assay and Mass Spectrometry

The number of times Eis_Ban acetylated each AG was determined though the use of two reactions per each AG. The first reaction contained 1 equiv of AcCoA for every equivalent of AG, and the second contained 5 equiv of AcCoA for 1 equiv of AG. Briefly, reaction mixtures (200 μL) containing AcCoA (100 μM, 1 equiv or 500 μM, 5 equiv), AG (100 μM), Tris (50 mM, pH 8.0), and DTNB (2 mM) were initiated by the addition of Eis_Ban (0.5 μM). Reaction progress at 25 °C was monitored at 412 nm as above until a plateau was achieved. Representative plots are presented in Figure S3, Supporting Information, and all data are summarized in Table 1. To confirm the findings of the UV–vis assay, a final reaction (20 μL) containing AcCoA (3.35 mM, 5 equiv), AG (0.67 mM, 1 equiv), Tris (50 mM, pH 8.0), and Eis_Ban (10 μM) was carried out at 25 °C. Reaction were quenched by precipitating the enzyme with an equal volume of ice-cooled MeOH. After 20 min at −20 °C, the precipitate was removed by centrifugation (13 000 rpm, room temperature, 10 min). The masses of the AG species in the supernatant in each sample were determined by LCMS in positive mode, after a 1:3 dilution in H₂O, by using H₂O with 0.1% formic acid as the eluent for the LC. Mass spectra are provided in Figure S4, Supporting Information and summarized in Table S1, Supporting Information.

Table 1.

Comparison of Levels of Acetylation of AGs by Eis from A. variabilis, B. anthracis, M. smegmatis, and M. tuberculosis

AG	B. anthracis	A. variabilisa	M. smegmatisb	M. tuberculosisc
AMK	di	tri	tri	tri
APR	di	mono	di	×d
HYG	di	mono	mono	di
KAN	di	tri	di	di
NEA	tri	tri	tri	tri
NEO	tri	tri	tri	tri
NET	mono	di	di	di
PAR	tri	tri	tri	di
RIB	tri	tri	tri	tri
SIS	di	di	tri	tri
SPT	×	×	×	×
STR	×	×	×	×
TOB	×	–e	tetra	tetra

^aThese data were previously reported.²⁶

^bThese data were previously reported.²⁸

^cThese data were previously reported.²⁷

^d×, indicates that the AG is not a substrate of the enzyme.

^e–, not reported.

Kinetic Characterization of Eis_Ban Activity

Michaelis–Menten kinetic parameters (K_m and k_cat) were determined for several AGs (AMK, APR, KAN, NEO, and SIS) and AcCoA with the Eis_Ban purified enzyme. To determine the kinetic parameters of the AGs, the concentration of AcCoA (500 μM) was held constant. To determine the AcCoA kinetic parameters, the concentration of KAN (500 μM) was held constant. In short, solutions containing AG (0–200 μM for KAN and PAR, or 0–2000 μM for AMK and NEO final concentrations) were used to initiate a premixed solution of AcCoA (500 μM), Eis_Ban (0.25 μM), DTNB (2 mM), and Tris (50 mM, pH 8.0). Reactions were set up independently and monitored as above, taking measurements every 15–20 s for 30 min, in the linear range of product accumulation over time. For determination of the AcCoA kinetic parameters, solutions of AcCoA (0–500 μM, final concentration) were used to initiate reactions containing KAN (500 μM), Eis_Ban (0.25 μM), DTNB (2 mM), and Tris (50 mM, pH 8.0). All experiments were performed at least in triplicate. Linear regression was used to obtain reaction velocities for each time course, with negligible uncertainly. These independently measured velocity values were plotted, and Michaelis–Menten parameters were determined without averaging or other data conversion by nonlinear regression with SigmaPlot 12.3 software (SysStat). Kinetic parameters are presented in Table 2, and representative curves can be seen in Figure S5, Supporting Information.

Table 2.

Apparent Steady-State Kinetic Parameters for [AG]-Dependent and [AcCoA]-Dependent Acetylation by Eis from Various Bacteria

(co)substrate	Km (μM)	kcat (s−1)	kcat/Km (M−1 s−1)
Eis_Ban
AMK	1070 ± 470	0.10 ± 0.02	92 ± 45
APR	275 ± 80	0.13 ± 0.01	469 ± 140
KAN	57 ± 9	0.135 ± 0.008	2370 ± 400
NEO	469 ± 70	0.186 ± 0.004	397 ± 60
PAR	44 ± 11	0.019 ± 0.001	430 ± 110
SIS	76 ± 31	0.017 ± 0.002	225 ± 95
AcCoAa	32 ± 8	0.241 ± 0.014	6700 ± 1700
Eis_Avab
AMK	–d	–	–
APR	340 ± 100	0.019 ± 0.002	57 ± 18
KAN	1000 ± 40	0.205 ± 0.004	205 ± 9
NEO	101 ± 16	0.106 ± 0.006	1050 ± 180
PAR	237 ± 36	0.18 ± 0.01	770 ± 130
SIS	–	–	–
Eis_Msmc
AMK	251 ± 55	0.034 ± 0.003	13 ± 3
APR	150 ± 43	0.019 ± 0.002	127 ± 39
KAN	655 ± 42	0.36 ± 0.01	541 ± 37
NEO	110 ± 14	0.148 ± 0.006	1345 ± 180
PAR	730 ± 160	0.24 ± 0.03	320 ± 80
SIS	82 ± 8	0.206 ± 0.006	2512 ± 256
AcCoA	39 ± 17	0.40 ± 0.06	10210 ± 4450
Eis_Mtbc
AMK	42 ± 9	0.103 ± 0.006	2450 ± 545
APR	×e	×	×
KAN	330 ± 40	0.53 ± 0.03	1590 ± 250
NEO	122 ± 23	0.61 ± 0.03	5000 ± 970
PAR	110 ± 21	0.14 ± 0.01	1,240 ± 260
SIS	166 ± 36	1.1 ± 0.1	6810 ± 1,650
AcCoA	10 ± 3	0.094 ± 0.004	9400 ± 2,850

^aThe kinetic parameters for the [AcCoA]-dependent acetylation by Eis_Ban, Eis_Msm, and Eis_Mtb were performed with KAN.

^bThese data were previously reported.²⁶

^cThese data were previously reported.²⁸

^d–, indicates this data was not available.

^e×, indicates the AG was not a substrate of the enzyme.

Determination of Positions Acetylated on NEA by Eis_Ban by TLC

To visualize which positions of NEA become acetylated by Eis_Ban, reactions (40 μL) containing Tris (50 mM, pH 8.0), AcCoA (4 mM, 5 equiv), NEA (0.8 mM, 1 equiv), and Eis_Ban (5 μM) were incubated at room temperature, and aliquots were spotted on TLC plates at various times (0, 1, 5, 10, 30 min, and overnight reaction). The mobile phase consisted of 3:0.8/MeOH:NH₄OH. The TLC plates were visualized with a cerium-molybdate stain. The R_f values found for the Eis_Ban reactions were in good agreement with those previously reported for Eis_Mtb.²⁷

Inhibition of Eis_Ban by Inhibitors of Eis_Mtb

The IC₅₀ values for chlorhexidine (1) and compounds 2–3 (Table 3 and Figure S6, Supporting Information) were determined in a manner similar to that previously reported.³⁵ Briefly, reactions (200 μL) contained compounds 1–3 dissolved in Tris-HCl buffer (50 mM, pH 8.0, 10% DMSO) with a 5-fold serial dilution, Eis_Ban enzyme (0.25 μM), and NEO (100 μM), AcCoA (0.5 mM), and DTNB (2 mM). Reagents were added in a stepwise manner, allowing the enzyme to incubate with the inhibitor and NEO, and the reactions were initiated with the AcCoA in solution with DTNB. The reactions were monitored as described for the kinetic analysis. The determination of IC₅₀ values was done with a Hill plot analysis using Kaleidagraph 4.1 software.

Table 3.

IC₅₀ Values of Eis_Mtb Inhibitors with Various Eis Homologues

compound	IC50 (μM)
	Eis_Ban	Eis_Avaa	Eis_Msma,b	Eis_Mtbc
chlorhexidine (1)	14 ± 4	20 ± 7	1.9 ± 0.4	0.19 ± 0.03
2	>200	>200	0.9 ± 0.2	1.09 ± 0.14
3	>200	>200	1.7 ± 0.3	1.24 ± 0.16

^aThese data previously reported.²⁶

^bThese data previously reported.^26,28

^cThese data previously reported.³⁵

Crystallization of Eis_Ban

Eis_Ban at 30 mg/mL was crystallized using sitting drop vapor diffusion at 297 K in a Crystal Quick VR 96-well round-bottomed plate (Greiner Bio-One North America). Eis_Ban (400 nL) was mixed with 400 nL of crystallization reagent using a Mosquito VR nanoliter liquid workstation (TTP LabTech) and allowed to equilibrate against 135 μL of crystallization reagent. Four different crystallization screens were used: MCSG-1, MCSG-2, MCSG-3, and MCSG-4 (Microlytic). The best crystals were obtained in the 93^th condition of MCSG-1 containing 0.1 M sodium HEPES pH 7.0, 0.2 M sodium chloride, and 20% polyethylene glycol 3350 at 24 °C. The crystals were cryoprotected using the solution prepared by adding 15% (v/v) glycerol to the crystallization condition and flash-frozen in liquid nitrogen. Single wavelength anomalous diffraction data at the selenium peak was collected from the SeMet-substituted protein. The data sets were collected on ADSC quantum Q315r charged coupled device detector at 100 K in the 19ID beamline of the Structural Biology Center at the Advanced Photon Source, Argonne National Laboratory. The crystal belongs to space group P2₁ with cell parameters of a = 79.4 Å, b = 176.9 Å, c = 110.0 Å, α = γ = 90°, β = 105.7°. The diffraction data were processed by using the HKL3000 suite of programs.³⁶ Data collection statistics are presented in Table S2, Supporting Information.

Structural Analysis of Eis_Ban

All procedures for SAD phasing, phase improvement by density modification, and initial protein model building were performed by the structure module of the HKL3000 software package.³⁶ The mean figure of merit of the phase set was 0.257 for 50–2.75 Å resolution data and improved to 0.899 after density modification (DM). The script build module using resolve in HKL3000 built 1781 out of 2310 residues, while side chain of 520 residues were placed. The initial model was rebuilt with the program Coot³⁷ by using electron density maps based on DM-phased reflection file. After each cycle of rebuilding, the model was refined by using REFMAC5 from the CCP4 suite with TLS refinement.^38,39 The geometrical properties of the model were assessed by the Coot and Molprobity software.⁴⁰

Construction of the Δeis Mutant of B. anthracis Sterne 34F2

A markerless, in-frame deletion of the eis allele that fused the first 10 codons of the gene with the final 10 codons (9 codons plus the native stop codon) was created using PCR. This mutant allele therefore had the internal 1098-bp of the gene removed. The PCR construct also consisted of 521-bp of upstream homology and 514-bp of downstream homology flanking the deleted gene and was generated by standard overlap extension PCR using the 5′ primer AAGGAAGATAAACGTAAGCCGTTTATTTATGATTTCT and the 3′ primer AAACGGCTTACGTTTATCTTCCTTTAATCGTATAACG. This PCR product was cloned into the allelic exchange vector pBKJ236 at the NotI and BamHI sites, and allelic exchange was performed as previously described⁴¹ with the following modification: plasmid pSS4332 (a kind gift from Dr. Scott Stibitz, FDA, Division of Bacterial, Parasitic, and Allergenic Products, Center for Biologics Evaluation and Research, Maryland, MD) was used in place of plasmid pBKJ223 to promote the second recombination event needed for allelic exchange. Successful exchange was determined using PCR with primers that anneal to sequences upstream and downstream of the cloned region (5′ primer CGCGCGGCCGCCCTGTAATTGCATTAGCGGCAACGC; 3′ primer CGCGGATCCCTAAACTGTTCACGAATATTTGCGT). The mutant allele yielded a ∼1.1-kb smaller product than the wild-type using this assay. The final mutant contained no added antibiotic resistance and should therefore be isogenic to the wild-type parent.

Determination of MIC and MBC Values for Various Antibiotics against the Δeis Ban Strain

MIC and MBC values were determined against 14 AGs (AMK, G418, GEN, HYG, KAN, NEA, NEO, NET, PAR, RIB, SIS, SPT, STR, and TOB), one β-lactam (ampicillin), two fluoroquinolones (ciprofloxacin and norfloxacin), one polyketide (erythromycin), a compound of nonribosomal peptide origin (chloramphenicol), and the antituberculosis drug isoniazid (Table S3, Supporting Information). Using the microdilution method,⁴² MIC values were determined in brain-heart infusion (BHI) broth. The MIC value was determined as the lowest concentration of the inhibitor that did not support bacterial growth, and this determination was confirmed for the respective well by staining with a solution of MTT (50 μL of 1 mg/mL). MBC values were determined by plating all cultures from all wells (20 μL of each) that showed no bacterial growth (prior to staining) on BHI-agar plates. The lowest concentration to show no additional growth after incubation at 37 °C was deemed the MBC value. All MIC and MBC tests were done at least in duplicate.

RESULTS

Comparison of Sequence Features

On the basis of its amino acid sequence, it is clear that Eis_Ban and Eis_Mtb belong to the same family of the GNAT superfamily of N-acyltransferases. However, the amino acid sequence identity is 18% between the Eis_Ban and Eis_Mtb, and the structure-based sequence alignment indicates that many of the AG binding pocket residues are distinctly different (Figure 1). This prompted us to investigate the substrate selectivity and recognition properties of Eis_Ban, both biochemically and structurally.

Figure 1.

Structure-based protein sequence alignment of Eis proteins from B. anthracis (Eis_Ban) and from M. tuberculosis H37Rv (Eis_Mtb). The two Eis homologues exhibit 18% sequence identity. The residues in the immediate active site of Eis_Mtb (also shown as sticks in Figure 4B) are marked by red ovals above the alignment and show stark dissimilarities with their counterpart in Eis_Mtb. Residues previously identified to bind to TOB in a structure of a TOB-CoA-Eis_Mtb structure (PDB ID: 4JD6)⁴⁴ are marked by yellow ovals. The residue numbering corresponds to accession code YP_029001.

Substrate Specificity and Multiacetylation Profiles of Eis Proteins

We recently reported the structural and biochemical features of Eis_Mtb,^{27,30,43–45} as well as biochemical characteristics of its close homologue Eis_Msm^28,46 and that of a less homologous protein Eis_Ava.²⁶ These studies revealed that all three enzymes function as aminoglycoside acetyltransferases (AACs) and are capable of multiacetylating AGs. In this work, we performed similar studies to investigate the number of acetylations catalyzed by Eis_Ban on a panel of 12 AGs (Table 1). The acetylation results of Eis_Ban along with those of other Eis homologues previously determined, for comparison, are presented in Table 1.

We performed an AcCoA titration assay and monitored the reaction products using a UV–vis assay (Figure S3) and mass spectrometry (Figure S4 and Table S1) to show that Eis_Ban catalyzed monoacetylation of NET, diacetylation of AMK, APR, HYG, KAN, and SIS, as well as triacetylation of NEA, NEO, PAR, and RIB. In addition, we found that SPT, STR, and TOB were not used as substrates by Eis_Ban.

Steady-State Kinetic Analysis of Eis Proteins

In an effort to further highlight and elucidate differences among Eis homologues, we determined Michaelis–Menten kinetic parameters (K_m and k_cat) of Eis_Ban toward various AGs with AcCoA as the cosubstrate. The kinetic parameters for Eis_Ban and that of other previously reported Eis homologues for comparison are presented in Table 2. The apparent K_m value of 32 ± 8 μM for AcCoA determined in the kinetic experiments where its concentration was varied was much smaller than the concentration of AcCoA (500 μM) used in experiments at varying concentrations of AGs. Therefore, the K_m values for different AGs, as with other Eis homologues,^28,43 should be interpreted as the K_m values for AGs interacting with the binary complex of Eis_Ban-AcCoA as a reactant. Moreover, given a similar random-sequential mechanism of acetylation with respect to the order of binding of AcCoA and AG,⁴³ these K_m values correspond to the apparent equilibrium constants for binding of AGs to the Eis_Ban-AcCoA complex. Further interpretation of these K_m and k_cat values in terms of microscopic mechanistic parameters depends on the number of acetylations per an AG-binding event and may be complicated by potentially different efficiencies of acetylations at different positions for a given AG. On the basis of our recent crystal structure of Eis_Mtb in complex with CoA and TOB, after the first acetylation, TOB would likely need to dissociate from the enzyme completely in order to rebind the enzyme for the second major acetylation, as there is not enough space in the binding site for its reorientation without a major protein conformational change. Therefore, acetylation of TOB and, similarly, other AGs would not be processive. If the individual acetylations are carried out with the same efficiency, then the K_m needs to be multiplied by the number of acetylations per AG to obtain the equilibrium constant for binding of an AG to Eis-AcCoA complex, and k_cat represents the true catalytic efficiency for an individual acetylation event. In the other extreme, when the mechanism is fully processive, i.e., an AG gets fully acetylated upon binding to the enzyme once, the K_m values reported in Table 2 are interpreted as true equilibrium constants for binding of an AG to Eis-AcCoA compex, whereas the k_cat is a rate constant averaged over all acetylation events for this AG.

TLC-Based Analysis of Positions of Acetylation on NEA by Eis_Ban

We previously demonstrated that both Eis_Mtb²⁷ and Eis_Msm²⁸ sequentially acetylate NEA at the 2′-, 6′-, and 1-positions. To test whether Eis_Ban also sequentially acetylates NEA at these positions, we performed TLC analysis of the reaction of AcCoA with NEA in the presence of Eis_Ban over a period of 24 h (Figure 2). The production of 2′,6′-diacetyl-NEA was observed after 1 min as determined by comparing the R_f of this compound to that of the previously reported reaction for Eis_Mtb under identical conditions. After 5 min, a product appeared whose R_f is congruent with that of 2′,6′,1-triacetyl-NEA. All NEA was consumed after a 30 min of incubation, and only a mixture of spots with similar retention factor (R_f) values to those of 2′,6′-diacetyl-NEA and 2′,6′,1-triacetyl-NEA, as generated by Eis_Mtb,²⁷ was observed. The reaction was incubated for an additional 16–20 h and a mixture of both triacetylated- and diacetylated-NEA remained. These data indicate that the 2′- and the 6′-acetylation events, whose order is unclear, are followed by the 1-acetylation of NEA by Eis_Ban.

Figure 2.

Multiacetylation by Eis_Ban of NEA over time (0, 1, 5, 10, and 30 min and overnight reactions) observed in TLC assay. The R_f values of 0.12, 0.30, and 0.42 represent NEA, 2′,6′-diacetyl-NEA, and 2′,6′,1-triacetyl-NEA, respectively (as these R_f values match those previously reported for these NEA derivatives generated by Eis_Mtb).²⁷

Inhibition of Eis Proteins

To further probe the differences in the active sites of Eis_Ban and other Eis homologues, we investigated inhibitory properties of three inhibitors of Eis_Mtb, chlorhexidine (1) and compounds 2 and 3, which we recently identified from a high-throughput screening study with Eis_Mtb (Figure S1C and Table 3).³⁵ Of the compounds tested, which were all good inhibitors of both Eis_Mtb and Eis_Msm, just as in the case of Eis_Ava, only chlorhexidine showed detectable inhibition of Eis_Ban with an IC₅₀ value of 14 ± 4 μM (Table 3 and Figure S6).

Overall Structure of Eis_Ban

In order to understand the interactions that dictate recognition and acetylation of AGs by Eis_Ban, we crystallized and determined a crystal structure of Eis_Ban (Figure 3). The crystals diffracted to 2.75 Å resolution and belonged to space group P2₁ with one hexamer in the asymmetric unit. The experimental data collection and refinement statistics are given in Table S2. A superposition of this structure with that of Eis_Mtb (PDB ID: 3R1K²⁷) showed that the monomers superpose well and display similar hexameric assemblies (Figure 4A).

Figure 3.

Structure of Eis_Ban. Left: A cartoon representation of the Eis_Ban monomer. The three domains are highlighted in different shades of red. Middle: surface representation of its trimeric and Right: hexameric assembly forms, with each monomer highlighted in a different color.

Figure 4.

(A) The superposition of Eis_Ban (brown) and Eis_Mtb (dark blue) showing overall structural similarity but drastic changes in loops around the active site (indicated with the black square). (B) A close-up cartoon representation view of the monomer of Eis_Ban with residues in its immediate active site pocket marked (Note: these residues are marked by red ovals in Figure 1). The CoA binding site is shown in orange.

Active Site Structure of Eis_Ban

The entire electron density map including that of the active site area was reasonably ordered for all six polypeptide chains in the asymmetric unit. As in other Eis homologues, each subunit of the 385-residue Eis_Ban contains three domains. The residue directly involved in the catalytic acetyl group transfer chemistry and proposed to be a general acid in the Mtb homologue, Tyr125, and the C-terminal residue as well as the location of the carboxyl terminal group of the protein, the last proposed to serve a general base, are conserved (Figure 1). The exact CoA position can also be predicted with high certainty due to the conserved nature of this cosubstrate-binding site and comparison to other CoA-bound structures of other Eis homologues. However, the backbones of the loops bearing residues lining the active site adopt considerably different conformations (Figures 1 and 4B). A close look at the active site indicates that the putative AG binding residues are also highly dissimilar (Figures 1 and 4B). Out of approximately 15 residues forming the substrate-binding site, only 3 residues are the same, 6 residues are similar, and 6 residues are very different. As a result, the shape and the surface-charge properties of the active site pocket of Eis_Ban are significantly different from those of Eis_Mtb (Figure 5). Overall, the substrate-binding pocket of Eis_Ban appears more open, and the face of the pocket formed by the N-terminal domain (Figure 5A,D) is less negatively charged than in Eis_Mtb (Figure 5B,E).

Figure 5.

Surface electrostatics of (A) Eis_Ban, (B) Eis_Mtb, and (C) AAC(2′)-Ic from M. tuberculosis; D, E, and F are magnified views of the immediate active site highlighting stark differences.

MIC and MBC Studies of Antibiotics against the Δeis Ban Strain

To investigate if Eis_Ban affects the antibacterial susceptibility of B. anthracis, we generated a mutant strain of B. anthracis 34F2 Sterne in which the eis gene was deleted (Δeis Ban). We determined MIC and MBC values for 20 antibiotics against wild-type and Δeis Ban (Table S3). The deletion strain did not have an apparent growth defect, indicating that the eis gene is not essential for growth of B. anthracis in rich media. Furthermore, MIC values for all AG and non-AG antibiotics were the same for the wild-type and Δeis Ban within a 2-fold dilution, indicating that basal levels of Eis_Ban do not contribute to antibiotic resistance.

DISCUSSION

The natural and engineered strains of B. anthracis pose a potential threat of rapid infection and global spread. Efforts to treat anthrax infections have primarily focused on developing effective vaccines. However, in the event a human or livestock becomes infected, effective antimicrobial agents will be necessary, and this raises an urgent call for effective antibiotics. GEN is a potent AG antibiotic that is used in combination therapy for inhalation anthrax disease.⁴⁷ Other AG antibiotics such as NEO and STR have also proven to be effective against B. anthracis,²⁰ and synthetic AGs have been shown to act as potent inhibitors of anthrax lethal factor.^20,25 To develop suitable AG antibiotics against presently found B. anthracis strains, we require a clear understanding of the resistance factors naturally present in these bacteria. Here, we focus on a highly potent and versatile acetyltransferase, Eis_Ban. We demonstrate here that Eis_Ban, similarly to its M. tuberculosis homologue, is neither essential for bacterial growth on rich media nor contributes to antibiotic resistance when expressed at basal levels. However, upregulation of Eis_Ban may cause resistance against chemotherapeutic agents of the AG class, as observed in mycobacteria.³¹ As in M. tuberculosis, Ban strains with upregulated Eis_Ban expression may arise through selective pressure of AG administration, as a result of promoter mutations. The biological need of acetylation function of Eis is still unclear. Little is known about the involvement of Eis in cellular metabolic pathways or its role in host invasion and infection, if any. Recent studies indicate that lysine or arginine residues of other proteins such as DUSP16/MKP7 are substrates of the Eis_Mtb homologue and Eis_Mtb might be involved in N-acetyltransferase activity against a plethora of other peptide substrates.²⁹ Nonetheless, this enzyme possesses a particularly unusual multiacetylating capability against AGs. Here, we investigated the structure–function relationship of Eis_Ban to gain a better understanding of its acetyltransferase specificity and to elucidate what type of AG modification would be significant, should resistance due to Eis_Ban arise. In addition, understanding of the specificity determinants in the active site could enable us to rationally modify AGs and generate new potential AG drugs against Eis_Ban or other pathogenic bacteria.

Resistance enzymes AG acetyltransferases (AACs) found in many pathogenic bacteria are known to monoacetylate the amine moiety at the 6′-, 2′-, or 3-position of various AGs. These regiospecific AACs are characterized by a negatively charged AG binding pocket formed in a single GNAT domain (Figure 5C). The small size of the binding pocket of these AACs likely dictates a unique orientation of a bound AG, in turn defining its regiospecificity. In contrast, the Eis family of proteins exhibits unusual regioversatility as acetyltransferases, modifying many AGs at multiple positions. This multiacetylation is enabled by a significantly larger negatively charged AG binding pocket formed by two conjoined GNAT domains.^27,44 Because the ability to acetylate a given AG and the regioversatility will primarily depend on the shape and the properties of residues in AG binding site of an Eis homologue,^26–28 we explored the biochemical properties and structural features of Eis_Ban.

For the panel of examined AGs, the Eis_Ban modifies the same number or fewer amines than does the Eis_Mtb homologue for all but two AGs, APR and PAR (Table 1). This lesser regioversatility is likely due to the somewhat less negatively charged AG binding pocket of Eis_Ban. APR, which is not acetylated by Eis_Mtb, is diacetylated by Eis_Ban. The ability to acetylate a relatively stiff APR molecule by Eis was previously shown to be dictated by the size of the AG binding pocket and to be specifically sensitive to the nature of the residue in place of Trp289 of Eis_Mtb.⁴⁶ The residue found in place of this Trp residue in Eis_Ban is a more flexible Glu275 (Figure 1); in addition, the backbone bearing this residue is shifted resulting in the more open pocket. We propose that this more open and malleable substrate binding pocket feature is what allows APR to be accommodated and acetylated by Eis_Ban. We speculate that the hydroxyl group of PAR at the 6′-position allows this AG to bind the active site pocket of Eis_Ban in an orientation that is disfavored in the more negatively charged pocket of Eis_Mtb, thus leading to an additional acetylation by Eis_Ban. Strikingly, TOB is not a substrate for Eis_Ban, while Eis_Mtb can acetylate TOB at up to four positions (Table 1).^27,44 Residues Asp26 and Glu401 of Eis_Mtb were recently demonstrated to interact directly with amino groups of the bound TOB in a crystal structure of Eis_Mtb-TOB complex.⁴⁴ In contrast, Eis_Ban contains Tyr29 and Phe389 in structurally analogous positions (Figure 1). The inability of these residues to form salt bridges with TOB amines likely prevents TOB from binding properly for its acetylation by Eis_Ban.

Of the tested inhibitors of Eis_Mtb only one displayed any activity with Eis_Ban, chlorhexidine (1). This inhibitor showed a 77-fold preference for Eis_Mtb (IC₅₀ = 0.188 μM) over Eis_Ban (IC₅₀ = 14 μM) further stressing the global differences in the charge and shape of the AG binding pockets of these enzymes.

While it is known that upregulation of Eis_Mtb causes resistance of Mtb to KAN, Eis_Mtb is not essential for bacterial growth in vitro.³¹ In order to better understand the physiological role of Eis protein in Ban, we decided to delete the eis gene from the bacterial genome and compare how the wild-type (B. anthracis 34F2 Sterne strain) and the mutant reacted to AG treatment. We successfully removed the gene encoding Eis from the B. anthracis genome using a markerless gene replacement protocol.⁴¹ This is corroborated by the MIC data (Table S3). All MIC values for the wild-type Ban and the Δeis Ban differ by no more than 2-fold. The MIC data indicate that, currently, the presence of Eis, in Ban, does not contribute to resistance to AGs. Mtb only becomes resistant to KAN and AMK through the upregulation of Eis_Mtb due to a mutation in its promoter region. Therefore, it is not surprising that wild-type Ban and Δeis Ban have similar MIC values. A difference between the two phenotypes was more apparent when we compared the bactericidal and bacteriostatic properties of the antibiotic compounds. With the exception of isoniazid, all compounds were bacteriostatic at or near the MIC concentration. While for most antibiotic compounds there was no difference between the MBC for the wild-type and Δeis Ban there are a few noteworthy exceptions. HYG was observed to be bactericidal with the wild-type Ban, but when tested with the Δeis Ban was bacteriostatic up to the highest concentration tested (188 μg/mL). This is the opposite of what we expected and may point to a distinct function of the enzyme in vivo. NEA was bactericidal for both phenotypes of Ban; however, the bactericidal dose of wild-type Ban was four times higher than that for the Δeis Ban.

In conclusion, we showed the Eis_Ban is a unique member of the Eis family with the ability to acetylate a variety of AGs generally at multiple positions. The unique range of multi-acetylation of the AGs exhibited by Eis_Ban must be a result of its divergently evolved AG binding pocket. Future research unveiling the in vivo function of Eis might shed light on these recognition signatures. Consistent with these observations are differences in kinetics parameters for AG acetylation, and the lack or significantly reduced inhibition of Eis_Ban by Eis_Mtb. This characterization improves our understanding of AG recognition by Eis and allows one to foresee the potential of B. anthracis to acquire resistance toward specific AGs upon Eis upregulation to optimize therapeutic regimens.

ABBREVIATIONS

AcCoA

acetyl-coenzyme A

AMK

amikacin

AG

aminoglycoside

AAC

aminoglycoside acetyltransferase

APH

aminoglycoside phosphotransferase

APR

apramycin

Ava

Anabaena variabilis

Ban

Bacillus anthracis

DTNB

5,5′-dithiobis(2-nitrobenzoic acid)

Eis

enhanced intracellular survival

equiv

equivalent

G418

geneticin

GEN

gentamicin

HYG

hygromycin

KAN

kanamycin A

LCMS

liquid chromatography mass spectrometry

MBC

minumum bactericidal concentration

MIC

minumum inhibitory concentration

Msm

Mycobacterium smegmatis

Mtb

Mycobacterium tuberculosis

NEA

neamine

NEO

neomycin B

NET

netilmicin

PAR

paromomycin

RIB

ribostamycin

SIS

sisomicin

SPT

spectinomycin

STR

streptomycin

TLC

thin-layer chromatography

TOB

tobramycin

TB

tuberculosis