Requirement for kasB in Mycobacterium mycolic acid biosynthesis, cell wall impermeability and intracellular survival: implications for therapy

Authors


Summary

Mycobacterium tuberculosis infects one-third of the world's population and causes two million deaths annually. The unusually low permeability of its cell wall contributes to the ability of M. tuberculosis to grow within host macrophages, a property required for pathogenesis of infection. Mycobacterium marinum is an established model for discovering genes involved in mycobacterial infection. Mycobacterium marinum mutants with transposon insertions in the β-ketoacyl-acyl carrier protein synthase B gene (kasB) grew poorly in macrophages, although growth in vitro was unaffected. Detailed analyses by thin-layer chromatography, nuclear magnetic resonance (NMR), matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, infrared spectroscopy, and chemical degradations showed that the kasB mutants synthesize mycolic acids that are 2–4 carbons shorter than wild type; the defect was localized to the proximal portion of the meromycolate chain. In addition, these mutants showed a significant (∼30%) reduction in the abundance of keto-mycolates, with a slight compensatory increase of both α- and methoxy-mycolates. Despite these small changes in mycolate length and composition, the kasB mutants exhibited strikingly altered cell wall permeability, leading to a marked increase in susceptibility to lipophilic antibiotics and the host antimicrobial molecules defensin and lysozyme. The abnormalities of the kasB mutants were fully complemented by expressing M. tuberculosis kasB, but not by the closely related gene kasA. These studies identify kasB as a novel target for therapeutic intervention in mycobacterial diseases.

Introduction

The global burden of Mycobacterium tuberculosis infection is overwhelming, with approximately one-third of the world's population infected and two million deaths each year (Dolin et al., 1994). High rates of disease and death from M. tuberculosis infection are primarily due to the bacterium's extreme infectivity and its ability to persist intracellularly and in other sites of the host poorly accessible to antibiotics. Resistance to commonly used antibiotics also has emerged as an increasing problem for control of the global epidemic. Mycobacterium tuberculosis has evolved sophisticated strategies to combat host defence mechanisms through co-evolution with humans and can modulate a number of host cell processes to benefit its intracellular parasitism (Gao and Kwaik, 2000; Russell, 2001). It also possesses a cell wall of unusually low permeability that resists an array of host defence mechanisms and chemotherapeutic agents (Brennan and Nikaido, 1995; Barry et al., 1998; Daffe and Draper, 1998; Jarlier and Nikaido, 1994). The primary constituents of this low–permeability barrier are mycolic acids (very long-chain fatty acids) that esterify either to cell wall arabinogalactan, leading to a continuous link with peptidoglycan, or to trehalose to yield a pool non-covalently associated with the outer layer of the cell envelope (Brennan and Nikaido, 1995; Barry et al., 1998; Daffe and Draper, 1998; Liu and Nikaido, 1999). It is apparent that a detailed understanding of the structure and function of mycolic acids and their biosynthetic pathway should lead to discovery of novel targets for the design of potent antimycobacterial drugs.

The structures of mycolic acids of several mycobacterial species have been determined using combinations of the technologies of thin layer chromatography (TLC), mass spectrometry (MS), nuclear magnetic resonance (NMR), and infrared spectroscopy (IR). The general structure of mycolic acids is R-CH(OH)-CH(R′)-COOH, where R is a ‘meromycolate’ chain (mero-chain) consisting of 50–60 carbons and R′ is a shorter aliphatic chain (α-branch) possessing 22–26 carbons. Structural variations in the functional groups of the mero-chain are used to classify mycolic acids into major groups, each of which shows microheterogeneity in carbon chain lengths. Mycobacterium tuberculosis produces three distinct mycolate groups: α-, methoxy-, and keto-mycolates (Fig. 1). α-mycolates contain two cis-cyclopropane rings, whereas in methoxy- and keto-mycolates the distal cyclopropane is replaced by a carbon-oxygen bond with an adjacent methyl branch, and the proximal cyclopropane has either a cis or trans configuration (Brennan and Nikaido, 1995; Barry et al., 1998; Daffe and Draper, 1998). These modifications of the mero-chain are introduced by a family of methyl-transferases (Dubnau et al., 2000; Glickman et al., 2001). A growing body of evidence suggests that these modifications play an important role in mycobacterial cell wall impermeability, cording, and virulence. For example, cyclopropanation of α-mycolates at the distal position enhances resistance of Mycobacterium smegmatis to oxidative killing (Yuan et al., 1995), whereas this modification at the proximal function enhances cell wall impermeability of M. smegmatis (George et al. 1995), as well as cording and persistence of M. tuberculosis in mouse tissues (Glickman et al., 2000). Similarly, oxygenation of mycolates (methoxy and keto) enhances the ability of M. tuberculosis to infect mice and cells (Yuan et al., 1998; Dubnau et al., 2000). In addition, markedly increased production of keto-mycolates has been observed for M. tuberculosis and Mycobacterium microti growing inside mouse macrophages (Yuan et al., 1998) and lung tissues (Davidson et al., 1982), suggesting the importance of this species of mycolates for bacterial survival during infection. How mycolate composition affects cell wall function is not understood, but a M. smegmatis mutant that accumulates incomplete meromycolate chains instead of mycolic acids exhibits marked increases in cell wall permeability and thermodynamic fluidity (Liu and Nikaido, 1999).

Figure 1.

Structures of the major mycolic acids of M. tuberculosis H37Rv (Laval et al., 2001). The main values of m1 and m3 are 15, 17, and 19; those of m2 depend on the presence (13, 15, and 17) or absence (12, 14, and 16) of a methyl branch at the proximal location. Full length mycolates have 78 and 80 carbons for α; 83, 85, and 87 carbons for methoxy; and 84 and 86 carbons for keto.

The mechanism of elongation of mycolates is not well understood. The genome sequence of M. tuberculosis (Cole et al., 1998) has predicted the presence of two fatty acid synthase (FAS) systems, FAS-I (a multifunctional single polypeptide) and FAS-II (a multiple enzyme complex). Past studies examining accumulation of mycolate precursors (Qureshi et al., 1998) or in vitroβ-ketoacyl-acyl carrier protein synthase (KAS) activities (Kremer et al., 2000; 2002; Schaeffer et al., 2001; Slayden and Barry, 2002) have suggested that FAS-I is involved in elongation of short-chain compounds to yield fatty acids of 16–26 carbons that are further extended to full lengths by FAS-II. Several putative FAS-II components are located in a gene cluster containing both kasA and kasB (Cole et al., 1998), the only two genes that encode enzymes that have been shown to possess FAS-II KAS activity in vitro. A previous study suggested that isoniazid (INH) inhibition of M. tuberculosis mycolate production might occur through formation of a complex with KasA (Mdluli et al., 1998). Recently, by using an in vitro fatty acid biosynthesis assay system, Schaeffer et al. (2001) and Kremer et al. (2002) showed that purified KasA or KasB of M. tuberculosis elongates fatty acids with substrate specificity of ≥14 carbons. Another recent study by Slayden and Barry (2002) showed that cell lysates of M. smegmatis expressing M. tuberculosis kasA could elongate fatty acids to ∼ 40 carbons, whereas expression of both kasA and kasB produced longer chain fatty acids (∼ 54 carbons) equivalent in length to the mero-chains of mycolates. Based on these in vitro biosynthesis and drug inhibition studies, it has been proposed that kasA is involved in the initial elongation of mycolates that are extended to their full lengths by kasB (Kremer et al., 2002; Slayden and Barry, 2002). However, direct in vivo tests of this hypothesis are lacking.

Mycobacterium marinum, a fish, amphibian and opportunistic human pathogen, has been developed recently as a model system for studying M. tuberculosis pathogenesis (Ramakrishnan et al., 1997; 2000; Talaat et al., 1998; Gao et al., 2003). Mycobacterium marinum is phylogenetically close to members of the M. tuberculosis complex (Tonjum et al., 1998), and the infection it causes in its natural hosts manifests pathological hallmarks of tuberculosis, including granulomas (Ramakrishnan et al., 1997; Talaat et al., 1998). These observations suggest that many of the basic mechanisms of disease initiation are conserved between M. marinum and M. tuberculosis. Furthermore, these two mycobacterial species have in common a number of lipids including phenolic glycolipids, phthiocerol diesters and mycolate profiles (Daffe and Draper, 1998). In addition, M. marinum grows significantly faster than M. tuberculosis and other slow-growing mycobacteria, with a generation time of 4 h, compared to 20 h for M. tuberculosis. Therefore, M. marinum offers an attractive and potentially valuable model for studying the molecular biology of mycobacterial pathogenesis. We have recently developed a highly efficient mariner-based transposon mutagenesis system in M. marinum (M4) for facile generation of random mutations and construction of a mutant library (Gao et al., 2003). This library has been used to identify mutants with distinct phenotypes potentially important for mycobacterial virulence. We recently reported identification of a M. marinum mutant attenuated for intracellular growth in macrophages that was fully complemented by a homologous M. tuberculosis locus (Gao et al., 2003). In the present work, we describe isolation of M. marinum kasB mutants and show that kasB is required for normal growth of M. marinum in macrophages. Whereas kasB is required for full elongation of mycolates to their mature forms, the defect in the lengths of the mycolates in these kasB mutants is rather small (∼ 2–4 carbon units). Despite the small difference in mycolate length, the kasB mutants have markedly increased permeability of their cell walls and consequently are severely defective in resisting host defence mechanisms and antibiotic action. Complete recovery of the kasB mutants to wild-type phenotypes is achieved by ectopic expression of the M. tuberculosis kasB gene, but not by kasA. These studies identify kasB as an important potential target for novel therapeutic intervention, independent of kasA.

Results

Identification of M. marinum kasB mutants with impaired growth within macrophages

Our screen of a M. marinum transposon mutant library developed by the M4 procedure has identified 24 mutants with impaired growth in J774 murine macrophage-like cells (our unpublished results). Sequencing of the transposon insertion junctions identified three mutants (4B8, 4E4 and 4E8) with independent insertions in a gene that is homologous to M. tuberculosis kasB (Fig. 2A) (Cole et al., 1998). Whereas the 4B8 and 4E8 mutants had transposon insertions that disrupted the coding sequence of kasB, the 4E4 mutant had a transposon insertion at the 3′ end of the gene, with a direction of transcription of the kanamycin-resistant gene (Kanr) opposite to kasB. Based on both our sequencing data and the assembled contig (mar821d11.p1n) of the incomplete M. marinum genome database at the Sanger Center, KasB of M. marinum is 89% identical and 93% homologous to M. tuberculosis KasB. In addition, both the sequences and genomic organization of genes surrounding kasB are conserved between M. marinum and M. tuberculosis. M. marinum KasA shows even higher similarity to M. tuberculosis KasA, sharing 92% identity and 97% similarity.

Figure 2.

Figure 2.

A. Partial map of M. marinum fasII locus and transposon insertions (vertical triangles) in the kasB gene at different orientations (black arrows indicate the direction of the transcription of the Kanr gene). B and C. Growth kinetics of M. marinum wild type and the kasB mutants in J774 cells and MDMs respectively.

Figure 2.

Figure 2.

A. Partial map of M. marinum fasII locus and transposon insertions (vertical triangles) in the kasB gene at different orientations (black arrows indicate the direction of the transcription of the Kanr gene). B and C. Growth kinetics of M. marinum wild type and the kasB mutants in J774 cells and MDMs respectively.

Figure 2.

Figure 2.

A. Partial map of M. marinum fasII locus and transposon insertions (vertical triangles) in the kasB gene at different orientations (black arrows indicate the direction of the transcription of the Kanr gene). B and C. Growth kinetics of M. marinum wild type and the kasB mutants in J774 cells and MDMs respectively.

The three M. marinum kasB mutants showed ∼10-fold fewer CFUs compared to wild type for infection of J774 cells at 4 days after infection (Fig. 2B). These mutants also showed reduced intracellular growth in murine bone marrow-derived macrophages (MDMs) (Fig. 2C). Coincident with the decreased rate of intracellular growth was a decrease in host cell cytotoxicity of the infection, as the J774 monolayer was lysed by wild-type M. marinum at 4 days but not until ≥6 days by the mutants (data not shown). The intracellular growth defect of the kasB mutants was not due to a general metabolic abnormality of these bacterial strains, as they grew similarly to wild type in bacterial culture medium (Fig. 3A and data not shown). These results demonstrate that kasB is required for a normal growth rate of M. marinum during macrophage infection.

Figure 3.

Figure 3.

Growth kinetic of M. marinum strains in 7H9 broth (A) or J774 cells (B). OD580, optical density at the wavelength of 580 nm.

Figure 3.

Figure 3.

Growth kinetic of M. marinum strains in 7H9 broth (A) or J774 cells (B). OD580, optical density at the wavelength of 580 nm.

Complementation of M. marinum kasB mutants by M. tuberculosis kasB gene

The high degree of similarity between KasB of M. marinum and M. tuberculosis provided an opportunity to assess the utility of M. marinum as a model organism for studying the function of M. tuberculosis genes. We expressed the coding sequence of the M. tuberculosis kasB gene in each of the three kasB mutants and examined their growth in J774 cells. Expression of M. tuberculosis kasB fully restored intracellular growth of all three kasB mutants (Fig. 3B and data not shown). In addition, the kinetics of cytotoxicity was also restored to wild type (data not shown). Expression of the downstream gene accD6 did not complement the kasB mutants, and co-expression of kasB and accD6 had a phenotype indistinguishable from expression of kasB alone (data not shown). Therefore, the growth defect of these three mutants was due to mutation of kasB and not a polar effect on downstream genes. Given the high degree of similarity between KasB and KasA, we examined whether expression of M. tuberculosis kasA could complement the kasB defect. Surprisingly, our results showed that M. tuberculosis kasA exacerbated the growth defect of the 4B8 mutant (Fig. 3B). The worsened defect of the kasA-expressing strain might be due in part to generally decreased hardiness, as the transformed strain also grew less well in vitro (Fig. 3A). To determine whether this general toxicity of expression of kasA was specific to the M. tuberculosis gene, M. marinum kasA gene also was expressed in each of the kasB mutants, and similar effects were observed (data not shown). Together, these results demonstrate that kasB has a unique and non-redundant role required for normal growth of mycobacteria in macrophages.

Cell wall mycolates are shortened in kasB mutants

Previous characterization of in vitro activity of KasA and KasB and inhibition of mycolate biosynthesis by INH have led to the hypothesis that both enzymes are required for elongation of mycolates during bacterial growth. To examine the effect of kasB mutation on mycolate biosynthesis, we examined mycolate profiles of M. marinum strains by high performance TLC (HPTLC). There was no difference among these M. marinum strains in the amount of total mycolates or mycolic acids either covalently bound to or non-covalently associated with the cell wall, as determined by quantification of 14C-acetate-labelled mycolates (data not shown). There was also no difference among these strains in the appearance and abundance of polar and apolar organic extractable cell wall lipids (data not shown). When separated by silica HPTLC plates, the mycolates of wild-type M. marinum showed three groups of mycolates (α, methoxy, and keto) (Fig. 4A), similar to those of M. tuberculosis (Dubnau et al., 2000) and in agreement with previous reports (Daffe et al., 1991; Chemlal et al., 2002). Although the 4B8 mutant had all three groups of the mycolates, the TLC mobility of each one was slightly but reproducibly reduced compared to wild type (Fig. 4A), suggesting the possibility of shortened carbon chain lengths. More prominent differences of migration were observed by reverse phase HPTLC for each of the three groups of mycolates isolated from wild type and the 4B8 mutant, under solvent conditions in which shorter chains run faster on the TLC plates (Fig. 4B). α-mycolates of wild-type M. marinum contained five major microheterogeneous species; in contrast, α-mycolates of the 4B8 mutant had only three major species equivalent to the three fastest migrating species of wild type (Fig. 4B). The 4B8 mutant likewise did not have the two slowest migrating species of the oxygenated mycolates (methoxy and keto) that were present in wild-type M. marinum (Fig. 4B). Whereas expression of M. tuberculosis kasB completely corrected these migration defects of the 4B8 mutant, expression of kasA did not.

Figure 4.

Figure 4.

A and B. Mycolate profiles of M. marinum labelled with 14C-acetate as revealed by silica and reverse phase HPTLC respectively.

C. Relative intensity of a mycolate type compared to the total mycolates.

Figure 4.

Figure 4.

A and B. Mycolate profiles of M. marinum labelled with 14C-acetate as revealed by silica and reverse phase HPTLC respectively.

C. Relative intensity of a mycolate type compared to the total mycolates.

Figure 4.

Figure 4.

A and B. Mycolate profiles of M. marinum labelled with 14C-acetate as revealed by silica and reverse phase HPTLC respectively.

C. Relative intensity of a mycolate type compared to the total mycolates.

In addition to these carbon chain length variations, there were differences in abundance of the three groups of mycolates between wild type and the kasB mutants. The percentage of α-, methoxy- or keto-mycolates in total mycolates for wild type was 41, 24 and 35% respectively (Fig. 4C). The 4B8 mutant showed a significant (∼ 30%) reduction in the abundance of keto-mycolates, with a slight compensatory increase of both α- and methoxy-mycolates. Complementation with M. tuberculosis kasB, but not kasA, corrected this abnormality in the relative abundance of these mycolate species.

To confirm shortened mycolates in the kasB mutant as suggested by TLC, mycolates from strains of M. marinum were analysed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Mass spectra matched well with the mycolate profiles revealed by HPTLC, demonstrating that the predominant chain lengths were 2–4 carbons shorter for oxygenated mycolates in the absence of kasB (Fig. 5 and Table 1). Oxygenated mycolates of wild type had chain lengths corresponding to 79–85 carbons (methoxy) and 80–82 carbons (keto), as opposed to 77–81 and 76–80 carbons, respectively, in the mutant (Fig. 5 and Table 1). Similarly, whereas α-mycolates of wild type had 73–80 carbons, those of the 4B8 mutant had 73–76 carbons (Table 1) (Laval et al., 2001). As already seen in other assays of complementation, M. tuberculosis kasB, but not kasA, reverted the mycolate length differences in the mutant to wild-type sizes.

Figure 5.

MALDI-TOF mass spectra of the keto-MAMEs of M. marinum strains. Values indicated the masses of the sodium adducts (M + 23).

Table 1. . MALDI-TOF MS data of the MAMEs of M. marinum strains.
MAtypeMmstrainTotal carbon number of free acids
73747576777879808182838485
  1. The pseudomolecular mass values [M + Na]+ of the major homologues are represented in italics and bold. Only molecular species representing> 20% of the mixture are indicated. MA, mycolic acid; M, methoxy; K, keto; Mm, M. marinum.

αWT1104 1118 1132 1146  1174 1202     
4B81104 1118  1146         
4B8 +  1118 1132 1146 1160 1174  1202     
pkasB             
4B8 +1104 1118  1146         
pkasA             
ΜWT      12061220 1234   1262  1290
4B8    1178  1206  1234    
4B8 +      12061220 1234 1248 1262  1290
pkasB             
4B8 +     1178   1206       
pkasA             
ΚWT        1218   1246    
4B8   1162  1190  1218     
4B8 +        1218   1246    
pkasB             
4B8 +    1162   1190  1218     
pkasA             

Detailed analysis of mycolates

Further analysis by NMR, IR and argentation TLC demonstrated that the global structures of the three types of mycolates from wild type and kasB mutants were similar. α-mycolates consisted primarily of dicyclopropanated compounds (IR bands at 1020 and 3060 cm−1). The cis configuration of cyclopropyl groups in the α-mycolates was established by 1H-NMR analysis (signals at – 0.3, 0.6 and 0.7 p.p.m.). Minor cis (1H-NMR signals at 5.35 p.p.m) and trans (IR bands at 970 cm−1 and 1H-NMR signals at 5.25 p.p.m.) ethylenic α-mycolates also occurred in both strains. In contrast to M. tuberculosis, the oxygenated mycolates of wild-type and 4B8 mutant M. marinum were composed almost exclusively of ethylenic forms, consistent with published data (Daffe et al., 1991); the configuration of the double bond was predominantly trans in both methoxy- and keto-mycolates (IR bands at 970 cm−1; 1H-NMR signals at 5.25 p.p.m.) with an adjacent methyl branch (1H-NMR doublets at 0.94 p.p.m.), although the presence of ethylenic oxygenated mycolates with a cis configuration (1H-NMR signals at 5.35 p.p.m.) was also detected.

To investigate the location of the 2–4 carbons in wild-type mycolates missing from the kasB mutants, mycolates purified from wild type and 4B8 were analysed by chemical degradation methods. Pyrolysis of the individual types of mycolates, a reaction that occurs during GC and GC-MS analyses of mycolates and cleaves the molecules into α-branches and mero-chains, showed that α-branches of the mycolates from both wild type and mutant contained mainly 24 carbons (m4 = 21, Fig. 6A). Therefore, the 2–4 carbon unit lacking in the mutant mycolates is in the mero-chains. The location of the shortening in the mero-chains can be determined by oxidative cleavage of mycolates, provided that the substances contain ethylenic bonds that are converted by permanganate into dihydroxylated compounds, which in turn are oxidized by periodate to yield long-chain fatty acids and dicarboxylic acids (van Rudloff, 1956). Accordingly, this reaction was applied to purified oxygenated mycolates of wild type and kasB mutant of M. marinum, which are most exclusively ethylenic substances, and the resulting fatty acids were esterified by diazomethane and purified. Analysis of the MALDI-TOF mass spectra of the monoesters derived from the mycolates of M. marinum showed no difference between the products from wild type and kasB mutants (data not shown), indicating that the 2–4 carbon unit difference is not located in the distal portion of the mycolates. This conclusion was supported further by analysis of the GC-MS spectra of the monoesters derived from the oxygenated mycolates of M. marinum, which allowed the definitive location of the keto and methoxy groups in the trans ethylenic methoxy- and keto-mycolates (m1 = 15,17; m2 = 13,15; Fig. 6A). On the other hand, MALDI-TOF mass spectra of the hydroxylated diesters clearly revealed size differences between oxygenated mycolates isolated from wild type and the kasB mutant (Fig. 6B and C). As pyrolysis showed that α-branches of all types of M. marinum mycolates had the same size (m4 = 21), it was deduced that m3 were 17 or 19 for wild type, but 13 or 15 for the mutant (Fig. 6A–C). Together, these data showed that the 2–4 carbon units lacking in the oxygenated mycolates of the kasB mutants are located in the proximal portion of the mero-chains.

Figure 6.

Figure 6.

A. Structures of the major species of α-, methoxy- and keto-mycolates of M. marinum and illustration of the strategies for determining m1 to m4 of purified methoxy- and keto-mycolates that contain a double bond, which can be cleaved by a permanganate/periodate reaction into analysable products. B and C. MALDI-TOF MS of the hydroxylated diesters of M. marinum wild type and the 4B8 mutant respectively.

Figure 6.

Figure 6.

A. Structures of the major species of α-, methoxy- and keto-mycolates of M. marinum and illustration of the strategies for determining m1 to m4 of purified methoxy- and keto-mycolates that contain a double bond, which can be cleaved by a permanganate/periodate reaction into analysable products. B and C. MALDI-TOF MS of the hydroxylated diesters of M. marinum wild type and the 4B8 mutant respectively.

Figure 6.

Figure 6.

A. Structures of the major species of α-, methoxy- and keto-mycolates of M. marinum and illustration of the strategies for determining m1 to m4 of purified methoxy- and keto-mycolates that contain a double bond, which can be cleaved by a permanganate/periodate reaction into analysable products. B and C. MALDI-TOF MS of the hydroxylated diesters of M. marinum wild type and the 4B8 mutant respectively.

KasB-dependent mycolate elongation is essential for cording and cell wall impermeability

A correlation between cording in culture and virulence has long been recognized for M. tuberculosis H37Rv and H37Ra strains (Middlebrook et al., 1947), and a change of morphology from granular, dry, and rough to viscous, moist, and smooth was observed by Calmette and Guerin during their serial passage of Mycobacterium bovis to create the BCG vaccine strain (Guerin, 1957). Recently, M. tuberculosis cording has been shown to require cyclopropanation of α-mycolates (Glickman et al., 2000). Wild-type M. marinum also forms cords in liquid culture (Fig. 7A) and shows a rough colony morphology on agar (Fig. 7E). In contrast, the 4B8 mutant showed strikingly altered morphology that appeared as non-ordered small aggregates in liquid (Fig. 7B) and smooth colonies on agar (Fig. 7F). The cording and colony morphology abnormalities of the mutant were completely recovered by expressing M. tuberculosis kasB (Fig. 7C and G), but not kasA (Fig. 7D and H). Expression of kasA actually made the mutant grow slower on agar (Fig. 7H), similar to its slow growth rate in liquid (Fig. 3A). As M. marinum wild type and the kasB mutants elaborated similar amounts of cell wall lipids and the trehalose mycolates, it follows that kasB-dependent elongation of mycolates is a second important factor for cording. The cording and colony morphology abnormalities of the 4B8 mutant did not, however, affect the ultrastructure of the mutant cell envelope, as assessed by electron microscopy (data not shown).

Figure 7.

A–D. Cording of M. marinum in 7H9 broth.
E–H. Colony morphology of M. marinum on 7H10 agar grown to 7 days (E, F and G) or 10 days (H). Colony size of 4B8 + pkasA at 7 days (image not shown) was ∼10-times smaller than wild type.

Given the importance of mycolates in maintaining cell wall integrity, we examined cell wall permeability for M. marinum wild type and the kasB mutants. As shown in Fig. 8A, whereas wild type and the kasB-complemented mutant were resistant to SDS at 0.01–0.5%, the 4B8 mutant (with or without kasA complementation) showed extreme sensitivity. As a second test of cell wall integrity, we examined the ability of M. marinum strains to resist loss of lipophilic stain, a well-known ‘acid fast’ characteristic of mycobacterial cell walls, presumed to depend on mycolates. As shown in Fig. 8B, the 4B8 mutant was much more vulnerable to destain than wild type or the kasB-complemented mutant.

Figure 8.

Figure 8.

Measurement of M. marinum cell wall permeability.

A. Sensitivity of M. marinum to SDS.

B. Resistance of M. marinum cell wall to acid destain following three destain procedures.

C. Kinetic of accumulation of [14C]-chenodeoxycholate by M. marinum.

Figure 8.

Figure 8.

Measurement of M. marinum cell wall permeability.

A. Sensitivity of M. marinum to SDS.

B. Resistance of M. marinum cell wall to acid destain following three destain procedures.

C. Kinetic of accumulation of [14C]-chenodeoxycholate by M. marinum.

Figure 8.

Figure 8.

Measurement of M. marinum cell wall permeability.

A. Sensitivity of M. marinum to SDS.

B. Resistance of M. marinum cell wall to acid destain following three destain procedures.

C. Kinetic of accumulation of [14C]-chenodeoxycholate by M. marinum.

To measure cell wall permeability directly, we examined diffusion of the lipophilic compound, [14C]-chenodeoxycholate, into M. marinum. As predicted, the 4B8 mutant accumulated this compound at a much higher initial rate (∼ threefold in the first 6 min) than wild type or the kasB-complemented mutant (Fig. 8C). Collectively, these results demonstrate that kasB-mediated elongation of mycolates (or the secondary effect on accumulation of keto-mycolates) is essential for normal mycobacterial cell wall barrier function.

KasB-dependent cell wall impermeability is essential for antibiotic resistance

To determine whether kasB–dependent cell wall barrier is essential for antibiotic resistance, we tested sensitivity of M. marinum strains to several antibiotics with different properties. As shown in Table 2, the 4B8 mutant was much more sensitive than wild type to lipophilic antibiotics such as rifampin (RIF), ciprofloxacin (CIP) and erythromycin (ERY), exhibiting 12- to 60-fold lower MIC (minimal inhibitory concentration), as tested by the agar method. In contrast, no differences were detected between wild type and the mutant in resistance to hydrophilic antibiotics such as INH and streptomycin (STR). Resistance to lipophilic antibiotics was fully restored for the 4B8 mutant by expressing M. tuberculosis kasB, but not kasA (Table 2). Activity of these lipophilic antibiotics has been shown to depend on the rate of penetration across mycobacterial cell wall (Brennan and Nikaido, 1995; Liu and Nikaido, 1999). The correlation between diffusion of 14C-chenodeoxycholate and sensitivity to lipophilic antibiotics strongly suggests that kasB-dependent cell wall impermeability is essential for mycobacterial resistance to these antibiotics.

Table 2. . Minimal inhibitory concentration (MIC) (µg ml−1) of M. marinum strains for different antibiotics tested by the agar method.
StrainsRIFCIPERYINHSTRCER
WT0.20.06303080.2
4B80.010.0050.53080.2
4B8 + pkasB0.20.06303080.2
4B8 + pkasA0.010.0050.53080.2

Cerulenin synergizes with RIF activity

Cerulenin (CER), a well-characterized potent inhibitor for KAS of both the FAS-I and FAS-II systems, has been shown to inhibit KasA and KasB activity in vitro and to affect mycolate biosynthesis (Kremer et al., 2000; 2002; Schaeffer et al., 2001). No differences in MIC (0.2 µg ml−1) for CER were detected between wild type and the 4B8 mutant (Table 2). To determine whether there was synergy between CER and RIF, we treated wild-type M. marinum with a subinhibitory concentration of CER (0.1 µg ml−1) together with varying concentrations of RIF and examined MIC by the broth culture method. As shown in Fig. 9, presence of CER significantly increased susceptibility of wild type to RIF, from an MIC of 0.3 µg ml−1 (no CER) to 0.04 µg ml−1, close to MIC (0.02 µg ml−1) for the 4B8 mutant (Fig. 9). In contrast, there was no synergy between CER and STR (data not shown), a hydrophilic antibiotic that presumably enters mycobacteria by a different route.

Figure 9.

Synergy between CER and RIF. M. marinum wild type or the 4B8 mutant was cultured in 7H9 broth in the presence or absence of CER (0.1 µg ml−1) together with RIF (0.005, 0.01, 0.02, 0.08, 0.32 µg ml−1) or in the absence and MIC was examined by the broth culture method.

KasB is essential for resistance to macrophage antimicrobial activity

We examined whether kasB is required for cell wall integrity of intracellular M. marinum in macrophages by acid-fast staining. As shown in Fig. 10A, intracellular wild type and the kasB-complemented mutant were more resistant to destain than the 4B8 mutant. To determine whether this cell wall defect could increase sensitivity of intracellular M. marinum to host defence molecules that might be encountered inside a phagosome or phagolysosome (Kisich et al., 2002; Duits et al., 2003), we compared susceptibility of M. marinum strains to the human neutrophil defensin peptide 1 (HNP-1) and the synthetic defensin-like compound protamine. The 4B8 mutant was significantly more sensitive than wild type to both HNP-1 (Fig. 10B) and protamine (data not shown). In addition, the 4B8 mutant was extremely susceptible to lysozyme at concentrations that had minimal effect on wild type (Fig. 10C). The sensitivity of the mutant to HNP-1, protamine, and lysozyme was completely reverted to wild type by expressing M. tuberculosis kasB (Fig. 10B and C, and data not shown). To determine whether the in vitro susceptibility of the kasB mutants to these antimicrobial molecules is relevant to the in vivo macrophage antimycobacterial activity, we examined maturation of the M. marinum-containing phagosome of MDM infected by wild type and the kasB mutants. Like M. tuberculosis, wild-type M. marinum inhibits fusion of the bacteria-containing phagosome with lysosomes, so that the bacteria can replicate within a vacuole protected from a variety of host defence molecules (Barker et al., 1997; Russell, 2001). As shown in Fig. 11A–D, the vast majority of the wild-type M. marinum bacteria (>95%) did not co-localize with the lysosomes, which were visualized with the tracer Lysotracker Red, which fluoresces only at acidic pH. In contrast, 30–40% of the kasB mutant bacteria co-localized with lysosomes, as indicated by overlap of the bacteria-containing phagosomes and the Lysotracker-labelled acidic compartments. The loss of inhibition of phagosome-lysosome fusion by the M. marinum kasB mutant was fully complemented by expressing M. tuberculosis kasB (data not shown). Together, these results suggest that kasB-dependent effects on cell wall mycolates are required for mycobacterial resistance to macrophage defence mechanisms.

Figure 10.

Figure 10.

A. Resistance of intracellular M. marinum in MDM (24 h post infection) to acid destain. B and C. Susceptibility of M. marinum to HNP-1 and lysozyme respectively.

Figure 10.

Figure 10.

A. Resistance of intracellular M. marinum in MDM (24 h post infection) to acid destain. B and C. Susceptibility of M. marinum to HNP-1 and lysozyme respectively.

Figure 10.

Figure 10.

A. Resistance of intracellular M. marinum in MDM (24 h post infection) to acid destain. B and C. Susceptibility of M. marinum to HNP-1 and lysozyme respectively.

Figure 11.

Inhibition of phagosome maturation by M. marinum. MDMs were infected by wild type (A–D) or the 4B8 mutant (E–H) at an MOI of 3. At 24 h post infection, MDMs were incubated with fluorescent Lysotracker Red for labelling of acidic lysosomal compartments. Fusion of M. marinum-containing phagosomes with lysosomes is indicated by overlap of the entire bacteria with the red fluorescence, as this demonstrates that the phagosome is at low pH (approximately pH 5.5). Several bacteria (D, wild type; H, 4B8) are indicated by arrowheads. D and H are enlarged sections of C and G, respectively, to better illustrate co-localization of kasB mutant bacteria with lysosomes (H) and lack of co-localization for wild type (D).

Discussion

In this work, we have analysed the role of kasB in mycolate biosynthesis in M. marinum and shown by HPTLC and MALDI-TOF MS that kasB is required for full elongation of each of the three types of mycolates by ∼ 2–4 carbon units. Detailed analyses by argentation TLC, infrared spectroscopy, MALDI-TOF MS, NMR and chemical degradation experiments showed that this kasB-mediated extension occurs in the proximal portion of the mero-chain of the mycolates (i.e. the end nearest to the arabinogalactan-peptidoglycan complex). In addition to their defect in mycolate chain length, the kasB mutants showed altered ratios among the three types of mycolates, with a significant reduction of keto-mycolates that was accompanied by a compensatory increase of α- and methoxy-mycolates. A current model for the structure of the mycobacterial cell wall suggests that the α-branches and mero-chains of mycolic acid residues align parallel to each other and perpendicular to the plane of the cell surface (Brennan and Nikaido, 1995; Barry et al., 1998; Daffe and Draper, 1998; Jarlier and Nikaido, 1994), with covalent linkage of mycolates to the arabinogalactan-peptidoglycan complex. We propose that the 2–4-carbon extension at the proximal portions of the mero-chains is necessary for the distal portions of the mero-chains to play their architectural role in the cell wall. An alternative, but not mutually exclusive, possibility is that distribution of the three types of mycolates on the cell wall at proper ratios is required for maintaining proper cell wall integrity. In any case, these small changes in mycolate length and composition must have a major effect on organization of the bacterial permeability barrier. We are currently investigating these possibilities for the permeability defect in the absence of kasB.

Our results are consistent with the hypothesis that kasA is involved in initial extension of the FAS-II-mediated growth of the mycolate chains, and kasB is involved primarily in extension to full lengths (Kremer et al., 2000; Slayden and Barry, 2002). It is, however, somewhat surprising that the mycolates in the M. marinum kasB mutant are only 2–4 carbons shorter than wild type. Based on in vitro assays of KAS enzyme activity in M. smegmatis cell lysates, it has been proposed that kasB may be required for addition of ∼ 14 carbons to the intermediate mycolate products to reach the final full-lengths (Slayden and Barry, 2002). It is, of course, possible that there is some redundancy in function between kasA and kasB, so that either enzyme is capable of adding carbons to the growing chain in vivo, but kasB is more efficient in the in vitro assay for chains of 40 or more carbons. However, it is clear that there is not total functional redundancy of the two enzymes in vivo, both because absence of kasB leads to specific shortening of the mero-chains by 2–4 carbons and because extra-chromosomal addition of kasB, but not kasA, can fix this defect. Thus, the last 2–4 carbon additions require kasB specifically, and not simply an increase in total FAS-II KAS activity.

Given the minimal effects of kasB disruption on mycolate length and composition and undetectable effects on other cell wall lipids and cell envelope ultrastructure, the extent of the changes in cell wall permeability in the kasB mutants is unexpected. The kasB mutants exhibited a ∼ threefold higher diffusion rate than wild type for the hydrophobic compound chenodeoxycholate to cross the cell wall, and this increased permeability made the mutant bacteria significantly more susceptible to lipophilic antibiotics, defensins and lysozyme. It is intriguing to speculate about the why these small changes of mycolate composition cause such major disruption of the overall physiology of the cell wall. One hypothesis is that appropriate packing of the mycolates covalently associated with the cell wall is essential to restriction of cell wall fluidity, as suggested by Liu and Nikaido (1999) while examining a M. smegmatis mutant that was blocked in the synthesis of mature mycolic acids. Changes in mycolate length or composition might significantly affect packing, resulting in increased fluidity and ultimately permeability of the cell wall. We have found that the kasB mutants exhibit much higher sensitivity than wild type to growth arrest at elevated temperatures (34–35°C, unpubl.), which is consistent with this hypothesis. Direct tests for mycobacterial cell wall fluidity have been performed using differential scanning calorimetry (Liu and Nikaido, 1999), and assay of the kasB mutant in comparison to wild type by this method should provide a direct test of this hypothesis.

The increased susceptibility of the kasB mutants to lipophilic antibiotics suggests that pharmacological inhibition of kasB could be extremely useful in therapeutic regimens in synergy with traditional lipophilic antibiotics such as RIF. As an initial approach to this clinically important problem, we showed that interference with KAS activity by a subinhibitory concentration of CER markedly synergized with the bactericidal activity of RIF for wild-type M. marinum. We suggest that this is a proof in principle of the therapeutic potential of kasB inhibition. M. marinum and M. tuberculosis are sufficiently genetically related that it is very likely that our findings are relevant to the treatment of tuberculosis. In this context, it is interesting to note that several derivatives of thiolactomycin (TLM, a potent and more specific inhibitor of KasA and KasB than CER) have recently been found to be more potent inhibitors of KasB than TLM itself (Kremer et al., 2000). Therefore, more extensive structure-activity studies of derivatives of these compounds, together with a better understanding of KasB catalysis, could lead to discovery of more potent and specific inhibitors for KasB with therapeutic potential. Because our work demonstrates that these studies can be done using the M. tuberculosis kasB gene to complement the kasB mutant of M. marinum, it is possible that novel antibiotic discovery could be carried out in an intact pathogenic organism capable of infecting macrophages and ectotherms, but with minimal risk to investigators.

Finally, this work has demonstrated that kasB is required for full virulence of M. marinum in macrophage infection. Likely, this is a result of the extreme sensitivity of the kasB mutants to innate antimicrobial mechanisms within macrophages, such as defensins and lysozyme. Defensins are potent antimicrobial peptides present in a variety of cell types including neutrophils and macrophages (Risso, 2000; Duits et al., 2002), and accumulation of defensins on the bacterial cell surface has been observed for M. tuberculosis inside the phagosomes of both macrophages and neutrophils (Kisich et al., 2002). Although the phagosome containing wild-type M. tuberculosis (Russell, 2001) or M. marinum (Barker et al., 1997) does not fuse efficiently with lysosomes, the M. marinum kasB mutant is inefficient at maintaining this block in phagosome maturation. Loss of kasB may result in damage of M. marinum by low concentrations of host defence molecules in the early phagosome or reduced ability for M. marinum to inhibit phagosome-lysosome fusion. In either case, the kasB mutant bacteria end up in a compartment in which they are exposed to higher concentrations of host defence molecules to which they are significantly more susceptible. We believe that these studies reinforce the utility of the M. marinum model of mycobacterial pathogenesis and of the M4 technique for creating, identifying, and studying mycobacterial genes required for survival of this important and complex pathogen family in their eukaryotic hosts.

Experimental procedures

Bacterial strains and media

Mycobacterium marinum strain M and the DH5α strain of Escherichia coli were cultured as described (Gao et al., 2003).

Isolation of M. marinum mutants with transposon insertions in the kasB gene

A M. marinum mutant library was generated by the M4 procedure exactly as previously described (Gao et al., 2003). Briefly, a mariner-based transposon mutagenesis plasmid was electroporated into log-phase M. marinum, and bacteria in which transposition occurred were identified by sequential selections on kanamycin (Kan)- and sucrose-containing plates. Kan- and sucrose-resistant colonies were repeatedly subcloned to insure that each colony arose from a single transposon insertion, and ∼ 1000 individual mutants were screened for defective intracellular growth in J774 cells (ATCC TIB67) and bone marrow-derived macrophages (MDMs) using our previously described method (Gao et al., 2003). A total of 24 mutants were isolated that exhibited more than fivefold reduction of CFUs for intracellular bacteria as compared to wild type after 4 days of infection. Recovery and sequencing of the transposon insertion sites were performed as described (Gao et al., 2003). Sequencing was performed by the Biomolecular Resource Center at the University of California, San Francisco. Similarity searches were performed using blastp and blastn at the NCBI site (http://www.ncbi.nlm.nih.gov/blast) and the incomplete M. marinum genome data base at the Sanger Center site (http://www.sanger.ac.uk/Projects/Mmarinum/).

Complementation of M. marinum kasB mutants

To complement the M. marinum kasB mutants by M. tuberculosis kasB gene, the coding sequence of M. tuberculosis kasB (Cole et al., 1998) was amplified by polymerase chain reaction (PCR) using primers Rv2246kasBU1 (5′-GGGTAC CACCACTTGCGGGGGCGAGTC-3′) and Rv2246kasBD1 (5′ GGCCCAAGCTTCCTGTCGCGTAACGC-3′) (Invitrogen) and cloned into pLYG206.Zeo, similarly as we previously described (Gao et al., 2003). pLGY206.Zeo was constructed by replacing the Kan-resistant gene (Kanr) (HpaI-SpeI fragment) of pMV261.Kan with the Zeocin-resistant gene (Zeor) (EcoRV-XbaI fragment) derived from pEM7.Zeo (Invitrogen). The vector pLYG206.Zeo alone or containing cloned genes was transformed into M. marinum strains by electroporation under optimal conditions as previously described (Gao et al., 2003). The PCR primers for M. tuberculosis kasA were Rv2245kasAU2 (5′-CCAGTCAGCCTTCCACCGCTAATG GCGG-3′) and Rv2245kasAD2 (5′-GGCAAGCTTCGTGCT TCAGTAACGCCCGAAG-3′). The primers for M. tuberculosis accD6 were Rv2247accD6U2 (5′-CTACAATCATGGC CCCCGAGGCGGTTGGC-3′) and Rv2247accD6D2 (5′-CCGAAGCTTGCTCGCGGTCAGAACTACAGCG-3′). The primers for M. marinum kasA were MmkasAU1 (5′-TGAC CAAGCCTTCCACTGCTAATGGCGGTTACCC-3′) and MmkasAD1 (5′-CGCAAGCTTTTAGTAGCGCCCGAAAGC GAGCGC-3′). Expression of the cloned genes was detected by SDS-PAGE of whole-cell lysates, which showed accumulation of proteins of predicted sizes.

Extraction and analyses of cell wall lipids

Extraction and separation of M. marinum cell wall lipids to polar lipids (PL), apolar lipids (AL), and mycolic acid methyl esters (MAMEs) were performed as described (Besra, 1998). Polar lipids and AL were resolved in one or two dimensions by high performance TLC (HPTLC) plates (EM Science) on either silica or reverse phase surface, yet no differences were detected between wild type and the kasB mutants. MAMEs were first separated to α-, methoxy- and keto-mycolates on silica HPTLC plates by hexanes:ethyl acetate (95:5, v/v), and the purified fractions from preparative plates were further resolved on reverse phase HPTLC plates by chloroform:methanol (40:60, v/v). Lipids were visualized by spraying with 20% sulphuric acid in ethanol and charring. For metabolic labelling of cell wall lipids, M. marinum were incubated with [14C]-acetic acid (1 µCi ml−1) for 2 h and the cell wall lipids were isolated and separated, exposed to a Cyclone phosphor screen, and analysed by the Cyclone Phosphor System software (Packard Instrument Company).

Argentation TLC analysis

To search for the presence of ethylenic compounds, purified mycolate fractions were resolved on AgNO3-impregnated silica HPTLC plates developed with CH2Cl2 or petroleum ether/diethylether (9:1, v/v, three runs).

Oxidative degradation of mycolates

Cleavage of double bonds in oxygenated mycolates was performed by permanganate-periodate oxidation (van Rudloff, 1956) at 30°C in tertiary butanol as solvent for 16 h. The reaction was stopped by addition of potassium pyrosulphite (Fluka) until decolorization of the solution. After acidification, the oxidative products were extracted with ether, washed with distilled water, and dried. The products were methylated using diazomethane and purified for further analyses.

NMR analysis

1H-NMR spectra of purified MAMEs were obtained in CDCl3 (100% D) using a Bruker AMX-500 spectrometer at 298 K. Chemical shifts values (in p.p.m.) were relative to the internal CHCl3 resonance (at 7.27 p.p.m.).

MALDI-TOF MS analysis

MALDI-TOF mass spectra in the positive mode were acquired on a Voyager-DE STR mass spectrometer (PerSeptive Biosystems, Framingham, MA) equipped with a pulsed nitrogen laser emitting at 337 nm. Samples were analysed in the Reflectron mode using an extraction delay time of 100 ns and an accelerating voltage of 20 kV. To improve the signal-to-noise ratio, 150 single shots were averaged for each spectrum and four individual spectra were accumulated to generate a summed spectrum. Calibration mixture 1 (SequazymeTM Peptide Mass Standards Kit, Perseptive Biosystems) was used to calibrate the spectra. Internal mass calibration was performed with sodiated synthetic corynomycolate ([C32643 + Na]+) and its methyl ester derivative ([C33663 + Na] +) or with sodiated MAME (C761483, [M + Na] +) (Laval et al., 2001). Mycolates dissolved in chloroform (1 mM) was applied as droplets of 1 µl, followed by 0.5 µl of the matrix solution [2,5-dihydroxybenzoic acid (10 mg ml−1) in CHCl3/CH3OH (1:1, v/v)].

Infrared spectroscopy analysis

IR spectra of samples as films on NaCl discs were recorded using a Perkin-Elmer FTIR 1600 spectrometer.

Gas chromatography

Fatty acid methyl esters (FAMEs) derived from saponification of M. marinum were analysed by gas chromatography (GC) on a Hewlett-Packard 5890 series II apparatus equipped with an OV1 capillary column (0.30 mm × 25 m) using helium gas. The temperature program involved an increase from 100 to 300°C (5°C min−1), followed by 10 min at 300°C. The temperature of the injection port used was 250°C, except of the analysis of purified α-, methoxy- and keto-mycolates in which a temperature of 350°C was used to allow the pyrolysis of mycolates and the identification of their α-branched chain. GC-MS analyses of FAMEs, purified mycolates and their oxidative products in the electron impact mode were performed on a HP 5889X mass spectrometer (electron energy, 70 eV) coupled to a HP 5890 series II gas chromatograph fitted with a column identical to that used for GC.

Measurement of cording and cell wall permeability

Cording or colony morphology of M. marinum was observed after growth in 7H9 broth or on 7H10 agar. Resistance to SDS was assayed by inoculating M. marinum (104 ml−1) in 7H9 broth containing SDS (0.01, 0.05, 0.1, or 0.5%, v/v), culturing for 4 days at 30°C, and enumeration of CFUs on 7H10 agar. Acid-fast stain and destain were performed using a TB fluorescent stain kit T (Difco). For in vitro-grown M. marinum, liquid culture at mid-log phase was centrifuged at 50 g for 5 min to remove large aggregates. The suspension was spread onto a microscope glass slide, dried, stained with the Auramine-Rhodamine T solution, destained with Decolorizer TM repetitively for 2 min in each time, and observed using a fluorescent microscope. For intracellular M. marinum in MDMs, the stain and destain procedure was similarly performed at 24 h after infection. Diffusion of [14C]-chenodeoxycholate (PerkinElmer) into M. marinum was measured as described (Jackson et al., 1999), with an incubation time of 0, 2, 4, 6, 8, 10, 15 or 20 min.

Susceptibility of M. marinum to antibiotics

Susceptibility of M. marinum to antibiotics was determined either on 7H10 agar or in 7H9 broth. For testing on agar, a suspension of single bacteria was spotted on 7H10 (102/spot) containing antibiotics at different concentrations, and the MIC was determined as the minimal concentration of an antibiotic that completely inhibited appearance of colonies on agar. For testing in broth culture, a suspension of single bacteria was inoculated to 7H9 (104 ml−1) containing antibiotics at different concentration and cultured for 4 days before enumeration of CFUs on 7H10. The MIC for tests in broth cultures was determined as the minimal concentration of an antibiotic that completely inhibited increase of CFUs.

Susceptibility of M. marinum to defensins and lysozyme

To examine susceptibility of M. marinum to human neutrophil defensin peptide-1 (HNP-1) or protamine, a suspension of single bacteria was inoculated to 7H9 (104 ml−1) containing either peptide at different concentrations and cultured for 4 days before enumeration of CFUs on 7H10. Susceptibility to lysozyme was similarly tested, except that 106 bacteria ml−1 were inoculated because of the extreme susceptibility for the kasB mutants.

Inhibition of phagosome maturation by M. marinum

Bone marrow-derived macrophages infected by M. marinum (MOI = 3) for 24 h were incubated with Lysotracker Red (Molecular Probes) at 100 nM for 40 min. Then, the monolayer was washed twice to remove residual Lysotracker Red and further incubated for 1 h before imaging.

Acknowledgements

The authors thank Robert Chalkley and A. L. Burlingame for assistance with lipid analysis and Luisa M. Stamm and M. A. Lanéelle for critical reading of the manuscript. This work was supported by grants from the NIH to E.J.B. and J.S.C. L.Y.G. is supported by T32 AI07334.

Ancillary

Advertisement