Specific Lipopolysaccharide Found in Cystic Fibrosis Airway Pseudomonas aeruginosa

NIH Consensus Statement 15, 1 (1997).

Govan J. R., Deretic V., Microbiol. Rev. 60, 539 (1996).

Guo L., et al., Science 276, 250 (1997);

Guo L., et al., Cell 95, 189 (1998);

Gunn J. S., Miller S. I., J. Bacteriol. 178, 6857 (1996) .

The laboratory CF strains used were PAK and PAO-1. Clinical isolates CF344, CF725, CF1153, CF1188, CF1212, CF1213, and CF1214 were collected as part of a multicenter study on infection and inflammation in young infants with CF. Non-CF clinical isolates (from patients with bronchiectasis, RKE024-026; from blood, RKE020) were from J. Burns and PathoGenesis Corporation. Cultures were grown at 37°C with aeration in either LB or N-minimal medium supplemented with 38 mM glycerol, 0.1% casamino acids, and either 8 μM or 1 mM MgCl₂ [

Garcia-Vescovi E., Soncini F. C., Groisman A. A., Cell 84, 165 (1996);

]. LPS was isolated by means of Mg²⁺ ethanol precipitation as described by Darveau and Hancock [

Darveau R. P., Hancock R. E., J. Bacteriol. 155, 831 (1983);

]; lipid A was isolated after hydrolysis in 1% SDS at pH 4.5 [

Caroff M., Tacken A., Szabo L., Carbohydr. Res. 175, 273 (1988);

]. LPS fatty acids were derivatized to fatty acid methyl esters and analyzed by GC [

Somerville J. E., Cassiano L., Bainbridge B., Cunningham M. D., Darveau R. P., J. Clin. Invest. 97, 359 (1996);

Darveau R. P., et al., Infect. Immun. 63, 1311 (1995)].

Negative-ion MALDI-TOF and electrospray experiments were performed as described, with the following modifications [Proceedings of the 46th ASMS Conference on Mass Spectrometry and Allied Topics, American Society for Mass Spectrometry (ASMS), Orlando, FL, 31 May to 4 June 1998 (ASMS, Santa Fe, NM, 1998)]. Lyophilized lipid A was dissolved with 5 μl of 5-chloro-2-mercaptobenzothiazole (CMBT) MALDI matrix in chloroform/methanol, 1:1 (v/v) and then applied (1 μl) onto the sample plate [

Xu N., Huang Z.-H., et al., J. Am. Soc. Mass Spectrom. 8, 116 (1997);

]. All MALDI-TOF experiments were performed with a Voyager DE mass spectrometer (PerSeptive Biosystems, Framingham, MA). The electrospray work was performed with either a TSQ 7000 triple quadrupole mass spectrometer or an LCQ ion trap (Finnigan/Thermoquest, San Jose, CA). Both electrospray instruments were equipped with an experimental low flow (nanoliters per minute) capacitive ion source [

Wang H., Hackett M., Anal. Chem. 70, 205 (1998)].

E. C. Yi, K. B. Lim, M. Hackett, unpublished data.

Bhat R., Marx A., Galanos C., Conrad R. S., J. Bacteriol. 172, 6631 (1990).

Kulshin V. A., et al., Eur. J. Biochem. 198, 697 (1991);

Karunaratne D. N., Richards J. C., Hancock R. E., Arch. Biochem. Biophys. 299, 368 (1992).

Nummila K., Kilpelainen I., Zahringer U., Vaara M., Helander I. M., Mol. Microbiol. 16, 271 (1995).

Based on the published lipid A structure for the penta-acylated form (7, 8), there are only two possible locations where the C16:0 can be attached to yield hexa-acylated lipid A, m/z 1685: (i) directly onto the 3-hydroxy position of the diglucosamine backbone or (ii) piggyback on the 3′ 3-oxo-C10:0 (see Fig. 1A for the numbering of the carbons on the sugar backbone). Fragmentation of the m/z 1685 precursor ion showed a peak at m/z 944 as a result of the loss of a phosphate, C10:0, 2-OH-C12:0, and C16:0 groups to yield triacyl monophosphate lipid A (Fig. 1, D and E). From the two possibilities discussed above, only the assignment of the C16:0 group on 3′ 3-oxo-C10:0 would yield the observed m/z 944 fragment. When similar logic was applied, with the C16:0 group arbitrarily assigned to the 3-hydroxy position of the diglucosamine backbone, mass peaks related to the triacyl and tetra-acyl monophosphate lipid A fragments were predicted to occur at m/z 926 and 1183 respectively, but were not observed in the product ion (MS²) spectra. Furthermore, the MS² spectra from the m/z 1447 precursor ion (with C16:0 absent) also yielded the m/z 944 fragment, which is the expected triacyl monophosphate product ion, and further confirmed that the location of C16:0 was not at the 3-hydroxy position. Finally, the presence of an ion at m/z 1948 (Fig. 1B) indicated that two aminoarabinose moieties could be added to the hexa-acylated form of lipid A. Because of the facile sequential loss of aminoarabinose residues during fragmentation in the mass spectrometer, our studies using electrospray tandem MS were unable to confirm the precise location on the molecule of the two amino sugar groups. We sometimes observe low–signal-to-noise ratio (S/N) ions in our MS work (m/z 1685), which are attributable to a C16:0-containing hexa-acylated lipid A in the PAK and PAO-1 reference strains. Such signals for m/z 1948 have not been detected in the reference strains.

Darveau R. P., et al., J. Clin. Invest. 90, 447 (1992);

. C18G was a gift from R. Darveau, Univ. of Washington, Seattle, WA.

Polymyxin B sulfate was purchased from USB (Cleveland, OH). Assays were performed as described [

Steinberg D. A., et al., Antimicrob. Agents Chemother. 41, 1738 (1997)].

Hancock R. E., Falla T., Brown M., Adv. Microb. Physiol. 37, 135 (1995).

Miller S. I., Kukral A. M., Mekalanos J. J., Proc. Natl. Acad. Sci. U.S.A. 13, 5054 (1989);

Garcia-Vescovi E., Soncini F. C., Groisman A. A., Cell 84, 165 (1996);

Gunn J. S., Hohmann E. L., Miller S. I., J. Bacteriol. 178, 6369 (1996).

Bell A., Hancock R. E., J. Bacteriol. 171, 3211 (1989).

The plasmid vector pBSphoP was constructed by cloning a 2.7-kb Bam HI–Mse I fragment [nucleotide 8054 through 10784 (nt 8054–10784)] containing the complete oprH1 (nt 8755–9354) and phoP (nt 9437–10110) genes and a partial phoQ (nt 10138–11454) gene into the vector pBluescript KS (+) digested with Eco RV. Nucleotide numbers are derived from contig 632 from the Pseudomonas Genome Project sequence database, release 3/15/98 (Pathogenesis Corporation, Univ. of Washington; available at www.pseudomonas.com). A 1.7-kb gentamicin resistance cassette with Eco RI ends was blunt-end cloned into a unique Sfi I (nt 9716) site located in the phoP gene. This plasmid, called pBSphoP-Gm, was used for the insertional inactivation of the phoP gene in PA strain PAK (RKE004).

Soncini F. C., Garcia Vescovi E., Solomon F., Groisman E. A., J. Bacteriol. 178, 5092 (1996).

R. K. Ernst, L. Guo, S. I. Miller, unpublished data.

Gunn J. S., Miller S. I., J. Bacteriol. 178, 6857 (1996);

Gunn J. S., et al., Mol. Microbiol. 27, 1171 (1998).

Basu S. S., White K. A., Que N. L., Raetz C. R., J. Biol. Chem. 274, 11150 (1999);

. PA lipid_IvA, a tetra-cylated lipid A precusor [C. R. H. Raetz, in Escherichia coli and Salmonella, F. C. Neidhardt, Ed. (ASM Press, Washington, DC, 1996), pp. 1035–1063] can be generated by the addition of C12 and 2-OH-C12 to generate the hexa-acylated species of lipid A, m/z 1616. The CF-specific dominant penta-acylated species m/z 1447 can be generated by the deacylation of the hexa-acylated species m/z 1616; then the species m/z 1447 can be palmitoylated to generate the hexa-acylated species at m/z 1685.

Because of the qualitative nature of our of main beam (MS¹) MALDI-TOF mass spectra, quantitative analyses of the lipid A signals at m/z 1447 (penta-acyl), 1685 (penta-acyl plus C16) 1816 (hexa-acyl plus one aminoarabinose residue), and 1948 (hexa-acyl plus two aminoarabinose residues) for each minimally passaged clinical isolate were performed on a Finnigan LCQ ion trap using isopentenyl pyrophosphate (IPP, Sigma) (5 pmol/μl) as an internal standard (m/z 245). Approximately 10 pmol of each sample per microliter was prepared in 1:1 chloroform/methanol, which also contained IPP. Samples were infused (at a rate of 0.7 μl/min) into the capacitive electrospray source. The spray voltage and heated capillary temperature were set at 2.4 kV and 250°C; 20 scans were acquired per analysis; and 10 replicate analyses were performed and averaged for each sample. The relative standard deviation values for replicate analyses were 5% or better. Signals from the ions given above were normalized to the signal for IPP and summed, and the percentage of the total signal due to C16:0-containing structures was calculated as follows: PAK, 8.8; CF1188, 33.7; CF344, 21.1; CF725, 11.2; CF1213, 6.6; CF1153, 13.1; CF1214, 18.9; and CF1212, 5.4.

Tandem MS data (6) from the m/z 1419 precursor were consistent with prior work [

Goldman R. C., Doran C. C., Kadam S. K., Capobianco J. O., J. Biol. Chem. 263, 5217 (1988)].

Growth media for the HUVECs consisted of Medium 199 (Gibco-BRL Gaithersburg, MD) supplemented with 4 mM l-glutamine, heparin (90 μg/ml), 1 mM Na pyruvate, endothelial cell growth stimulant (30 μg/ml) (Biomedical Products, Bedford, MA), and 20% fetal bovine serum (Summit Biotechnology, Fort Collins, CO). Stimulation medium consisted of Medium 199 plus 4 mM l-glutamine, heparin (90 μg/ml), 1 mM Na pyruvate, human serum albumin (1 mg/ml), and 5% pooled human serum (Gemini Bioproducts, Calabasas, CA).

HUVEC stimulation assays were performed as described [

Darveau R. P., et al., Infect. Immun. 63, 1311 (1995);

], with the following modifications. IL-8 samples (75 μl) were removed after the stimulation interval (22 hours) and stored at −20°C until assayed. IL-8 immunodetection was performed as follows: Microtiter plates (Immulon II, Dynex Technologies, Chantilly, VA) were coated using 50 μl of IL-8 capture monoclonal antibody (2.0 μg/ml) (M-801, Endogen, Woburn, MA) in phosphate-buffered saline (PBS) overnight at 4°C. Free binding sites were blocked with 200 μl of 2% bovine serum albumin in PBS at room temperature for 1 hour. Washes were performed with PBS containing 0.2% Tween-20 between incubations. Fifty microliters of sample and 50 μl of IL-8 biotinylated monoclonal antibody (0.2 μg/ml) (M-802-B, Endogen, Woburn, MA) were incubated at room temperature. After being shaken for 2 hours, wells were washed and were incubated with 50 μl Vectastain (PK6100, Burlingame, CA) at 37°C for 1 hour. Plates were developed according to manufacturers' recommendations using EIA chromogen reagent (R6, Redmond, WA) and were read with a Molecular Devices Thermomax microplate reader (Sunnyvale, CA) at an optical density of 450 nm. MY4, the monoclonal antibody to human CD14, was obtained from Coulter Immunology (Hialeah, FL).

MS² refers to a single stage of the sequential product ion experiment in which a precursor ion (for example, m/z 1616) is isolated from others present in the main beam spectrum (MS¹) and fragmented by colliding the precursor ion with a neutral target gas, such as argon (∼3 mTorr) in the triple quad collision cell [

Yost R. A., Boyd R. K., Methods Enzymol. 193, 154 (1990);

] or the helium bath gas (∼1 mTorr) in the ion trap [K. R. Jonscher and J. R. Yates 3rd, Anal. Biochem. 244, 1 (1997);

Wong P. S. H., Cooks R. G., Curr. Sep. 16, 85 (1997);

; M. E. Bier and J. C. Schwartz, in Electrospray Ionization Mass Spectrometry, R. B. Cole, Ed. (Wiley, New York, 1997), pp. 235–289]. MS³ refers to an additional stage of this same process, typically performed in an ion trap, in which a structurally informative fragment found in the MS² spectrum is in turn isolated and subjected to fragmentation.

We thank S. Lory and S. Moskowitz for critically reading the manuscript, K. Walsh and L. Ericsson for use of their MALDI-TOF mass spectrometer, and W. N. Howald for capillary GC/MS. Support was provided by grants from the Cystic Fibrosis Foundation [grants CFF 97 Z0 (S.I.M.), CFF R565 (M.H.), and CFF A922, a postdoctoral research fellowship (R.K.E.), and grant R565 (J.L.B.)]. Additional support was provided by NIH [grants R21 R13400 (M.H.) and R55 HL 48888 (J.L.B.)], Finnigan/Thermoquest, Michrom Bioresources, and the University of Washington Office of Technology Transfer Prototype Fund (M.H.).