Differential roles of the hemerythrin-like proteins of Mycobacterium smegmatis in hydrogen peroxide and erythromycin susceptibility

Hemerythrin-like proteins are oxygen-carrying non-heme di-iron binding proteins and their functions have effect on oxidation-reduction regulation and antibiotic resistance. Recent studies using bioinformatic analyses suggest that multiple hemerythrin-like protein coding sequences might have been acquired by lateral gene transfer and the number of hemerythrin-like proteins varies amongst different species. Mycobacterium smegmatis contains three hemerythrin-like proteins, MSMEG_3312, MSMEG_2415 and MSMEG_6212. In this study, we have systematically analyzed all three hemerythrin-like proteins in M. smegmatis and our results identified and characterized two functional classes: MSMEG_2415 plays an important role in H2O2 susceptibility, and MSMEG_3312 and MSMEG_6212 are associated with erythromycin susceptibility. Phylogenetic analysis indicated that these three proteins have different evolutionary origins, possibly explaining their different physiological functions. Here, combined with biological and phylogenetic analyses, our results provide new insights into the evolutionary divergence of the hemerythrin-like proteins in M. smegmatis.

Hemerythrin-like proteins are non-heme, di-iron and O 2 -binding proteins that are ubiquitous from bacteria to mammals and function in oxygen storage and transport. Bioinformatic evidence indicates that prokaryotic genomes collectively encode hundreds of hemerythrin-like proteins [1][2][3] . Based on the structural characterization of hemerythrin-like proteins, their functions are correlated with redox regulation in bacteria. One study showed that the hemerythrin-like protein in Methylococcus capsulatus functions as an oxygen-carrier 4 , while the hemerythrin-like protein in Campylobacter jejuni acts to protect iron-sulfur cluster enzymes from oxidative damage 5 . Although hundreds of hemerythrin-like proteins have been predicted in bacteria, studies on the biological functions of hemerythrin-like proteins are few. One study showed that the multi domain protein VcBhr-DGC (with a hemerythrin domain and a diguanylate cyclase GGDEF domain) functions as a regulatory oxygen sensor for switching between reducing or anaerobic environments in Vibrio cholerae 6 . This is the first demonstration of a regulatory function for a hemerythrin domain and hints that hemerythrin-like proteins might have other unannotated functions.
The number of hemerythrin-like proteins differs from strain to strain and this variation is predicted to be related to differences in the oxygen concentration of the environment 2 . A number of hemerythrin-like proteins have been found in magnetotactic bacteria and their functions are predicted to be correlated with bacterial physiological conditions (survival under certain oxygen tensions or high concentrations of iron in vivo) 2 . However, the functions predicted for many hemerythrin-like proteins are simply based on their molecular sequences. The functions of multiple hemerythrin-like proteins in one organism have not previously been identified. Multiple homologs are common in bacteria and understanding the functional divergence of paralogs in one organism is a challenge in biology, as the multiplicity of genes for hemerythrin-like proteins is an obstacle to study their distinctive individual functions.
The genus Mycobacterium is comprised of a number of Gram-positive bacteria, including both pathogens, such as Mycobacterium tuberculosis and Mycobacterium leprae, and nonpathogens, such as the soil microorganism Mycobacterium smegmatis, which is commonly used in laboratory experiments as a model organism for M. tuberculosis 7 . Mycobacterium is capable to survive under environmental stresses, such as oxidative stress, hypoxia and exposure to multiple antimicrobial agents 8,9 . The identification of undefined proteins and pathways involved into oxidative stress and antimicrobial response might give new insights to understanding the pathogenesis of M. tuberculosis and response to antibiotic exposure in mycobacteria 10,11 . Mycobacteria are predicted to contain many hemerythrin-like proteins. For example, M. tuberculosis (NC_000962.3) has been predicted to contain three hemerythrin-like proteins. Five genes have been predicted to encode hemerythrin-like proteins in Mycobacterium avium (NC_008595). M. smegmatis possesses three hemerythrin-like proteins, MSMEG_3312, MSMEG_2415 and MSMEG_6212. In this study, we sequentially overexpressed and deleted each of the three genes encoding hemerythrin-like proteins in M. smegmatis. We showed that MSMEG_6212 and MSMEG_3312 modulated erythromycin susceptibility and that the resistance of msmeg_3312 and the msmeg_6212 double-knockout strain, mc 2 155:Δ 3312-6212, was similar to single-knockout strains mc 2 155:Δ 3312 and mc 2 155:Δ 6212. MSMEG_2415 plays a major role in H 2 O 2 susceptibility but not in erythromycin susceptibility. MSMEG_3312 exhibited only a mild H 2 O 2 response in mc 2 155:Δ 2415.
In addition, MSMEG_6212 was not associated with H 2 O 2 susceptibility; overexpression of msmeg_6212 in both mc 2 155:Δ 2415 and the mc 2 155:Δ 2415-3312 double-knockout strain did not influence H 2 O 2 susceptibility relative to the corresponding parental strains. Phylogenetic analysis of bacterial hemerythrin-like proteins showed that three mycobacterial hemerythrin-like proteins are likely derived from different lineages, possibly explaining their different biological functions. Here, combined with analyses of biological function and phylogenetic analyses our results provide new insights into the evolutionary divergence of the hemerythrin-like proteins in M. smegmatis.

Results
MSMEG_6212 is associated with erythromycin susceptibility. To investigate whether the three M. smegmatis hemerythrin-like proteins, MSMEG_3312, MSMEG_2415 and MSMEG_6212, have distinct or overlapping functions, we used a series of strains overexpression individual genes and knockout mutants. The specialized transduction strategy for the sequential deletion of the three genes encoding hemerythrin-like proteins in M. smegmatis, and the overexpression of individual genes encoding hemerythrin-like proteins is shown in Fig. 1. We have previously shown that MSMEG_2415 is involved in the SigF-mediated H 2 O 2 pathway and that MSMEG_3312 is associated with erythromycin susceptibility 12,13 . Here, to characterize the function of MSMEG_6212, we constructed a msmeg_6212 knockout strain (mc 2 155:Δ 6212). Knockout mutants were confirmed by PCR analysis (Fig. 2B). A msmeg_6212 gene fragment was not amplified and no msmeg_6212 mRNA was detected in an assay of its mRNA expression (data not shown). The msmeg_6212 mutant, mc 2 155:Δ 6212, was complemented with a single integrated copy using pMV361-6212. The constructed M. smegmatis mutant strain mc 2 155:Δ 6212 was initially tested for growth in rich 7H9 medium and defined Sauton medium. Growth of mc 2 155:Δ 6212 appeared to have no discernable phenotypic difference from the wild type strain mc 2 155, in either rich (Fig. 3A) or defined media (data not shown). These results indicate that msmeg_6212, like the previous investigated msmeg_3312 and msmeg_2415, is not an essential gene for M. smegmatis growth in either 7H9 rich medium or Sauton defined medium. In order to characterize the potential roles of MSMEG_6212, we compared the minimum inhibitory concentrations (MICs) of eleven antibiotic drugs, and H 2 O 2 in the msmeg_6212 knockout strain mc 2 155:Δ 6212 and wild type strain mc 2 155 (Table S1). Surprisingly, a difference in MIC values was detected only for the macrolides erythromycin and azithromycin (AZM) ( Table S1). To clarify the effect of MSMEG_6212 on erythromycin susceptibility, we performed drug exposure experiments to compare the growth rates of wild-type strain mc 2 155, msmeg_6212 knockout strain mc 2 155:Δ 6212, and the complemented strain pMV361-6212/mc 2 155:Δ 6212 in the presence of 1.56mg/L erythromycin (Fig. 3B). The strain mc 2 155:Δ 6212 showed a growth advantage compared with wild type mc 2 155, which was partially reversed in the complemented strain pMV361-6212/ mc 2 155:Δ 6212 in the presence of erythromycin (Fig. 3B). Furthermore, we compared the survival of various M. smegmatis strains every few hours under treatment with 31.2 mg/L (10x MIC) erythromycin. As shown in Fig. 3C and Fig. S1, the percentage survival of mc 2 155:Δ 6212 was greater than that of wild type mc 2 155, whereas the complemented strain pMV361-6212/mc 2 155:Δ 6212 did not grow well and its survival was partially reversed to that of the wild-type. As overexpression of MSMEG_6212 increased susceptibility to erythromycin, we used 15.6 mg/L (5 × MIC) to perform the killing experiment: overexpression of msmeg_6212 caused greater susceptibility to erythromycin and lower survival than in wild type mc 2 155 under the same treatment ( Fig. 3D and Fig. S1). Taken together, these results show that, like MSMEG_3312, MSMEG_6212 negatively impacts erythromycin resistance.
Both MSMEG_3312 and MSMEG_6212 affect MtrA-mediated erythromycin susceptibility and MSMEG_3312, but not MSMEG_6212, has a redundant role in the H 2 O 2 response. Both mc 2 155:Δ 3312 and mc 2 155:Δ 6212 mutant cells were found to be slightly resistant than the wild type strain mc 2 155 under erythromycin treatment (Table S1). When we compared the mRNA levels of transcriptional regulators MtrA and WhiB7, known to affect erythromycin susceptibility [14][15][16] , among mc 2 155, mc 2 155:Δ 6212 and mc 2 155:Δ 3312, we found no difference in the level of WhiB7 mRNA between the mc 2 155, mc 2 155:Δ 6212 and mc 2 155:Δ 3312 strains with or without erythromycin treatment (data not shown). This result suggests that WhiB7 responds to erythromycin independent of MSMEG_3312, and MSMEG_6212. In contrast, knockout of msmeg_3312 led to a 2.19 ± 0.07 fold increase in the mRNA level of mtrA relative to wild type mc 2 155, and a 1.94 ± 0.05 fold increase in mtrA mRNA in mc 2 155:Δ 6212 (Fig. 4A,D). These increases were partially reversed in the complemented strains pMV361-3312/mc 2 155:Δ 3312 and pMV361-6212/mc 2 155:Δ 6212 (Fig. 4A,D). These results suggest that both msmeg_3312 and msmeg_6212 affect the mRNA level of mtrA. We then examined the influence of MSMEG_3312/MSMEG_6212 on the MtrA-mediated erythromycin response pathway. We In addition, we measured changes in the mRNA levels of MtrA regulon genes msmeg_1875 (encoding sensor histidine kinase MtrB) and msmeg_0637 (encoding iron-sulfur binding oxidoreductase) 17 in response to erythromycin treatment in mc 2 155, mc 2 155:Δ 6212, mc 2 155:Δ 3312 and complemented strains pMV361-3312/mc 2 155:Δ 3312 and pMV361-6212/mc 2 155:Δ 6212. The level of msmeg_1875 and msmeg_0637 mRNA increased in mc 2 155 in response to erythromycin, but induction of msmeg_1875 and msmeg_0637 was abrogated in both mc 2 155:Δ 6212 and mc 2 155:Δ 3312 in response to erythromycin ( Fig. 4B,C,E,F). Correspondingly, induction of msmeg_1875 and msmeg_0637 was restored in both the complemented strain pMV361-3312/ mc 2 155:Δ 3312 and pMV361-6212/mc 2 155:Δ 6212 in response to erythromycin (Fig. 4B,C,E,F). Taken together, those results indicate that both MSMEG_6212 and MSMEG_3312 are required for the MtrA-mediated erythromycin response. To characterize the relationship between MSMEG_3312 and MSMEG_6212, we constructed a double knockout mutant strain mc 2 155:Δ 3312-6212 and assayed its resistance to erythromycin. As shown in Figs 5A and 6A, the resistance of the double-knockout mutant mc 2 155:Δ 3312-6212 to erythromycin appeared to be comparable to mc 2 155:Δ 3312. Moreover, a significant increase of mtrA mRNA in mc 2 155 was observed in response to erythromycin (Fig. 5B), while no significant changes of mRNA level in mc 2 155:Δ 3312-6212 was observed in response to erythromycin (Fig. 5B) The unchanged level of erythromycin resistance in the double-knockout mutant strain mc 2 155:Δ 3312-6212 suggested that there was no cumulative effect in the double-knockout mutant mc 2 155:Δ 3312-6212. Our results thus indicate that MSMEG_3312 and MSMEG_6212 fall in the same pathway. To determine the order of MSMEG_3312 and MSMEG_6212, we analyzed the level of msmeg_6212 mRNA expression in mc 2 155:Δ 3312 and of msmeg_3312 expression in mc 2 155:Δ 6212. As shown in Fig. 5C, the level of msmeg_6212 mRNA decreased 4.86 ± 0.70 fold in mc 2 155:Δ 3312 compared with the wild type strain mc 2 155, while the level of msmeg_3312 mRNA was not significantly different in the msmeg_6212 knockout strain mc 2 155:Δ 6212 (Fig. 5D). We thus reason that MSMEG_3312 is upstream of MSMEG_6212. Knockout of msmeg_3312 decreased the level of msmeg_6212 mRNA and thus dysregulated mtrA mRNA.
As the remaining M. smegmatis hemerythrin-like protein MSMEG_2415 is involved in the H 2 O 2 stress response 13 , we also examined the effects of MSMEG_3312 and MSMEG_6212 on H 2 O 2 susceptibility. Cells of strain mc 2 155:Δ 3312 were treated with 5 mM H 2 O 2 for 3 hour then spotted onto 7H10 medium. No growth defects were observed after H 2 O 2 treatment relative to the wild type strain mc 2 155 (Fig. 5E). In addition, no differences in growth were observed between the wild type strain mc 2 155 and the msmeg_3312 overexpression strain after H 2 O 2 treatment (Fig. 5E). Increased susceptibility to H 2 O 2 was detected only when msmeg_3312 was overexpressed in mc 2 155:Δ 2415 (Fig. 5E). In contrast, we did not detect any growth difference between the wild type strain mc 2 155, the msmeg_6212 knockout strain mc 2 155:Δ 6212 and the msmeg_6212 overexpression strain pMV261-6212/mc 2 155 after H 2 O 2 treatment. Moreover, overexpression MSMEG_6212 in mc 2 155:Δ 2415 and the mc 2 155:Δ 2415-3312 double-knockout strain did not influence H 2 O 2 resistance relative to the corresponding parental strains (Fig. S2A). As MSMEG_2415 is involved in the SigF-mediated H 2 O 2 response 13 , we also measured changes in the mRNA level of sigF and SigF regulon components msmeg_1782 (encoding oxidoreducatse) and msmeg_4753 (encoding antioxidant) in mc 2 155:Δ 6212 and mc 2 155:Δ 3312 relative to that in mc 2 155. Consistent with results for H 2 O 2 survival assays, there were no significant changes in the levels of sigF, Results are shown as the mean ± standard deviation of three replicates (*p < 0.05). The same wild type was used for evaluation the relative expression levels of mtrA and its regulon genes.
Scientific RepoRts | 5:16130 | DOi: 10.1038/srep16130 msmeg_1782 and msmeg_4753 mRNA (Fig. S2B). Taken together, our results show that MSMEG_3312 has no effect on the H 2 O 2 response in the presence of MSMEG_2415, and that MSMEG_3312 contributes to the mild H 2 O 2 susceptibility in mc 2 155:Δ 2415. MSMEG_6212 is not involved in the H 2 O 2 response. MSMEG_2415 is not associated with erythromycin susceptibility. To further examine the role of MSMEG_2415 in erythromycin susceptibility, we overexpressed msmeg_2415 in the mc 2 155:Δ 3312-6212 strain and examined its susceptibility to erythromycin. We incubated 15.6 mg/L erythromycin with the msmeg_3312 knockout strain harboring a pMV261 empty vector (pMV261/ mc 2 155:Δ 3312-6212) or the msmeg_3312 knockout strain harboring pMV261-2415 (pMV261-2415/ mc 2 155:Δ 3312-6212) for 3h and then spotted the cells on 7H10 media. We did not detect any difference in growth between the pMV261/ mc 2 155:Δ 3312-6212 strain and the pMV261-2415/ mc 2 155:Δ 3312-6212 strain under drug treatment (Fig. S3A). Moreover, no differences in growth were observed between the msmeg_3312 knockout strain (mc 2 155:Δ 3312) and the msmeg_3312 and msmeg_2415 double-knockout strain (mc 2 155:Δ 3312-2415) when treated with erythromycin (Fig. S3A). We also measured the level of mtrA mRNA in mc 2 155:Δ 2415, relative to that in mc 2 155. Unlike MSMEG_3312 and MSMEG_6212, knockout of msmeg_2415 did not affect the level of mtrA mRNA (Fig. S3B) and no change in the mRNA levels of MtrA regulon genes msmeg_1875 and msmeg_0637 was detected. Taken together, these results suggest that MSMEG_2415 is not associated with erythromycin susceptibility.
The triple hemerythrin-like gene knockout strain mc 2 155:Δ3312-6212-2415 exhibits comparable erythromycin susceptibility to that of mc 2 155:Δ3312 and comparable H 2 O 2 susceptibility to that of mc 2 155:Δ2415. We next constructed a triple hemerythrin-like proteins knockout strain, mc 2 155:Δ 3312-6212-2415 and confirmed it by PCR (Fig. S4). We also compared erythromycin susceptibility and H 2 O 2 susceptibility in mc 2 155:Δ 3312-6212-2415 with that in wild type mc 2 155, mc 2 155:Δ 3312-6212, and mc 2 155:Δ 2415. The erythromycin susceptibility of the triple mutant strain mc 2 155:Δ 3312-6212-2415 appeared to be indistinguishable from that of the double mutant strain (mc 2 155:Δ 3312-6212) (Fig. 6A). The level of mtrA mRNA and that of its regulon (msmeg_1854 and msmeg_0637) in the triple mutant were comparable to that in mc 2 155:Δ 3312 (Fig. 6B). As expected, the H 2 O 2 susceptibility of mc 2 155:Δ 3312-6212-2415 was comparable to that of mutant mc 2 155:Δ 2415 (Fig. 6C). The level of sigF and SigF regulon (msmeg_4753 and msmeg_1782) mRNA in Δ 3312-6212-2415 was comparable to that in mc 2 155:Δ 2415 (Fig. 6D).  Table S2. Interestingly, MSMEG_3312, MSMEG_2415 and MSMEG_6212 are present in 3 different clades (Fig. 7). All the proteins in the MSMEG_2415 cluster belong to the genus Mycobacterium, while those in the MSMEG_3312 cluster were derived from Mycobacterium, Rhodococcus and Sciscionellas. A large portion of the hemerythrin-like proteins belonging to the Streptomyces, Saccharomonospora and Amycolatopsis were present in the MSMEG_6212 cluster, suggesting that msmeg_6212 may have an independent origin. Taken together, this phylogenetic data suggests that the origins and evolution of MSMEG_3312, MSMEG_2415 and MSMEG_6212 are different. Differences in origins may explain the differences in their physiological functions.

Discussion
In this study, we have systematically evaluated the roles of multiple hemerythrin-like proteins (MSMEG_3312, MSMEG_2415 and MSMEG_6212) on erythromycin and H 2 O 2 susceptibility in M. smegmatis. This study is the first to analyze the function and relationship between multiple hemerythrin-like proteins within one organism.
We showed that MSMEG_6212 is associated with erythromycin susceptibility but not susceptibility to the other drugs tested, including isoniazid (INH), ciprofloxacin (CIP) and rifampin (RFP) ( Table S1). Erythromycin is a macrolide, a class of molecules which targets the 50S ribosome and inhibits bacterial protein synthesis 19 . WhiB7, a transcription factor for the Fe-S cluster, has been shown to be involved in inherent resistance to erythromycin, but not INH 20 . The MtrA-MtrB system has been confirmed as an essential two-component system in mycobacteria 21,22 . Several previous studies have shown that MtrA modulates M. tuberculosis proliferation by binding to the dnaA promoter 23,24 . MrtA has also been shown to be related to antimicrobial resistance in mycobacteria [15][16][17]25 . MtrA has been predicted to target 264 genes, including ABC transporters, ribosomal proteins, and a methyltransferase, all of which are related to drug resistance 17 . MSMEG_3312 and MSMEG_6212 are required for MtrA-mediated erythromycin susceptibility, but not for WhiB7-mediated erythromycin resistance (Fig. 4). Our results are the first to show that MSMEG_3312 and MSMEG_6212 impact the mRNA level of MtrA and thus affect its regulon, causing drug resistance. It will be interesting to explore the relationship between the WhiB7-mediated and MSMEG_6212-involved-MtrA-mediated pathways in erythromycin susceptibility.
Multiple homologs are common in mycobacteria. For example, M. tuberculosis contains five resuscitation-promoting factor (Rpf)-like proteins and ten universal stress proteins (USPs) [26][27][28] . It was not possible to determine the exact function of USP proteins Rv1996, Rv2005c, Rv2026c and Rv2028c by knockout of individual usp genes, suggesting that USP proteins in M. tuberculosis have redundant functions 26 . Similarly, reports on the M. tuberculosis rpf genes indicate that they have redundant roles 27 . M. smegmatis possesses three hemerythrin-like proteins, MSMEG_3312, MSMEG_2415 and MSMEG_6212 29 . We defined the hierarchy of biological functions among hemerythrin-like proteins (Fig. 8). The double-knockout mutant strain mc 2 155:Δ 3312-6212 showed comparable erythromycin resistance to that of the single-knockout mutant mc 2 155:Δ 3312 (Figs 5A and 6A). In addition, msmeg_3312 influenced the level of msmeg_6212 mRNA, but not vice versa (Fig 5C,D). We thus reasoned that MSMEG_3312 and MSMEG_6212 are in same pathway for erythromycin susceptibility mediated by MtrA. Knockout of the msmeg_3312 gene leads to a decrease in the level of msmeg_6212 mRNA and upregulation of MtrA expression (Fig. 4A). Similarly, knockout of msmeg_6212 disrupted the regulation of MtrA (Fig. 4D). MSMEG_2415 has previously been shown to be important for H 2 O 2 susceptibility 13 . The contribution of MSMEG_3312 was minor; the overexpression of MSMEG_3312 only slightly increased H 2 O 2 susceptibility in mc 2 155:Δ 2415 (Fig. 5E). Taken together, these results indicate that MSMEG_2415 and MSMEG_6212 are exclusively associated with H 2 O 2 susceptibility and erythromycin susceptibility, respectively, while MSMEG_3312 is associated with both H 2 O 2 and erythromycin susceptibility (Fig. 8).
Phylogenetic analysis of respiratory hemerythrin-like proteins and hemocyanins shows that the distribution of analyzed respiratory proteins may partially explain physiological adaptions 3 . Comparative genomic analysis of M. indicus prannii suggested that multiple hemerythrin-like protein coding sequences might have been acquired by lateral gene transfer and these proteins help mycobacteria survival in different oxygen concentrations of the environment 30 . In this study, we showed that the three mycobacterial hemerythrin-like proteins have different functions and that phylogenetic analysis of hemerythrin-like proteins in M. smegmatis indicates that the three proteins are distributed within three distinct clades (Fig. 7). Interestingly, all the proteins in the MSMEG_2415 cluster belong to the genus Mycobacterium, suggesting that MSMEG_2415-like hemerythrin proteins are more conserved in mycobacteria. The role of MSMEG_2415 in H 2 O 2 susceptibility might be an inherent function in Mycobacterium. In contrast, some of the hemerythrin-like proteins in the MSMEG_6212 cluster belong to the genus Mycobacterium, and several belong to Saccharomonospora and Amycolatopsis, suggesting that msmeg_6212 might be of independent origin. Strikingly, the MSMEG_3312 cluster included representatives of the genera Streptomyces, Saccharomonospora, Saccharopolyspora, Nocardiopsis, and Amycolatopsisand a few Mycobacterium species. Of interest, Saccharopolyspora erythraea, an environmental soil actinomycete, can produce the natural antimicrobial agent erythromycin 31 . Soil is a highly complex and competitive environment, in which interactions between diverse organisms occur. Evolutionary pressures in soil are high due to competition for resources have significant selective advantages over competitors. M. smegmatis is also an environmental soil strain, the acquisition of antimicrobial-related regulatory proteins (MSMEG_3312 and MSMEG_6212) might have a selective advantage facilitating its survival. This independent phylogenetic clade may explain the different roles of MSMEG_2415 (in H 2 O 2 susceptibility), MSMEG_6212 (in erythromycin susceptibility) and MSMEG_3312 (in both erythromycin and H 2 O 2 susceptibility). Further work to identify the functions of M. tuberculosis hemerythrin-like proteins and a comparison of their hemerythrin-like protein with those in M. smegmatis would provide insight into the evolution and selection of virulence and antibiotic susceptibility.
In summary, we have systematically analyzed all three hemerythrin-like proteins in M. smegmatis and our results indicated that the three members of this protein family possess overlapping and distinct functions: MSMEG_2415 plays an important role in H 2 O 2 susceptibility, and MSMEG_3312 and MSMEG_6212 can modulate erythromycin resistance. Phylogenetic analysis indicated that these three proteins have different evolutionary origins, possibly explaining their different physiological functions. The functional and phylogenetic analyses of hemerythrin-like proteins in M. smegmatis would provide insight into the evolutionary selection of antimicrobial resistant traits.
Construction of knockout hemerythrin-like protein MSMEG_6212 strain and corresponding complemented and overexpression strains. Mycobacteriophage-based specialized transduction was used to generate hemerythrin-like gene knockout strains 13,32 . Briefly, the 5' and 3' sequences flanking the msmeg_6212 gene were amplified from M. smegmatis genomic DNA using the following PCR conditions: 98 °C for 3 min, 32 cycles of 98 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s, and 72 °C for 10 min. The primers for msmeg_6212 knockout are listed in Supplemental Table S3, and the corresponding positions are indicated in Fig. 2A. The primers used had a PflMI site on the 5' end to allow insertion into the pYUB1471 vector, and the temperature sensitive phage phAE159 and pYUB1471-6212 were then digested with PacI and ligated using T4 DNA ligase to create a shuttle plasmid. MaxPlax packaging extract (Epicenter Biotechnologies, USA) was used for phage packaging and transformed into E. coli HB101 cells according to the manufacturer's instructions. Successful phasmids were transduced into M. smegmatis strain mc 2 155 at 30 °C, which allowed replication and amplified high titer phages. Transduction into M. smegmatis was then performed at 37 °C with the high titer lysate at an MOI of 10:1. Gene knockout was confirmed by PCR screening using primers outside the upstream and downstream flanking regions and the corresponding vector primers. Complemented strain of msmeg_6212 was constructed by cloning the full-length genes into the integrating vector pMV361 to yield pMV361-6212/ mc 2 155:Δ 6212.
To obtain the msmeg_3312 and msmeg_6212 double knockout strain, we unmarked the mc 2 155:Δ 3312 strain used in previous study 12 according to a previously published method 33 . Briefly, plasmid pJH532 was transformed into the mc 2 155:Δ 3312 strain by electroporation and plated onto 7H10 media containing 25 mg/L kanamycin. Kanamycin resistant colonies were screened by a pick-and-patch method for hygromycin sensitivity, streaked on 7H10 media alone and on 7H10 media with 50 mg/L hygromycin. The hygromycin-sensitive colonies were then plated onto 7H10 media with 5% sucrose. The selected colonies were spread on 7H10 media supplemented with 5% sucrose to obtain kan S hyg S colonies. The unmarked mc 2 155:Δ 3312 strain was used for construction of the msmeg_6212 knockout strain, yielding the mc 2 155:Δ 3312-6212 double-knockout strain. The triple mutant mc 2 155:Δ 3312-6212-2415 was generated from the double-knockout progenitor mc 2 155:Δ 3312-6212.
To overexpress hemerythrin-like proteins, the corresponding full-length coding genes, msmeg_3312, msmeg_2415 and msmeg_6212, were sub-cloned into pMV261 to yield pMV261-3312, pMV261-2415 and pMV261-6212 for transformation into the corresponding M. smegmatis strains (All strains used in this study are listed in Table S4).
Drug and H 2 O 2 susceptibility testing. The killing curve under erythromycin treatment was determined as indicated. Logarithmic phase cultures (OD 600 ~ 0.3) were treated with erythromycin at the indicated concentrations, aliquots (~50 μ l) were removed at the indicated times and spread onto 7H10 medium. Colony Forming Units (CFUs) were counted after 3 days of incubation. Experiments were repeated at least 3 times.
Survival under erythromycin and H 2 O 2 treatment was determined as indicated. Logarithmic phase cultures (OD 600 ~ 0.3) were treated with erythromycin (15.6 mg/L) or H 2 O 2 (5 mM) at the indicated concentrations for 3 h, serially diluted (1:10) and spotted (3 μ l) onto 7H10 medium. Photographs were taken after three days of incubation at 37 °C. Experiments were repeated at least 3 times.
Quantitative real-time PCR analysis. Logarithmic phase (OD 600 ~ 0.3)cultures of the corresponding M. smegmatis strains treated with 0 or 3.125 mg/L erythromycin for 30 min were collected. After resuspending in TRIzol (Invitrogen), RNA was purified according to the manufacturer's instructions. The SuperScriptTM III First-Strand Synthesis System (Invitrogen) was used to synthesize the corresponding cDNA. qRT-PCR was performed on a Bio-Rad iCycler. M. smegmatis rpoD (the coding sequencing of RNA polymerase sigma factor SigA) was used to normalize gene expression. The relative ratio was calculated using the 2 −ΔΔCT method 34 . Experiments were repeated at least 3 times. Primers used for qRT-PCR are listed in Table S3.
Statistical method. Statistical analysis was performed with GraphPad Prism 5.0c software. Significant differences in the data were determined using t-tests. P values of < 0.05 were considered significant.