Two methionine aminopeptidases from Acinetobacter baumannii are functional enzymes

Abstract

Drug resistance in Gram-negative bacteria, such as Acinetobacter baumannii, is emerging as a significant healthcare problem. New antibiotics with a novel mechanism of action are urgently needed to overcome the drug resistance. Methionine aminopeptidase (MetAP) carries out an essential cotranslational methionine excision in many bacteria and is a potential target to develop such novel antibiotics. Two putative MetAP genes were identified in A. baumannii genome, but whether they actually function as MetAP enzymes was not known. Therefore, we established an efficient E. coli expression system for their production as soluble and metal-free proteins for biochemical characterization. We demonstrated that both could carry out the metal-dependent catalysis and could be activated by divalent metal ions with the order Fe(II) ≈ Ni(II) > Co(II) > Mn(II) for both. By using a set of metalloform-selective inhibitors discovered on other MetAP enzymes, potency and metalloform selectivity on the A. baumannii MetAP proteins were observed. The similarity of their catalysis and inhibition to other MetAP enzymes confirmed that both may function as competent MetAP enzymes in A. baumannii and either or both may serve as the potential drug target.

Acinetobacter baumannii is an aerobic Gram-negative bacillus, which exists ubiquitously in nature (soil, water and animals).¹ It has emerged globally as one of the most troublesome pathogens, especially over the last 20 years.^{2, 3} The bacterium can infect various parts of the body including lung, urinary tract and bloodstream and cause both community and nosocomial infections.⁴ A. baumannii infections have also been identified in American soldiers injured in Afghanistan and the Iraq/Kuwait region.⁵ The significant capacity of this bacterium to survive in hospital environment for a long period makes the infection even more problematic, because the transmission rate between patients is high.⁴ Treatment of infection caused by this bacterium is often difficult, because of the widespread resistance of this bacterium to most classes of antibiotics in use. Only a few “abandoned” antibiotics, such as colistin and polymyxin B, are effective, but they can also cause severe side-effects. New drugs to treat this deadly infection with a novel mechanism of action are in urgent need to overcome the drug resistance.³

Methionine aminopeptidase (MetAP) is ubiquitously found in both prokaryotic and eukaryotic cells.⁶ It is responsible for removing the N-terminal methionine from 50–70% of nascent proteins, which is required for localization, activation and degradation.⁶ There are two subtypes of MetAP (type 1 and type 2). Most bacteria and achaea have a single MetAP gene (either type 1 or 2), while eukaryotic cells express both type 1 and type 2 MetAP enzymes. Lethal deletion of this gene has been shown in Escherichia coli ⁷ and Salmonella typhimurium.⁸ Therefore, MetAP is a potential target for developing novel antibiotics. We demonstrated that inhibitors of E. coli MetAP (EcMetAP) displayed antibacterial activity on E. coli, Bacillus megaterium, and Bacillus subtilis.^9–11 Liu and colleagues also reported that inhibitors of Mycobacterium tuberculosis MetAPs displayed antitubercular activity,¹² and inhibitors of Plasmodium falciparum MetAPs displayed antimalarial activity.^{13, 14}

Multiple putative MetAP genes have been identified in a small number of bacteria, when their genomic sequences became known. For example, in genome of A. baumannii ATCC 17978,¹⁵ two genes for putative A. baumannii MetAP (AbMetAP) were annotated. To investigate whether they code for functional MetAP enzymes and whether the MetAPs can serve as drug targets, we cloned and expressed the two MetAP proteins from A. baumannii ATCC 17978 in E. coli and named them as AbMetAPx (gene locus A1S_1940) and AbMetAPy (gene locus A1S_2324). All bacterial MetAPs, such as EcMetAP, belong to type 1 MetAP that contains only a catalytic domain. Both AbMetAPx (264 residues) and AbMetAPy (275 residues) are either identical to or slightly longer than EcMetAP (264 residues) in length, and when the three sequences were aligned, high sequence homology was observed, with 119 identical residues (Fig. 1). Importantly, all five residues (D97, D108, H171, E204, and E235 in EcMetAP) for ligation to two catalytic metal ions ¹⁶ are conserved in both AbMetAPs, as well as the catalytically critical residues C70,¹⁷ H79,¹⁸ and H178 ¹⁹ in EcMetAP. AbMetAPy is longer than the other two by nine residues, and an insert of six residues (GRPAFQ) is possibly located in a loop region (between P66 and K67 in EcMetAP), according to the structure-based sequence alignment. The existence of every key structural element for structure and function in the two AbMetAP sequences suggests that both may be functional MetAP enzymes.

Figure 1. Sequence alignment between EcMetAP, AbMetAPx and AbMetAPy.

The secondary structures (H, helix; S, sheet) are based on our EcMetAP structure (1XNZ) ²¹ and shown above the sequences. Identical residues are shaded, and the five conserved metal-coordinating residues are in bold face and underlined.

The genes for AbMetAPx and AbMetAPy were amplified by PCR, using primers designed with ligation independent cloning (LIC) specific 5’ overhangs (EMD, Gibbstown, NJ) and according to the gene sequences in A. baumannii ATCC 17978 genome (795 nucleotides and 828 nucleotides, respectively). The PCR fragments were cloned into pET-30 Ek/LIC vector using LIC procedures with His-tags at both N and C termini. The recombinant plasmids were transformed into E. coli BL21(DE3) strain for an IPTG-inducible expression. For purification of AbMetAPx, the E. coli strain was cultured in LB media, containing kanamycin, and expression of recombinant AbMetAPx was induced by adding IPTG. The harvested cells were lysed by passing through a French press. The supernatant of the lysate was separated on an HisTrap column. The protein was treated with Chelex-100 resin to remove divalent metal ions and concentrated to 33.5 mg/mL. AbMetAPy was prepared in a similar manner, except that the Chelex 100 resin treatment was replaced by EDTA. The purified AbMetAPy was concentrated to 11.0 mg/mL. The constructed E. coli expression systems produced both AbMetAP proteins in high yield in soluble form (65 mg AbMetAPx or 45 mg AbMetAPy per 500 mL of E. coli cell culture), and both were purified as metal-free preparations. Although the proteins were not analyzed for metal content after treatment with either Chelex-100 (for AbMetAPx) or EDTA (for AbMetAPy), AbMetAPx showed only very weak residual activity, and AbMetAPy showed no detectable activity, indicating most of the metal ions were removed.

All MetAPs studied so far require divalent metal ions for catalysis as purified enzymes,¹⁶ but the actual metal ion used by MetAP in cells is largely unknown.^{9, 20} We have previously shown the importance of inhibiting the MetAP enzyme in its native metalloform for antibacterial activity ⁹ and discovered MetAP inhibitors that can distinguish different metals at the enzyme active site (i.e., metalloform-selective inhibitors).^{10, 21} Therefore, we used the two metal-free AbMetAPs to study their activation by divalent metal ions and inhibition by a set of metalloform-selective inhibitors.

Metal ions directly participate in MetAP catalysis,²² and EcMetAP can be activated by Co(II), Mn(II), Fe(II), Ni(II) and Zn(II).^23–25 To test activation of the AbMetAP proteins by different divalent metal ions, we used a fluorescence assay with the fluorogenic methionyl aminomethylcoumarin (Met-AMC) as the substrate.²⁶ The formation of 7-amino-4-methylcoumarin (AMC) was monitored continuously for 30 min at room temperature by fluorescence (λ_ex = 360 nm, λ_em = 460 nm). While no activation was observed for Zn (II) in this assay, Co(II), Mn(II), Fe(II) and Ni(II) were found to activate both AbMetAPs. The metal activation was instant, and the fluorescence increased linearly for at least 15 min once the metal-free proteins were mixed with the metals. Activity of AbMetAPx or AbMetAPy (both at 1 µM) increased along with increasing amounts of metals until peak values were reached. Inhibition of the enzymatic activity by excessive metal ions at high concentrations was observed, which has been observed in other MetAP enzymes as well.^{25, 27, 28} For AbMetAPx, Ni(II) and Fe(II) were the best activators (Fig. 2A). Both Co(II) and Mn(II) showed much weaker but detectable activation. The metal activation profile of AbMetAPy was similar, except that Fe(II) was found to be the best, and the activation by Fe(II) required much lower metal concentration than that by Ni(II) (Fig. 2B). Using the observable activity as an indication of metal binding to the AbMetAP proteins, we determined the affinity of the activating metals [Co(II), Mn(II), Fe(II) and Ni(II)] to either AbMetAPx or AbMetAPy (Table 1), by monitoring the activity with increasing amounts of the metal, and calculated the K_d values.²⁹ The affinities distributed within a narrow range of 0.14 µM to 1.84 µM, in a good agreement with reported affinities for EcMetAP [0.3 µM, 0.2 µM and 6 µM for Co(II), Fe(II) and Mn(II), respectively].^{30, 31} The K_d value for Ni(II) binding to AbMetAPy at 1.84 µM is significantly higher than other K_d values for the two AbMetAP proteins, and the weak binding is consistent with the shift of Ni(II) activation profile to higher metal concentrations for AbMetAPy (Fig. 2B).

Figure 2. Activation of AbMetAPx (A) and AbMetAPy (B) by divalent metals.

Co(II), closed squares; Mn(II), open squares; Fe(II), closed diamonds; and Ni(II), open diamonds.

Table 1.

Binding and kinetic parameters of AbMetAPx and AbMetAPy ^a

	AbMetAPx				AbMetAPy
	Co(II)	Mn(II)	Fe(II)	Ni(II)	Co(II)	Mn(II)	Fe(II)	Ni(II)
Kd, µM	0.45	0.57	0.14	0.26	0.39	0.33	0.54	1.84
Km, µM	168	154	155	111	130	115	161	144
kcat, s−1 × 104	1.59	0.508	2.63	3.05	1.37	0.550	4.49	3.12
kcat/Km, M−1s−1	0.948	0.330	1.70	2.75	1.05	0.479	2.78	2.17

^aK_d is the dissociation constant. K_m and k_cat are the Michaelis–Menten constants.

Conversion of Met-AMC to fluorescent AMC is easily monitored kinetically by fluorescence, and we used this fluorescence assay to investigate the catalytic efficiencies of AbMetAPx and AbMetAPy, when they were activated by Co(II), Mn(II), Fe(II), and Ni(II). The kinetic constants k_cat and K_m were determined at the optimal concentration of activating metals (for AbMetAPx: 2.5 µM of FeCl₂, 5 µM of NiCl₂, 20 µM of CoCl₂, 20 µM of MnCl₂; for AbMetAPy: 2.5 µM of FeCl₂, 20 µM of NiCl₂, 20 µM of CoCl₂, 20 µM of MnCl₂) (Table 1). For AbMetAPx, the Ni (II)-form was the most efficient metalloform, followed by the Fe(II)-form. The Co (II) and Mn(II) forms were relatively less efficient. The efficiencies of different metalloforms of AbMetAPy are similar to those of AbMetAPx, except that the Fe(II)-form is noticeably more efficient than the Ni(II)-form. These parameters are consistent with the metal titration curves (Fig. 2).

To inhibit the growth of A. baumannii, it is important to know the native metal used by AbMetAPs in cells and to effectively inhibit AbMetAPs in their physiologically relevant metalloform. We have discovered unique metalloform-selective inhibitors of EcMetAP ^{10, 21} and a M. tuberculosis MetAP,²⁸ and we selected a few of them and tested them on the newly purified AbMetAPs for metalloform-selective inhibition. Based on our previous studies, the catechol compounds 1 and 2 were the Fe(II)-form selective, 3 and 4 were the Mn(II)-form selective, and 5–7 were the Co(II)- and Ni(II)-forms selective. When tested against AbMetAPs, the test compounds showed both good potency on some of the metalloforms and similar metalloform selectivity as we have observed before (Table 2).^{10, 21, 28} The metalloform-selective inhibition by these inhibitors further confirms the structural and functional similarity of the two AbMetAPs to other known MetAP enzymes. The similarity of their catalysis and inhibition to other MetAP enzymes confirms that both may function as competent MetAP enzymes in A. baumannii.

Table 2.

Inhibition of AbMetAPx and AbMetAPy by metalloform-selective inhibitors ^a

Compounds	AbMetAPx				AbMetAPy
	Co(II)	Mn(II)	Fe(II)	Ni(II)	Co(II)	Mn(II)	Fe(II)	Ni(II)
1	39	79	32	72	24	34	0.95	56
2	33	22	0.65	27	21	30	0.51	55
3	494	7.1	>500	251	>500	6.1	355	286
4	394	13	358	188	381	0.92	288	186
5	0.88	303	>500	23	0.64	34	>500	2.2
6	2.1	145	>500	44	3.8	103	>500	4.4
7	0.78	64	128	4.9	0.38	14	94	0.40

^aIC₅₀ values are expressed in µM.

^bPurified metal-free AbMetAP proteins were activated by the metals at the optimal activation concentration [for AbMetAPx: 2.5 µM Fe(II), 5 µM Ni(II), 20 µM Co(II) or Mn(II); for AbMetAPy: 2.5 µM Fe(II); 20 µM Co(II), Ni(II) or Mn(II)].

MetAP exists in every cell. Mycoplasma genitalium has one of the smallest genomes with only 521 genes, and MetAP gene is one of its 382 essential genes.³² The recently synthesized 1.08 Mbp Mycoplasma mycoides JCVI-syn1.0 genome has the MetAP gene.³³ The ubiquitous presence of this enzyme, coupled with the essential nature revealed in the gene deletion studies,^{7, 8, 34} suggests that MetAP is a potential target for discovering broad-spectrum antibacterial agents, and its novelty as a drug target may help to develop antibiotics that can overcome the current problematic drug-resistance in bacteria.

Most bacteria have only a single MetAP, but A. baumannii is among a small number of bacteria that have multiple MetAP genes. Two homologous MetAP isozymes in B. subtilis were isolated and investigated.³⁵ Although both showed enzymatic activity, only one of them was shown to be essential for growth, and the other one was concluded as non-essential for reason of low expression.³⁵ Protozoan parasite Plasmodium falciparum has four MetAP proteins, and inhibitors were discovered and characterized on one of the four and showed antimalarial activity.¹³ Therefore, although both AbMetAPx and AbMetAPy are competent MetAP enzymes, their relative biological significance in A. baumannii remains to be elucidated, and either or both may serve as the drug target.

Bacterial MetAPs, including AbMetAPx and AbMetAPy, are also homologous to human type 1 MetAP, and to a lesser extent, to human type 2 MetAP. Because of the significant role of human MetAPs in cell proliferation and angiogenesis,^{36, 37} potentially, antibacterial MetAP inhibitors may exhibit toxicity due to inhibition of human MetAPs. However, the redundant human MetAPs have overlapping substrate specificity,³⁸ which may alleviate some of the potential toxicity of AbMetAP inhibitors. In addition, EcMetAP was found to use Fe(II) as the native cofactor,⁹ and other bacterial MetAPs may use Fe(II), too, due to its availability. On the other hand, Fe(II) is tightly regulated in mammalian cellular environment. The ability for human hosts to sequester iron is a defense mechanism against bacterial infection, and the amount of free iron (10⁻¹⁸ M) is well below the levels required to support the growth of most bacteria (10⁻⁶ to 10⁻⁷ M).^{39, 40} It was reported that human type 2 MetAP uses Mn(II) as its physiological metal,²⁰ and the native metal for human type 1 MetAP is currently unknown but could be Mn(II) due to the low Fe(II) abundance. Therefore, our metalloform-selective inhibitors, especially the Fe(II)-form selective inhibitors, could be the promising leads with such selectivity for bacterial MetAPs.

Abbreviations

MetAP

methionine aminopeptidase

AbMetAP

A. baumannii methionine aminopeptidase

EcMetAP

E. coli methionine aminopeptidase

Met-AMC

methionyl aminomethylcoumarin