Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (17)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Tuffal, G.
Right arrow Articles by Puzo, G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tuffal, G.
Right arrow Articles by Puzo, G.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Glycobiology Pages 675-684  

Newly found 2-N-acetyl-2,6-dideoxy-[beta]-glucopyranose containing methyl glucose polysaccharides in M.bovis BCG: revised structure of the mycobacterial methyl glucose lipopolysaccharides
Introduction
Results
Discussion
Materials and methods
Acknowledgments
Abbreviations
References

Newly found 2-N-acetyl-2,6-dideoxy-[beta]-glucopyranose containing methyl glucose polysaccharides in M.bovis BCG: revised structure of the mycobacterial methyl glucose lipopolysaccharides

Newly found 2-N-acetyl-2,6-dideoxy-[beta]-glucopyranose containing methyl glucose polysaccharides in M.bovis BCG: revised structure of the mycobacterial methyl glucose lipopolysaccharides

Gilles Tuffal, Renaud Albigot, Michel Rivière1, Germain Puzo

Institut de Pharmacologie et de Biologie Structurale du Centre National de la Recherche Scientifique, 205 Route de Narbonne, 31077 Toulouse Cedex, France

Received on September 30, 1997; revised on December 19, 1997; accepted on December 30, 1997

The specific mycobacterial methyl polysaccharides 3-O-methyl mannose polysaccharide (MMP) and the 6-O-methyl glucose lipopolysaccharides (MGLPs) were shown to modulate the fatty acid biosynthesis by the mycobacterial fatty acid synthetase I (FAS I). This activity is attributed to their fatty acid complexing properties allowing the release of the neosynthesized fatty acyl chain from the enzyme and probably their transport in the cell. To elucidate, at a molecular level, the mechanism of this unusual kind of polysaccharide-lipid biological interaction, we first analyzed, by mass spectrometry and proton nuclear magnetic resonance (1H NMR) spectroscopy, the structure of the polysaccharidic backbone (MGPs) of the MGLPs from Mycobacterium bovis BCG. This work reveals that this strain produces a new kind of MGP containing an unusual monosaccharide never described in the mycobacterial genus: a 2-N-acetyl-2,6-dideoxy-[beta]-glucopyranosyl. In addition, 1H NMR data afforded evidence for the revision of three glycosidic linkages described previously. These modifications affect mainly the reducing end tetrasaccharide and have great consequences on the previously proposed molecular model of the MGP.

Key words: mycobacterium/methyl glucose lipopolysaccharides/2-N-acetyl-[beta]-quinovosamine/NMR

Introduction

Among infectious diseases, tuberculosis remains the leading cause of mortality with an estimated 2.9 million deaths and 8 million new cases each year (Murray et al., 1993). Moreover, during the past decade, its incidence has dramatically increased in the developed countries, renewing interest in this almost-forgotten disease (Bloom and Murray, 1992). This reversal of the tuberculosis situation is mainly attributed to the appearance of multi-drug-resistant strains of M.tuberculosis as a consequence of carelessness infection control policies (Weiss, 1992; Nowak, 1995). Thus, the challenge today for combating the outbreak of tuberculosis relies on the development of new antituberculous agents. The efforts made in this way are mainly addressed toward the search for biosynthesis inhibitors of the major components of the mycobacterial cell wall since this structure is thought to act as a protective barrier (Jarlier and Nikaido, 1994; Nikaido, 1994). The specific peptidoglycan-arabinogalactan-mycoloyl complex, which is the covalently linked backbone of the mycobacterial cell wall, represents a potential target for the development of new drugs. However, its biosynthesis is still poorly understood and real efforts are necessary in this field. Recently, though, two major advances were achieved with (1) the development of a d-arabinosyl transferase assay (Lee et al., 1995) allowing the study of biosynthesis of the d-arabinan portion of the arabinogalactan domain of the cell wall, and (2) the finding of the molecular basis of the Isoniazid (INH) (most widely used antituberculous drug) resistance (Dessen et al., 1995). This latest work shows that INH target was a 2-trans-enoyl reductase (encoded by the mycobacterial inhA gene) involved in the mycolic acid biosynthesis and more precisely in the elongation of the mero mycolic chain from a fatty acyl precursor: the enoyl-ACP. Further kinetic study indicates that the optimum chain length precursor is the hexadecenoyl-CoA (C16-CoA) derivative (Quemard et al., 1995). The origin of those precursors is still not known, although one can expect that the fatty acyl chain is synthesized in the cytoplasm by the mycobacterial FAS I complex described by Bloch (Vance et al., 1973; Bloch, 1975). The activity of this multienzymatic complex (Odriozola et al., 1977) as well as the length of the released neosynthesized fatty acyl CoA derivative (Peterson and Bloch, 1977) were shown to be modulated by specific mycobacterial methyl polysaccharides (Ilton et al., 1971; Gray and Ballou, 1975; Banis et al., 1977). These polysaccharides-the 3-O-methyl mannose polysaccharide (MMP) (Gray and Ballou, 1971) and the 6-O-methyl glucose lipopolysaccharides (MGLP) (Lee and Ballou, 1964)-raise the rate of overall synthesis and shift the neosynthesized chain length pattern from the long to the short one, increasing the latter's proportion (mainly hexadecanoate) from 25% to 85% (Flick and Bloch, 1974). These effects are thought to result from their ability to complex acyl CoA (the higher affinity being for the C16-CoA; Machida and Bloch, 1973), thus facilitating the release of the neosynthesized chain from the enzyme. Moreover, due to their complexing properties, these methyl polysaccharides were also proposed to act as lipid carrier in the cell (Bloch, 1977). These properties suggest that these molecules could play a key role in the synthesis of the fatty acyl precursor of the mycobacterial cell wall mycolic acids (Goren, 1972). Thus, in order to confirm this role and to further understand the molecular basis of this unique specific recognition between fatty acid and polysaccharide, we undertook a reinvestigation of the 6-O-methyl glucose lipopolysaccharides (MGLP). The structure of the MGLP from Mycobacterium smegmatis has been the focus of extensive work from Ballou's group (Saier and Ballou, 1968a-c; Forsberg et al., 1982; Dell and Ballou, 1983). However, recently, our reinvestigation of the MGLP from a slow growing pathogenic mycobacteria, i.e., M.xenopi, led us to show that the polysaccharidic core (MGP) presents a great heterogeneity of length, methylation degree, and side chain glucosyl substitution (Tuffal et al., 1995a,b). Since the functional properties of these molecules may be affected by such heterogeneity, we analyzed the MGLPs from the vaccinating M.bovis BCG (strain Pasteur), to confirm that this finding is not specific of M.xenopi but a general feature from mycobacterial MGLPs.

Here we report the purification and the structural elucidation of MGPs from M.bovis BCG by mass spectrometry and 2D 1H NMR spectroscopy. We identify a new molecule containing a 2-N-acetyl-2,6-dideoxy-[beta]-Glcp described for the first time in the mycobacterial genus, and we propose to revise the general structure of the mycobacterial MGLPs polysaccharide backbone.

Results

MGPs characterization

The MGLPs from M.bovis BCG were purified as previously described by affinity chromatography on an octadecyl silica bound gel and deacylated by mild alkali treatment to lead to the MGPs, analyzed by HPAEC. The chromatogram obtained (Figure 1) shows a complex profile similar to the one reported for M.xenopi. The quantitatively major peaks (A, B, C, D, and E) were tentatively attributed to MGP15,10, MGP19,12, MGP20,13, MGP20,12, and MGP20,11, on the basis of their retention time and by coinjection with already identified MGP purified from M.xenopi. To further confirm the attribution, these peaks were purified by preparative HPAE chromatography and analyzed by LSIMS (results are presented in Table 1). The mass spectrum of the major compound D shows in its high mass range, two peaks at m/z = 3537 and 3559 attributed to sodium cationized pseudomolecular ions of acidic (RCOOH) or sodium saline (RCOONa) forms in agreement what expected for the MGP20,12 (Tuffal et al., 1995a). Beside these two major ions, abundant fragment ions arising from the glycosidic bound cleavage are also observed allowing a partial sequencing of the polysaccharide backbone (Table 1). The highest mass fragment, noted C17, is found at m/z = 2963 and arises from the loss of the aglycon linked trisaccharide (574 amu). The next one, C16, is shifted by 162 amu at m/z = 2801 in agreement with the loss of an anhydro Glcp. The difference of 180 amu between this latter and the following one, namely C[prime]15 at m/z = 2621, is attributed to a double cleavage process induced by the presence of side chain Glcp residue, thus characterizing a ramification. From this latest ion, the spectrum shows a series of ions (from C14 to C6) distant by 176 amu corresponding to the successive loss of anhydro mono-O-Me-Glcp units. Thus, taken together, HPAEC and LSIMS data agree with the proposed structure of the MGP20,12. In the same way, the compounds B, C, and E were unambiguously identified as the MGP19,12, MGP20,13, and MGP20,11, respectively (Table 1).


Figure 1. HPAEC analysis of the purified M.bovis BCG MGPs.Eluent A was 0.1 M NaOH and eluent B was 0.5 M CH3COONa in 0.1 M NaOH. Elution was performed at 1 ml/min and started with a step of 38% of eluent B for 2 min followed by a concave gradient from 38% to 48% of eluent B in 60 min. Shaded peaks were collected.

As for others, peak A was tentatively attributed on basis of its retention time, to a short chain MGP (MGP15,10 or MGP16,11) similar to those previously described in M.xenopi (Tuffal et al., 1995b). However, coinjection with these latter in more resolutive chromatographic conditions reveals a lower retention time (Figure 2a), suggesting that it must correspond to a shorter MGP. Although its mass spectrum shows two high mass ion at m/z = 3562 and 3584 attributed to the pseudo molecular ions (M+Na)+ and (M-H+2Na)+, respectively, revealing an unexpectedly high molecular mass of 3539 amu (Figure 2b). More interesting, the uneven molecular weight strongly suggests the presence of one amino sugar. Moreover, the excess of 25 amu compared to the mass of the MGP20,12 appears to be restricted to the pseudomolecular ions since the spectrum exhibits the same sequence ions than those observed for the MGP20,12. This observation indicates that the differences are localized on the glyceric acid linked trisaccharide.

Table I. . M.bovis BCG MGPs analysis by HPAEC and LSIMS
Peak A B C D E
% amount Absolute 5 18 6 37 16
  Relativea 14 49 16 100 43
Molecular ions
  {RCOOH+Na}+ 3562 3375 3551 3537 3523
  {RCOONa+Na}+ 3584 3397 3573 3559 3545
Fragment ions
  C17 2963 2949
  C16 2801 2787
  C[prime]15 2621 2607
  [le]C14   C14: 2463, C13: 2287, C12: 2111, C11: 1935, C10: 1759,
    C9: 1583, C8: 1407, C7: 1231, C6: 1055.		  
MGP   Unknown 19,12 20,13 20,12 20,11
aPercentages were calculated relative to the MGP20,12.


Figure 2. HPAEC and positive LSIMS analysis of the MGP A. Inset a: HPAEC chromatogram of the coinjected MGP A and purified M.xenopi MGP15,10 and MGP16,11 (elution was performed with 38% of eluent B in A); b, positive LSIMS spectrum of purified MGP A (masses are reported in Table I).

Thus, in order to gain information on its structure, this MGP was analyzed by 1H NMR spectroscopy. However, due to the small amount of purified compound, some signals, particularly those of the glyceric acid, could not be detected and the complete structural elucidation of this new MGP has first required a complete 1H NMR analysis of the quantitatively major MGP20,12.

1H NMR study of the MGP20,12

The 1D 1H NMR spectrum (Figure 3) recorded in D2O is similar to the one reported by Forsberg et al. for the MGPs from M.smegmatis (Forsberg et al., 1982). The anomeric proton resonance zone is dominated by a broad intense signal centered at 5.491 ppm surrounded by four well resolved signals at 5.552 (I), 5.089(IX), 5.00 (X), and 4.97 (XI, XII) ppm. Their integration relative to the signal X taken as the unity is consistent with the presence of 20 monosaccharides as expected from the MS data. The three resonances noted I, IX, and X integrating each for a single proton appear as doublet with a coupling constant of 3.5 Hz typifying three [alpha]Glcp residues. Based on their chemical shifts, they were previously tentatively assigned to the anomeric protons of Glcp [alpha](1->3), Glcp [alpha](1->2)Glycerate, and [alpha](1->6) residues, respectively (Forsberg et al., 1982). The highest field doublet of doublet at 4.97 ppm integrating for two protons (XI, XII) with a large coupling constant (3JH1-H2= 9 Hz) is in agreement with the presence of two [beta] Glcp. From 2D HOHAHA experiments (Figure 3), seven supplementary spin systems are easily defined (from II to VIII). Moreover, analysis of their relative intensity and their cross peaks pattern suggests that six of them (II, IV, V, VI, VII, and VIII) can be attributed to a single monosaccharide residue. Thus, by difference, the resonance III was assigned to the anomeric proton signal of the remaining nine residues which can be reasonably assigned to 6-O-Me-Glcps.


Figure 3. 500 MHz partial 1H NMR 30 ms mixing time HOHAHA spectrum of the MGP20,12 in D2O. (f2: [delta] 3.31 - 4.35 / f1: [delta] 4.92 - 5.6) showing connectivities involving anomeric protons. Cross peaks are labeled with a roman numeral identifying the residue followed by the number assigning the proton atom. Shaded background spots are observed in the 10ms mixing time HOHAHA spectrum and corresponds to the H-1-H-2 correlations.

Partial attribution of those spin systems was achieved by performing several 2D HOHAHA experiments with increasing mixing time in order to assign sequentially the different resonances. In Figure 3 are reported the anomeric proton correlations (shaded background) observed with a 10 ms mixing time experiment. Under these conditions a first step magnetization transfer occurs almost exclusively between the H-1s and the H-2s, allowing the localization of these latter. Increasing the mixing time from 10 to 30 ms allows a second magnetization transfer to occur and thus the localization of the H-3s (Figure 3). Among the 11 spin systems defined, two (XI and XII) are very characteristic and can readily be attributed to the two [beta]-d-Glcps. Moreover, six of them (I, II, III, VI, VII, and VIII) are quite similar and present homogeneous proton chemical shifts with their anomeric protons resonating in a quite narrow zone between 5.449 ppm (VIII1) and 5.552 ppm (I1) as well as their respective H2 (3.578 < [delta]H2 < 3.727 ppm, [Delta][delta] = 0.149 ppm) and H3 (4.0 < [delta]H3 < 4.078 ppm, [Delta][delta] = 0.078 ppm). These latter resonate at a quite low field in agreement with the [beta] deshielding effect of glycosylation on C4. This indicates that all those residues must be (1->4) linked as described previously. It is unlikely that the H3 chemical shifts of the four remaining spin systems differ greatly from the preceding and allow their individual assignment. The first of them, the relatively shielded H3 signal of unit IX (3.87 ppm), signals the absence of glycosylation on the C4 and thus can be assigned to the 6-linked reducing end [alpha]-Glcp. In the same way, the characteristic chemical shift of the V3 signal at 3.525 ppm (resonating at higher field than its respective H2 ) allows its attribution to the nonreducing end 3-O-Me-[alpha]-d-Glcp. Finally, and more interesting, are the patterns of systems IV and X in which H2, H3, and H4 are strongly downfield shifted in agreement with glycosylation on both C3 and C4. Thus, these systems were assigned to the two expected residues bearing the side chain Glcp units.

With longer mixing time (100 or 175 ms) the spectrum becomes much more complex, although some additional correlation can be assigned unambiguously (Table 2). Moreover, from these latest spectra, the proton resonances of the aglycon moiety (i.e., glyceric acid) were localized, respectively, at 4.262 ppm for the H-2 and 3.952, 3.92 for the H-3, H-3[prime] (Hunter et al., 1979).

Table II. MGP20,12 1H NMR chemical shifts in D20 at 313 K
Residue H-1 H-2 H-3 H-4 H-5 H-6
I 5.552 3.578 4.029 3.645 3.871  
II 5.508 3.714 4.023      
III 5.491 3.688 4.00      
IV 5.477 3.889 4.28 3.832    
V 5.469 3.727 3.535      
VI 5.458 3.608 4.078      
VII 5.452 3.708 4.02      
VIII 5.449 3.716 4.012      
IX 5.089 3.613 3.87 3.588 4.07 4.03/3.83
X 5.00 3.86 4.30 3.83 3.961 3.512
XI 4.971 3.361 3.575 3.48 3.838 4.021
XII 4.969          
Aglycon   4.262 3.95/3.92    

To establish the linkage and the ramification sites, the MGP20,12 was analyzed by 1H NMR dipolar correlation spectroscopy (ROESY experiment). From Figure 4, the linkage between the residue IX and the aglycon moiety is unambiguously deduced from the intense cross peak at [delta](f1,f2)= (5.089, 4.262) between the H-2 of the glyceric acid and the unit IX anomeric proton. Moreover, glycosylation of this residue IX on its C-6 is confirmed by the magnetization transfer observed between its exocylic protons H-6 and H-6[prime] ([delta] = 4.03, 3.83 ppm) and the anomeric proton of the unit X. From MS data, this latter residue X is expected to be diglycosylated. This is supported by a first correlation plot noted {XI1-X3} found at [delta](f1,f2) = (4.971, 4.30), indicating that the [beta]-Glcp XI is linked to the C-3 of unit X. Beside this first branching, careful inspection of the ROESY spectrum reveals a second cross peak {VI1-X4} indicating that this residue X is also substituted on its C4 by the [alpha]-Glcp VI. From these observations, we can propose the following partial structure for the reducing end tetrasaccharide which differs from the previous one either by the anomeric configuration and the linkage sites of the two substituent of the residue X:


Figure 4. 500 MHz partial 1H NMR ROESY spectrum of the MGP20,12 in D2O (f2: [delta] 3.31 - 4.35 / f1: [delta] 4.92 - 5.6). Shaded background cross peaks are discussed in the text.

->GlcpVI[alpha](1->4){ GlcpXI[beta](1->3)}GlcpX[alpha] (1->6)GlcpIX[alpha](1->2)Glyceric acid

In a same way, proton chemical shift analysis of the different residues (Table 2) reveals a strong deshielding of proton H-2 and H-3 of unit IV, suggesting that this residue must be the second disubstituted monosaccharide. This is confirmed by the presence of the two correlation spots noted {XII1-IV3} and {I1-IV4}, indicating that the second [beta]-Glcp is linked to the C-3 while unit I substitutes the C-4 of residue IV. Finally, the weak correlation at [delta]f1 = 5.477 (IV1), [delta]f2 = 4.078 (VI3) suggests that unit IV glycosylates residue VI. However, the unambiguous determination of this latest linkage position was hindered by the complexity of the spectrum resulting from a strong overlapping. However, similar 1H NMR investigation performed in DMSO-d6 instead of D2O (unpublished observations) afforded evidence for an [alpha](1->4) linkage between those two residues and confirm the above data. Thus, from this study we can propose the following new structure for the M.bovis BCG MGP20,12 (Figure 5).


Figure 5. Proposed structure of the MGP20,12. a, Glcp in the MGP20,11; b, 6-O-Me-Glcp in the MGP20,13; c, absent from the MGP19,12. Brackets indicate tentative attribution.

Structure of the MGP A

Comparative analysis of 1H 1D NMR spectrum of the MGP A in D2O (Figure 6a,b) shows that it is quite different than the MGP20,12 one (Figure 3). Integration of the anomeric proton region reveals the presence of 20 monosaccharidic units as for the MGP20,12. The two [beta] anomeric proton resonances are not superimposed anymore since one of them has been slightly shifted at lower field close to the [alpha] anomeric proton signal. Moreover, the additional [alpha] anomeric proton signal at [delta]= 5.42 ppm seems to have been upfield shifted from the intense signal III. However, more interesting are the signals present at high field in the aliphatic proton resonance zone between 1.5 and 2.3 ppm (Figure 6c). Besides the signal ([delta] = 2.09 ppm) of the contaminating sodium acetate resulting from the HPAE chromatography, two additional resonances are observed at [delta] = 2.23 and 1.57 ppm integrating each for three protons. The latter one appears as a doublet with a coupling constant of 5.3 Hz signaling the presence of a 6-deoxy hexose residue. The second signal at 2.23 ppm, resonating as a singlet close to the sodium acetate one, has a chemical shift characteristic of methyl acetoxy protons and is in agreement with the presence of an acetamido substituent. It is noteworthy that the simultaneous presence of those two chemical groups is in good agreement with the MS results since the C-6 deoxy methyl results in a difference of minus 16 amu while substitution of a hydroxyl by an acetamido function is responsible of an increase of 41 amu ([Delta] = 25 amu). Thus, in order to confirm the presence of an acetamido function, the spectrum was recorded in H2O (80% in D2O) at pH 4.8 (Figure 6d). Under these conditions we could observe at 8.23 ppm a doublet (3J = 7.5 Hz) typifying an amide proton and thus proving the presence of a HexNAc. Further 2D 1H NMR experiments were performed in DMSO-d6 (Figure 7) since this solvent greatly enhanced resolution of the four more shielded anomeric proton resonances allowing the discrimination between the two [beta] anomers. The HOHAHA spectrum (Figure 7b) ([tau]m = 100 ms) reveals that the amide proton and the 6-deoxy-methyl group signals belong to the same spin system characterized by the [beta] anomeric proton resonance at [delta]DMSO = 4.92 ppm. Thus, starting from this H-1 signal, the sequential attribution of the different proton resonances of this residue was achieved by performing various HOHAHA experiments using increasing mixing time: [tau]m = 10, 30, 65, and 100ms. With [tau]m = 10 ms, a first cross peak appears at [delta]DMSO = 3.41 ppm, attributed to the correlation with the H-2. Increasing the mixing time to 30 ms, two additional correlations involving the H-1 are observed. The first one at [delta]DMSO = 3.24 ppm was thus assigned to the H-3, while the second at [delta]DMSO = 7.55 ppm with the amide proton resonance strongly suggests that this group is born by the C-2. In the same way, with a mixing time [tau]m = 65 ms the H-4 resonance was localized at [delta]DMSO = 2.92 ppm and was shown to resonate as a well resolved triplet (3J3-4 = 3J4-5 = 8.5Hz) in the 1D spectrum, thus indicating a trans di-axial relative configuration of the H-3, H-4, and H-5 protons. However, with the longer mixing time, [tau]m = 100 ms, the only supplementary cross peak observed at the frequency of the anomeric proton signal is the correlation with the C-6 methyl group while the H-5 resonance could not be localized. This latter was thus identified on the COSY spectrum (not shown), which shows a strong correlation at [delta]DMSO = (3.24, 1.47) between the H-6 and a signal attributed to the H-5 suggesting that this latter is superimposed with the H-3. Moreover, analysis of the different proton signal coupling constants determined either on the 1D or 2D spectra allows the attribution of this spin system to a monosaccharide having a gluco hexo pyranose configuration and more precisely to a 2-N-acetyl-2,6-dideoxy-[beta]-Glcp (QuiNAc) residue. To further localize this new monosaccharide, MGP A was thus analyzed by 1H NMR ROESY experiments as above in D2O since no dipolar magnetization transfer (NOE) could be observed in DMSO-d6. Under these conditions, the QuiNAc anomeric proton resonance (attributed on basis of subsequent HOHAHA experiments) was shown to exhibit a strong correlation plot with the H-3 of an [alpha]-Glcp unit identified as the reducing end closer di branched Glcp residue (noted X in the MGP20,12). Thus, these data, in agreement with the MS results presented above, clearly demonstrate that the MGP A differs from MGP20,12 by the presence of a branched [beta](1->3) linked 2-N-acetyl-2,6-dideoxy-[beta]-Glcp in place of Glcp residue.


Figure 6. 500 MHz 1H NMR spectrum of the MGP A. (a) Complete spectrum in D2O; (b) and (c) enlargement of the anomeric and aliphatic proton resonance zone; (d) spectrum obtained in H2O (20% D2O, pH 4.5) revealing the amide proton.

Figure 7. 500 MHz partial 1H NMR 100 ms mixing time HOHAHA spectrum of the MGP A in DMSO-d6 (f2: [delta] 4.62 - 5.21 / f1: [delta] 1.38 - 7.62). a, 1D spectrum; b, 2D spectrum.


Discussion

This work has extended our knowledge on the mycobacterial MGLPs in two important ways. First, the observation of a new type of methyl glucose polysaccharidic core containing an unusual carbohydrate residue: a 2-N-acetyl-2,6-dideoxy-[beta]-Glcp. This was first detected by anion exchange chromatography and mass spectrometry analysis. HPAEC shows that M.bovis BCG produce a complex mixture of MGP similar to the one previously reported for M.xenopi (excepted the lack of longer chain polysaccharides MGP21,12 and MGP21,11 observed in the two studied strains of M.xenopi; Tuffal et al., 1995a,b). Moreover, in agreement with all the mycobacterial strains analyzed, the two most abundant component of the mixture are the polysaccharides containing 20 and 19 residues (MGP20,11, MGP20,12, MGP20,13, and MGP19,11, MGP19,12). They account for only 59% and 18% ([Sigma] = 77%) of the MGP mixture, respectively, while in M.smegmatis they represent 95% (83% and 12%, respectively). Besides those, M.bovis BCG also produce some quantitatively minor MGPs with lower retention times found only in the slow growing mycobacterial strain. In the case of M.kansasii or the two studied strains of M.xenopi, these were shown to correspond to shorter MGP containing 15 or 16 Glcp residues. It is unlikely that the low retention time of MGP A results only from the presence of the 2-N-acetyl-2,6-dideoxy-[beta]-Glcp instead of a [beta]-Glcp. It is noteworthy that the change of this single monosaccharide among 20 affects the chromatographic behavior of the polysaccharide so much. However, this result is consistent with the previously observed effect of the presence of hexosamines, which are much less retained than hexoses in these chromatographic conditions.

Finally, the complete structural elucidation of MGP A has been achieved thanks to the preliminary 1H NMR analysis of the MGP20,12. Taken together, mass spectrometry and NMR spectroscopy data indicate that the MGP A differs from the quantitatively major MGP20,12 by the presence of a branched 2 NAc quinovosamine in place of Glcp residue on the reducing end trisaccharide. The role or the effect of this sugar still remains undetermined. However, it has been proposed previously that the Glcp residue found in this position can be succinylated on its primary hydroxyl group (OH-6) in the native MGLP. Thus, the presence of 6-deoxy hexose prevents such modifications and thus may have a role in the interaction between this compound and the fatty acyl ligand.

The second major finding of this work is the revision of the general structure of the polysaccharidic backbone of the mycobacterial MGLPs. 1H NMR dipolar coupling (ROESY) suggested that the reducing end closer branched Glcp residue is bound by a [beta](1->3) linkage instead of [beta](1->4) as described previously. Unfortunately, it has been shown that this kind of experiment could lead to incorrect linkage assignment since the dipolar exchange depends on proton proximity rather than on bond connectivity. Indeed, in some cases, the relative conformation of the sugar rings may results in bringing the anomeric proton closer to a proton other than the one involved in the glycosidic linkage. Such a situation may thus result in the presence of cross peaks in the spectrum, which can leads to false conclusion. Moreover, since our proposal differs from the previous structure only by a positional isomery, this could not be confirmed by more primary linkage analysis methods such as methylation analysis since the two structures will give the same results.

Nevertheless, some other evidences allow us to propose this different structural model. First of all are the chemical shifts of the two [beta]-Glcp ring protons (strictly superimposed) which are in agreement with the absence of any substitution. This first observation is inconsistent with the previous model and strongly suggests that those two residues (XI and XII) are located at nonreducing ends of the molecule and must correspond to the two ramifications. Consequently, the anomery of Glcp VI (substituting the residue X) must be [alpha] , as confirmed by the pattern of its H1-H2 cross peak observed in the HOHAHA spectrum. The second convincing point regarding the linkage position of the main chain on residue X arises from the comparative structural study of the MGP19,12. We previously determined, on the basis of FAB/MS and methylation analysis data, that this compound differs from the quantitatively major MGP20,12 by the absence of residue XI (Tuffal et al., 1995a). Moreover, the absence of the expected 2,4,6 tri-O-Me, 3-O-Ac-glucitol strongly suggested a different linkage type. This was latter confirmed by 1H NMR analysis (unpublished observations), which shows an upfield shift of the H3 proton signal of residue X confirming the absence of substitution at this position and consequently that residue VI is linked to the Xth by a 1->4 linkage.

Since we can reasonably imagine that the MGP20,12 arises from the same metabolic pathway and derives from the MGP19,12 by action of [beta]3glucosyltransferase on residue X, all these arguments tend to confirm the correctness of the ROESY spectrum interpretation and lead us to propose the following new revised structure of the mycobacterial MGP (Figure 5).

The consequences of these structural revisions are quite important for the proposed model of interaction between these molecules and their putative ligand (acyl-CoA derivatives) since they indicate that the main polysaccharidic chain is almost exclusively constituted by [alpha](1->4) linked Glcp residues with the exception of the reducing end terminal monosaccharide. Indeed, by analogy with the amylose V type structure, it has been previously proposed that MGLP and their deacylated counterparts MGP could adopt a helical conformation generated as a consequence of the stereochemical constraints due to the [alpha](1->4) linkages (Bloch, 1977). Such conformation therefore results in the formation of a cylindrical hydrophobic cavity in which the fatty acid derivative is included. However, molecular modeling reveals that the presence of a [beta]1->3 linkage in the main chain of the previously proposed structure leads to an important break point at the reducing end side of the helix (Figure 8). This bump constituted by reducing end trisaccharide is generated by steric hindrances preventing a proper folding of this part of the molecule in an helical conformation. In the same way, as also shown by the molecular model of the previous structure, due to its [alpha] anomeric configuration, the second branched Glcp residue is located outside perpendicularly to the main axis of the helix. In contrast, a model of the revised structure shows that this latter can be easily oriented along the longitudinal axis of the helix, thus prolonging the cylindrical cavity. Likewise, modifications of the anomeric configurations and the linkage site of the two residues linked on the penultimate residue totally change the conformation of the reducing end trisaccharide, allowing it to extend the helix.


Figure 8. Structural model of the MGP20,12. (a) Revised structure; (b) previous structure; glyceric acid and the three modified linkages are shown (boldface type corresponds to the modified main chain linkage).

This precise structure determination is a necessary step for understanding the mechanism of this unique biological interaction between lipids and polysaccharide. Moreover, preliminary data, using capillary zone electrophoresis coupled to mass spectrometry, have confirmed the formation of stoichiometric complex between MGLP and their putative biological ligand: fatty acyl-coenzyme A. A more exhaustive study using different polysaccharidic or lipidic partners of various lengths will be necessary in characterizing more precisely the specificity and recognition mechanism. From a biological point of view, these molecules have been shown to have a key role in the fatty acid biosynthesis and thus indirectly on the cell wall mycoloyl-arabinogalactan backbone formation. Moreover, since they are very specific of the mycobacterial genus, search of specific inhibitors of their biosynthesis would be of great value for the development of new antituberculous drugs.

Materials and methods

Purification of MGLPs

MGLPs were purified as described previously (Hindsgaul and Ballou, 1984; Tuffal et al., 1995a) from 15 day old culture of M.bovis BCG strain Pasteur grown as pellicle on Sauton medium.

Briefly, the cells (100 g wet weight) were extracted thoroughly with a mixture of CHCl3/CH3OH (2/1). The dried crude lipidic extract (55 g) in chloroform (10 ml/g) was extensively washed with water. The different water washing fraction were pooled and the MGLPs were recovered from the aqueous phase by reverse phase-FPLC on a 60 ml column (Merck) packed with HPLC reverse phase (C18, 25-60 µm, Merck). Elution was run in steps of 200 ml of mixtures containing increasing amounts of methanol in water (0%, 20%, 40%, 70%, and 100% (v/v)) at 2.5 ml/min. Fractions from each step were pooled and analyzed by GC and GC/MS after methanolysis and trimethylsilyl (TMS) derivatization. As previously the MGLPs were eluted with 70% methanol in water and were further purified on a G50 Sephadex gel filtration column eluted with 30 mM ammonium acetate. Elution was monitored by refractive index and sulfuric anthron carbohydrate assay (OD 630 nm). Fractions containing MGLPs (20 mg) were pooled and lyophilized.

HPAEC analysis of MGPs

Prior HPAEC analysis, the MGLPs were deacylated with 0.1 M NaOH for 15 min at room temperature yielding MGPs. The mixture was neutralized with 1 M HCl, extracted with chloroform and desalted using reverse phase C18 T Sep Pak (Millipore) chromatography with step elution with water and 70% methanol in water. Salts were eluted in the water fraction and MGPs in the 70% methanol eluate.

HPAEC analysis was conducted with ~1 µg of MGPs using a DX 300 Dionex system equipped with an eluent degas module, a gradient pump and a Carbopac PA1 column (4 × 250 mm). Eluents were made using (18 M[Omega]) deionized water (Millipore), 50% dilute sodium hydroxide (Baker), and ACS grade sodium acetate (Merck). Eluent A was 0.1 M NaOH and eluent B was 0.5 M CH3COONa in 0.1 M NaOH. Elution was performed at 1 ml/min and started with a step of 38% of eluent B for 2 min followed by a concave gradient (no. 7) from 38% to 48% of eluent B for 60 min. Detection was monitored by a PED detector (Pulsed Electrochemical Detector) with a standard wave program for carbohydrate (gold electrode without pH reference, E1 = 0.1 V (0.5s), E2 = 0.6V (0.1s), E3 = -0.6V (0.05s)).

Purification of MGPs for LSIMS analysis was performed with the same system but using a semipreparative column (9 × 250 mm), eluted at 5 ml/min.

Desalting of MGPs after HPAEC purification

HPAEC fractions were neutralized using HCl (1 M) and applied to a C18 T Sep-Pak cartridge (Millipore) eluted with water and mixtures of 20 and 70% (v/v) of methanol in water. Salts were eluted in the water fraction and in the 20% methanol in water fraction. MGPs were recovered in the 70% of methanol fraction.

LSIMS analysis of MGPs

MS spectra were recorded on a two-sector instrument ZAB-SE (VG Analytical) in the positive and negative modes. The cesium beam energy was 35 kV and the accelerating voltage was 8 kV. Full spectra were recorded over a mass range of 800-4000 and a resolution of [Delta]M/M of 800. Data were processed on a Vax 2000 station (Digital). Sample (1 µl of MGPs at 5 µg /µl) was mixed with thioglycerol (10% of acetic acid) and deposited on a probe tip.

NMR analysis

MGPs were exchanged twice with D2O, dried and dissolved in D2O (100% Atom, Sigma) or DMSO-d6 (100% Atom, Sigma) at final concentration of 6 mg/ml for the MGP20,12 and less then 2 mg/ml for the MGP A. NMR spectroscopy was performed on a Brucker AMX 500 spectrometer operating at 500 MHz for 1H and equipped with a X32 computer. Spectra were recorded at 313K and chemical shifts were referenced indirectly to tetramethylsilane setting the HOD peak at 4.8 ppm.

The one-dimensional proton (1D 1H) spectrum was measured using a 90° tipping angle for the pulse and 1 s as a recycle delay between each of the 1024 acquisitions of 2 s. The spectral width of 3915 Hz was digitized on 8192 points that were multiplied by a Lorentz-Gauss function (LB = -3 Hz,GB = 0.4) prior to processing to 16384 real points in the frequency domain.

All 2D NMR data sets were recorded without sample spinning. The data were acquired in the phase-sensitive mode using the time-proportional phase increment (TPPI) method (Marion and Wüthrich, 1983), except for COSY experiments.

COSY 45 were performed according to the standard pulse sequences supplied by Bruker. A spectral width of 3508 Hz in both dimensions was used to collect a 4096 × 1024 data matrix with four transients for each t1 increment. The data matrix was multiplied by an unshifted sine-bell function in both dimensions before Fourier transformation.

The homonuclear Hartmann-Hahn (HOHAHA) spectra (Bax and Davies, 1985) was acquired with a MLEV-17 sequence, increasing mixing time from 10 to 175 ms and a spectral width of 3915 Hz. The relaxation delay was 1 s and the data matrix was 4096 × 1024 points, with 8 scans × t1 values. For processing, a sine-bell window shifted by [pi]/2 was applied in both dimensions.

ROESY spectra were acquired with the Bruker standard pulse sequence using 120 ms continuous wave spin lock time. 4096 × 1024 points, with 16 scans × t1 values were collected over a sweep width of 2000 Hz. The final data were multiplied by a [pi]/2 shifted square sine bell prior Fourier transformation (4K × 1[Kgr]).

Analytical methods

Routine gas chromatography (GC) was performed on a Girdel series 30 chromatograph equipped with an OV1 wall-coated open tubular capillary column (0.32 mm × 25 m, Spiral France) using nitrogen gas at a flow of 2.5 ml/min with a flame ionization detector set at 310°C. A temperature program from 100°C to 250°C at a speed of 3°C/min was used for trimethylsilyl methyl glycoside analysis. GC/MS experiments were done on a Hewlett Packard 5889X mass spectrometer working in electron impact (EI) and chemical ionization (CI) (NH4+) modes.

Molecular modeling

Graphics were performed on a ELAN Silicon Graphics workstation using INSIGHT and Discover program from Biosym (San Diego, CA) molecular modeling package.

The MGPs model was generated in two steps (Tuffal et al., 1995b) according to the amylose V structure (Zugenmaier and Sarko, 1976) with six residues per turn.

Acknowledgments

We thank C. Ponthus (Sanofi Recherche, Toulouse, France) for mass spectrometry analysis. This work was supported by grants from Ministère de l' Education Nationale, De l' Enseignement Superieur, de la Recherche et de l'Insertion Professionnelle ACC SV6/9506005 and from the region Midi-Pyrénées, RECH/9407528 .

Abbreviations

INH, soniazid; Enoyl-ACP, enoyl-acyl carrier protein; MGP, methyl glucose polysaccharides; MGPm,n, methyl glucose polysaccharide composed of m glucose residues n of which are mono-O-methylated; MGLP, methyl glucose lipopolysaccharides; FAS I, fatty acid synthetase, type 1; HPAEC, high pH anion exchange chromatography; LSIMS, liquid secondary-ion mass spectrometry; Glcp, d-glucopyranose.

References

Banis ,R.J., Peterson,D.O. and Bloch,K. (1977) Mycobacterium smegmatis fatty acid synthetase. Polysaccharide stimulation of the rate-limiting step. J. Biol. Chem., 252, 5740-5744. MEDLINE Abstract

Bax ,A. and Davies,D.G. (1985) MLEV-17 based two-dimensional homonuclear magnetization transfer spectroscopy. J. Magn. Reson., 65, 355-360.

Bloch ,K. (1975) Fatty acid synthases from Mycobacterium phlei. Methods Enzymol.,35, 84-90. MEDLINE Abstract

Bloch ,K. (1977) Control mechanisms for fatty acid synthesis in Mycobacterium smegmatis. Adv. Enzymol., 45, 1-84. MEDLINE Abstract

Bloom ,B.R. and Murray,C.J. (1992) Tuberculosis: commentary on a reemergent killer. Science, 257, 1055-1064. MEDLINE Abstract

Dell ,A. and Ballou,C.E. (1983) Fast-atom-bombardment, negative-ion mass spectrometry of the mycobacterial O-methyl-d-glucose polysaccharide and lipopolysaccharides. Carbohydr. Res., 120, 95-111. MEDLINE Abstract

Dessen ,A., Quemard,A., Blanchard,J.S., Jacobs,W.R.,Jr. and Sacchettini,J.C. (1995) Crystal structure and function of the isoniazid target of Mycobacterium tuberculosis. Science, 267, 1638-1641. MEDLINE Abstract

Flick ,P.K. and Bloch,K. (1974) Vitro alterations of the product distribution of the fatty synthetase from Mycobacterium phlei. J. Biol. Chem., 249, 1031-1036. MEDLINE Abstract

Forsberg ,L.S., Dell,A., Walton,D.J. and Ballou,C.E. (1982) Revised structure for the 6-O-methylglucose polysaccharide of Mycobacterium smegmatis. J. Biol. Chem., 257, 3555-3563. MEDLINE Abstract

Goren ,M.B. (1972) Mycobacterial lipids: selected topics. Microbiol. Rev., 36, 33-64.

Gray ,G.R. and Ballou,C.E. (1971) Isolation and characterization of a polysaccharide containing 3-O- methyl-d-mannose from Mycobacterium phlei. J. Biol. Chem., 246, 6835-6842. MEDLINE Abstract

Gray ,G.R. and Ballou,C.E. (1975) Methylated polysaccharide activators of fatty acid synthase from Mycobacterium phlei. Methods Enzymol., 35, 90-95. MEDLINE Abstract

Hindsgaul ,O. and Ballou,C.E. (1984) Affinity purification of mycobacterial polymethyl polysaccharides and a study of polysaccharide-lipid interactions by 1H NMR. Biochemistry, 23, 577-584. MEDLINE Abstract

Hunter ,B.K., Mowbray,S.L. and Walton,D.J. (1979) A compound representing the d-glycerate terminus of the methylglucose- containing polysaccharide of Mycobacterium smegmatis. Biochemistry,18, 4458-4465. MEDLINE Abstract

Ilton ,M., Jevans,A.W., McCarthy,E.D., Vance,D., White,H.B.D. and Bloch,K. (1971) Fatty acid synthetase activity in Mycobacterium phlei: regulation by polysaccharides. Proc. Natl. Acad. Sci. USA, 68, 87-91. MEDLINE Abstract

Jarlier ,V. and Nikaido,H. (1994) Mycobacterial cell wall: structure and role in natural resistance to antibiotics. FEMS Microbiol. Lett., 123, 11-18. MEDLINE Abstract

Lee ,R.E., Mikusová,K., Brennan,P.J. and Besra,G.S. (1995) Synthesis of the mycobacterial arabinose donor [beta]-d-arabinofuranosyl-1-monophosphoryldecaprenol, development of a basic arabinosyl-transferase assay, and identification of ethambutol as an arabinosyl-transferase inhibitor. J. Am. Chem. Soc., 117, 11829-11832.

Lee ,Y.C. and Ballou,C.E. (1964) 6-O-Methyl-d-glucose from mycobacteria. J. Biol. Chem., 239, 3602-3603.

Machida ,Y. and Bloch,K. (1973) Complex formation between mycobacterial polysaccharides and fatty acyl-CoA derivatives. Proc. Natl. Acad. Sci. USA, 70, 1146-1148. MEDLINE Abstract

Marion ,D. and Wüthrich,K. (1983) Application of phase sensitive two dimensional correlated spectroscopy (COSY) for measurements of 1H-1H spin-spin coupling constants in proteins. Biochem. Biophys. Res. Commun., 113, 967-974. MEDLINE Abstract

Murray ,C.J.L., Stylbo,K. and Rouillon,A. (1993) In Jamison,D.T. and Mosley,W.H. (eds.), Disease Control Priorities in Developing Countries. Oxford University Press: New York, pp. 233-259.

Nikaido ,H. (1994) Prevention of drug access to bacterial targets: permeability barriers and active efflux. Science, 264, 382-387. MEDLINE Abstract

Nowak ,R. (1995) WHO calls for action against TB. Science, 267, 1763. MEDLINE Abstract

Odriozola ,J.M., Ramos,J.A. and Bloch,K. (1977) Fatty acid synthetase activity in Mycobacterium smegmatis. Characterization of the acyl carrier protein-dependent elongating system. Biochim. Biophys. Acta, 488, 207-217.

Peterson ,D.O. and Bloch,K. (1977) Mycobacterium smegmatis fatty acid synthetase. Long chain transacylase chain length specificity. J. Biol. Chem.,252, 5735-5739.

Quemard ,A., Sacchettini,J.C., Dessen,A., Vilcheze,C., Bittman,R., Jacobs,W.R.,Jr. and Blanchard,J.S. (1995) Enzymatic characterization of the target for isoniazid in Mycobacterium tuberculosis. Biochemistry, 34, 8235-8241. MEDLINE Abstract

Saier ,M.H.,Jr. and Ballou,C.E. (1968a) The 6-O-methylglucose-containig lipopolysaccharide of Mycobacterium phlei. Complete structure of the polysaccharide. J. Biol. Chem., 243, 4332-4341. MEDLINE Abstract

Saier ,M.H.,Jr. and Ballou,C.E. (1968b) The 6-O-methylglucose-containing lipopolysaccharide of Mycobacterium phlei. Identification of d-glyceric acid and 3-O-methyl-d-glucose in the polysaccharide. J. Biol. Chem., 243, 992-1005. MEDLINE Abstract

Saier ,M.H.,Jr. and Ballou,C.E. (1968c) The 6-O-methylglucose-containing lipopolysaccharide of Mycobacterium phlei. Structure of the reducing end of the polysaccharide. J. Biol. Chem., 243, 4319-4331. MEDLINE Abstract

Tuffal ,G., Albigot,R., Monsarrat,B., Ponthus,C., Picard,C., Rivière,M. and Puzo,G. (1995a) Purification and LSIMS analysis of methylglucose polysaccharides from Mycobacterium xenopi, a slow growing species. J. Carbohydr. Chem.,14, 631-642.

Tuffal ,G., Ponthus,C., Picard,C., Riviere,M. and Puzo,G. (1995b) Structural elucidation of novel methylglucose-containing polysaccharides from Mycobacterium xenopi. Eur. J. Biochem., 233, 377-383. MEDLINE Abstract

Vance ,D.E., Mitsuhashi,O. and Bloch,K. (1973) Purification and properties of the fatty acid synthetase from Mycobacterium phlei. J. Biol. Chem., 248, 2303-2309. MEDLINE Abstract

Weiss ,R. (1992) On the track of 'killer" TB. Science, 255, 148-150. MEDLINE Abstract

Zugenmaier ,P. and Sarko,A. (1976) Packing analysis of carbohydrates and polysaccharides. IV. A new method for detailed crystal structure refinement of polysaccharides and its application to V-amylose. Biopolymers, 15, 2121-2136. MEDLINE Abstract

1To whom correspondence should be addressed

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 10 Jun 1998
Copyright©Oxford University Press, 1998.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 J. Biol. Chem.Home page
M. Jackson and P. J. Brennan
Polymethylated Polysaccharides from Mycobacterium Species Revisited
J. Biol. Chem., January 23, 2009; 284(4): 1949 - 1953.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
G. Stadthagen, T. Sambou, M. Guerin, N. Barilone, F. Boudou, J. Kordulakova, P. Charles, P. M. Alzari, A. Lemassu, M. Daffe, et al.
Genetic Basis for the Biosynthesis of Methylglucose Lipopolysaccharides in Mycobacterium tuberculosis
J. Biol. Chem., September 14, 2007; 282(37): 27270 - 27276.
[Abstract] [Full Text] [PDF]

Home page
 GlycobiologyHome page
S. Berg, D. Kaur, M. Jackson, and P. J Brennan
The glycosyltransferases of Mycobacterium tuberculosis--roles in the synthesis of arabinogalactan, lipoarabinomannan, and other glycoconjugates
Glycobiology, June 1, 2007; 17(6): 35R - 56R.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (17)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Tuffal, G.
Right arrow Articles by Puzo, G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tuffal, G.
Right arrow Articles by Puzo, G.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?