Skip to main content
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2021 Apr 20;118(17):e2023663118. doi: 10.1073/pnas.2023663118

An ABC transporter Wzm–Wzt catalyzes translocation of lipid-linked galactan across the plasma membrane in mycobacteria

Karin Savková a,1, Stanislav Huszár a,1, Peter Baráth b, Zuzana Pakanová b, Stanislav Kozmon b, Marie Vancová c, Martina Tesařová c, Jaroslav Blaško d, Michal Kaliňák e, Vinayak Singh f,g, Jana Korduláková a, Katarína Mikušová a,2
PMCID: PMC8092408  PMID: 33879617

Significance

The cell envelope of Mycobacterium tuberculosis serves as a primary protective barrier of the pathogen, which claims more than a million lives each year. Its basis, the unique mycobacterial cell wall core, is composed of covalently linked peptidoglycan, branched heteropolysaccharide arabinogalactan, and mycolic acids. Construction of this complex structure located on the bacterial surface requires an array of enzymes acting on both sides of the plasma membrane, as well as translocation of metabolic intermediates from the cytoplasm to the periplasmic space. In this work, we characterized an ATP-binding cassette (ABC) transporter involved in the export of galactan polymer produced by cytoplasmic enzymes across the plasma membrane, providing an important insight into the biogenesis of a structure critical for pathogen survival.

Keywords: mycobacterium, cell wall, transport, galactan

Abstract

Mycobacterium tuberculosis, one of the deadliest pathogens in human history, is distinguished by a unique, multilayered cell wall, which offers the bacterium a high level of protection from the attacks of the host immune system. The primary structure of the cell wall core, composed of covalently linked peptidoglycan, branched heteropolysaccharide arabinogalactan, and mycolic acids, is well known, and numerous enzymes involved in the biosynthesis of its components are characterized. The cell wall biogenesis takes place at both cytoplasmic and periplasmic faces of the plasma membrane, and only recently some of the specific transport systems translocating the metabolic intermediates between these two compartments have been characterized [M. Jackson, C. M. Stevens, L. Zhang, H. I. Zgurskaya, M. Niederweis, Chem. Rev., 10.1021/acs.chemrev.0c00869 (2020)]. In this work, we use CRISPR interference methodology in Mycobacterium smegmatis to functionally characterize an ATP-binding cassette (ABC) transporter involved in the translocation of galactan precursors across the plasma membrane. We show that genetic knockdown of the transmembrane subunit of the transporter results in severe morphological changes and the accumulation of an aberrantly long galactan precursor. Based on similarities with structures and functions of specific O-antigen ABC transporters of gram-negative bacteria [C. Whitfield, D. M. Williams, S. D. Kelly, J. Biol. Chem. 295, 10593-10609 (2020)], we propose a model for coupled synthesis and export of the galactan polymer precursor in mycobacteria.


Mycobacterium tuberculosis (Mtb) infects about one-quarter of the global human population and is a causative agent of one of the deadliest infectious diseases in the history of mankind (1). The most recent report by the World Health Organization estimated 10 million cases of tuberculosis (TB) and more than 1.4 million deaths due to this disease in 2019 (2). Although these numbers have been slowly but steadily decreasing over the last years, the situation is expected to become even more severe due to the crisis caused by the COVID-19 pandemic (2, 3). Today, it is largely acknowledged that among difficulties encountered in the efforts to defeat TB, numerous strategies used by the pathogen to evade the immune system of the human host play an important role (4, 5). One of these is the construction of an intricate cell wall, which forms a highly impermeable protective coat of a TB bacillus (6, 7). Interference with its biosynthesis is lethal for the pathogen, as exemplified by the actions of the first-line TB drugs isoniazid and ethambutol—key components of the four-drug regimen currently in use for the treatment of drug-sensitive TB (2).

The core of the mycobacterial cell wall is a covalent complex of peptidoglycan, the branched heteropolysaccharide arabinogalactan, and mycolic acids, which are extremely long (C70 to C90) α-alkyl β-hydroxy fatty acids forming a basis for the mycobacterial outer membrane (8). Attachment of arabinogalactan to peptidoglycan is accomplished by the “linkage region” composed of rhamnose-N-acetyl glucosamine-1-phosphate (Rha-GlcNAc-1-P) (9), which also serves as an initiation point for the synthesis of arabinogalactan (10). The principal steps in the construction of this large and rather complicated structure counting around 100 monosaccharide units are well understood. Synthesis of arabinogalactan begins with the transfer of GlcNAc-1-P from UDP-GlcNAc to decaprenylphosphate (decaprenyl-P) catalyzed by the WecA enzyme, followed by addition of Rha from dTDP-Rha by rhamnosyltransferase WbbL, giving rise to the decaprenol-linked “linkage region” (decaprenyl-P-P-GlcNAc-Rha, glycolipid 2, GL2) (11, 12). Synthesis of galactan is catalyzed by two bifunctional galactofuranosyltransferases, GlfT1 (rv3782 gene product in Mtb H37Rv) and GlfT2 (rv3808c gene product in Mtb H37Rv). GlfT1 attaches the first two galactofuranoses (Galf) to GL2, forming Galf-β(1→4)-Rhap and Galf-β(1→5)-Galf glycosidic bonds. The processive galactosyltransferase GlfT2 then extends the product of GlfT1 reaction, glycolipid 4 (GL4), by adding alternating β(1→5) and β(1→6) linked Galfs in a linear fashion up to about 30 galactose units, producing lipid-linked galactan polymer (decaprenyl-P-P-GlcNAc-Rha-Gal∼30) (13). Galf residues for transferase reactions are donated by UDP-D-Galf, which is produced from UDP-D-Galp by the action of UDP-galactopyranose mutase (Glf, rv3809c gene product in Mtb H37Rv) (14, 15). Branched arabinan domains are built by an array of arabinosyltransferases AftA, AftB, AftC, AftD, EmbA, and EmbB, which utilize decaprenylphosphoryl arabinose as an activated sugar donor (16). While all these arabinosyltransferases belong to the so-called GT-C family of enzymes with multiple transmembrane domains (17, 18), galactosyltransferases GlfT1 and GlfT2 are considered to be soluble enzymes. However, their acceptor substrates have lipidic nature, and the association of GlfT2 with a plasma membrane was proposed based on its crystal structure (19).

It is presumed that arabinogalactan synthesis is initiated on the cytosolic face of the plasma membrane by enzymes relying on sugar nucleotide donors, thus at some point, the arabinogalactan precursor must be translocated across the plasma membrane before its attachment to peptidoglycan by the recently identified ligase(s) (20, 21). Previously, we and others proposed an ATP-binding cassette (ABC) transporter from the conserved mycobacterial arabinogalactan biosynthetic cluster (22) to carry out this function (17, 23, 24). In Mtb, as well as in other pathogenic or environmental mycobacteria, the gene region encoding the transporter has an unusual structure, in that the genes encoding a nucleotide-binding subunit (rv3781 in Mtb H37Rv) and a membrane-spanning subunit (rv3783 in Mtb H37Rv) are separated by the gene encoding the galactosyltransferase GlfT1 (rv3782 in Mtb H37Rv) (Fig. 1A), which pointed out to a role of this transporter in galactan biosynthesis (17, 23, 24). We showed that orthologs of these three genes in Mycobacterium smegmatis mc2155 (Msm) are transcriptionally linked and that the gene encoding the nucleotide-binding subunit is essential (24). Next, we prepared a mutant strain with a disrupted ortholog of the membrane-spanning subunit and performed its phenotypic characterization, but these experiments did not bring conclusive results regarding the precise function or substrate of this ABC transporter (24). Availability of the CRISPR-Cas technology for regulated silencing of the genes in mycobacteria (25) encouraged us to revisit our investigation of the function of this ABC transporter in Msm. We provide evidence that genetic knockdown of the transmembrane subunit, MSMEG_6369, as well as the nucleotide-binding subunit, MSMEG_6366, resulted in the accumulation of lipid-linked galactan polymer (LLG) in the cells and propose that this is a substrate of the studied ABC transporter. Given the similarities of mycobacterial galactan synthesis and transport with those of specific O-antigen polysaccharides in gram-negative bacteria (26, 27), we propose to adopt the terminology from this field (28) and name the nucleotide-binding subunit (Rv3781 orthologs) Wzt and the transmembrane subunit (Rv3783 orthologs) Wzm.

Fig. 1.

Fig. 1.

Conservation of wzt and wzm gene loci in selected mycobacterial genomes and evaluation of their silencing in Msm. (A) Structure of the wzt-glfT1-wzm operons in selected mycobacteria. Wzt homologs are colored in green, GlfT1 homologs are colored in yellow, and Wzm homologs are colored in blue. (B) TLC analysis of lipids extracted from Msm PLJR962 (CON), Msm PLJR962-wztSM (WZT), and Msm PLJR962-wzmSM (WZM) strains grown in the absence or presence of ATc (100 ng/mL). The lipids were separated on silica gel plates in CHCl3/CH3OH/H2O (20:4:0.5, vol/vol), and visualized with cupric sulfate. TDM—trehalose dimycolates, TMM—trehalose monomycolates, and PE—phosphatidyl ethanolamine. (Shown is a representative image corresponding to Experiment 2 specified in SI Appendix, Fig. S1C). (C) DdPCR analysis of wzt, glfT1, and wzm expression in each of the strains Msm PLJR962 (CON), Msm PLJR962-wztSM (WZT), and Msm PLJR962-wzmSM (WZM) strains grown in the absence or presence of ATc (100 ng/mL). The gene expression is normalized to sigA. The error bars are SDs of three technical replicates.

Results

Silencing of wzm and wzt in M. smegmatis Leads to Growth Defects and Morphological Changes, and Points to an Interference with Arabinogalactan Synthesis.

For silencing of the genes encoding Wzt (MSMEG_6366, wztSM) and Wzm (MSMEG_6369, wzmSM) in Msm, we constructed plasmids derived from PLJR962, which contained fragments of the target genes (SI Appendix, Table S1) and allowed CRISPR interference (CRISPRi) inducible by anhydrotetracycline (ATc) (25). The culturing of the Msm transformants carrying these constructs on solid media confirmed that these were severely compromised for growth in the presence of ATc (SI Appendix, Fig. S1A).

For the initial phenotypic characterization of the effects caused by silencing of the target genes, we cultivated the CRISPRi strains in liquid media in the absence or presence of 100 ng/mL ATc. Monitoring of the OD600 indicated severe growth inhibition of ATc-treated Msm PLJR962-wztSM and Msm PLJR962-wzmSM strains compared with the untreated samples, especially at the later time points (SI Appendix, Fig. S1 B and C). Light microscopy revealed the changes in cell shapes of the CRISPRi strains upon ATc-induced silencing of the target genes. In particular, we observed the lemon-shaped structures which corroborate the involvement of these genes in the biosynthesis of the cell wall (29) (SI Appendix, Fig. S2). We next analyzed the lipids obtained from the CRISPRi strains cultivated with or without ATc. The ATc-induced transcriptional repression of wztSM and wzmSM caused a major accumulation of trehalose dimycolates (TDM) and trehalose monomycolates (TMM) in both strains (Fig. 1B and SI Appendix, Fig. S3). Such a phenotype was initially observed after the treatment of mycobacteria with ethambutol, which interferes with the biosynthesis of arabinan components of the mycobacterial cell wall, including the core mycolyl-arabinogalactan-peptidoglycan complex (3032). Effects similar to this phenotype could also indicate the disruption of a galactan component to which the arabinan chains are attached.

Before further characterization of the CRISPRi strains, we decided to evaluate silencing of the target genes in our experimental conditions (SI Appendix, Fig. S1B) and examine the possible polar effects within the operon by measuring the messenger RNA copies using droplet digital PCR (ddPCR). The ATc addition reduced the wztSM expression to almost negligible levels in Msm PLJR962-wztSM, while the repression of the target gene in Msm PLJR962-wzmSM was less severe, reaching ∼46% (Fig. 1C). However, in Msm PLJR962-wztSM, transcriptional repression of the target gene caused a considerable reduction in the expression of the downstream genes (Fig. 1C). Therefore, for the functional characterization of the ABC transporter encoded by wztSM and wzmSM, we decided to use Msm PLJR962-wzmSM, where expression of the critical upstream genes of the operon was not inhibited, and phenotypic changes characteristic for arabinogalactan inhibition after silencing of the target gene were obvious. These were further exemplified by examination of this strain using scanning electron microscopy. The images obtained after 24 and 48 h of ATc treatment revealed severe morphological changes of the Msm PLJR962-wzmSM cultures, particularly at the longer exposure. The most frequent phenotype was swelling of the classically rod-shaped bacteria, predominantly at the septum (Fig. 2).

Fig. 2.

Fig. 2.

Genetic knockdown of wzmSM causes severe morphological changes of Msm. Scanning electron microscopy images of Msm PLJR962 (CON) and Msm PLJR962-wzmSM (WZM) grown in the absence or presence of ATc (100 ng/mL) for indicated times (Scale bar, 1 μm).

Genetic Knockdown of wzm in M. smegmatis Triggers the Accumulation of LLG.

To get a better insight into metabolic changes caused by the genetic knockdown of wzmSM, we performed [14C]-glucose labeling of the CRISPRi strain and the control strain containing the empty PLJR962 plasmid grown with and without ATc. Given the hypothesis that the studied ABC transporter is involved in the translocation of LLG across the plasma membrane, inhibition of wzmSM expression could cause the accumulation of these metabolic intermediates in the cells. In search of such molecules in the radiolabeled cells, we developed the extraction procedure inspired by the experimental scheme utilized for obtaining LLG synthesized by the cell-free system (10, 33). The first step included incubation of the washed harvested cells in hot ethanol (34). The pellets obtained by centrifugation of the ethanolic cell suspensions were extracted with hot solvents CHCl3/CH3OH/H2O (10:10:3) and E-soak. Intriguingly, in this experiment, silencing of wzmSM caused about two to threefold accumulation of the radiolabeled material in both CHCl3/CH3OH/H2O (10:10:3) and E-soak extracts. Since radioactive LLG produced by cell-free reactions can be analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), followed by the transfer of the separated molecules to nitrocellulose and their visualization by autoradiography, we examined the material extracted from the radiolabeled cells in the same way. In addition, we tested the sensitivity of the extracted products toward mild acid and mild alkali hydrolyses. Stability in mild alkali conditions and lability in mild acid conditions are indicative of the polyprenylphosphate-based molecules, and these properties were confirmed for LLG produced in the cell-free reactions (33). These analyses point to the presence of presumably two different polymers extracted by CHCl3/CH3OH/H2O (10:10:3) and E-soak, respectively, which have properties resembling lipid-linked oligosaccharides (Fig. 3A). Next, we analyzed the monosaccharide composition of these extracts and discovered that the major radiolabeled monosaccharide in all extracts is [14C]-glucose, except for the E-soak fraction obtained from Msm PLJR962-wzmSM grown in the presence of ATc. In this case, radioactive galactose was the main monosaccharide (Fig. 3B). Traces of arabinose in this sample could be related to the presence of lipoarabinomannan, which was found in E-soak extract (Fig. 4A). To ensure that the observed changes were due to the on-target effects, we examined two additional CRISPRi strains, bearing PLJR962-wzmSM plasmids with different single guide RNAs for the target gene. We found that genetic knockdown of wzmSM in all tested strains caused overproduction of LLG, as well as of TDM and TMM, as described above (SI Appendix, Fig. S4). Collectively, these data suggest that wzmSM silencing leads to the accumulation of LLG in mycobacteria.

Fig. 3.

Fig. 3.

Radiolabeling reveals the accumulation of LLG in Msm with genetically repressed wzm. Msm PLJR962 (CON) and Msm PLJR962-wzmSM (WZM) strains were grown in the absence or presence of ATc (100 ng/mL). (A) Bacteria were radiolabeled with [14C]-glucose for 3 h and subjected to stepwise extractions. Aliquots of deproteinized CHCl3/CH3OH/H2O (10:10:3) and E-soak extracts were subjected to mild acid (H+) and mild alkali (OH) hydrolyses, neutralized, and separated on SDS-PAGE followed by transfer to nitrocellulose and autoradiography. UN—untreated samples. (B) Monosaccharide composition of CHCl3/CH3OH/H2O (10:10:3) and E-soak extracts. Aliquots of the extracts were dried, hydrolyzed with 2 M trifluoroacetic acid, and separated by TLC in pyridine/ethyl acetate/glacial acetic acid/H2O (5:5:1:3). TLC plates were subjected to autoradiography and detection with α-naphthol. Ara—arabinose, Glc—glucose, Gal—galactose. (C) Cell-free synthesis of LLG by cell envelope enzyme fractions. Reactions were carried out for 2 h and contained UDP-GlcNAc, TDP-Rha, and UDP-[14C]-Gal as a tracer. Tunicamycin was used at 50 μg/mL (lanes labeled with *). LLG were extracted with CHCl3/CH3OH/H2O (10:10:3) and E-soak and subjected to SDS-PAGE analysis followed by transfer to nitrocellulose and autoradiography. Shown are representative images of at least two independent experiments.

Fig. 4.

Fig. 4.

Characterization of LLG from Msm with genetically repressed wzm. (A) SDS-PAGE and Western blot analyses of deproteinized E-soak extracts subjected to mild acid (H+) and mild alkali (OH) hydrolyses. The gel was stained with silver; the blot was probed with anti-LAM antibodies CS-35 and visualized with secondary antibodies conjugated to alkaline phosphatase. The red arrow points to the accumulated LLG. UN—untreated samples. (B) MALDI-TOF MS analysis of fractions 46 to 51 obtained by gel filtration of mild acid–treated LLG. Square—GlcNAc; triangle—Rha; circle—Gal. (C) GC/MS analyses of alditol acetates prepared from the combined galactose-rich fractions from Bio-Gel P-100 (Left) and cell walls obtained from Msm PLJR962-wzmSM (WZM) grown in the absence and presence of ATc (100 ng/mL) (Right). (D) gHSQCAD spectrum of the combined galactose-rich fractions from Bio-Gel P-100 showing the assignments of 5-Galf (black) and 6-Galf (red), respectively. For chemical shifts, see SI Appendix, Fig. S10A.

Our previous examination of the Msm with disrupted wzmSM revealed increased cell-free production of LLG compared with the control strain (24). We thus decided to subject Msm PLJR962-wzmSM to a similar study, along with the control strain containing the empty plasmid PLJR962. Cell envelope fractions prepared from these strains grown with and without ATc served as a source of enzymes and decaprenyl-P, the acceptor for galactan polymerization. Cell-free production of LLG was initiated by the addition of sugar nucleotides necessary for the synthesis of the linkage region, i.e., UDP-GlcNAc and TDP-Rha, and UDP-[14C]-Gal as a tracer for monitoring of galactan polymerization. We observed almost three to seven times higher incorporation of [14C]-Gal into LLG in the reactions containing the enzymes prepared from ATc-treated Msm PLJR962-wzmSM compared with the untreated cells (SI Appendix, Fig. S5A). We then asked whether the increase is due to more efficient de novo synthesis of these metabolic intermediates or if it could be attributed to a mere extension of the preformed LLG, which accumulates in bacteria as a result of wzmSM knockdown. To answer this, we included tunicamycin in the reaction mixtures. This drug inhibits WecA, an enzyme initiating the synthesis of LLG by attaching GlcNAc-1-P from UDP-GlcNAc to decaprenyl-P (10, 12). As shown in Fig. 3C, tunicamycin almost completely inhibited the radiolabeled LLG synthesis when the enzyme fraction isolated from the untreated Msm PLJR962-wzmSM strain was used. On the other hand, the cell envelope fraction prepared from ATc-treated Msm PLJR962-wzmSM strain produced radiolabeled LLG very efficiently (Fig. 3C and SI Appendix, Fig. S5A). We thus concluded that the increase in the [14C]-Gal incorporation into LLG, in this case, is due to the extension of galactan precursors, which are stalled from further processing because of the missing functional transport system. Furthermore, we observed decreased cell-free production of [14C]-GlcNAc-labeled GL1 and GL2 in reactions using the enzyme fractions prepared from ATc-treated Msm PLJR962-wzmSM compared with the bacteria cultivated without ATc. This observation can be explained by the lack of decaprenyl-P, which is bound in the arrested LLG (SI Appendix, Fig. S5B).

M. smegmatis with Genetically Silenced wzm Produces Oversized Galactan Precursors.

Accumulation of galactose-containing material with properties corresponding to LLG in the CRISPRi strain with the genetically silenced expression of wzmSM (Fig. 3) encouraged us to obtain this substance from these cells and to subject it to structural characterization. To this end, we treated the E-soak extracts obtained from the cell envelope fractions of Msm PLJR962-wzmSM grown with and without ATc with proteinase K and subjected this material to mild acid and mild alkali hydrolyses. E-soak efficiently extracted carbohydrate-positive material, which was separated by SDS-PAGE and visualized using silver staining (Fig. 4A). In the ATc-treated culture, we observed accumulation of the material that appeared to be resistant to mild alkali hydrolysis. However, the majority of the extracted carbohydrate matter was labile under these conditions and resistant to mild acid, which resembles the properties of lipoarabinomannan (LAM)/lipomannan (35). This assumption was confirmed by Western blotting and probing the nitrocellulose membrane with anti-LAM–specific antibodies CS-35 (36, 37) (Fig. 4A). Next, mild acid hydrolysis was applied to release the oligosaccharide part from the predicted LLG extracted from ATc-treated Msm PLJR962-wzmSM, which was then separated by gel filtration on Bio-Gel P-100, as described previously (33). Based on the elution volume, the size of this material corresponded to the radiolabeled galactan polymer released from the similarly treated E-soak extract obtained from [14C]-glucose–labeled ATc-treated Msm PLJR962-wzmSM (SI Appendix, Fig. S6A). Matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) of the individual nonradioactive carbohydrate–containing samples revealed the presence of polysaccharides from almost 10 to about 6 kDa in the fractions with most carbohydrates (Fig. 4B and SI Appendix, Fig. S6B). To verify their internal structure, we performed a fragmentation MS/MS experiment on precursor signals (e.g., m/z 8,149). The results showed the presence of oligohexose structures. Terminal N-acetyl hexosamine (in its anhydro form) and deoxyhexose structures were further verified in pseudo-MS3 (T3) experiment that uses fragmentation of the fragment signals generated by the in-source decay (38). Among dominant oligohexose fragments found at the low m/z range of the MS spectrum, we were able to detect minor fragments with a mass corresponding to the HexNAc-DeoxyHex-Hex series. Fragmentation of such ions (e.g., 1,182) showed the presence of HexNAc-DeoxyHex-Hex5 structure (SI Appendix, Fig. S7), which corresponds to the expected GlcNAc-Rha-Galn oligosaccharide. Intriguingly, the sum of the presumed galactosyl monomers, as established by MALDI-TOF MS, ranged from ∼40 to more than 50 residues, which largely exceeds the number of Galf units in the mature mycolyl-arabinogalactan complex recently established to contain ∼35 Galf residues in Msm (39). On the other hand, an examination of the E-soak extract from Msm PLJR962-wzmSM grown without ATc showed almost negligible amounts of the α-naphthol–positive material in the fractions from gel filtration. Their MALDI-TOF MS analysis revealed that in this case, no hexose-containing polymer is present (SI Appendix, Fig. S8).

To verify the nature of the monosaccharides in the Bio-Gel P-100-isolated fractions, we subjected the selected samples to gas chromatography-MS (GC/MS). Indeed, Gal was the major monosaccharide in the fractions containing the predicted galactan polymer, and we also confirmed the presence of the expected Rha (SI Appendix, Fig. S9A). Traces of arabinose, mannose, and glucose, which were found in these fractions, as well as in a few of those preceding the galactan-rich fractions, are very likely derived from arabino-mannans and glucans, which are common components in mycobacterial cell envelope (8) (SI Appendix, Fig. S9A). Precise GC/MS-based quantification of Gal residues relative to Rha in the combined galactan-rich fractions purified by gel filtration revealed the ratio of 46 Gal per 1 Rha, which corresponds to an average galactan size in this material established by MALDI-TOF MS (Fig. 4C). Analysis of insoluble cell wall pellets purified from Msm PLJR962-wzmSM grown with and without ATc disclosed a slightly higher size of galactan after wzmSM silencing (Fig. 4C and SI Appendix, Fig. S9B). However, even in this case, the average estimated number of Gal units per one Rha in the mature cell wall galactan was ∼34, considerably lower than the size of the above-mentioned galactan portion of isolated LLG. Also, the genetic knockdown of wzmSM resulted in a decreased ratio of arabinose to Gal, as well as in the reduction of the glucan content in the purified cell wall (Fig. 4C and SI Appendix, Fig. S9B).

Finally, we assessed the linkages in the Bio-Gel P-100–purified galactan by one-dimensional and two-dimensional (2D) NMR (Fig. 4D). The recorded chemical shifts corresponded well with the reported values for 5-β-Galf and 6-β-Galf (40) (SI Appendix, Fig. S10A). We then confirmed the presence of 5-β-Galf and 6-β-Galf in the ratio of ∼1:1, whereas minor peaks were attributed to the methyl of Rha (1.26 ppm) and acetyl of GlcNAc (2.08 ppm), corresponding to ∼1 molecule of each per 46 Gal. Anomeric region (SI Appendix, Fig. S10B) contains several minor signals that were not assigned but could belong to the above-mentioned Rha/GlcNAc or arabinose, mannose, and glucose-containing impurities.

Genetic Knockdown of wztSM Phenocopies Transcriptional Silencing of wzmSM.

The distinct phenotype of the Msm strain with transcriptionally silenced wzmSM encouraged us to search for the identified changes in the strain targeting the nucleotide-binding subunit of the studied transporter, Msm PLJR962-wztSM. To overcome the polar effect of ATc-induced silencing of wztSM on the downstream gene glfT1SM (Fig. 1), we complemented Msm PLJR962-wztSM with the plasmid pCG76-glfT1SM. Both the parent strain transformed with empty plasmid pCG76 as well as the complemented strain were severely inhibited in the growth in the presence of ATc (SI Appendix, Fig. S11A). Proteomic analysis showed that under these conditions, the amounts of Wzt were drastically reduced in both strains. On the contrary, the GlfT1 protein was hardly detectable in the parent strain but was produced efficiently in the complemented strain (SI Appendix, Table S2). We next examined the lipid composition of these two strains grown with and without ATc, as well as their ability to produce LLG. Thin layer chromatography (TLC) analysis of the organic extracts revealed accumulation of TDM and TMM in both ATc-treated strains (SI Appendix, Fig. S11B). Radiolabeling of the cultures with [14C]-glucose and analysis of the material released with CHCl3/CH3OH/H2O (10:10:3) and E-soak disclosed profiles similar to those obtained with CRISPRi strain targeting wzmSM (SI Appendix, Fig. S11C), although it appeared that LAM was extracted in a relatively higher proportion (SI Appendix, Fig. S11D). As expected, in the cultures grown with ATc, the amount of [14C]-LLG was lower in the parent strain (SI Appendix, Fig. S11 C and E) due to the lack of GlfT1, which synthesizes GL4 serving as an acceptor for galactan polymerization. The structure of the putative glucose-containing polymer found in CHCl3/CH3OH/H2O (10:10:3) extracts of all analyzed CRISPRi strains grown with ATc (Fig. 3 and SI Appendix, Figs. S4 and S11) is currently unknown—its amount is much lower than that of LLG, which precluded its structural characterization. This is exemplified by no detectable signals in the area corresponding to [14C]-labeled CHCl3/CH3OH/H2O (10:10:3) extract on silver-stained SDS-PAGE gel (SI Appendix, Fig. S12).

Molecular Modeling of WzmSM–WztSM Suggests Structural Features Similar to those of a Channel-Forming O-Antigen Polysaccharide ABC Transporter from Aquifex aeolicus.

The crystallographic structure of the Wzm–Wzt domains from A. aeolicus (PDB ID: 6M96) (41) was used as a template for WzmSM–WztSM homology model preparation (Fig. 5A). The alignment of primary sequences of the two proteins shows reasonable identities and homologies for both domains, reaching 30 and 55% for Wzt, and 17 and 42% for Wzm, respectively (Fig. 5B). These comparisons proved that sequence homologies between the Wzt and Wzm from Msm and A. aeolicus are sufficient to produce an accurate homology model, which was generated using the MODELER program (42). Models of WztSM and WzmSM domains were generated as one heterodimeric complex. Five models and additional five loop models for each model, altogether 30 models, were produced and their probability density function (PDF) energies were calculated (SI Appendix, Table S3). The WzmSM–WztSM model with the lowest PDF energy was chosen for further studies (MSMEG6366_MSMEG6369.BL00030001 model) and its alignment with the template X-ray structure is presented in Fig. 5A.

Fig. 5.

Fig. 5.

Molecular modeling of mycobacterial WzmSM–WztSM. (A) Superimposition of the prepared WzmSM–WztSM homology model with the Wzm–Wzt from the template (PDB ID: 6M96) (41). Wzm domains are colored in shades of blue and Wzt domains in shades of green. The lighter color represents the WzmSM–WztSM model, and the darker color represents the Wzm–Wzt template. Several important Wzm–Wzt regions are labeled and highlighted in red or blue colors. PG1 and PG2—periplasmic gate helices 1 and 2; IF—interface helix; GH—gate helix. (B) Sequence and secondary structure assessment (SSA) alignments of the modeled WzmSM (MSMEG_6369) and WztSM (MSMEG_6366) domains with the template Wzm (6M96_b) and Wzt (6M96_a) domains. The sequence similarities and identities are highlighted in the shades of red color. The purple cylinders and khaki arrows represent the helix and sheet secondary structure elements.

The WzmSM homology model structure consists of six transmembrane helixes TM1 to TM6, two periplasmic gate helixes PG1 and PG2, and an interface helix (IF). The alignment of the Wzm secondary structures from Msm and A. aeolicus showed a very good correlation (Fig. 5 A and B), with only a few small differences. Namely, WzmSM TM2 helix is shorter, and TM4 and TM6 helixes are split into two shorter helixes. In case of the WztSM model, the secondary structure alignment showed an almost identical overlay with the Wzt 6M96 domain from A. aeolicus. All secondary structure elements followed the template structure with minimal differences. The important motifs, such as beta-sheet structures β1 and β2, cytosolic gate helix GH, Walker A and Walker B domains, and H-loop and Signature loops have the same lengths (Fig. 5) in both parent structure and the model. Moreover, these structural elements also show high sequence identities.

The three-dimensional (3D) structure alignment of the WzmSM–WztSM complex with 6M96 template structure displays a high level of similarity (Fig. 5A). The measured RMSD of atomic positions on the common backbone atoms revealed a favorable value of 0.4 Å (SI Appendix, Table S3). The main difference can be seen for the TM2 helix, which is shorter compared with the template. However, TM1b and TM2 interconnecting loop is long enough to place the TM2 in the same space as in the 6M96 template. The quality of the 3D structure of the WzmSM–WztSM complex homology model was also evaluated by the Ramachandran plot, and nine amino acid residues were found in the unfavorable regions (SI Appendix, Fig. S13). Considering that the template 6M96 X-ray structure contains three amino acid residues in the unfavorable regions, this result is in accordance with a very good quality of the produced homology model.

Discussion

Biosynthesis of mycobacterial arabinogalactan shares certain features with the assembly of specific O-antigen polysaccharides (O-PS) in gram-negative bacteria (27), as well as of cell wall teichoic acids (WTA) in gram-positive bacteria (43). Probably the closest example is the O2-antigen of Klebsiella pneumoniae. It is characterized by the presence of D-galactan I, containing the alternating D-Galf and D-Galp in the repeating unit [→3)-β-D-Galf-(1→3)-α-D-Galp-(1→]n (44). Its biosynthesis is initiated on an undecaprenylphosphate by the WecA-catalyzed transfer of GlcNAc-1-P from UDP-GlcNAc (45). This step is followed by activities of galactosyltransferases WbbN and WbbO, forming an adapter region, which is extended by bifunctional galactosyltransferase WbbM (46). The genes for the synthesis of the three galactosyltransferases form a cluster with the wzm and wzt genes encoding the subunits of an ABC transporter responsible for translocation of the O2-antigen across the plasma membrane (26, 47). Among the intriguing features of this transporter is its proposed capacity to control the length of the D-galactan I; however, the mechanistic principles of this regulation are not currently known (26, 27). Yet, the recently solved structures of the Wzm–Wzt homolog from A. aeolicus substantially contribute to the understanding of the long polysaccharide translocation across the bacterial plasma membrane (41, 48). The versions of the ABC transporter used for crystallization lack the carbohydrate-binding domain (CBD) in Wzt protein, which is responsible for the recognition of the terminal structures of the translocated O-antigen polymers. Examples of such capping residues include methyl and phosphate groups in Escherichia coli O9a (49, 50) or 3-keto-D-manno-oct-2-ulosonic acid in K. pneumoniae O12 (51), respectively. Nevertheless, as noted by Bi et al. (48) even such truncated forms of Wzt were shown to be completely functional in both of these strains if the CBD domain was supplemented in trans (52, 53), so the obtained structures are relevant for the functional studies. Intriguingly, CBD is naturally missing in the Wzt proteins of both O2-antigen transporter of K. pneumoniae and the mycobacterial ABC transporter of our investigation, which correlates with the lack of termination signals in the respective galactan polymers (26, 54, 55). These common features open the possibility that, similar to O2 antigen, the control of mycobacterial galactan length is linked to translocation of its precursor across the plasma membrane. Such a mechanism for O-antigen ABC transporter was described for the first time in 2009 by Whitfield and coworkers (26). They showed that in K. pneumoniae O2 overexpression of wzm and wzt resulted in the production of shorter forms of D-galactan I–containing smooth LPS (S-LPS) compared with the strain with natural levels of this transporter. On the other hand, lower expression of wzmwzt genes from K. pneumoniae in the host recombinant strain of E. coli, relative to the expression of the gene cluster responsible for D-galactan synthesis, resulted in the production of S-LPS forms with longer galactan polymers. Importantly, the D-galactan I synthesis and its transport were temporally coupled [i.e., no export of the galactan polymer was observed if the expression of ABC transporter genes was induced only after galactan was already produced (26)]. The discovery of the abnormally long LLG in Msm as a result of silencing of wzmSM suggests a similar mechanism, i.e., the studied ABC transporter is very likely involved in both galactan length control and transport in mycobacteria.

The question of mycobacterial galactan length regulation was addressed in several studies. In 2009, Kiessling and coworkers studied the galactan-polymerizing enzyme, GlfT2, and proposed a tethering mechanism for the length control of its product (56). They suggest that a specific acceptor lipid anchoring site on the enzyme is important for the regulation of the extent of polymerization. In this model, both ends of the GL4 acceptor substrate, i.e., the decaprenyl portion, as well as the oligosaccharide part, need to directly interact with GlfT2 to achieve efficient galactan polymerization (56). Subsequent resolution of the GlfT2 structure by Ng and Lowary with collaborators proposed that most of the decaprenyl chain is embedded in the plasma membrane and that GL4 diffuses freely in the membrane to deliver its nonreducing end to the active site of GlfT2 (19). The enzyme is assembled as a tetramer associated with the plasma membrane through the part of its surface, which contains hydrophobic and positively charged amino acid residues. The distinct feature of the GlfT2 homotetramer is the presence of a hollow funnel-shaped pore with a diameter over 40 Å at the part interacting with the membrane and of less than 10 Å at the opposing face. It was suggested that the volume of the cavity regulates the size of the produced galactan, as it can accommodate about 100 to 150 residues, corresponding to four chains each containing about 30 Galf residues (19). In a recent study focusing on galactan length control, Kiessling and coworkers compared the enzymatic products of GlfT2 from Corynebacterium diphtheriae and Mtb and concluded that GlfT2 itself can regulate the galactan length, as the former enzyme produced shorter galactan polymer than the latter one, which appears to be in accordance with the galactan sizes in the two bacterial species (57). This idea was further corroborated in a follow-up work by the same group, pointing out different sizes of galactan polymer in Msm strains in which native glfT2 was replaced with orthologs from C. diphtheriae or Nocardia brasiliensis (39). Our isolation of LLG from Msm cells after transcriptional repression of wzmSM and its thorough chemical analysis not only provided evidence about the substrate of the studied ABC transporter but also led us to suggest an alternative mechanism for the galactan length control linked to its transport. Indeed, the size of the isolated LLG was unexpectedly large, reaching more than 50 Galf residues, far exceeding the number of ∼30 Galf residues found in the mature cell wall galactan. However, an aberrantly long galactan polymer reaching up to 48 units was also produced by purified recombinant mycobacterial GlfT2 when a specific long (C19) alkene phenoxy acceptor substrate [19-phenoxy-nondec-2-enyl β-D-galactofuranosyl-(1→6)-β-D-galactofuranoside] was used (56). Similarly, the polymerizing galactosyltransferase WbbM from K. pneumoniae was found to generate abnormally long products in the cell-free system (46), as well as in the bacteria in the absence of the transporter (26). Thus, the chain-length distribution in both O2a O-PS biosynthesis and in the mycobacterial cell wall galactan precursor production appears to be influenced by the stoichiometry between export and biosynthesis proteins.

It should be noted that until now, LLG molecules were not isolated from mycobacteria, although their involvement in the biosynthesis of the mycobacterial cell wall is widely accepted. Indeed, the key enzymes responsible for their production are known, and they were thoroughly studied (58). However, initially, the pathway was proposed based merely on the characterization of the radioactive enzymatically produced metabolic intermediates, in which structural features typical for mycobacterial arabinogalactan were recognized (10, 33). This included confirmation of the presence of the linkage region monosaccharides GlcNAc and Rha, as well as Gal and even arabinose in the LLG. In fact, the occurrence of arabinose in the enzymatically produced galactan intermediate was unexpected because polymerization of galactan and arabinan were predicted to be spatially separated—the former taking place in the cytoplasm and the latter in the periplasm. However, methylation analysis of the radioactive LLG produced in cell-free reaction revealed the presence of 5,6-linked [14C]-Gal, confirming the presence of arabinosyl residues (33). Our current findings, particularly the MALDI-TOF MS analysis and 2D NMR analyses do not support the presence of arabinose in LLG produced in the cells with transcriptionally repressed wzmSM. This is in accordance with the distinct and well-separated sites of galactan and arabinan build-up in the intact cells.

Molecular modeling suggests a very good correlation of Wzm–Wzt from Msm with the resolved structures of the channel-forming O-antigen polysaccharide ABC transporter from A. aeolicus (41, 48). Accordingly, we propose a similar hypothetical mechanism of LLG translocation across the plasma membrane (Fig. 6). In this model, synthesis of the adapter region by the action of WecA, WbbL, and GlfT1 (Step 1) is followed by GlfT2-catalyzed polymerization (Step 2). The GlcNAc pyrophosphate group at the carbohydrate-reducing end in LLG is recognized by the transporter (59) and it is engaged in translocation. The forward movement of the polymer is assured by ATP hydrolysis, inducing conformational changes in Wzm–Wzt. The structural analysis of the GlfT2 homotetramer indicates the position of the groove responsible for the entrance of the GlfT2 acceptor substrate, GL4. It is located at the membrane-interacting bottom side of GlfT2 between its monomeric units (SI Appendix, Fig. S14). The groove is ∼12 Å wide in the middle, providing enough space for the entry of the GL4 substrate. Moreover, three positively charged amino acid residues that are present in the groove, Lys500, Lys510, and Arg543, can possibly interact with the diphosphate moiety of GL4 and thus be responsible for bringing the substrate into the inner space of the GlfT2 tetramer. We propose that the grooves between the subunits of the tetramer serve not only as the entry points for the GL4 substrates but also as the exit gates for LLG, which is produced by the enzyme (Step 3). This arrangement would allow initiation of the galactan translocation during its synthesis and thus perhaps also regulation of its length.

Fig. 6.

Fig. 6.

Hypothetical model of the coupled galactan synthesis and transport in mycobacteria. Adapter synthesis ensured by the stepwise actions of N-acetyl glucosaminyl-1-P transferase WecA, rhamnosyltransferase WbbL, and galactosyltransferase GlfT1 (Step 1) is followed by galactan polymerization catalyzed by galactosyltransferase GlfT2 (Step 2). Translocation of galactan polymer across the plasma membrane through the Wzm–Wzt transporter starts from the reducing end of the polymer and is driven by ATP hydrolysis occurring in the Wzt subunit (Step 3). Lack of Wzm–Wzt transporter results in excessively long galactan intermediates. The proposed protein–protein interactions are only hypothetical, and they were not experimentally confirmed. The image was created with BioRender.com.

Historically, the determination of the galactan size in Mtb cell wall was challenging, and the reported values differ by about 10 Galf units, ranging from 22 to ∼30 Galf residues (54, 55). Indeed, the presumed heterogeneity of both galactan and arabinan chains was mentioned as one of the limitations of the proposed mycolyl-arabinogalactan-peptidoglycan model (54). In the view of our current findings, such heterogeneity seems to be the intrinsic property of the mature galactan. This conclusion is supported by the precise evaluation of the structure of arabinogalactan from a mutant of Corynebacterium glutamicum with a disrupted emb gene, in which only three arabinoses attached by the initiating arabinosyltransferase AftA were present (60). MALDI-TOF MS examination of the methylated cell wall from this mutant revealed arabinogalactan species containing increasing numbers of Gal residues: Ara3Gal13Rha, Ara3Gal15Rha, Ara3Gal17Rha, and Ara3Gal19Rha, pointing to heterogeneity of the galactan domain size in the intact cells (60).

The microscopic inspection of the Msm CRISPRi strain with the silenced expression of wzmSM disclosed severe morphological changes consistent with the inhibition of the cell wall in a subset of the affected cells. Indeed, similar “lemon-shape” phenotypes were observed in Msm strains after transcriptional repression of the genes involved in the terminal cytosolic steps of cell wall synthesis—murG, encoding glycosyltransferase producing the peptidoglycan precursor lipid II, galactosyltransferase glfT2, and pks13 that is necessary for the synthesis of mycolic acids (29). These enzymes, as well as the studied ABC transporter Wzm–Wzt, play critical roles in mycobacterial cell wall core biosynthesis. It is expected that the discovery of transporters important for biogenesis and maintenance of mycobacterial cell wall will open further possibilities for the TB drug development (61). Recently, cryogenic electron microscopy of an ABC transporter TarGH from Alicyclobacillus herbarius, which translocates WTA precursors from the cytoplasm into extracellular space, revealed the binding site and an inhibitory mechanism of Targocil, a drug active against methicillin-resistant Staphylococcus aureus (62). We believe that elucidation of the function of the Wzm–Wzt transporter from mycobacteria will lead to the exploration of its potential in finding new medicines against TB.

Materials and Methods

The CRISPRi strains derived from M. smegmatis mc2155, which allow ATc-induced silencing of wzm and wzt expression, were prepared by the procedure described by Rock et al. (25) using the plasmid PLJR962 (Addgene). Gene expression analysis in the constructed strains was performed by ddPCR. Cultivation of the strains in liquid media was performed by shaking in glycerol–alanine–salts medium supplemented with tyloxapol (0.025%) and kanamycin (25 μg/mL), with and without ATc (100 ng/mL) at 37 °C. Radiolabeling of the CRISPRi and control strains with [14C] D-glucose was performed essentially as described (30). The radiolabeled cells were extracted step-wise with 96% ethanol for 20 min at 70 °C; CHCl3/CH3OH/H2O (10:10:3) and E-soak [H2O/C2H5OH/diethyl ether/pyridine/NH4OH (15:15:5:1:0.017)] both for 30 min at 70 °C. The latter two extracts were subjected to mild acid and mild alkali hydrolyses as described (33), analyzed by SDS-PAGE followed by blotting to nitrocellulose, and autoradiography. LLG was identified in the E-soak extract and isolated in the same fashion from the cell envelope fraction corresponding to 100,000 × g pellet of the cell lysate obtained by sonication (33). Oligosaccharide part of the lipid-linked galactan was isolated from the deproteinized, mild acid–treated E-soak fraction extracted from ATc-treated M. smegmatis PLRJ962-wzmSM by gel filtration on Bio-Gel P100 (BioRad) in 50 mM ammonium bicarbonate (33). The carbohydrate-positive fractions were analyzed by MALDI-TOF MS, NMR, and GC analyses. Cell-free reactions using cell envelope fraction as an enzyme source and analysis of the reaction products were performed as described previously (33). Full details of the methods are described in SI Appendix.

Supplementary Material

Supplementary File
pnas.2023663118.sapp.pdf (17.3MB, pdf)
Supplementary File

Acknowledgments

This work was supported by the Slovak Research and Development Agency (Grant APVV-15-0515, Grant APVV-18-0282); by the Ministry of Education, Science, Research and Sport of the Slovak Republic (Grant VEGA 2/0137/20 and Grant KEGA 037STU-4/2020); by the Ministry of Education, Youth and Sports of the Czech Republic (Grant MEYS CR -LM2018129 Czech-BioImaging); and by the Operational Program Integrated Infrastructure for the project: Center for Biomedical Research - BIOMEDIRES - II. stage, ITMS: 313011W428, cofinanced by the European Regional Development Fund. V.S. acknowledges support from the South African Medical Research Council, South Africa. The following reagent was obtained through BEI Resources, NIAID, NIH: monoclonal anti-Mycobacterium tuberculosis LAM, clone CS-35 (produced in vitro), NR-13811. PLJR962 was a gift from Sarah Fortune (Addgene plasmid #115162; http://n2t.net/addgene:115162; RRID: Addgene_115162).

Footnotes

The authors declare no competing interest.

This article is a PNAS Direct Submission. C.E.B. is a guest editor invited by the Editorial Board.

This article contains supporting information online at https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.2023663118/-/DCSupplemental.

Data Availability

All study data are included in the article and/or supporting information.

References

  • 1.Houben R. M., Dodd P. J., The global burden of latent tuberculosis infection: A re-estimation using mathematical modelling. PLoS Med. 13, e1002152 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.“Global tuberculosis report 2020” (Geneva, World Health Organization, 2020); Licence: CC BY-NC-SA 3.0 IGO.
  • 3.Togun T., Kampmann B., Stoker N. G., Lipman M., Anticipating the impact of the COVID-19 pandemic on TB patients and TB control programmes. Ann. Clin. Microbiol. Antimicrob. 19, 21 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Baer C. E., Rubin E. J., Sassetti C. M., New insights into TB physiology suggest untapped therapeutic opportunities. Immunol. Rev. 264, 327–343 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Abreu R., Giri P., Quinn F., Host-pathogen interaction as a novel target for host-directed therapies in tuberculosis. Front. Immunol. 11, 1553 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Garcia-Vilanova A., Chan J., Torrelles J. B., Underestimated manipulative roles of Mycobacterium tuberculosis cell envelope glycolipids during infection. Front. Immunol. 10, 2909 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Batt S. M., Minnikin D. E., Besra G. S., The thick waxy coat of mycobacteria, a protective layer against antibiotics and the host’s immune system. Biochem. J. 477, 1983–2006 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Daffe M., Marrakchi H., Unraveling the structure of the mycobacterial envelope. Microbiol. Spectr. 7, GPP3-0027-2018 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.McNeil M., Daffe M., Brennan P. J., Evidence for the nature of the link between the arabinogalactan and peptidoglycan of mycobacterial cell walls. J. Biol. Chem. 265, 18200–18206 (1990). [PubMed] [Google Scholar]
  • 10.Mikusová K., Mikus M., Besra G. S., Hancock I., Brennan P. J., Biosynthesis of the linkage region of the mycobacterial cell wall. J. Biol. Chem. 271, 7820–7828 (1996). [DOI] [PubMed] [Google Scholar]
  • 11.Mills J. A., et al., Inactivation of the mycobacterial rhamnosyltransferase, which is needed for the formation of the arabinogalactan-peptidoglycan linker, leads to irreversible loss of viability. J. Biol. Chem. 279, 43540–43546 (2004). [DOI] [PubMed] [Google Scholar]
  • 12.Huszár S., et al., N-acetylglucosamine-1-phosphate transferase, WecA, as a validated drug target in Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 61, e01310-17 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Belánová M., et al., Galactosyl transferases in mycobacterial cell wall synthesis. J. Bacteriol. 190, 1141–1145 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Weston A., et al., Biosynthetic origin of mycobacterial cell wall galactofuranosyl residues. Tuber. Lung Dis. 78, 123–131 (1997). [DOI] [PubMed] [Google Scholar]
  • 15.Soltero-Higgin M., Carlson E. E., Gruber T. D., Kiessling L. L., A unique catalytic mechanism for UDP-galactopyranose mutase. Nat. Struct. Mol. Biol. 11, 539–543 (2004). [DOI] [PubMed] [Google Scholar]
  • 16.Jankute M., Cox J. A., Harrison J., Besra G. S., Assembly of the mycobacterial cell wall. Annu. Rev. Microbiol. 69, 405–423 (2015). [DOI] [PubMed] [Google Scholar]
  • 17.Berg S., Kaur D., Jackson M., Brennan P. J., The glycosyltransferases of Mycobacterium tuberculosis–Roles in the synthesis of arabinogalactan, lipoarabinomannan, and other glycoconjugates. Glycobiology 17, 35–56R (2007). [DOI] [PubMed] [Google Scholar]
  • 18.Rahlwes K. C., Sparks I. L., Morita Y. S., Cell walls and membranes of actinobacteria. Subcell. Biochem. 92, 417–469 (2019). [DOI] [PubMed] [Google Scholar]
  • 19.Wheatley R. W., Zheng R. B., Richards M. R., Lowary T. L., Ng K. K., Tetrameric structure of the GlfT2 galactofuranosyltransferase reveals a scaffold for the assembly of mycobacterial Arabinogalactan. J. Biol. Chem. 287, 28132–28143 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Harrison J., et al., Lcp1 is a phosphotransferase responsible for ligating arabinogalactan to peptidoglycan in Mycobacterium tuberculosis. mBio 7, e00972-16 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Grzegorzewicz A. E., et al., Assembling of the Mycobacterium tuberculosis cell wall core. J. Biol. Chem. 291, 18867–18879 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Belanger A. E., Inamine J. M., “Genetics of cell wall biosythesis” in Molecular Genetics of Mycobacteria, Hatfull G. F., Jacobs W. R. Jr, Eds. (ASM Press, Washington, DC, 2000), pp. 191–202. [Google Scholar]
  • 23.Cuthbertson L., Kos V., Whitfield C., ABC transporters involved in export of cell surface glycoconjugates. Microbiol. Mol. Biol. Rev. 74, 341–362 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Dianišková P., et al., Investigation of ABC transporter from mycobacterial arabinogalactan biosynthetic cluster. Gen. Physiol. Biophys. 30, 239–250 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Rock J. M., et al., Programmable transcriptional repression in mycobacteria using an orthogonal CRISPR interference platform. Nat. Microbiol. 2, 16274 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Kos V., Cuthbertson L., Whitfield C., The Klebsiella pneumoniae O2a antigen defines a second mechanism for O antigen ATP-binding cassette transporters. J. Biol. Chem. 284, 2947–2956 (2009). [DOI] [PubMed] [Google Scholar]
  • 27.Whitfield C., Williams D. M., Kelly S. D., Lipopolysaccharide O-antigens-bacterial glycans made to measure. J. Biol. Chem. 295, 10593–10609 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Reeves P. R., et al., Bacterial polysaccharide synthesis and gene nomenclature. Trends Microbiol. 4, 495–503 (1996). [DOI] [PubMed] [Google Scholar]
  • 29.Meniche X., et al., Subpolar addition of new cell wall is directed by DivIVA in mycobacteria. Proc. Natl. Acad. Sci. U.S.A. 111, E3243–E3251 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Mikusová K., Slayden R. A., Besra G. S., Brennan P. J., Biogenesis of the mycobacterial cell wall and the site of action of ethambutol. Antimicrob. Agents Chemother. 39, 2484–2489 (1995). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Belanger A. E., et al., The embAB genes of Mycobacterium avium encode an arabinosyl transferase involved in cell wall arabinan biosynthesis that is the target for the antimycobacterial drug ethambutol. Proc. Natl. Acad. Sci. U.S.A. 93, 11919–11924 (1996). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Zhang L., et al., Structures of cell wall arabinosyltransferases with the anti-tuberculosis drug ethambutol. Science 368, 1211–1219 (2020). [DOI] [PubMed] [Google Scholar]
  • 33.Mikusová K., et al., Biosynthesis of the galactan component of the mycobacterial cell wall. J. Biol. Chem. 275, 33890–33897 (2000). [DOI] [PubMed] [Google Scholar]
  • 34.Wolucka B. A., McNeil M. R., de Hoffmann E., Chojnacki T., Brennan P. J., Recognition of the lipid intermediate for arabinogalactan/arabinomannan biosynthesis and its relation to the mode of action of ethambutol on mycobacteria. J. Biol. Chem. 269, 23328–23335 (1994). [PubMed] [Google Scholar]
  • 35.Hunter S. W., Brennan P. J., Evidence for the presence of a phosphatidylinositol anchor on the lipoarabinomannan and lipomannan of Mycobacterium tuberculosis. J. Biol. Chem. 265, 9272–9279 (1990). [PubMed] [Google Scholar]
  • 36.Hunter S. W., Gaylord H., Brennan P. J., Structure and antigenicity of the phosphorylated lipopolysaccharide antigens from the leprosy and tubercle bacilli. J. Biol. Chem. 261, 12345–12351 (1986). [PubMed] [Google Scholar]
  • 37.Kaur D., Lowary T. L., Vissa V. D., Crick D. C., Brennan P. J., Characterization of the epitope of anti-lipoarabinomannan antibodies as the terminal hexaarabinofuranosyl motif of mycobacterial arabinans. Microbiology (Reading) 148, 3049–3057 (2002). [DOI] [PubMed] [Google Scholar]
  • 38.Suckau D., et al., A novel MALDI LIFT-TOF/TOF mass spectrometer for proteomics. Anal. Bioanal. Chem. 376, 952–965 (2003). [DOI] [PubMed] [Google Scholar]
  • 39.Justen A. M., et al., Polysaccharide length affects mycobacterial cell shape and antibiotic susceptibility. Sci. Adv. 6, eaba4015 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Lee R. E., Li W., Chatterjee D., Lee R. E., Rapid structural characterization of the arabinogalactan and lipoarabinomannan in live mycobacterial cells using 2D and 3D HR-MAS NMR: Structural changes in the arabinan due to ethambutol treatment and gene mutation are observed. Glycobiology 15, 139–151 (2005). [DOI] [PubMed] [Google Scholar]
  • 41.Caffalette C. A., Corey R. A., Sansom M. S. P., Stansfeld P. J., Zimmer J., A lipid gating mechanism for the channel-forming O antigen ABC transporter. Nat. Commun. 10, 824 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Sali A., Blundell T. L., Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779–815 (1993). [DOI] [PubMed] [Google Scholar]
  • 43.Brown S., Santa Maria J. P. Jr, Walker S., Wall teichoic acids of gram-positive bacteria. Annu. Rev. Microbiol. 67, 313–336 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Whitfield C., Perry M. B., MacLean L. L., Yu S. H., Structural analysis of the O-antigen side chain polysaccharides in the lipopolysaccharides of Klebsiella serotypes O2(2a), O2(2a,2b), and O2(2a,2c). J. Bacteriol. 174, 4913–4919 (1992). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Clarke B. R., et al., Role of Rfe and RfbF in the initiation of biosynthesis of D-galactan I, the lipopolysaccharide O antigen from Klebsiella pneumoniae serotype O1. J. Bacteriol. 177, 5411–5418 (1995). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Clarke B. R., et al., A bifunctional O-antigen polymerase structure reveals a new glycosyltransferase family. Nat. Chem. Biol. 16, 450–457 (2020). [DOI] [PubMed] [Google Scholar]
  • 47.Bronner D., Clarke B. R., Whitfield C., Identification of an ATP-binding cassette transport system required for translocation of lipopolysaccharide O-antigen side-chains across the cytoplasmic membrane of Klebsiella pneumoniae serotype O1. Mol. Microbiol. 14, 505–519 (1994). [DOI] [PubMed] [Google Scholar]
  • 48.Bi Y., Mann E., Whitfield C., Zimmer J., Architecture of a channel-forming O-antigen polysaccharide ABC transporter. Nature 553, 361–365 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Clarke B. R., Cuthbertson L., Whitfield C., Nonreducing terminal modifications determine the chain length of polymannose O antigens of Escherichia coli and couple chain termination to polymer export via an ATP-binding cassette transporter. J. Biol. Chem. 279, 35709–35718 (2004). [DOI] [PubMed] [Google Scholar]
  • 50.Hagelueken G., et al., A coiled-coil domain acts as a molecular ruler to regulate O-antigen chain length in lipopolysaccharide. Nat. Struct. Mol. Biol. 22, 50–56 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Vinogradov E., et al., Structures of lipopolysaccharides from Klebsiella pneumoniae. Eluicidation of the structure of the linkage region between core and polysaccharide O chain and identification of the residues at the non-reducing termini of the O chains. J. Biol. Chem. 277, 25070–25081 (2002). [DOI] [PubMed] [Google Scholar]
  • 52.Cuthbertson L., Powers J., Whitfield C., The C-terminal domain of the nucleotide-binding domain protein Wzt determines substrate specificity in the ATP-binding cassette transporter for the lipopolysaccharide O-antigens in Escherichia coli serotypes O8 and O9a. J. Biol. Chem. 280, 30310–30319 (2005). [DOI] [PubMed] [Google Scholar]
  • 53.Mann E., Mallette E., Clarke B. R., Kimber M. S., Whitfield C., The Klebsiella pneumoniae O12 ATP-binding cassette (ABC) transporter recognizes the terminal residue of its O-antigen polysaccharide substrate. J. Biol. Chem. 291, 9748–9761 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Bhamidi S., et al., The identification and location of succinyl residues and the characterization of the interior arabinan region allow for a model of the complete primary structure of Mycobacterium tuberculosis mycolyl arabinogalactan. J. Biol. Chem. 283, 12992–13000 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Bhamidi S., et al., Detailed structural and quantitative analysis reveals the spatial organization of the cell walls of in vivo grown Mycobacterium leprae and in vitro grown Mycobacterium tuberculosis. J. Biol. Chem. 286, 23168–23177 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.May J. F., Splain R. A., Brotschi C., Kiessling L. L., A tethering mechanism for length control in a processive carbohydrate polymerization. Proc. Natl. Acad. Sci. U.S.A. 106, 11851–11856 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Wesener D. A., Levengood M. R., Kiessling L. L., Comparing galactan biosynthesis in Mycobacterium tuberculosis and Corynebacterium diphtheriae. J. Biol. Chem. 292, 2944–2955 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Konyariková Z., Savková K., Kozmon S., Mikušová K., Biosynthesis of galactan in Mycobacterium tuberculosis as a viable TB drug target? Antibiotics (Basel) 9, 20 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Schirner K., Stone L. K., Walker S., ABC transporters required for export of wall teichoic acids do not discriminate between different main chain polymers. ACS Chem. Biol. 6, 407–412 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Alderwick L. J., et al., Deletion of Cg-emb in corynebacterianeae leads to a novel truncated cell wall arabinogalactan, whereas inactivation of Cg-ubiA results in an arabinan-deficient mutant with a cell wall galactan core. J. Biol. Chem. 280, 32362–32371 (2005). [DOI] [PubMed] [Google Scholar]
  • 61.Jackson M., Stevens C. M., Zhang L., Zgurskaya H. I., Niederweis M., Transporters involved in the biogenesis and functionalization of the mycobacterial cell envelope. Chem. Rev., 10.1021/acs.chemrev.0c00869 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Chen L., et al., Cryo-electron microscopy structure and transport mechanism of a wall teichoic acid ABC transporter. mBio 11, e02749-19 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary File
pnas.2023663118.sapp.pdf (17.3MB, pdf)
Supplementary File

Data Availability Statement

All study data are included in the article and/or supporting information.


Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES