Mab_3168c, a Putative Acetyltransferase, Enhances Adherence, Intracellular Survival and Antimicrobial Resistance of Mycobacterium abscessus

Mycobacterium abscessus is a non-tuberculous mycobacterium. It can cause diseases in both immunosuppressed and immunocompetent patients and is highly resistant to multiple antimicrobial agents. M. abscessus displays two different colony morphology types: smooth and rough morphotypes. Cells with a rough morphotype are more virulent. The purpose of this study was to identify genes responsible for M. abscessus morphotype switching. With transposon mutagenesis, a mutant with a Tn5 inserted into the promoter region of the mab_3168c gene was found to switch its colonies from a rough to a smooth morphotype. This mutant had a higher sliding motility but a lower ability to form biofilms, aggregate in culture, and survive inside macrophages. Results of bioinformatic analyses suggest that the putative Mab_3168c protein is a member of the GCN5-related N-acetyltransferase superfamily. This prediction was supported by the demonstration that the mab_3168c gene conferred M. abscessus and M. smegmatis cells resistance to amikacin. The multiple roles of mab_3168c suggest that it could be a potential target for development of therapeutic regimens to treat diseases caused by M. abscessus.


Introduction
Mycobacterium abscessus is a rapid growing mycobacterium. It has emerged as an important pathogen of soft tissue, pulmonary, and disseminated infections in both immunocompromised and immunocompetent patients [1,2,3,4]. The soft tissue infections are mainly due to penetrating trauma or surgery. A study of 86 nontuberculous mycobacterial infections of surgical wound and tympanic membrane in central Taiwan found that 100% of these cases were caused by M. abscessus [5,6,7,8].
M. abscessus is one of the most drug-resistant, rapid-growing mycobacteria [2,9,10]. Like other mycobacteria [11], M. abscessus has a complex hydrophobic cell wall that constitutes an efficient permeability barrier. Based on analyses of genomic sequences, M. abscessus is predicted to produce b-lactamases, aminoglycoside phosphotransferases, and aminoglycoside acetyltransferases that may confer multiple drug resistance [12]. M. abscessus is an intracellular pathogen [13,14]. In culture, M. abscessus exhibits two different colony morphology types referred to as rough and smooth morphotypes [13,15]. These morphotypes correlate with the virulence of M. abscessus, and cells with a rough morphotype are more virulent.
The mmpL4b gene in the glycopeptidolipid biosynthesis pathway has been shown to be responsible for switching M. abscessus colonies from a smooth to a rough morphotype [16]. In this study, we identified a gene designated mab_3168c, whose function was unknown, and found that mab_3168c controlled the switching of M. abscessus colony morphology from a rough to a smooth morphotype. We also found that mab_3168c played a role in biofilm formation, intracellular survival, and resistance to antimicrobial agents.

Screening and Identification of M. abscessus Mutants with Colony Morphotype Switching
In order to identify the genes involved in M. abscessus colony morphotype switching, Tn5 transposon mutagenesis was performed. A mutant designated mab_3168c::Tn (Fig. 1A) that switched its colonies from a rough to a smooth morphotype was identified. Characterization of the genome of this mutant revealed that the transposon was inserted into a place 56 bp upstream from the initiation codon (GTG) of mab_3168c (GenBank accession no. NC_010397) and 76 bp downstream from the stop codon of ispG (mab_3169c) (Fig. 1B). To confirm that this morphotype switching was due to the defect in mab_3168c, the intact mab_3168c gene was cloned into the E. coli/mycobacterium shuttle vector pYUB412A to generate pYUB412A-mab_3168c and then introduced into the mab_3168c::Tn mutant. This complementation was found to almost completely convert the colonies of the mab_3168c::Tn mutant from a smooth back to a rough morphotype (Fig. 1A), suggesting that the mab_3168c gene conferred M. abscessus the ability to form rough colonies. Since this complementation was not complete (Fig. 1A), RT-PCR was performed to determine mRNA levels of mab_3168c. No mab_3168C mRNA band was detected in the samples from the mab_3168c::Tn mutant (Fig. 1C), and mab_3168c mRNA levels in the mab_3168c complemented mutant were approximately 60% that of the wild type (Fig. 1D). This result suggested that the incomplete complementation of morphotype was due to sub-optimal expression of the mab_3168c gene introduced into the mutant.
As the Tn5 was inserted 76 bp downstream of the ispG gene, ispG mRNA levels were also determined and found to be the same in the wild type, mab_3168c::Tn mutant, and the complemented strains (Fig. 1C). These results indicated that the Tn5 insertion inactivated mab_3168C but did not affect the expression of its neighboring gene, ispG which encodes 4-hydroxy-3-methylbut-2en-1-yl diphosphate synthase [12,17]. For simplicity, the mab_3168c::Tn mutant and mab_3168c-complemented mab_3168c::Tn mutant will be referred to as the mutant and the complemented mutant, respectively, hereafter.

Increased Sliding Motility of the mab_3168c Mutant
Previous studies [18] have shown that the smooth strains of M. smegmatis and M. avium have a higher sliding ability. Therefore, the motility of the mutant cells on agar plates was examined. As shown in Fig. 2, the wild type M. abscessuss cells were non-motile (0.0760.02 mm), whereas the mutant cells were highly motile (1.5560.13 mm). The complemented mutant cells regained the non-motile phenotype (0.2560.05 mm). These results indicated that mab_3168c played a major role in inhibiting the motility of M. abscessus.

Decreased Cell Surface Hydrophobicity, Biofilm Formation and Lysozyme Susceptibility of the mab_3168c Mutant
Since the hydrophobicity of cell surface is associated with the sliding activity of mycobacteria [19], experiments were performed to investigate whether the hydrophobicity of the mutant was altered. M. abscessus cells of the wild type, mutant, and complemented mutant were grown in liquid Middlebrook 7H9 medium without Tween 80 for 3 days. The mutant exhibited a homogenously dispersed culture, whereas the cultures of both the wild type and complemented mutant had a clear supernatant with cells aggregated in the bottom of the culture tube (Fig. 3A). To adjust for possible variations in growth rates of the 3 strains, an aggregation index of each culture, which is the value of the number of aggregated cells divided by that of dispersed cells, was calculated. The mutant culture was found to have an aggregation index less than 2.5, but the wild type and complemented mutant cultures had aggregation indices of 1261.2 and 760.9, respectively (Fig. 3B). These results indicated that the mutant had a greatly reduced ability to aggregate.
As cell surface hydrophobicity is a determinant of adhesion [20,21], the ability of the mutant to form biofilms was examined. Cells were grown in wells of a 96-well polyvinylchloride plate for 6 days. As shown in Fig. 4A and 4B, substantial formation of biofilms was observed in the wells of wild type M. abscessus cultures (OD 595 = 0.43060.03). In contrast, biofilm formation by the mutant was diminished (OD 595 = 0.11360.005), and the complemented mutant regained the biofilm-forming ability (OD 595 = 0.37860.03). These results indicated that the mab_3168c gene conferred M. abscessus cells the ability to form biofilms by increasing the cell surface hydrophobicity. The reduced ability of the mutant to form biofilm was not due to decreased growth rate because the mutant cells grew equally well as the wild type cells in Middlebrook 7H9 medium (Fig. 4C).
The susceptibility of the mutant to lysozyme was then assessed. Cells were grown in 7H9 broth containing varying amounts (0, 0.5, and 2.5 mg/ml) of lysozyme. The results showed that the mutant cells were much more susceptible to 0.5 mg/ml of lysozyme than the wild type cells (19% vs. 51% survival) (Fig. 5). Cells of the complemented mutant were found to be almost as resistant to lysozyme as those of the wild type (59% vs. 51% at 0.5 mg/ml lysozyme) (Fig. 5).

Decreased Intracellular Survival of the mab_3168c Mutant in Macrophages
The correlation between lysozyme susceptibility and intracellular survival of M. abscessus cells was then evaluated. THP-1 macrophages were infected with wild type, mutant, and complemented mutant at a multiplicity of infection (MOI) of 1.  (Fig. 6A). To confirm this result, confocal microscopy was performed to enumerate intracellular mycobacteria. The result showed that the number of mab_3168c::Tn mutant was significantly lower (average 6 vs. 20 organisms per cell) than that of wild type and complemented (approximately 16 organisms per cell) strains at 72 hours post infection (Fig. 6B). To show that this lower intracellular organism count was not due to death of infected THP-1 cells, the viability of uninfected and infected cells was assessed by determining LDH levels in culture supernatants. As shown in Fig. 6C, similar levels of LDH were observed in the culture supernatants of THP-1 cells infected with wild type, mutant, and complemented mutant (approximately 10% of that of total cell lysate). Very little LDH was detected in the culture supernatant of uninfected cells. Taken together, these results demonstrated that mab_3168c was required for intracellular survival of M. abscessus in macrophages.

Bioinformatics Analysis of mab_3168c
Bioinformatic analyses of the Mab_3168c protein were performed to investigate its possible functions (data not shown). A large number of proteins with a significant homology to the 270aa Mab_3168c protein were found by BLASTP. Most of these proteins were members of the GCN5-like N-acetyltransferase (GNAT) superfamily. Reversed Position Specific Blast (RPS-BLAST) analyses revealed the presence of an acetyltransferase domain of the pfam00583 family. This domain extends from residues 205 to 259 of the putative Mab_3168c protein and contains the consensus sequence V/I-x-x-x-x-Q/R-x-x-G-x-G/A of acetyltransferases [22]. Results of 3D-PSSM prediction also showed that Mab_3168c bears a strong structural similarity to several acetyltransferases of the GNAT family, especially to the aminoglycoside 69-N-acetyltransferase of Enterococcus faecium. These

Increased Susceptibility of the Mutant to Amikacin
As the aminoglycoside 69-N-acetyltransferase of Enterococcus faecium contributes to its aminoglycoside resistance [23], the possibility that mab_3168c conferred M. abscessus antibiotic resistance was examined. Cells of the wild type, mutant, and complemented mutant were grown on 7H11 agar plates with or without rifampin, ciprofloxacin, or amikacin. Other aminoglycosides such as kanamycin, neomycin, paromomycin, ribostamycin, and gentamycin B were not tested because the transposon used for mutagenesis contains the kanamycin-resistance gene, which also confers resistance to these aminoglycosides. No difference in susceptibility to rifampin and ciprofloxacin was observed among the three different strains (data not shown). However, cells of the mutant were more sensitive to amikacin than those of the wild type and the complemented mutant (Fig. 7A). In an overnight culture inoculated with 10 7 or 10 6 CFU/ml, the growth of mutant cells was completely inhibited by 20 mg/ml of amikacin, and the growth of the culture inoculated with 10 8 was inhibited by 6.164.4% with a survival rate of (4.163.0)610 25 , which was calculated as the CFUs on the amikacin plate divided by those on the plate without amikacin. Cells of the complemented mutant were almost as resistant as those of the wild type to 20 mg/ml of amikacin with a survival rate of (6.760.5)610 24 and (7.862.0)610 24 , respectively (Fig. 7A). To confirm amikacin susceptibility, the E test was carried out. As shown in Table 1, the amikacin MICs of both the wild type and complemented strains were 4 mg/ml, but the MIC of the mab_3168c::Tn mutant was only 2 mg/ml.
To further confirm that the mab_3168c gene conferred resistance to amikacin, it was introduced into a different mycobacterium, M. smegmatis, and then assayed the transformants for their susceptibility to amikacin. The MIC value of M. smegmatis cells containing pYUB412A-mab_3168c was two-fold higher than those containing the vector pYUB412A (0.25 vs. 0.12 mg/ml) ( Fig. 7B and Table 1).

Discussion
In this study, we showed that loss of mab_3168c expression resulted in alterations in colony morphology, cell surface hydrophobicity, sliding motility, biofilm forming ability, amikacin and lysozyme resistance, and intracellular survivability of M. abscessus. Although bioinformatic analyses revealed the presence of a sequence motif characteristic of the GCN5-related N-acetyltransferase (GNAT), it remains to be determined whether the putative Mab_3168c protein is an acetyltransferase as our repeated attempts to express the mab-3168c gene have not been successful.
The relationship between an acetyltransferase and colony morphology was first determined in M. smegmatis [24]. In this organism, disruption of the atf1 gene was found to cause its colonies to switch from a smooth to a rough morphotype [24]. The atf1 gene encodes an O-acetyltransferase which is believed to acetylate glycopeptidolipids (GPLs) [24]. However, we did not observe any differences in the lipid profiles of both the wild type and mab_3168c mutant cells by thin-layer chromatography (TLC) and MALDI-TOF analysis (data not shown).
It was unexpected to observe that the cell wall lipid profiles of the wild type with a rough morphotype and the mutant with a smooth morphotype had the same lipid profiles as GPL is believed to make mycobacterial colonies smooth [16,19,24,25]. One  possibility is that other cell wall components also affect colony morphology. In mycobacteria, UDP-N-acetylglucosamine (UDP-GlcNAc), a major component of peptidoglycan, is synthesized by the GlmU protein, which is a glucosamine-1 phosphate acetyltransferase. Mutation of glmU has been shown to impair the synthesis of peptidoglycan and reduce the growth of M. smegmatis [26]. It is possible that instead of affecting GPL production, Mab_3168c affects the synthesis of other cell wall components, similar to the GlmU protein.
Another gene that has been shown to be associated with colony morphotype of M. abscessus is mmpL4b, which encodes a membrane protein [16]. Deletion of this gene renders M. abscessus unable to produce GPL and to form smooth colonies. The DmmpL4b mutant also loses the sliding motility and the ability to form biofilms.
However, it survives better in macrophages. In this study, we found that the mab_3168c mutant formed smooth colonies, gained the sliding motility, but lost the ability to survive inside macrophages. These three properties are opposite to those of the DmmpL4b mutant. However, similar to the DmmpL4b mutant, the mab_3168c mutant also lost the ability to form biofilms. These results suggest that biofilm formation is controlled by multiple mechanisms, and both mmpL4b and mab_3168c genes regulate biofilm formation. Further support of this hypothesis is the finding that inactivation of the lsr2 gene renders M. smegmatis hyper motile and unable to form biofilms [27], very similar to the mab_3168c mutant. The lsr2 mutant also has no changes in cell wall lipid profiles [27,28]. In M. tuberculosis, Lsr2 is a nucleoid-associated protein, similar to the histone-like nucleoid structural protein H-NS [29,30] and is involved in the regulation of cell wall synthesis [31] as well as transcription suppression of many genes [30]. Although lsr2 and mab_3168c mutants are phenotypically similar, lsr2 and mab_3168c genes are distinct with no significant sequence homologies.
The mab_3168c gene was shown to confer M. abscessus resistance to amikacin which is a semisynthetic aminoglycoside derived from kanamycin. Many bacteria that are resistant to gentamicin and tobramycin are sensitive to amikacin. Therefore, amikacin is often used to treat M. abscessus infections [32]. Aminoglycoside resistance may be due to decreased cell permeability, alterations in ribosome binding, or inactivation by aminoglycoside modifying enzymes [33]. It is likely that mab_3168c is involved in cell wall synthesis making M. abscessus cells less permeable to amikacin. It is also possible that Mab_3168c acetylates amikacin rendering it inactive. In addition to Mab_3168c, M. abscessus may also produce other enzymes that can inactivate antibiotics as analyses of genomic sequences revealed its potential to produce b-lactamases and aminoglycoside converting enzymes including aminoglycoside Ophosphotransferases and aminoglycoside N-acetyltransferases [12,34,35]. This property could explain the multiple drug resistance of M. abscessus.
We also found that inactivation of mab_3168c decreased the ability of M. abscessus to survive inside macrophages. This defect is likely due to increased susceptibility to lysozyme. This possibility is supported by the finding that disruption of the aminoglycoside 29-  N-acetyltransferase gene, acc(29)-Id, renders M. smegmatis susceptible to lysozyme [36].
In conclusion, compared to the known characteristics of different members of the GNAT superfamily, we predict that Mab_3168c is an N-acetyltransferase. Inactivation of mab_3168c may cause changes in the structure of the cell wall, resulting in a pleiotropic phenotype of M. abscessus with altered colony morphotype, increased sliding motility, reduced cellular aggregation and ability to survive inside macrophages, and increased susceptibility to amikacin. Since mab_3168c plays a role in many different cellular functions, it could be a good target for development of drugs against M. abscessus.

Transposon Mutagenesis and Genetic Analysis of Mutants
The M. abscessus transposon mutant library was generated using the EZ-Tn5 TM ,KAN-2.Tnp Transposome TM Kit (EPICEN-TRE, USA). Approximately 2000 mutants were screened to detect the ones with altered colony morphology. The Tn5 insertion site in the chromosome was identified by inverted PCR using KAN-2 FP-1 forward and KAN-2 RP-1 reverse primers (Fig. 1B) and by DNA sequencing.

Complementation of Mutants with mab_3168c
The mycobacterial shuttle vector pYUB412 [37] was used as a backbone in which the hygromycin-resistance gene was replaced by an apramycin-resistance gene to construct pYUB412A. Genomic DNA of M. abscessus was used as template for PCR to amplify the region encompassing the whole-length mab_3168c  gene and 1 kb upstream of the gene, using the primer pair mab_3168c-F (59-AGGTATACCATCTTCGCGGCGAT-39) and mab_3168c-R (59-AGCTCGAGTTAGCTGACGGGGA-39), containing BstZ17I and XhoI sites (underlined), respectively. The resulting 1.8-kb DNA fragment was cloned into pYUB412A between ZraI and SalI sites, generating plasmid pYUB412A-mab_3168c. For complementation, pYUB412A-mab_3168c was introduced into mab_3168c::Tn mutant by electroporation. Electrocompetent M. abscessus cells were prepared by growing them to mid-log phase. The cells were harvested, washed in 10% glycerol and then resuspended in cold 10% glycerol at a concentration of 10 7 cells/ml.

Sliding Motility Test
One colony of each mycobacterium was inoculated in the center of a motility plate, consisting of Middlebrook 7H9 with 0.3% agar. The inoculated plates were incubated at 37uC for 5 days [18]. The sliding distance was measured in millimeters.

Aggregation Capability Assay
Mycobacterial cells were incubated in a tube containing 5 ml of Middlebrook 7H9 broth at a concentration of OD 600 = 0.1 and incubated on a shaker at 37uC for 3 days. After allowing the culture tube to stand still for 3 hours, the upper portion of the culture containing dispersed cells was removed, and its OD 600 value was determined. The OD 600 of the bottom portion of culture was measured after the aggregated cells had been completely suspended by vortexing with glass beads of 4.5 mm in diameter (Biospec, USA) as described previously [13,38,39]. The aggregation index was calculated as the ratio of optical density of aggregated cells to that of dispersed cells.

Biofilm Formation
Mycobacterial cells at a concentration of OD 600 = 0.1 were inoculated in 100 ml of Middlebrook 7H9 broth in each well of a sterile 96-well, flat-bottom polyvinylchloride plate (BD, USA). After 6 days of incubation, the medium in each well was removed, and the wells were washed with sterile PBS to remove nonadherent cells. The wells were then stained with 0.5% (wt/vol) crystal violet for 1 hour. After washing with PBS, the stained biofilms were photographed. To quantitate cells, cells in the biofilm were suspended in 100% ethanol, and the OD 595 value of the cell suspension was determined [40].

Lipid Extraction and Analysis
Total lipids from mycobacterial cells of plate-grown cultures were extracted with chloroform/methanol (2:1, v/v) at 56uC for 60 min with sonication. The extracted lipids were spotted on an aluminum-backed silica gel 60 TLC plate (MERCK, German) and resolved with a solvent containing chloroform and methanol at a ratio of 90:10 (v/v) or chloroform, methanol, and water at a ratio of 100:16:2 (v/v) or 60:16:2 (v/v) as previously described [41,42]. To visualize lipids, the plate was sprayed with 1% 1-naphthol, 5% H 2 SO 4 in ethanol and then charred with a heat gun until spots with hues characteristic of different lipid classes appeared. The mass of each lipid species was determined by matrix-assisted laser desorption ionization-time-of-flight (MALDI-TOF) spectrometry with a pulse laser emitting at 337 nm. Samples were mixed with 2,5-dihydroxybenzoic acid as the matrix and analyzed in reflectron mode with an accelerating voltage of 25 kV.

Lysozyme Susceptibility Assay
M. abscessus cells (10 7 /ml) were inoculated into 7H9 broth containing various concentrations of lysozyme (0.5 mg/ml and 2.5 mg/ml) and incubated at 37uC for 24 hours, followed by enumeration of CFU on 7H11 agar plates. Infection of Human Macrophages THP-1 cells, a human acute monocytic leukemia cell line, were obtained from the American Type Culture Collection (ATCC) and cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS, GIBCO) at 37uC in a humidified CO 2 incubator. THP-1 cells were differentiated into adherent macrophages by adding 500 ng/ml of phorbol-12-myristate-13-acetate (PMA) to the culture. Two days after addition of PMA, the cells were infected with mycobacteria at a multiplicity of infection (MOI) of 1 for 2 hours at 37uC. The infected macrophages were washed with sterile PBS to remove extracellular mycobacteria, lysed with 1% Triton X-100, and then plated on 7H11 agar plates to determine the colony forming unit of intracellular mycobacteria as describe previously [43,44,45].
Viability of M. abscessus-infected THP-1 macrophages was evaluated by measuring the levels of lactate dehydrogenase (LDH) in culture supernatants. THP-1 macrophages were infected with wild type, mab_3168c::Tn mutant, and complemented mutant. The levels of LDH activity in the culture supernatants were determined using a CytoTox 96 assay kit (Promega, USA) according to manufacturer's protocol.

Confocal Microscopy
Infected macrophages were fixed with 4% paraformaldehyde for 15 min and permeabilized with 0.1% Triton-X 100 for 20 min. The intracellular mycobacteria were stained with auramine (Sigma, USA) for 20 min at 25uC, treated with 0.5% acid alcohol for 3 min, and then examined under a confocal laser scanning microscope (Leica SP5 confocal Microscopy equipped with a 100X NA1.4 objective lens).

Antimicrobial Susceptibility Test
Susceptibility to amikacin was determined by the E test. M. abscessus and M. smegmatis cells were incubated in 7H9 medium until their culture turbidity reached McFarland standard of 1.0 (,3610 8 CFU/ml). This cell suspension was then spread on a 7H11 agar plate (10 cm) supplemented with 10% OADC using a cotton swab. An amikacin E test strip (Oxoid) was then placed on the plate, and the plate was incubated for 3-5 days until the MIC was read. The value shown on the strip at the place where the strip intersected the growth inhibition zone was the amikacin MIC of the organism tested. The MIC data presented were average of duplicate determinations.