JBC Papers in Press. Published on August 29, 2017 as Manuscript M117.804377 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M117.804377 LmeA is required for mycobacterial lipomannan elongation The Cell Envelope-Associated Phospholipid-Binding Protein LmeA is Required for Mannan Polymerization in Mycobacteria Kathryn C. Rahlwesa, Stephanie A. Haa, Daisuke Motookab, Jacob A. Mayfieldc, Lisa R. Baumoela, Justin N. Stricklanda, Ana P. Torres-Ocampoa, Shota Nakamurab, and Yasu S. Moritaa1 a. Department of Microbiology, University of Massachusetts, Amherst, MA, 01003, USA b. Department of Infection Metagenomics, Research Institute for Microbial Diseases, Osaka University, Osaka, 565-0871, Japan c. Division of Rheumatology, Immunology and Allergy, Brigham and Women’s Hospital, Harvard Med- ical School, Boston, MA 02445, USA Running title: LmeA is required for mycobacterial lipomannan elongation To whom correspondence should be addressed: Yasu S. Morita, Department of Microbiology, University of Massachusetts, 639 North Pleasant Street, Amherst MA 01003, USA. Tel: +1-413-545-4604, Fax: +1- 413-545-1578, E-mail: [email protected] D o Keywords: bacterial genetics, carbohydrate metabolism, cell envelope, glycolipid, glycosyltransferase, w n lipomannan, Mycobacterium smegmatis, PimE loa d e d fro m ABSTRACT mannan as observed in the suppressor mutants. http The integrity of the distinguishing, multilaminate LmeA carries a predicted N-terminal signal pep- ://w cell envelope surrounding mycobacteria is critical tide, and density gradient fractionation and deter- w w to its survival and pathogenesis. The prevalence of gent extractability experiments indicated that .jb c phosphatidylinositol mannosides in the cell enve- LmeA localizes to the cell envelope. Using a lipid .org lope suggests an important role in the mycobacte- ELISA assay, we found that LmeA binds to plas- b/ y rial life cycle. Indeed, deletion of the pimE gene ma membrane phospholipids such as phosphati- gu e s (ΔpimE) encoding the first committed step in dylethanolamine and phosphatidylinositol. LmeA t o n phosphatidylinositol hexamannoside biosynthesis is widespread throughout the Corynebacteriales; A p in Mycobacterium smegmatis results in the for- therefore, we concluded that LmeA is an evolu- ril 4 mation of smaller colonies than wildtype colonies tionarily conserved cell-envelope protein critical , 2 0 on Middlebrook 7H10 agar. To further investigate for controlling the mannan chain length of lipo- 19 potential contributors to cell-envelope mannan mannan/lipoarabinomannan. biosynthesis while taking advantage of this colony morphology defect, we isolated spontaneous sup- pressor mutants of ΔpimE that reverted to wildtype colony size. Of 22 suppressor mutants, 6 accumu- The rise in multi-drug resistant Mycobacterium lated significantly shorter lipomannan or lipoarab- tuberculosis (Mtb) is a global concern. One avenue inomannan. Genome sequencing of these mutants to search for novel drug targets is through the bio- revealed mutations in genes involved in the lipo- synthetic pathway of the mycobacterial cell enve- mannan/lipoarabinomannan biosynthesis, such as lope components (1). Several enzymes in biosyn- those encoding the arabinosyltransferase EmbC thetic pathways of the key cell envelope glycoli- and the mannosyltransferase MptA. Furthermore, pids, such as phosphatidylinositol mannosides we identified three mutants carrying a mutation in (PIMs), lipomannan (LM), and lipoarabinomannan a previously uncharacterized gene, MSMEG_5785, (LAM), are shown or predicted to be essential in that we designated lmeA. Complementation of the- Mtb (2, 3). Moreover, alterations in the LM/LAM se suppressor mutants with lmeA restored the orig- structures make both Mtb and nonpathogenic My- inal ΔpimE phenotypes, and deletion of lmeA in cobacterium smegmatis (Msmeg) susceptible to β- wildtype M. smegmatis resulted in smaller lipo- lactam antibiotics, which otherwise cannot cross 1 Copyright 2017 by The American Society for Biochemistry and Molecular Biology, Inc. LmeA is required for mycobacterial lipomannan elongation the mycobacterial cell wall efficiently (4). These cell envelope. In the current study, our initial ob- observations suggest that PIMs/LM/LAM play servation that ΔpimE shows smaller colony mor- important structural roles within the cell envelope phology when grown on Middlebrook 7H10 agar to maintain the permeability barrier. Furthermore, led us to isolate spontaneous suppressors of PIMs/LM/LAM are critical for virulence, with ΔpimE, which restored the small colony morphol- recognized host factor interactions (5-7). ogy to that of WT. In some of these suppressor mutants, we found structural alterations in The biosynthesis of PIMs/LM/LAM begins with LM/LAM, which were caused by mutations in a phosphatidylinositol (PI) and is mediated by se- previously uncharacterized protein, quential actions of two mannosyltransferases: Pi- MSMEG_5785. Characterization of mA and PimB′, followed by the acylation of the MSMEG_5785 reveals this protein as a cell enve- mannose residue by an acyltransferase, PatA, to lope-associated phospholipid binding protein, in- produce AcPIM2 (Fig. 1A) (8-11). Sequential ad- volved in the mannan elongation of LM/LAM. dition of mannoses to AcPIM2 is mediated by an undetermined mannosyltransferase(s), resulting in RESULTS the production of AcPIM4 (12). The AcPIM4 in- Isolation of ΔpimE suppressor mutants. When the termediate serves as a branch point for ΔpimE Msmeg mutant was grown on Middlebrook PIMs/LM/LAM biosynthesis, feeding into either 7H10 agar plates for 4 days at 37°C, we noticed D o LM/LAM formation or AcPIM6 production (Fig. that the colony size of the mutant was significantly w n 1A). The α1–2 mannosyltransferase PimE trans- smaller than WT (WT, 4.17 ± 0.91 mm; ΔpimE, lo a d fers a mannose residue from polyprenol- 1.31 ± 0.45 mm; average ± standard deviation, n = ed phosphate-mannose (PPM) to AcPIM4, commit- 50). Interestingly, we noticed occasional appear- fro m ting the pathway to the production of AcPIM6 ances of spontaneous large colony mutants after h ttp (13). LM/LAM biosynthesis is mediated by α1–6 sub-culturing (Fig. 1B). We isolated a total of 22 ://w mannosyltransferase MptA (MSMEG_4241) and spontaneous large colony revertants. Because w w the α1–2 mannosyltransferase MptC ΔpimE was generated by replacing the central 213 .jb c (MSMEG_4247) (14-18). MptA is a PPM- bp of the pimE gene with kanamycin-resistant .o rg dependent polymerase of the α1–6 mannose back- gene (13), it seemed unlikely that the pimE gene b/ y bone, and the deletion of mptA results in the ac- was restored. Indeed, all 22 isolates remained kan- g u e cumulation of an immature LM intermediate that amycin-resistant (not shown) and failed to produce st o carries 5-20 mannose residues instead of 21-34 AcPIM6 (Figs. 1C and S1). These results indicated n A p mannose residues found in mature LM/LAM (14). that these revertants are suppressor mutants rather ril 4 MptC is another PPM-dependent mannosyltrans- than same site revertants of the pimE gene. , 2 0 ferase, involved in mono-mannose side chain addi- 1 9 tion. While the functional roles of mono-mannose Suppressor mutants show changes in LM/LAM side chains remain obscure, we have previously structure. As described in the Introduction, the suggested that mannan chain biosynthesis requires growth defect observed in ΔlpqW was restored by a balance between the enzymatic activities of suppressor mutations in pimE (19). Inspired by MptA and MptC (18). these studies, we examined if our suppressor mu- tants of ΔpimE show structural changes in LpqW is a possible regulator of LM/LAM biosyn- LM/LAM. Various changes in both LM/LAM size thesis at the AcPIM4 branching point (19-21). Mu- and abundance were noted in several of the sup- tants lacking lpqW show defective LM/LAM bio- pressor mutants (Figs. 1D and S2). For example, synthesis and colony size smaller than wildtype the suppressor mutant S4 shows smaller LAM (WT) (19). These phenotypic defects were sponta- while LM appears comparable to that of the WT. neously resolved in suppressor mutants of ΔlpqW, Another prominent example is found in S1, S10, and mutations in pimE identified in these mutants S20, S21, and S22, where both LM and LAM ap- suggest that alterations in AcPIM6 biosynthesis peared smaller. We also found mutants with can phenotypically compensate for defects in changes in the amount of LM/LAM (e.g. S5, S9, LM/LAM. These studies highlight the complex S12, S18, and S19). While there are a number of interplay of PIMs/LM/LAM to create a functional suppressor mutants that show no apparent changes 2 LmeA is required for mycobacterial lipomannan elongation in LM/LAM, these initial observations suggested restored in each mutant when complemented with that one way for the ΔpimE mutant to revert back the P -lmeA-HA vector (Fig. 3B). Additionally, hsp60 to the WT colony morphology is to alter the abun- the complemented strains restored the ΔpimE dance or structures of LM/LAM. small colony morphology (Fig. 3C). Taken togeth- er, three out of the twenty-two suppressor mutants Whole genome sequencing reveals mutations in of ΔpimE carried a loss-of-function mutation in genes involved in LM/LAM biosynthesis. We used the novel gene lmeA. whole genome sequencing of the parental ΔpimE mutant and the subset of suppressor mutations LmeA is critical for LM/LAM mannan chain length with apparent structural changes in LM/LAM to maturation. The LM/LAM phenotype of the sup- identify genetic alterations consistent with altered pressor mutants suggested a potential role of LM/LAM biosynthesis (Dataset S1). Table 1 LmeA in LM/LAM biosynthesis. However, the summarizes the mutations found in the sequenced phenotype could be dependent on the ΔpimE mutant strains. In the mutant S4, where smaller background. To test a direct role of LmeA in LAM was observed, we found a point mutation in LM/LAM biosynthesis, we generated a strain car- embC, an arabinosyltransferase involved in LAM rying a markerless deletion of lmeA (ΔlmeA) in the biosynthesis, being consistent with the LAM- WT background (Fig. S3A). The deletion of lmeA specific defect. For other mutants accumulating was confirmed by PCR (Fig. S3B). The ΔlmeA D o smaller LM and LAM, such as S1, S10, S20, S21, mutant showed the production of smaller LM and w n and S22, we suspected that there might be a muta- more disperse LAM, which corroborates the phe- lo a d tion in mptA gene, which encodes α1–6 mannosyl- notype of the suppressor mutants (Fig. 4A). Im- ed transferase involved in LM/LAM mannan elonga- portantly, there were no apparent changes in the fro m tion. We therefore PCR-amplified mptA gene and biosynthesis of PIMs or other phospholipids (Figs. h ttp determined the gene sequence by the Sanger 4B and S4A). Moreover, the ΔlmeA mutant colony ://w method. As suspected, we identified a mutation in morphology was similar to WT (Fig. 4C). By w w the mptA gene amplified from S20 and S21 (Table complementing with either Phsp60-lmeA-HA or Pna- .jbc 1). These two mutants have the same 24 bp dele- -lmeA-HA, mature LM/LAM production was .o tive rg tion, suggesting that they originated from the same restored (Fig. 4A). We also complemented the de- b/ y parental mutant. In contrast, we found no mutation letion mutant with lmeA carrying the point muta- g u e in the mptA gene from S1, S10, and S22. There- tions found in the suppressor mutants S1 and S10. st o fore, we subjected S1 and S10 to whole genome Neither of the lmeA mutants was able to restore n A p sequencing, and found missense mutations mature LM/LAM (Fig. S4B), further supporting ril 4 (G170D and V181G, respectively) in the gene our notion that the mutations are loss-of-function. , 2 0 MSMEG_5785, which has no previously assigned Interestingly, the expression levels of the mutated 1 9 function (Fig. 2). We asked whether the third sup- LmeA proteins were lower than that of WT protein pressor mutant (S22) had mutations in and were undetectable by western blot (Fig. S4C), MSMEG_5785 by PCR amplification (see Table indicating that the mutant LmeA may be unstable S1 for primer sequences). Interestingly, S22 had and susceptible to protein degradation. These re- an insertion of a 2,276 bp transposon (TnpR) after sults indicate that LmeA is involved in LM/LAM the first 14 bp of the MSMEG_5785 gene (Fig. 2). biosynthesis and the lmeA mutant phenotype is The TnpR insertion into the near 5’ end of the independent of the pimE deletion. gene suggested that the phenotypic changes in LM/LAM are likely due to the loss of the gene LmeA is a cell envelope protein. LmeA is a con- function. Given these mutant characteristics and served protein of unknown functions widely pre- additional features described below, we termed sent in the order Corynebacteriales (Fig. S5A), but MSMEG_5785 as lipomannan elongation factor A are apparently absent in other bacteria and other (LmeA). To confirm loss of LmeA function, we domains of life. Its ortholog in Mtb, Rv0817c, is introduced a P -lmeA-HA expression vector into 60% identical to Msmeg LmeA at the amino acid hsp60 each suppressor mutant. No changes in PIMs were level, and the missense mutations found in the detected in the complemented suppressor mutants suppressor mutants S1 and S10 are both conserved (Fig. 3A). Importantly, mature LM and LAM were in Mtb LmeA (Fig. S5B). LmeA has a conserved 3 LmeA is required for mycobacterial lipomannan elongation hydrophobic region near the amino terminus, and either the unknown polymerase or MptA. In order SignalP 3.0, an algorithm effective for the predic- to determine the genetic interaction of LmeA with tion of mycobacterial signal peptides (22), predict- either polymerase, we introduced a tet-off mptA ed both Rv0817c and MSMEG_5785 to carry a knockdown system (27) into the ΔlmeA mutant, signal peptide with high probabilities (0.998- and suppressed the expression of mptA by anhy- 1.000). Indeed, Mtb LmeA is a periplasmic protein drotetracycline (atc). As shown in Fig. 6A, the identified by a recent secretome analysis (23). Fur- small LM accumulating in ΔlmeA (-atc, lane 3) thermore, Mtb LmeA is found in the proteome of became even smaller on SDS-PAGE upon sup- Mtb cell lysate but not in that of cell filtrate (24), pression of the mptA expression (+atc, lane 4). The implying cell envelope association. In elucidating small LM accumulating in ΔlmeA strain under the accessibility of Msmeg LmeA-HA, we noticed mptA knockdown condition was apparently no that the protein could not be immunoprecipitated different from the small LM accumulating in the from crude cell lysate using anti-HA antibody mptA knockdown under the WT background (Fig. 5A). This is in contrast to GlnA-HA, a cyto- (compare lanes 2 and 4). To confirm further, we plasmic protein readily immunoprecipitated using analyzed the size distribution of LM by MALDI- anti-HA antibody. When we added Triton X-100, TOF using a previously established method (27). a mild detergent, to the cell lysate, we were able to We observed that the ΔlmeA LM carried 10-22 pull down the LmeA-HA protein (Fig. 5A). These mannose residues, being larger than the LM pro- D o data also implied that LmeA-HA is a cell envelope duced under mptA knockdown, which carried 11- w n protein and the HA epitope is not exposed for the 17 mannoses (Fig. 6B). Nevertheless, ΔlmeA LM lo a d antibody recognition when detergent is not pre- was still significantly smaller than WT LM (21-34 ed sent. To confirm further, we performed sucrose mannose residues). Therefore, these analyses by fro m density gradient sedimentation to fractionate the MALDI-TOF were consistent with the SDS- h ttp cell envelope, intracellular membrane domain PAGE migration patterns. Importantly, the LM ://w (IMD), and cytoplasmic fractions (25, 26). Mark- intermediates accumulating in the mptA knock- w w ers, MptC, PimB′, and Mpa, respectively, showed down strains were identical regardless of lmeA .jb c separation of these three subcellular fractions, and deletion. Taken together, these data suggest that .o rg demonstrated that LmeA-HA co-localizes to the MptA is epistatic to LmeA. b/ y cell envelope fraction (Fig. 5B). These data pro- g u e vide further evidence that LmeA is an extra- The hypostatic nature of LmeA suggested that it st o cytoplasmic cell envelope-associated protein. might facilitate the function of MptA or other n A p downstream enzymes. We therefore wondered if ril 4 LmeA acts on a biosynthetic step genetically hypo- the over-expression of LmeA-HA could restore the , 2 0 static to MptA-mediated LM mannan elongation. small colony morphology in other suppressor mu- 1 9 Small intermediates of LM have been previously tants such as S4 and S21, which have mutations in identified in the knockdown or knockout of MptA, embC (S4) and mptA (S21), respectively. Howev- the mannosyltransferase that elongates the mannan er, the additional expression of LmeA had no ob- chain during LM/LAM biosynthesis (14, 27). This vious impact on the colony size (Table S2), indi- immature LM intermediate is known to carry 5-20 cating that LmeA does not have dominant effects mannoses. While LmeA has no homology to on these other mutants. known mannosyltransferases, the small LM pro- duced in the ΔlmeA mutant suggested that LmeA LmeA binds to phospholipids. While LmeA is plays a role in the elongation of LM. At least two widely conserved in Corynebacteriales, standard mannosyltransferases are thought to be involved in BLAST analyses did not reveal any homologs the elongation step: an unknown α1–6 mannosyl- with a protein of known function. Using RaptorX transferase(s) that extends α1–6 mannose chain on Structure Prediction algorithm (28), we were able AcPIM4 to produce the LM intermediates seen in to predict the β-rich structure of LmeA with a P- mptA deletion or knockdown strains, and MptA, value of 4.90 x 10-4 using several templates includ- the α1–6 mannosyltransferase that extends the ing the bactericidal/permeability-increasing pro- chain generated by the unknown polymerase(s) tein (BPI). BPI is a lipid-binding protein produced (18). We hypothesize that LmeA may facilitate by lymphocytes and specifically binds to the lipid 4 LmeA is required for mycobacterial lipomannan elongation portion of bacterial lipopolysaccharides (29). Giv- monophosphate (GGP) was minimal with an ap- en this homology, we were prompted to test the parent Kd value of 8.65 µM (Fig. S6I). Finally, we possibility that LmeA binds the plasma membrane tested if LmeA binding to PE could be inhibited phospholipids or LM intermediates directly to ex- by the addition of soluble mannose 1-phosphate or ert its function in the LM biosynthesis. We created GDP-mannose. However, up to 10 mM of these an Escherichia coli strain transformed with an compounds had no effects on the binding of LmeA IPTG-inducible His-LmeA expression vector and to PE, implying that mannose-containing mole- prepared a cell lysate of E. coli heterologously cules are not involved in the substrate recognition expressing His-LmeA. We tested the E. coli lysate of LmeA (Fig. S7). Taken together, we conclude by anti-His western blotting and found a single that the binding of LmeA is specific to glycer- band with the expected molecular weight of 29 ophospholipids. kDa (Fig. 7A). We modified an established LAM- binding assay (30) and designed a lipid ELISA DISCUSSION assay to test if His-LmeA binds to any lipids. In this study, we revealed that the previously un- Commercially available lipids were first separated characterized protein LmeA encoded by by TLC and stained by cupric acetate to evaluate MSMEG_5785 is involved in LM/LAM biosyn- their purities (Fig. S6A-C). We also purified the thesis through a forward genetic screen of the small LM intermediates from the mptA- ΔpimE mutant. We propose that LmeA is a cell D o knockdown strain. We found that the binding of envelope-associated phospholipid-binding protein w n His-LmeA to the LM intermediates was minimal that facilitates the maturation of mannan chain lo a d (Figs. 7B and S6D). In contrast, we observed more length. ed robust binding of His-LmeA to phospholipid spe- fro m cies such as PI and phosphatidylethanolamine Several pieces of evidence suggest that LmeA is h ttp (PE), (Fig. 7B), which are major structural com- directly involved in mannan elongation of ://w ponents of Msmeg plasma membrane (31, 32). The LM/LAM biosynthesis. First, the deletion of the w w binding of His-LmeA to PI and PE was dose- lmeA gene resulted in the production of LM/LAM .jb c dependent and saturated at ~2 µM (Fig. S6E-F). with shorter chain length. Second, LmeA is an .o rg Although our current assay system has limitations extra-cytoplasmic cell envelope-associated pro- b/ y that make the accurate determination of Kd values tein. Third, LmeA is genetically hypostatic to g u e difficult, we determined the apparent Kd values MptA. Together, these observations are consistent st o from the available data shown in Fig. S6 to com- with the idea that LmeA regulates the mannan n A p pare the relative binding affinities of LmeA to dif- chain polymerization by controlling the catalytic ril 4 ferent lipid species. The Kd values for PI and PE activity or processivity of MptA (Fig. 8). Howev- , 2 0 were 0.242 µM and 0.487 µM, respectively, show- er, the precise function of LmeA remains un- 1 9 ing that the binding affinity of LmeA to PI was known. One possibility is that LmeA controls the comparable but slightly higher than that to PE. availability of the mannose donor substrate PPM Because LmeA bound both PI and PE effectively, or polyprenol-phosphate-based carbohydrate do- we used phosphatidic acid (PA), a “headless” nors in general. Such a function of LmeA could phospholipid, to examine if the phospholipid head rescue other suppressor mutants such as S4 and group is relevant to the binding (Fig. 7B and S6G). S21 by increasing the general availability of poly- The binding of LmeA to PA was comparable to PI prenol-phosphate-based carbohydrate donors. with the Kd value of 0.297 µM, indicating that the However, we did not observe such effects of phospholipid head groups do not affect the binding LmeA on these other suppressor mutants. Fur- affinity. We also tested triacylglycerol (TAG), but thermore, lipid binding assays suggested that LmeA did not bind to this glycerolipid (Figs. 7B LmeA preferentially binds glycerophospholipids and S6H), indicating that the phosphate residue is over geranylgeranyl-phosphate and that the bind- critical for the binding. Because mannan polymer- ing of LmeA was not competitively inhibited by ization for LM biosynthesis utilizes PPM as the GDP-mannose or mannose 1-phosphate. There- mannose donor, we tested if LmeA binds to poly- fore, we suggest that LmeA is not involved in the prenol-phosphate with a comparable affinity to metabolism and trafficking of polyprenol- PA. However, binding of LmeA to geranylgeranyl phosphate-based substrates. 5 LmeA is required for mycobacterial lipomannan elongation that the lack of one particular glycan structure in Based on the fact that LmeA is extra-cytoplasmic mammalian cells is compensated by the produc- and binds to phospholipids such as PI and PE, we tion of bioequivalent glycans in the Golgi appa- propose that LmeA is a protein peripherally bound ratus (39). Such principles of glycan homeostasis to the plasma membrane. Nikaido and colleagues may apply to the maintenance of cell surface gly- have taken the approach of reverse micelle extrac- cocalyx in mycobacteria as well. However, com- tion to show that major phospholipid species such pletely different scenarios are also possible. For as PI and PE are found predominantly in the plas- ma membrane and relatively depleted in the my- example, it has not been explored if PIMs play a cobacterial or corynebacterial outer membrane role as signaling molecules to facilitate colony (33, 34). Therefore, the ability of LmeA to bind to growth. We have previously reported that Msmeg phospholipid species suggests that LmeA is a as well as Corynebacterium glutamicum produce periplasmic plasma membrane-associated protein PI 3-phosphate (40), but nothing is known if inosi- (Fig. 8). In Mycobacterium marinum, the LmeA tol polyphosphates is released by a phospholipase. ortholog (MMAR_4866) was relatively resistant to Similarly, it is completely unknown if there are a differential detergent extraction, which was ef- phospholipases that can act on PIMs to release the fective in selectively extracting known outer glycan head groups and if such released molecules membrane proteins (35), further supporting the D periplasmic location of LmeA. Nevertheless, Mtb can function as signaling molecules. Interestingly, ow n homolog of LmeA (Rv0817c) is predicted to be an a recent report suggested that synthetic lipid- lo a d outer membrane β-barrel (36). These contradictory linked arabinomannan heptasaccharide can effec- ed observations highlight that further studies are tively inhibit biofilm formation in Msmeg (41), fro m needed to determine functional and structural fea- implying that comparable molecules like phospho- http tures of LmeA. Although our attempts have so far lipase-digested AcPIM6 may have similar biologi- ://w been unsuccessful, an important next step is the w cal activities. Finally, lipid-free D-arabino-D- w purification of LmeA in its active form. .jb mannan, D-mannan, and their phosphorylated c.o Complex interplay of PIMs/LM/LAM in the integ- counterparts are found in the extracellular capsules brg/ y rity of the mycobacterial cell envelope is an of Mtb and other mycobacteria, implying that there gu e may be phospholipases that can act on larger PI- s emerging theme. Previous studies showed that mu- t o tations in pimE can compensate for the growth anchored glycolipids such as LM and LM (42-46). n A p defect of the lpqW mutant that produces reduced We speculate that such phospholipases may pro- ril 4 miscuously act on PIMs as well. The ΔpimE sup- , 2 levels of LM/LAM (19, 21). Our current study 0 1 pressor mutants that we identified includes many 9 showed an opposite biological response where the additional mutants that do not show changes in growth defect of ΔpimE was compensated by addi- LM/LAM structures, suggesting that there are tional structural changes in LM/LAM. It appears multiple pathways to rescue the small colony mor- that the balance between the levels of AcPIM6 and phology of ΔpimE. Analysis of other suppressor LM/LAM is critical for fitness, and mycobacteria mutants may further reveal the potential molecular regain optimal growth by genetic changes that mechanisms behind the biological responses of warrant the homeostasis of the glycocalyx. A con- PIMs/LM/LAM mutants. ceptually similar glycan compensation is well- known as chitin emergency response in Saccha- The enzymes for the synthesis of AcPIM2, such as romyces cerevisiae (37, 38). In this stress re- PimA and PimB′, are essential for the viability of sponse, compromised cell wall integrity due to Msmeg (8, 9). AcPIM2 is a mature product, but defects in surface glycans such as b-glucan, man- also serves as a precursor for the synthesis of Ac- nan, O-linked glycans, or glycosylphosphatidylin- PIM6, LM, and LAM. Therefore, it remains un- ositol anchors is compensated by the upregulation known if AcPIM2 or any of the downstream prod- ucts are essential. Interestingly, a recent study re- of chitin synthesis. More recently, it was reported vealed that a corynebacterial membrane protein 6 LmeA is required for mycobacterial lipomannan elongation NCgl2760 is involved in LM biosynthesis likely at cin (Fisher Scientific), and 100 µg/mL hygromy- a step prior to the MptA-mediated mannan elonga- cin (Wako). tion (47). The orthologous gene is predicted to be essential in Mtb and cannot be deleted from the DNA purification, whole genome sequencing, and endogenous locus in Msmeg (MSMEG_0317) un- mutation analysis less an extra copy of the gene is present (47). The- Genomic DNA from select suppressor mutants se data imply that defects in the early stage of was purified as previously described (49). Whole- LM/LAM biosynthesis might be lethal to Msmeg. genome sequencing was performed on the Illumi- In contrast, mild structural defects in LM/LAM, na MiSeq platform with 251 bp paired-end se- such as the one caused by the mptA depletion, can quencing. Each genomic DNA (300 ng) was be tolerated in Msmeg. Therefore, the non- sheared to an average size of 600 bp with the Co- essential nature of LmeA further supports our hy- varis S220 (Covaris). The DNA library was pre- pothesis that this protein acts on the downstream pared using the KAPA Library Preparation Kit of MptA-mediated mannan elongation. (Kapa Biosystems) and TruSeq adapters (Illumina) according to manufacturer's instructions. Sequenc- In contrast to Msmeg, the requirement of es were filtered and trimmed based on quality PIMs/LM/LAM appears to be more stringent in score using Quick Read Quality Control (Buffalo Mtb. EmbC, an arabinosyltransferase involved in V. 2011. Quick read quality control. Bioconductor, D o LAM biosynthesis, has been experimentally Seattle. Available at w n shown to be essential for viability in Mtb (3). Ge- www.bioconductor.org/packages/release/bioc/html lo a d nome-wide transposon mutagenesis studies further /qrqc.html. Accessed October 17, 2013.), Sickle ed predict that many genes involved in (Joshi N. 2011. Sickle—A windowed adaptive fro m PIMs/LM/LAM biosynthesis, such as pimA, trimming tool for FASTQ files using quality. h ttp pimB′, pimE , and mptA are essential in Mtb (2). GitHub, San Francisco. Available at ://w We have previously created a mutant strain of Mtb https://github.com/najoshi/sickle. Accessed Octo- w w over-expressing MptC (Rv2181) (27). This mutant ber 17, 2013.) and Scythe (Buffalo V. 2011. .jb c produces aberrant LM/LAM with truncated man- Scythe—A very simple adapter trimmer. GitHub, .o rg nan chain, and such mild modifications were suffi- San Francisco. Available at b/ y cient to make Mtb defective in establishing infec- https://github.com/ucdavis-bioinformatics/scythe. g u e tion in mice. Indeed, the lmeA ortholog (Rv0817c) Accessed October 17, 2013.), were aligned to the st o is predicted to be essential in Mtb (2). The cell reference Msmeg NC_008596.1 genome using n A p envelope localization of LmeA, its predicted es- Bowtie2 (50), and variant calls were made using ril 4 sentiality in Mtb, and the absence of its homologs Samtools (51, 52). The sequence data were ana- , 2 0 in human make this protein a potentially attractive lyzed by the Integrative Genomics Viewer (53). 1 9 drug target. To this end, we are currently generat- Additional suppressor mutants were analyzed spe- ing lmeA mutant in Mtb to demonstrate that lmeA cifically for pimE (A115 & A118), mptA (A193 & is an essential gene. A194), and MSMEG_5785 (A197 & A198) by amplifying the genes by PCR (see Table S1) and EXPERIMENTAL PROCEDURES sequencing the amplified products by standard Sanger sequencing. Mycobacterial growth conditions Msmeg mc2155 and derived mutants were grown Lipid extraction and analysis at 37°C on Middlebrook 7H10 agar (BD) supple- Crude lipids were extracted as described and the mented with 0.2% glucose (w/v) and 15 mM NaCl delipidated pellet was incubated with phenol/water as described (48). Liquid cultures were at 30°C in (1:1) for 2 hours at 55°C to extract LM/LAM (18). Middlebrook 7H9 broth (BD) supplemented with PIMs were separated by high performance thin 0.2% glycerol (v/v), 0.05% Tween 80 (v/v), 0.2% layer chromatography (TLC) silica gel 60 (EMD glucose (w/v) and 15 mM NaCl (48). Knockdown Merck) using chloroform/methanol/13 M ammo- of mptA was induced with 40 ng/mL of atc (Acros) nia/1 M ammonium acetate/water for 48 hours. Antibiotic concentrations used were (180:140:9:9:23) as a mobile phase and visualized 20 µg/mL kanamycin (MP), 50 µg/mL streptomy- by orcinol staining as described (18). LM/LAM 7 LmeA is required for mycobacterial lipomannan elongation samples were separated by SDS-PAGE (15% gel) and visualized using ProQ Emerald 488 glycan Bead-beating cell lysis staining kit (Life Technologies). For the mass Cell pellets were washed in 50 mM HEPES/NaOH spectrometric analysis and lipid binding assay (see (pH 7.4) twice and resuspended in a lysis buffer below), LM/LAM were purified using an octyl- containing 25 mM HEPES/NaOH (pH 7.4), 15% Sepharose column (GE Healthcare) as before (18). glycerol, 2 mM EGTA and a protease inhibitor mix. Four times the pellet weight of acid-washed Construction of plasmids glass beads (Millipore-Sigma) was added, and Knockout and expression vectors were constructed cells were lysed by a BeadBug Microtube Homog- as detailed below. enizer (Benchmark Scientific) at 4°C with beating pMUM57: The upstream and downstream at 4000 rpm for 30 sec. Bead-beating was repeated region of lmeA were amplified using A217/A218 for 5 times with 1 min interval on ice. Beads and and A219/A220 primer sets (Table S1), respec- cell debris were removed by centrifugation. tively. These two fragments were then digested with Van91I and ligated into Van91I-digested SDS-PAGE and western blotting pCOM1 (26), resulting in the MSMEG_5785 Protein samples (12 µl) were mixed with reducing knockout construct. sample loading buffer, denatured on ice for 30 min pMUM54: lmeA was amplified using pri- or boiled for 5 min, and separated on SDS-PAGE D o mers A215 and A216. The fragment was then di- (12% gel). After western blot transfer, the PVDF w n gested by NdeI and ScaI and ligated into membrane was incubated with a primary antibody lo a d pMUM12 (26), which was digested with NdeI and at 1:2000 dilution (mouse anti-HA (Millipore- ed ScaI, to generate pMUM54, an expression vector Sigma), rabbit anti-MptC (18), mouse anti- fro m for P -lmeA-HA. Penta·His (Qiagen), rabbit anti-PimB′ (18), or rab- h hsp60 ttp pMUM107, pMUM125 and pMUM126: bit anti-Mpa (54)), followed by incubation with a ://w Primers A470 and A472 were designed to amplify horseradish peroxidase-conjugated secondary an- w w lmeA including 165 bp of upstream native promot- tibody, either anti-rabbit or anti-mouse IgG (GE .jb c er region from either WT (pMUM107), S1 Healthcare), at a 1:2000 dilution. Bands were vis- .o rg (pMUM125) or S10 (pMUM126). The PCR frag- ualized by chemiluminescence and recorded using b/ y ment was then digested by KpnI and XbaI and li- ImageQuant LAS 4000mini (GE Healthcare). g u e gated into pMV306 that was digested by KpnI and st o XbaI, resulting in pMUM107, pMUM125, and Sucrose density gradient sedimentation and frac- n A p pMUM126, expression vectors for Pnative-lmeA- tionation ril 4 HA, Pnative-lmeA(G170D)-HA, and Pnative- Sucrose density gradient fractionation was per- , 2 0 lmeA(V181G)-HA, respectively. formed as before (25, 26). 1 9 pMUM121: Primers A552 and A553 were designed to amplify lmeA, excluding the sequence Immunoprecipitation coding for the predicted N-terminal signal se- Anti-HA agarose beads (Millipore-Sigma) were quence. A TOPO cloning kit (Invitrogen) was used washed with a buffer containing 25 mM HEPES- to insert the fragment into pET100, creating an NaOH (pH 7.4), 2 mM EGTA, and 150 mM NaCl IPTG inducible His-LmeA expression construct. (HES) or another buffer containing 25 mM HEPES-NaOH (pH 7.4), 2 mM EGTA, 150 mM MALDI-TOF MS analysis NaCl, and 1% Triton-X100 (HEST). Bead beating Octyl-Sepharose-purified LM/LAM preparation cell lysate (36 µl) was added to the pre-washed (0.5 µl) was mixed with an equal volume of matrix beads (10 µl bed volume), and incubated with solution (20 mg/ml sinapinic acid (Millipore- HES or HEST buffer at 4°C overnight under gen- Sigma), 30% acetonitrile (Fisher Scientific), 1% tle rotation at 5 rpm. Beads were washed with 1 ml trifluoroacetic acid (Fisher Scientific) in water). of HES or HEST buffer prior to elution with 10 µl Samples were analyzed on a Bruker Microflex of 1 mg/ml HA peptide (AnaSpec) twice at 30°C. MALDI-TOF instrument (Bruker Daltonics) using Eluates were separated on SDS-PAGE and visual- linear mode and positive ion detection. The data ized via western blot as described above. were analyzed using the Microflex software. 8 LmeA is required for mycobacterial lipomannan elongation E. coli cell transformation and lysis ed in isopropanol by two-fold from an initial E. coli BL21 strain was transformed with amount of 0.4 nmol/well and the solvent was pMUM121 by heat shock. The transformed strain evaporated in 96-well immunoplates (Brand was grown in terrific broth to log phase and in- GmbH). Once the wells were completely dried, duced with 1 mM IPTG (Fisher Scientific) for 3 they were blocked at 4°C overnight with 200 µl of hours at 37°C. Untransformed E. coli was grown 5% milk in phosphate buffer saline with 0.05% under the same condition. Cells were incubated Tween-20 (PBST20). A mixture of 8 µl cell lysate with 1 mg/mL lysozyme (Fisher Scientific) for 20 and 32 µl of PBST20 was added to each well and min at room temperature in a buffer containing 50 incubated at 37°C for 2 hours. After washing with mM HEPES-NaOH (pH 7.4) 200 mM NaCl, and PBST20, 40 µl of mouse anti-Penta·His IgG 1x FastBreak Cell Lysis Reagent (Promega) and (1:2000 dilution, Qiagen) was added and incubat- was lysed by sonication. The lysate was centri- ed at 4°C overnight. After washing with PBST20, fuged and the supernatant was used for all binding 40 µl of horseradish peroxidase-conjugated anti- assays or western blot analysis. mouse IgG (1:4000 dilution, GE Healthcare) was added and incubated at room temperature for 1 Lipid binding assay hour. After washing with PBST20, 100 µl of L-α-phosphatidylinositol (PI) from soybean, L-α- 3,3’,5,5’-tetramethylbenzidine (BD) was added phosphatidylethanolamine (PE) from egg yolk, and the colorimetric changes were read at 650 nm D o and glycerol trioleate (TAG) were purchased from after 1 hour incubation at room temperature. Kd w n Millipore-Sigma. L-α-phosphatidic acid from values were calculated using a nonlinear regres- loa d chicken egg was purchased from Avanti. Geranyl- sion function of Prism 7 (GraphPad Software), ed geranyl monophosphate was purchased from La- assuming that LmeA has one lipid binding site. fro m rodan. PI was a crude preparation with many con- For competition assays, microtiter plates were h ttp taminants, and was further purified by preparative coated with 0.05 nmol of PE (corresponding to ://w TLC. Purity of each lipid was examined by TLC 1.25 µM). E. coli lysate was pre-incubated with 10 w w using two solvent systems: hexane / diethyl ether / mM mannose 1-phosphate (Millipore-Sigma) or .jb c formic acid (40:10:1) for TAG or chloroform / GDP-Mannose (Millipore-Sigma) at room temper- .o rg methanol / 13 M ammonia / 1 M ammonium ace- ature for 30 min before being added to the PE- b/ y tate / water (180:140:9:9:23) for PI, PE, PA and coated well. g u e GGM. For lipid ELISA, lipids were serially dilut- st o n A p Acknowledgements: This work was supported by a Biomedical Research Grant (RG-414805) from the ril 4 American Lung Association and a Research Grant from the Pittsfield Anti-Tuberculosis Association to , 2 0 YSM. APTO was a summer research student supported by the UMass Amherst PREP Program. We thank 1 9 Dr. Stephen Eyles (Institute for Applied Life Sciences, University of Massachusetts Amherst) for help with mass spectrometry and Dr. Heran Darwin (New York University) for the gift of anti-Mpa antibody. We also thank Julia Puffal, William Eagen, and Sarah Osman for discussion and critical reading of the manuscript. Conflicts of interest: The authors declared no conflicts of interest with the content of this article. Author contributions: KCR conducted most of the experiments, analyzed the results, and wrote the pa- per. SAH conducted a part of the experiments shown in Figs. 5 and 6. DM, JAM and SN sequenced and analyzed the whole genome of the suppressor mutants (Fig. 2 and Table 1). LRB and APTO conducted some initial experiments shown in Fig. 1. JNS contributed to the experiments shown in Fig. 7. YSM con- ceived the idea, designed the study and wrote the paper. REFERENCES 1. Kaur, D., Guerin, M. E., Škovierová, H., Brennan, P. J., and Jackson, M. (2009) Chapter 2: Bio- genesis of the cell wall and other glycoconjugates of Mycobacterium tuberculosis. Adv. Appl. Mi- crobiol. 69, 23–78 9 LmeA is required for mycobacterial lipomannan elongation 2. Griffin, J. E., Gawronski, J. D., DeJesus, M. A., Ioerger, T. R., Akerley, B. J., and Sassetti, C. M. (2011) High-resolution phenotypic profiling defines genes essential for mycobacterial growth and cholesterol catabolism. PLoS Pathog. 7, e1002251 3. Goude, R., Amin, A. G., Chatterjee, D., and Parish, T. (2008) The critical role of embC in Myco- bacterium tuberculosis. J. Bacteriol. 190, 4335–4341 4. Jarlier, V., and Nikaido, H. (1990) Permeability barrier to hydrophilic solutes in Mycobacterium chelonei. J. Bacteriol. 172, 1418–1423 5. Hmama, Z., Peña-Díaz, S., Joseph, S., and Av-Gay, Y. (2015) Immunoevasion and immunosup- pression of the macrophage by Mycobacterium tuberculosis. Immunol. Rev. 264, 220–232 6. Vergne, I., Gilleron, M., and Nigou, J. (2014) Manipulation of the endocytic pathway and phago- cyte functions by Mycobacterium tuberculosis lipoarabinomannan. Front Cell Infect Microbiol. 4, 187 7. Ishikawa, E., Mori, D., and Yamasaki, S. (2017) Recognition of mycobacterial lipids by immune receptors. Trends Immunol. 38, 66–76 8. Korduláková, J., Gilleron, M., Mikušová, K., Puzo, G., Brennan, P. J., Gicquel, B., and Jackson, M. (2002) Definition of the first mannosylation step in phosphatidylinositol mannoside synthesis. PimA is essential for growth of mycobacteria. J. Biol. Chem. 277, 31335–31344 9. Guerin, M. E., Kaur, D., Somashekar, B. S., Gibbs, S., Gest, P., Chatterjee, D., Brennan, P. J., and D o Jackson, M. (2009) New insights into the early steps of phosphatidylinositol mannoside biosynthe- w n sis in mycobacteria: PimB' is an essential enzyme of Mycobacterium smegmatis. J. Biol. Chem. lo a d 284, 25687–25696 ed 10. Lea-Smith, D. J., Martin, K. L., Pyke, J. S., Tull, D., McConville, M. J., Coppel, R. L., and Crel- fro m lin, P. K. (2008) Analysis of a new mannosyltransferase required for the synthesis of phosphatidyl- h ttp inositol mannosides and lipoarbinomannan reveals two lipomannan pools in Corynebacterineae. J. ://w Biol. Chem. 283, 6773–6782 w w 11. Korduláková, J., Gilleron, M., Puzo, G., Brennan, P. J., Gicquel, B., Mikušová, K., and Jackson, .jb c M. (2003) Identification of the required acyltransferase step in the biosynthesis of the phosphati- .o rg dylinositol mannosides of Mycobacterium species. J. Biol. Chem. 278, 36285–36295 b/ y 12. Mishra, A. K., Alderwick, L. J., Rittmann, D., Wang, C., Bhatt, A., Jacobs, W. R., Takayama, K., g u e Eggeling, L., and Besra, G. S. (2008) Identification of a novel alpha(1-->6) mannopyranosyltrans- st o ferase MptB from Corynebacterium glutamicum by deletion of a conserved gene, NCgl1505, af- n A p fords a lipomannan- and lipoarabinomannan-deficient mutant. Mol. Microbiol. 68, 1595–1613 ril 4 13. Morita, Y. S., Sena, C. B. C., Waller, R. F., Kurokawa, K., Sernee, M. F., Nakatani, F., Haites, R. , 2 0 E., Billman-Jacobe, H., McConville, M. J., Maeda, Y., and Kinoshita, T. (2006) PimE is a poly- 1 9 prenol-phosphate-mannose-dependent mannosyltransferase that transfers the fifth mannose of phosphatidylinositol mannoside in mycobacteria. J. Biol. Chem. 281, 25143–25155 14. Kaur, D., McNeil, M. R., Khoo, K.-H., Chatterjee, D., Crick, D. C., Jackson, M., and Brennan, P. J. (2007) New insights into the biosynthesis of mycobacterial lipomannan arising from deletion of a conserved gene. J. Biol. Chem. 282, 27133–27140 15. Mishra, A. K., Alderwick, L. J., Rittmann, D., Tatituri, R. V. V., Nigou, J., Gilleron, M., Eggeling, L., and Besra, G. S. (2007) Identification of an alpha(1-->6) mannopyranosyltransferase (MptA), involved in Corynebacterium glutamicum lipomanann biosynthesis, and identification of its orthologue in Mycobacterium tuberculosis. Mol. Microbiol. 65, 1503–1517 16. Kaur, D., Berg, S., Dinadayala, P., Gicquel, B., Chatterjee, D., McNeil, M. R., Vissa, V. D., Crick, D. C., Jackson, M., and Brennan, P. J. (2006) Biosynthesis of mycobacterial lipoarabinomannan: role of a branching mannosyltransferase. Proc Natl Acad Sci USA. 103, 13664–13669 17. Kaur, D., Obregón-Henao, A., Pham, H., Chatterjee, D., Brennan, P. J., and Jackson, M. (2008) Lipoarabinomannan of Mycobacterium: mannose capping by a multifunctional terminal mannosyl- transferase. Proc Natl Acad Sci USA. 105, 17973–17977 18. Sena, C. B. C., Fukuda, T., Miyanagi, K., Matsumoto, S., Kobayashi, K., Murakami, Y., Maeda, Y., Kinoshita, T., and Morita, Y. S. (2010) Controlled expression of branch-forming mannosyl- 10
Description: