Wacker et al. 10.1073/pnas.0509207103. |
Supporting Figure 5
Supporting Text
Supporting Figure 6
Fig. 5. Relevant area of the positive ion mode reflectron MS spectrum of the tryptic digest of AcrA purified from EVV11 cells carrying PglB and WbbL, showing the glycopeptide ion selected for TOF/TOF MS/MS analysis, the result of which is shown (Lower).
Fig. 6. (Top) Relevant area of the positive ion mode reflectron MS spectrum of the tryptic digest of AcrA purified from CWG44 (O9a-/K30) cells carrying PglB, showing a peptide ion series of which the clusters are separated by 162 m/z, suggestive of a family of compounds differing from each other by one hexose residue. The ions at m/z 3,445.7 and 3,607.8 were selected for TOF/TOF MS/MS analysis, the result of which is shown in Middle and Bottom. The MS/MS data confirm that both ions correspond to the peptide at 2,755 m/z (which is known to contain an AcrA N- glycosylation site) that carries a linear oligohexoside on a peptide-proximal N-acetylhexosamine and differ only in the presence of an additional hexose residue. Considering the genotype of this strain, the oligohexoside most likely derives from the O9-antigen biosynthetic machinery that is mutated, but at an unknown site. Thus, we propose the structures in Fig. 2C for these AcrA-linked glycans.
Supporting Text
Construction of EVV11 Polymerase-Defective Strain.
A deletion of the wzy gene was constructed by using the method described by Datsenko and Wanner (1). Escherichia coli K-12 W3110 was transformed with pKD46, a temperature-sensitive plasmid carrying the Red recombinase system from the l bacteriophage under the control of the arabinose-inducible PBAD promoter. The Red recombinase system mediates the replacement of the target chromosomal sequence with an antibiotic resistance cassette obtained by PCR amplification using primers that carry homologies to the vicinity on the gene targeted for disruption. A kanamycin-resistance cassette flanked by homologies upstream and downstream of the wzy gene from E. coli K-12 W3110 was obtained by PCR amplification using pKD4 as a template (1), with primers 709 (5'-ATAATGAGTACGGATTAATGATCTATCTTGTAATTAGTGGTGTAGGCTG-GAGCTGCTTCG-3') and 710 (5'-CAGCATCGCGTCTAGAGAAATTTAAATCATTCAAAAAATACATATGAATATCCTCCTTAG-3'). The region aligning with plasmid pKD4 is underlined.E. coli
W3110 carrying pKD46 was transformed by electroporation with this PCR product, and transformants were selected on LB agar containing 50 mg/ml kanamycin. After confirming the replacement of the wzy gene by PCR, plasmid pKD46 was cured from the strain by growing it at 42°C, plating on LB, and screening for ampicillin sensitivity. The kanamycin-resistance cassette was then excised from the strain as follows: the mutant strain was transformed with the temperature-sensitive plasmid pCP20 (ampR) (1) that constitutively expresses the FLP recombinase from yeast. This recombinase allows the excision of the kanamycin resistance, due to recognition sites at both ends of the cassette. Excision of the kanamycin resistance cassette was confirmed by plating the strain on LB and screening for kanamycin sensitivity and PCR. Finally, pCP20 was cured from strain EVV11, by growing it at 42°C, plating on LB, and screening for ampicillin sensitivity.MALDI-TOF MS Details
MALDI-TOF MS in the positive ion reflectron mode was performed on an ABI 4700 instrument, using a laser fluency tuned for an optimal balance between sensitivity and spectral resolution for this analyte-matrix combination. External on-plate calibration in the m/z range 750-4,250 was performed using a peptide calibration standard (Applied Biosystems). The spectra shown are averaged over at least 2,500 laser shots and are smoothed by a five-point Gaussian transformation. Tandem TOF/TOF MS/MS) was carried out at low-collision gas pressure (5.10E7 millibar), and at least 5,000 laser shots were averaged. The spectra were processed by using algorithmic resolution reduction to 2,000 in case of the very low-intensity MS/MS spectra from the O9-oligosaccharide modified AcrA glycopeptides, and by five-point Gaussian transformation for the AcrA glycopeptide modified with the O16 subunit.
MS Analysis of Tryptic Glycopeptides Containing One O16 Subunit
Mono- and di-glycosylated AcrA from EVV11 strain transformed with pMF19 (Fig. 1B, lane 1) were digested with trypsin, and the product mixture was analyzed by MALDI-TOF/TOF MS and MS/MS to determine the structure of the short oligosaccharide (Fig. 5). As a result of spectral crowding in the low m/z region, we detected only the largest glycopeptide (expected peptide m/z 2,755.4 + carbohydrate) in the trypsin digestion mixtures of both the mono-and di-glycosylated AcrA bands. A single peak at m/z 3,632.1 was detected in this case, accompanied by its sodiated analogue (Fig. 5 Upper). A smaller peak at around m/z 3,470 was also present, with the hallmarks of metastable decay (unfocused ions toward the higher m/z side of a focused isotope pattern). Upon TOF/TOF MS/MS fragmentation of the analyte at 3,632.1 (ion selector focused on the most abundant isotope at 3,634.1), we obtained the spectrum shown in Fig. 5 Lower. The pattern could be interpreted as containing mainly the y-ions resulting from fragmentation of the postulated O16 repeating subunit, with a hexose on the peptide-proximal HexNAc residue. The glycosidic linkages of the two outermost residues appeared to be rather labile under MALDI conditions, because their fragmentation also occurred via metastable decay (which accounts for the "double" peaks around 3,470 and around 3,308, and confirming the preliminary interpretation of the MS spectrum). Again, a strong ion at m/z 1,567 was present in the MS/MS spectrum (not shown), typical for the TOF/TOF fragmentation pattern of the AcrA peptide expected at 2,755.4, which contains one N-glycosylation consensus sequence. Based on the known structure and genetic data, the only candidate for this hexose residue was the branching Glc at the C-6 position of the reducing GlcNAc residue of the O16 antigen repeating unit. Thus, we concluded that a substitution at the C-6 position in the reducing end of the saccharide did not prevent its transfer by PglB to the protein acceptor AcrA.
Glycopeptide Analysis Shows That the K30 Antigen Is Not Transferred to AcrA
To identify the nature of the AcrA modification in CWG44 cells (Fig. 2B, lane 5), we carried out MS analysis of peptides resulting from trypsin digestion of purified protein. After induction of PglB, periplasmic AcrA was extracted, purified by affinity chromatography using a NTA-agarose column, separated by SDS/PAGE, and stained by Coomassie. Glycosylated AcrA was digested with trypsin, and the product mixture was analyzed by MALDI-TOF/TOF MS and MS/MS (Fig. 6). In both positive and negative ion mode MS, a series of low-abundant analytes in the m/z range >3,000 were observed, in a region devoid of reproducibly detectable peaks for nonglycosylated AcrA (largest bona fide tryptic peptide is detected at m/z = 2,755) (Fig. 6 Top). The series was spaced by 162 m/z intervals, indicative of one hexose unit. Both the protonated and sodiated ions were detected (m/z difference of 22). Despite the extremely low abundance (and/or ionization efficiency) of these analytes and their relatively high m/z (both of which factors negatively affect MS/MS spectral quality), we attempted TOF/TOF MS/MS. Averaging 10,000 laser shots, we obtained the MS/MS spectra in Fig. 6 Middle and Bottom. As expected, spectral quality was poor, but a clear series of y-ions deriving from an oligosaccharide fragmentation was evident, all the way down to the unmodified peptide (predicted at m/z 2,755.4). A peptide fragment ion at m/z = 1,567 was also present, characteristic for the TOF/TOF fragmentation pattern of one of the two N-glycosylation consensus-site containing tryptic glycopeptides of AcrA (data not shown). The y-ion series strongly suggested the presence of a linear oligo-hexose on a peptide-proximal HexNAc residue. There appears to be microheterogeneity on this N-glycosylation site by virtue of a different length of the oligohexose. The most abundant glycoform contains three hexose residues, whereas the ones with four and five hexose residues are present in an abundance that appears to follow an exponentially decaying function. For the glycopeptide with five hexose residues, no informative MS/MS data were obtained. It should be noted that the abundance observed here might be skewed by the possibility of rapidly decreasing ionization efficiency with increasing oligosaccharide length. This analysis showed that AcrA indeed was glycosylated in the CWG44 strain. The oligosaccharides covalently linked to protein were most likely incompletely assembled O9 antigen, because they consist of HexNAc and additional Hex residues (Fig. 2C).
1. Datsenko, K. A. & Wanner, B. L. (2000) Proc. Natl. Acad. Sci. USA 97, 6640-6645.