immunochemical and biochemical analysis of the polyvalent ... · polyvalent vaccine pevwas derived...

INFECTION AND IMMUNITY, Feb. 1986, p. 675-6860019-9567/86/020675-12$02.00/0Copyright © 1986, American Society for Microbiology

Immunochemical and Biochemical Analysis of the PolyvalentPseudomonas aeruginosa Vaccine PEVSHEILA MACINTYRE,t TREASA McVEIGH, AND PETER OWEN*

Department of Microbiology, Trinity College, Dublin, Ireland

Received 3 July 1985/Accepted 4 October 1985

The Pseudomonas aeruginosa polyvalent vaccine PEV and its 16 constituent monovalent extracts fromInternational Antigenic Typing System serotypes 1 through 13 and 15 through 17 (J. J. Miler, J. F. Spilsbury,R. J. Jones, E. A. Roe, and E. J. L. Lowbury, J. Med. Microbiol. 10:19-27, 1977) were subjected tobiochemical analysis and to detailed immunochemical analysis with rabbit anti-PEV immunoglobulins. Theresults of chemical analysis, of analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresisperformed in coajunction with silver staining, and of analysis by crossed immunoelectrophoresis, sodiumdodecyl sulfate-polyacrylamide gel-crossed immunoelectrophoresis, and Western blotting showed clearly thatlipopolysaccharide was a major constituent of each monovalent extract and that it was probably the dominantantigen present in at least 15 of the 16 monovalent extracts. A 16.2-kilodalton protein, which was pronaseresistant and nonsedimentable at 105,000 x g and which appeared to be biochemically and antigenicallyunrelated to pili, was a common although minor antigen for all extracts. Several other proteins, some of outermembrane origin, were also detected in unformalinized extracts, but these were also minor antigenicconstituents of the vaccine. Neither pilin nor flagellin appeared to be major protein constituents of testedmonovalent extracts, although anti-flagella antibodies could be demonstrated in rabbit anti-PEV by Westernblotting. Preliminary analysis by crossed immunoelectrophoresis of serum raised in volunteers to PEV alsoindicated the presence therein of antibodies to lipopolysaccharide antigens.

Pseudomonas aeruginosa is an opportunistic pathogenwhich is frequently associated with nosocomially acquiredinfections. Certain groups of patients, for example, thosesuffering from cystic fibrosis or those who are immunocom-promised as a result of severe burn wounds, malignancy, orimmunosuppressive therapy, appear to be particularly sus-ceptible to infection by this organism (3). Moreover, evenwith the availability of improved antibiotic therapy, themortality rate of patients succumbing to infection by P.aeruginosa is still relatively high (see reference 9 for re-view). Immunotherapy is one promising method of treatmentof patients at high risk to infection by P. aeruginosa.Therefore, in the past decade, much attention has beenfocused on the protective properties of the surface antigensof this bacterium and on the development of a safe andeffective vaccine.

Lipopolysaccharide (LPS) has probably received the mostattention as a protective surface antigen of P. aeruginosa. Ithas been clearly established that antibody directed againstthis outer membrane component protects mice against infec-tion by homologous strains of P. aerluginosa (10, 11). Anti-LPS antibody appears to be important also in the protectionof humans against infection by this organism. High anti-LPStiters have been correlated with increased rates of survivalamong patients with pseudomonal bacteremia (50), and inaddition, during a 5-year clinical trial, the rate of Pseudomo-nas-related mortality among severely burned patients wasfound to be greatly reduced after immunization with aheptavalent LPS-based vaccine called Pseudogen (Parke,Davis & Co., Detroit, Mich. [1, 23]). Severe adverse reac-tions to this vaccine, however, were common (1, 45). The

* Corresponding author.t Present address: Max-Planck Institut fur Biologie, Corren-

strasse 38, 7400 Tubingen 1, Federal Republic of Germany.

major outer membrane pore-forming protein, protein F, isanother surface-expressed antigen of P. aeruginosa forwhich the ability to elicit a protective antibody response inmice has been demonstrated (18, 38). In contrast to LPS,protein F is antigenically related in all 17 serotypes of P.aeruginosa (39). In addition, at least three other surfacecomponents of P. aeruginosa have been reported to functionas protective antigens. These include flagella (27, 37), a toxicglycolipoprotein (2, 14, 54), and a high-molecular-weightpolysaccharide (47, 49, 51) which shares immunologicalidentity with the 0-side chains of homologous LPS (46, 48).PEV-01 (Wellcome Biotechnology Ltd., Beckenham,

United Kingdom) is a polyvalent vaccine which is composedof pooled surface antigen extracted, under mild conditions,from viable cells of 16 different serotypes of P. aeruginosa(36). This vaccine has been tested in volunteers (29) and hasundergone preliminary clinical trials, notably at SafdarjangHospital, New Delhi, India (28, 52). The results indicate thatimmunization of burn patients with either the polyvalentvaccine or human anti-PEV immunoglobulins provides pro-tection, with few adverse side effects, against fatal infectionby P. aeruginosa (52). Despite these encouraging results, theserologically active component of the PEV vaccine remainsunidentified. However, in view of the fact that the Pseu-domonas cells retain viability after extraction (with EDTA-glycine buffer [36]), and in view of the well-recognizedsensitivity of the outer membrane of this organism to chelat-ing agents (25), one might suspect that surface componentswould be important constituents of PEV.

Bearing in mind the well-documented role of antibody inprotection against Pseiudomonas infection (9), we have un-dertaken, as a preliminary step towards identifying theprotective antigen(s) in PEV-01, a high-resolution immuno-chemical analysis of the vaccine and its componentmonovalent extracts. We report here on the results of this

675

Vol. 51, No. 2

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

http://iai.asm.org/

676 MACINTYRE ET AL.

investigation and show that LPS is a major constituent of thevaccine.

(Preliminary accounts of this work have been presented atthe 100th and 103rd Ordinary Meetings of the Society forGeneral Microbiology, Warwick, United Kingdom, 1984,P30 and 1985, P33, respectively, and at the LunturenLectures on Molecular Genetics, The Netherlands, 1984,p. 30.)

MATERIALS AND METHODS

Bacterial strains and growth conditions. The 16 strains ofP.aeruginosa that were used to prepare the original vaccinePEV-01 (36) were obtained from Wellcome BiotechnologyLtd. They correspond to virulent strains of the InternationalAntigenic Typing System (IATS) serotypes 1 through 13 and15 through 17. For preparation of monovalent extracts, P.aeruginosa was grown as described (36) in 4-liter fermentorscontaining minimal medium supplemented where appropriatewith [35S]methionine (1.4 to 1.7 1Ci/ml). For the preparationof some cellular components, bacteria were grown at 37°C in500 ml of minimal medium (36) in 2-liter Erlenmeyer flasks,with agitation at 200 rpm. Cells were harvested at earlystationary phase (A6w, approximately 2.5) and were either (i)washed once in distilled water, lyophilized, and stored at-70°C until required for isolation of LPS or (ii) washed oncein 30 mM Tris hydrochloride (pH 8.0) and used immediatelyfor preparation of membrane fractions. A multipiliated strain(DB2) of P. aeruginosa PAO (IATS serotype 2) was used forthe preparation of pili and was grown on Trypticase soy agar(BBL Microbiology Systems, Cockeysville, Md.) as de-scribed (44).Monovalent extracts and polyvalent vaccine PEV. The

polyvalent vaccine PEV and its component monovalentextracts were prepared at Wellcome Biotechnology Ltd. aspreviously described (36). Briefly, bacteria from 16 serotypestrains were extracted at 3 x 1010 to 5 x 10'° viableorganisms per ml in EDTA-glycine buffer at 37°C, and thecell-free supernatant fractions were sterilized by filtrationand by the addition of Formalin to 0.3% (vol/vol). Thepolyvalent vaccine PEV was derived by pooling equal vol-umes of all 16 monovalent extracts (36). The number as-cribed to each monovalent extract denotes the IATSserotype of the strain from which the extract was prepared,with the exception that extract 14 is derived from IATSserotype 17. (IATS serotype 14 is rarely isolated and is notused in the production of PEV). For chemical analyses andanalysis by sodium dodecyl sulfate (SDS)-polyacrylamidegel electrophoresis, PEV and the monovalent extracts weredialyzed against 10 mM Tris hydrochloride (pH 7.5) for 36 hat 4°C. Undialyzed extracts were used for analyses bycrossed immunoelectrophoresis (CIE). The routine treat-ment of monovalent extracts with formaldehyde was omittedin experiments designed to assess the effect of this reagenton the resolution of constituent antigens.

Preparation of LPS. LPS was extracted from whole lyoph-ilized cells by the hot aqueous phenol procedure of Westphaland Jann (57). After extensive dialysis against water, theaqueous phase was lyophilized and then suspended in 0.5 MNaCl to a final concentration of 0.75% (wt/vol). Nucleic acidwas removed by fractional precipitation with cetyl-trimethylammonium bromide (57), and the detergent waseliminated by twice precipitating LPS from 0.5 M NaCl with10 volumes of ethanol at 4°C for 2 h. The final precipitate wassuspended in a small volume of distilled water and dialyzedextensively against distilled water at 4°C. Yields of LPS

isolated in this manner from P. aeruginosa serotypes weregenerally in the range of 3.2 to 8.9% of the cell dry weightand contained less than 1% protein or nucleic acid.

Isolation of cell fractions. Outer and inner membranes wereisolated from P. aeruginosa IATS serotype 6 by Frenchpress lysis, followed by centrifugation of the lysate on twoconsecutive sucrose density gradients, as described byHancock and Nikaido (22). The procedure yielded an innermembrane fraction (density, 1.17 g/ml) together with twoouter membrane fractions at densities of 1.23 g/ml (OM2) and1.248 g/ml (OM,). Both OM1 and OM2 displayed the sameprotein profiles on analysis by SDS-polyacrylamide gel elec-trophoresis. 1

Flagella were sheared from cells of the same strain byblending in a Silverson V5099 mixer-emulsifier at high speedwith 10 1-min bursts. After whole cells were removed bycentrifugation at 12,000 x g for 20 min at 4°C, flagella wereharvested from the supernatant fraction by centrifugation at105,000 x g for 3 h at 4°C. The pelleted material gave onemajor protein band with a molecular weight of 53,000 whenanalyzed by SDS-polyacrylamide gel electrophoresis, andthe material revealed characteristic flagellar structures whenviewed by negative staining in the electron microscope.

Pili were isolated from P. aeruginosa PAO/DB2 essen-tially as described by Paranchych et al. (44), but with theomission of the final centrifugation steps on CsCl gradients.Pilin isolated in this manner gave a major band at 17kilodaltons, which reacted strongly in immunoblotting ex-periments with anti-PAO pili serum. When analyzed bySDS-polyacrylamide gel electrophoresis, the pilin prepara-tion showed a contaminant at 53 kilodaltons attributable toflagellin.Anti-PEV immunoglobulins. Six New Zealand White rab-

bits were initially injected intradermally with the polyvalentvaccine (54 Ftg of protein, 4.2 ,ug of 2-keto-3-deoxyoctonicacid [KDO]) emulsified in Freund complete adjuvant.Booster injections with Freund incomplete adjuvant weregiven both subcutaneously and intradermally on days 14 and28 and then at monthly intervals. Serum was collected andpooled at fortnightly intervals over a period of 5 months.Human anti-PEV serum raised in volunteers was kindlysupplied by Wellcome Biotechnology Ltd. ImmunoglobulinsG and M (IgG and IgM) were isolated and concentrated10-fold with respect to the original serum concentration aspreviously described (42a, 55).

Adsorption of rabbit anti-PEV serum with LPS derivedfrom serotype 6 (LPS-6) was achieved by mixing 250 pul ofpurified and concentrated immunoglobulins with 50 RI ofLPS-6 (2 mg/ml) and 150 pI of phosphate-buffered saline (pH7.4). Incubation was continued for 1 h at 37°C, and the turbidsuspension was then held at 20°C for 18 h. Precipitated LPSwas removed by centrifugation at 10,000 x g for 5 min. Thesupernatant fraction was removed and readsorbed with anadditonal 100 ,ug of LPS-6 as above. Negligible precipitationwas observed, and residual LPS was finally removed bycentrifugation at 41,000 x g and 4°C for 1 h.

Electrophoretic techniques. CIE was performed at 20°C in1% agarose (Seakem LE; relative mobility [-mr], 0.01 to0.15; Miles Laboratories, Stoke Poges, United Kingdom)dissolved in CIE running buffer (barbital hydrochloride [pH8.6] containing 1% [vol/vol] Triton X-100), as previouslydescribed (42a, 55). Samples, suspended in 1% TritonX-100-50 mM Tris hydrochloride-5 mM EDTA (pH 8.6),were routinely subjected to electrophoresis at 5.5 V/cm for75 min in the first dimension and at 2 V/cm overnight (18 to22 h) in the second dimension. Unless otherwise indicated,

INFECT. IMMUN.

on Novem


.org/D

ownloaded from

http://iai.asm.org/

ANALYSIS OF PSEUDOMONAS VACCINE PEV 677

the second-dimension gel contained 4.72 pl1 of rabbit anti-PEV immunoglobulin per ml (equivalent to 0.78 mg ofprotein per ml). Gels were washed, pressed, and dried, andthe immunoprecipitates were stained with Coomassie bril-liant blue as described previously (42a, 55). The method ofprecipitate excision and analysis used to determine thenature of protein antigens contained within radiolabeled CIEimmunoprecipitates has been detailed elsewhere (42, 42a).

SDS-polyacrylamide gel CIE was performed essentially asdescribed by Chua and Blomberg (8) with the followingmodifications. In the first dimension, duplicate samplescontaining 75 ng of KDO were subjected to electrophoresisuntil the dye front reached 9.0 cm into the separating gel.Half of the polyacrylamide gel was then fixed and stained forcarbohydrate (16). The other half was washed three times in200 ml of distilled water (3 times for 15 min each) and oncein 200 ml of CIE running buffer (15 min) before excision ofthe individual lanes (width, 6 mm) for immunoelectrophore-sis into antibody. In the second dimension, immunoelectro-phoresis was performed on glass plates (106 by 80 by 2 mm)by using the Triton X-100 barbital buffer system routinelyused for CIE (see above). Polyethylene glycol was omittedfrom the antibody gel, which contained 11.32 p.l of rabbitanti-PEV immunoglobulins per ml (1.9 mg of protein per ml).After electrophoresis at 4 V/cm for 16 to 18 h, gels weresoaked for several hours in distilled water before pressing,washing, drying, and staining with Coomassie brilliant blue(42).SDS-polyacrylamide gel electrophoresis with 12.5%

(wt/vol) polyacrylamide separating gels (32) and 15 proteinmolecular weight standards (43) and procedures forfluorography (4) and for silver staining for protein (41) andcarbohydrates (16) were performed as described previously.Western blotting was done by the procedure of Burnette (5)with BA85 nitrocellulose membrane filters (pore size, 0.45p.m; Schleicher & Schuell, Dassel, Federal Republic ofGermany) and either standard-sized or mini gels (MarysolInd. Co. Ltd., Tokyo, Japan). Peroxidase-conjugated goatanti-rabbit IgG (Miles Laboratories) was used as secondantibody, and reacting antigens were visualized by incubat-ing blots in freshly prepared solutions containing 18 ml of 50mM Tris hydrochloride (pH 7.5), 2 ml of 0.6% (wt/vol)4-chloro-1-napthol in methanol, and 10 p.l of 30% (100volumes) H202.

Electron microscopy. Samples were negatively stainedwith 1% (wt/vol) potassium phosphotungstate (pH 7.0) andexamined in a Hitachi HU-12A electron microscope.

Analytical procedures. Protein content was determined bya modified Lowry procedure (35) with bovine serum albuminas the standard. KDO was assayed by the thiobarbituric acidmethod after hydrolysis of samples in 0.1 N sulfuric acid for30 min (30). Carbohydrate was estimated by the phenol-sulphuric acid procedure with a glucose standard (15).Nucleic acid was estimated from A2%0 by using an extinctioncoefficient (E260) for a 1-mg/ml solution of 50.8/cm (13).

RESULTS

Biochemical analysis of monovalent extracts. Chemicalanalyses revealed the presence of various amounts of totalcarbohydrate, KDO, and protein in all 16 extracts (Table 1).Calculations based on KDO content indicate that about 20%of the cellular LPS was removed from IATS serotype 6 bythe extraction procedure. The presence of high levels of LPSwas further substantiated by analysis of the extracts bySDS-polyacrylamide gel electrophoresis. When gels were

TABLE 1. Carbohydrate and protein contents of the polyvalentvaccine and monovalent extracts

Extract Protein Total carbohydrate KDO ([ug/ml(pLg/mi) (Rg/mi)'1 38.5 16.0 4.42 71.5 19.0 5.63 105.6 27.6 6.64 51.0 29.5 3.85 55.1 15.5 5.26 50.9 21.6 4.17 41.1 18.2 4.38 73.8 19.3 5.89 58.3 17.6 3.510 95.8 25.6 6.711 97.5 33.8 5.112 127.9 21.0 8.813 60.8 29.0 3.214 107.5 21.9 4.415 59.3 41.7 7.316 89.8 11.4 2.7PEV 72.2 24.26 5.6

" This value does not include amino sugars, which are not detected by theprocedure used to quantitate total carbohydrate (15).

stained for carbohydrate, banding patterns characteristic ofheterogeneously sized LPSs (19) were observed with 15 ofthe 16 extracts (Fig. 1A). The remaining extract (extract 15)contained only two resolvable species. The species close tothe track dye was presumed to correspond to core-lipid A.Whether the other high-molecular-weight component corre-sponds to uniformly sized LPS, free 0-antigen chains, or apolymer unrelated to LPS is unclear at present. The bandingpatterns observed for most of the extracts were unique, andseveral extracts (e.g., extracts 2, 5, 10, and 13) evidentlycontained groups of LPS species which fell into two or mnorediscrete size ranges. These highly characteristic profilesappeared to be true reflections of the sizes of the 0-antigenrepeats (19) and not artifacts caused by LPS oligomers or bythe presence of nucleic acid, since they were unaffected bypretreatment of the extracts with 40 mM EDTA or withnucleases or by increasing the concentration of SDS in thepolyacrylamide gels from 0.1 to 0.5% (wt/vol). Furthermore,for those serotypes examined (IATS serotypes 1, 2, 5, 6, 8through 10, 13, 15, and 17), the characteristic banding profileobserved for each monovalent extract correlated almostprecisely with that obtained for LPS purified from lyophi-lized cells of the corresponding serotype (not shown). Formost extracts, the dominant species of LPS had between 10and 35 0-antigen repeating units. The similarities in thebanding profiles for extracts 7 and 8 may in part reflect theclose chemical and antigenic structure of the LPS of thesetwo serotypes (7, 34; see the legend to Fig. 5).The protein profiles of all 16 monovalent extracts were

also analyzed by SDS-polyacrylamide gel electrophoresis(Fig. 1B). A major protein with a molecular weight of 16,200was resolved for 15 of the 16 extracts. It could also bedetected in the remaining extract (extract 7) at increasedloadings of sample. At least half of the extracts also con-tained a second protein with a molecular weight of approx-imately 21,400, and most extracts contained several otherminor proteins. From a consideration of protein- and carbo-hydrate-stained gels (Fig. 1A and B), neither the 16.2- northe 21.4-kilodalton protein appeared to be glycosylated. Aband located close to the dye front was also consistentlystained by the protein silver staining procedure (Fig. 1B).

VOL. 51, 1986

on Novem


.org/D

ownloaded from

http://iai.asm.org/


6

Ia

- 21 4

* 1Iw Jll ..--16 2

FIG. 1. Carbohydrate and protein profiles of the monovalent extracts. Dialyzed monovalent extracts 1 through 16 were subjected toSDS-polyacrylamide gel electrophoresis, and the fixed gels were stained with silver for carbohydrate (A) and protein (B). For carbohydrateanalysis, the applied samples contained 0.7 ,ug of carbohydrate (0.09 to 0.29 Fg of KDO); for protein analysis, samples contained 2.0 ,ug ofprotein. Extracts 1 through 16 are analyzed in numerical order as indicated. The banding profile characteristic of smooth LPS becomes moreevident for extract 16 (panel A, lane 16) at higher loadings of carbohydrate. The apparent sizes of salient polypeptides are indicated inkilodaltons at the side of the figure.

1 2 3 4 56

616:55.4

485_...

41-

30-*25-o..

20.*

141 ''4 AS

1.3-.;S,g,<-

FIG. 2. Comparison by SDS-polyacrl_r _-- ___ I 1 le .r a la

However, as this band was stained much more intensely forcarbohydrate (Fig. 1A) and was not stained by Coomassie

7 8 9 10 11 brilliant blue, it was assumed more likely to be carbohydrateor lipid in nature than protein.

In addition to the above well-resolved components, themonovalent extracts also contained various amounts of a

em +*..........high-molecular-weight component which remained unre-x<^ * ...solved towards the top of the separating gel. This material- ^,,,, _..................wasstained for both protein (Coomassie blue and silver) and

i3 .... carbohydrate (Fig. 1A and B) and, unlike the 16.2-kilodalton,,m X.protein, could be degraded in part by pronase (data not

_a ;............shown). To ascertain whether this unresolved material wasrelated to treatment of vaccine extracts with Formalin and toassess the likely identity of proteins, unformalinized extractswere prepared from serotype 6 organisms, and their poly-peptide profiles were compared with those of formalinized

,11101111W preparations and with those of fractionated membranes,flagella, and pili (Fig. 2). Freshly prepared extract 6 dis-played a polypeptide profile considerably more complex

*. Fthan that observed for similar extracts that were stored inFormalin, and the unresolved material observed for thelatter preparation can be attributed to Formalin-inducedpolymerization of most of these proteins (with the apparentexception of the 16.2- and 21.4-kilodalton proteins; Fig. 2,lanes 3 and 4). Similar results were observed for otherunformalinized extracts examined (data not shown). The

ylamide gel electrophoresis dominant 16.2-kilodalton polypeptide of monovalent extractof formalinized and unformalinized extract 6 and of varioussubcellular fractions derived from P. aeruginosa IATS serotype 6.Lane 1, Molecular weight standards; lane 2, supernatant fractionobtained after Silverson blending of cells and centrifugation at105,000 x g for 3 h (35 jLg of protein); lanes 3 and 8, unformalinizedextract 6 (20 p.g of protein); lane 4, formalinized extract 6 (19 ,ug ofprotein); lane 5, formalinized extract 6 (19 ,g of protein) plus crudepili preparation (10 p.g of protein); lane 6, crude pili preparation (10,ug of protein); lane 7. flagella (2.5 ,ug of protein); lane 9,

purified outer membranes (20 p.g of protein); lane 10, purified outermembranes (20 ,ug of protein) plus unformalinized extract 6 (20 ,ug ofprotein); lane 11, purified inner membranes (20 ,ug of protein). Sizesof the protein standards are indicated in kilodaltons at the side of thefigure.

INFECT. IMMUN.

O.. ..,

1 13 14 15 16

f,Idt: :.. Vi

i.lif.T?..I i:

i

.ii

on Novem


.org/D

ownloaded from

http://iai.asm.org/


. 10

A480

r 160

1Ju / 1U

FIG. 3. Analysis by CIE of the polyvalent P. aeruginosa vaccinePEV. A sample of PEV, 15 [LI containing 0.07 ,ug of KDO and 0.89,ug of protein, was subjected to CIE as described in Materials andMethods. Immunoprecipitates are numbered (in multiples of 10) inorder of decreasing electrophoretic mobility. Visualization of im-munoprecipitates formed by minor antigens 20, 40, and 41 is difficultat these loadings of antigen and antibody. The immunoprecipitatesformed by antigens designated 130/140 were heterogeneous anddifficult to resolve by CIE. Therefore, for the purpose of this study,they were considered as one (no. 130/140). The anode is to the leftand top of the gel.

6 was not sedimentable by centrifugation at 105,000 x g for3 h (Fig. 2, lane 2) and did not correspond to pilin, as judgedby coelectrophoresis experiments (Fig. 2, lanes 4 through 6)and by Westem blotting experiments conducted with anti-PAO or anti-PAK pili antibody (data not shown). In addi-

1 2

'-"N5

9

13

6

10

14

tion, flagella did not appear to be a major constituent of theextracts (Fig. 2, lanes 3, 4, and 7), although anti-flagellaantibody could be demonstrated in anti-PEV serum byWestern blotting (see Fig. 11). Comparison of the polypep-tide profiles of unformalinized extracts and isolated outermembrane fractions (Fig. 2, lanes 8 through 10) and studiesof heat modifiability (not shown) suggested that some vac-cine proteins, notably those with apparent molecular weightsof 74,000, 46,000, and 24,500, were probably outer mem-brane in origin and that the 24,500-molecular-weight proteinprobably corresponded to protein G* described by Hancockand Carey (20).

Analyses of PEV and monovalent extracts by CIE. Theantigenic compositions of PEV and all 16 monovalent ex-tracts were analyzed by CIE. The standard pattern routinelyobtained with the polyvalent vaccine, together with thenumber assigned to each resolved immunoprecipitate, isshown in Fig. 3. Thus, 10 major antigens (no. 50 through 140)and 5 minor antigens (no. 10 through 41) were identified inPEV. Few additional antigens were detected by manipula-tion of the antibody concentration over an eightfold range,by using agarose of lower endoosmotic flow (-mr < 0.1) orby screening for cathodally migrating antigens (not shown).In contrast to PEV, the monovalent extracts had extremelysimple CIE profiles, displaying, under standard conditions,only one or two immunoprecipitates (Fig. 4). The immuno-precipitates formed by the major antigen in at least 10 of theextracts (extracts 1 through 3, 5 through 8, 10, 13, and 16)had similar overall morphologies in their possession of ananodal shoulder or double peak. After application of four to

3

7

11

15

4

8

12

16

/ I%...FIG. 4. CIE profiles of monovalent extracts 1 through 16 (panels 1 through 16, respectively). Volumes of antigen applied were as follows:

15 pJ, extracts 4, 9, and 11; 12 ,ul, extract 14; 10 p.J, extracts 12 and 16; 2 ,ul, all other extracts. Loadings corresponded to 6.4 to 88 p.g of KDOand 0.08 to 1.1 mg of protein. Each extract number corresponds to the IATS serotype from which the extract is derived, with the exceptionthat extract 14 is obtained from IATS serotype 17. The anode is to the left and top of all immunoplates.

VOL. 51, 1986

s.E

on Novem


.org/D

ownloaded from

http://iai.asm.org/


C110

60

50

110

130/140

FIG. 5. Origin of antigens 50 through 140 established by coelectrophoresis. Monovalent extracts 1, 2, 10, 11, and 16 (2 to 10 ,ul) wereanalyzed by CIE either alone (Fig. 4) or mixed with 13 Kl of PEV (B through F, respectively). The resulting immunoprecipitin profiles werethen compared to the control ClE profile of the polyvalent vaccine alone (A). In each case, after coimmunoelectrophoresis, the peak heightof a single (arrowed) immunoprecipitate of PEV increased, thus identifying the origin of the corresponding antigen. After similar experiments(data not shown) were conducted with the remaining monovalent extracts, the origins of all major resolved antigens were established asfollows: antigen no. 50, extracts 2 and 16; antigen no. 60, extract 10; antigen no. 70, extract 13; antigen no. 80, extract 6; antigen no. 90, extract5; antigen no. 100, extracts 7 and 8; antigen no. 110, extract 1; antigen no. 120, extract 3; antigen no. 130/140, extracts 4, 9, 11, 12, 14,and 15.

eight times more antigen, several additional immunoprecip-itates, notably one designated no. 41, could be detected withall 16 extracts (see Fig. 9B). The profiles of CIE immuno-precipitates resolved for unformalinized extracts resembledthose of the corresponding formalinized vaccine extract andwere dominated by the same analogous major immunopre-cipitin band (see Fig. 9A).The origins of major antigens in PEV were determined by

coimmunoelectrophoresis of PEV with each individual ex-tract. A few representative examples are shown in Fig. 5,and the results for all 16 extracts are summarized in thelegend to that figure. Thus, it was demonstrated that six ofthe extracts (extracts 1, 3, 5, 6, 10, and 13) each contributedto PEV an immunologically distinct antigen. In contrast, themajor antigenic components derived from extracts 2 and 16cross-reacted, as did those from extracts 7 and 8. Interest-ingly, the major antigen of extracts 2 and 16 not onlycross-reacted with antigen no. 50 of PEV but also affectedthe height of the immunoprecipitate (no. 90) formed byextract 5 (Fig. 5). This may indicate the existence of com-mon determinants between these components. Antigen no.130/140 was apparently derived from six different extracts,i.e., extracts 4, 9, 11, 12, 14, and 15. This immunoprecipi-tate, however, was heterogeneous in nature and difficult toresolve with clarity. Further analysis of pairs of theseextracts by coimmunoelectrophoresis and immunodiffusionindicated, at best, reactions of only partial identity betweentheir major antigens.To ascertain whether a similar antibody response is elic-

ited against PEV in humans, the polyvalent vaccine wasanalyzed by CIE with an intermediate gel containing humananti-PEV immunoglobulins. All of the major antigens pre-cipitated by rabbit anti-PEV immunoglobulins also reactedto some extent with human immunoglobulins (Fig. 6).

Identification and properties of antigens resolved by CIE.The effect of protease, detergent, and heat treatment on the

CIE profile of PEV is shown in Fig. 7. Analysis indicatedthat antigens 50 through 120 were largely resistant to Strep-tomyces griseus protease (Fig. 7B). The anodic shoulders ofthe immunoprecipitates did disappear after protease treat-ment, but this was paralleled by an increase in the peakheights of the main immunoprecipitin arcs and may reflect anassociation between the major antigens and some minorprotease-sensitive vaccine component. In contrast, the anti-gen complex designated no. 130/140 was apparently affectedby protease (Fig. 7B). When Triton X-100 was omitted fromthe buffer system, individual immunoprecipitates were notresolved (Fig. 7C), indicating that these antigens are likely tobe membrane derived. In addition, antigens 50 through 120were shown to be remarkably resistant to heat. After exten-

A B

FIG. 6. Analysis of human anti-PEV immunoglobulins by CIEwith intermediate gel. PEV (0.07 pLg of KDO, 0.89 p.g of protein) wasanalyzed by CIE with a main reference gel containing rabbitanti-PEV immunoglobulins (4.72 pl/ml) and either no antibody (A)or 125 ,ul of human anti-PEV immunoglobulins per ml (B). Allimmunoprecipitates resolved in panel A are depressed to someextent in panel B, indicating their reaction with the human immu-noglobulins. The anode is to the left and top of both gels.

A B

D E

A ^

F

130/1440.._.x.0

INFECT. IMMUN.

on Novem


.org/D

ownloaded from

http://iai.asm.org/


sive boiling, they still retained their ability to bind antibody,although they evidently underwent some physical or chem-ical alteration which affected their electrophoretic mobilitiesin agarose (Fig. 7D). Furthermore, this ability to bindantibody was unaffected by extended incubation (18 h at20°C) of boiled samples with any one of three differentproteases, viz., S. griseus protease, staphylococcal V8 pro-tease, and trypsin (not shown).As the properties of antigens 50 through 120 (viz., their

serotype specificity [Fig. 5] and heat and pronase resistance[Fig. 7]) are all reminiscent of LPS, the relationship betweenresolved antigens and LPS purified from selected serotypeswas examined by CIE. LPS preparations from serotypes 1,2, 5, 6, 8, 10, and 13 were each shown to contain a majorantigenic component which shared immunological identitywith antigens no. 110, 50, 90, 80, 100, 60, and 70, respec-tively (Fig. 8). The anodic shoulders of the immunoprecipi-tates, however, were consistently less prominent for theLPS preparations than for the vaccine extracts. Outer mem-branes isolated from P. aeruginosa serotype 6 were alsoanalyzed by CIE with rabbit anti-PEV immunoglobulins.One major antigen was precipitated, and this was alsoidentified as LPS (not shown).The identity of the minor antigen (no. 41) detected for all

monovalent extracts at elevated loadings (Fig. 9A and B)was established by the method of precipitate excision andanalysis (42, 42a). Thus, CIE analysis of unformalinized[35S]methionine-labeled extract 6, followed by excision ofthe relevant immunoprecipitate (Fig. 9B and C) and analysisby SDS-polyacrylamide gel electrophoresis andfluorography (42, 42a), identified a polypeptide with a mo-lecular weight of 16,200 as the sole protein constituent of theantigen (Fig. 9D, lane 2). Identical results were obtained forthe corresponding antigen present in unformalinized[35S]methionine-labeled extract 12 (not shown). By the sameprocedures, two other minor antigens (x and y) present inextract 6 (Fig. 9B) were also identified. Antigen x containeda 42-kilodalton polypeptide (Fig. 9D, lane 3), the presence of

A B

C D

FIG. 7. Effect of protease, detergent, and heat treatment on theCIE profiles of PEV. The polyvalent vaccine was analyzed by CIEin the presence (A, B, and D) or absence (C) of Triton X-100 andafter either treatment with S. griseus protease (20 ,ug/ml) at 37°C for2 h and subsequent dialysis for 4 days against 10 mM Tris hydro-chloride (pH 7.5) containing 0.02% sodium azide (B) or heating at100°C for 2.5 h (D). The arrows indicate regions of the immunopre-cipitates affected by protease. The anode js to the left and top ofall gels.

1A 2A

1B 2B

AV110 #50

5A

5B,#90

_/6A 8A

6B 8B

Jr80

10A

10Bt100 ,060

FIG. 8. Identity of antigens 50 and 60 through 110 as LPS.Purified LPSs from IATS serotypes 1, 2, 5, 6, 8, and 10 wereanalyzed by CIE at loadings of 185, 187, 161, 90, 246, and 91 ng,respectively (panels A) and by coelectrophoresis with the corre-sponding monovalent extract (panels B) with the same loadings ofLPS and the following loadings of the monovalent extracts: 2, 2, 1.7,1, 2.7, and 0.9 Jl, respectively. Individual monovalent extracts wererun in parallel at identical loadings (data not shown; see Fig. 4 forgeneral profiles). Loadings of monovalent extract were adjusted togive peak areas of 50 to 100% of those generated by the correspond-ing LPS serotype. The appearance, after coelectrophoresis (panelsB), of single immunoprecipitates of increased area over that ob-served after CIE of LPS (panels A) thus establishes the identity ofnumbered antigens in PEV with LPSs of different serotypes. Thepanel number indicates the extract and, thus, the IATS serotypeunder investigation. Similar results (not shown) indicated that LPSpurified from IATS serotype 13 corresponded to antigen no. 70 ofextract 13. The standard loading of rabbit anti-PEV immunoglobulin(4.72 ,ud/ml) was doubled in the case of serotype 6 which generateda somewhat fainter LPS immunoprecipitate (Fig. 4). The anode is tothe left and top of each immunoplate.

the 16.2-kilodalton protein in this preparation being attrib-uted to antigen entrapment (42, 42a). Antigen y, whichformed an incomplete immunoprecipitate, contained a singleprotein (probably protein G*) with a molecular weight of24,500 (Fig. 9D, lanes 5 and 6).

Analysis of monovalent extracts by SDS-polyacrylamide gelCIE. Antigens which retain or regain their antigenicity afterSDS-polyacrylamide gel electrophoresis may be detected byelectrophoresis in a second dimension into an antibody-containing gel. By using this technique it was demonstratedthat the rabbit anti-PEV serum contained high levels ofantibody directed against the major carbohydrate compo-nent of each monovalent extract tested (viz., extracts 2, 4, 6,

VOL. 51, 1986

1.

on Novem


.org/D

ownloaded from

http://iai.asm.org/


A D 1 2 3 4

80

B80o'

41O

x

'-IIc i'\%I

:1

z-- '%-|lI I

I-

-

FIG. 9. Minor antigens of extract 6 and their iprecipitate excision and polypeptide analysis.[35S]methionine-labeled extract 6 (1-,ul [A] and 16-was analyzed by CIE against rabbit anti-PEV immunand 4.72 pJl/ml, respectively). Immunoplates in panestained with Coomassie brilliant blue. Panel C shovgram of the immunoplate in panel B. Antigen noclearly the dominant antigen for the unformalinizeA). Minor antigens (no. 41, x, and y), detected at elof extract (panel B), were located by autoradiograpdried gels and marked for excision (panel C). Patanalysis of excised immunoprecipitates by SDS-polelectrophoresis and fluorography. Lanes 1 and 6,labeled extract 6 (0.3 p.l); lane 2, immunoprecipitatimmunoprecipitate x; lane 4, immunoprecipitateimmunoprecipitate y. Note that labeled polypeptdetected for immunoprecipitate 80 (LPS). Theproteins are indicated in kilodaltons at the side of t

7, 10, 11, and 13 through 15). The profiles obtof these extracts are shown in Fig. 10. In eposition and height of the immunoprecipitateprecisely to that predicted from the position a

the polyacrylamide gel of the bands of smoothpolysaccharide in the case of extract 15).immunoprecipitate corresponding to rough LIthe protein components could be detectedemphasizing that these results demonstrate th(viz., extracts 4, 11, 14, and 15) of the siextracts which contributed to the ill-resolvedprecipitate no. 130/140 (Fig. 3) possessed LPcharides which readily formed immunopagarose.

Analysis by Western blotting. Analysis ofvalents extracts by Western blotting with ra

serum generated a profile almost identical toafter silver staining of the extracts for carb1A). For each serotype, by far the most domicould be attributed to heterogeneously sized Lbands could be discerned among the intenselspecies (not shown). Since it was import-establish whether antibodies to other Pseudon

were present, anti-PEV serum was exhaustively adsorbed5 6 with LPS-6. The adsorbed serum was then used in Western

blotting experiments to detect reactive antigens in severalcellular fractions derived for IATS serotype 6 (Fig. 11).Strong reactions were observed for flagella (Fig. 11, lane 6),the protein with a molecular weight of 16,200 (lane 8), andfor three outer mermbrane components with molecularweights of approximately 38,000, 19,000, and 10,000 (lane 4).The two outer membrane proteins of highest molecular

-o-42 weight we assume to be proteins F and H1/H2 (20). Theidentity of the 10-kilodalton component is unclear, but it maycorrespond to protein I or to glycolipid. Several otherreacting proteins in the 53- to 42-kilodalton range weredetected for the Silverson supernatant fraction (Fig. 11, lane

_.-24-5 8). Some of these may be related to flagella and to degrada-tion products thereof. It is notable that reactions attributableto pili (Fig. 11, lane 7) or to inner membrane components(lane 6) could not be discerned. Furthermore, analysis of alysis supernatant containing soluble components of the cell

4j+416.2 revealed that, of over 55 proteins resolved in this fraction(data not shown), only about 9 reacted with antibody (Fig.11, lane 9). Of these, only the three with molecular weightsin excess of 53,000 could not be attributed to either an outermembrane (lane 4) or a surface origin (lane 8).

It is notable that, of the reacting species documentedidentification by above, only proteins with molecular weights of 42,000,Unformalinized 16,200, and 10,000 featured prominently in reactions for,ul [B] samples) extract 6 (Fig. 11, lanes 2 and 3). Significantly, both the 42-oglobulins (18.9 and 16.2-kilodalton proteins fractionated into surface ex-Is A and B were tracts and not with outer membrane (Fig. 2 and 11). Thevs an autoradio- diffuse staining reactions observed for formalinized but noti.80 (LPS) was unformalinized extract 6 probably represent interaction of

ldeatedct (padingl antibody with polymerized components of the extract.hy on unstained DISCUSSIONnel D shows anIyacrylamide gel For the following reasons, there seems little doubt that[35S]methionine- LPS is the major antigenic constituent of most if not all of thee no. 41; lane 3, 16 monovalent extracts which together constitute the Pseu-no. 80; lane 5, domonas vaccine PEV. First, as judged by their KDOtides cannot be content (Table 1) and carbohydrate profile (Fig. 1A), allsizes of salient extracts contained roughly comparable concentrations ofthe fluorogram. LPS. Second, the major antigen observed for each of the

monovalent extracts 1, 2, 5, 6, 8, 10, and 13 shared immu-ained for eight nological identity with LPS purified from the corresponding-ach case, the IATS serotype (Fig. 8). Third, the results of SDS-corresponded polyacrylamide gel CIE indicated that LPS was a majorLnd intensity in antigenic constituent for extracts 2, 4, 6, 7, 10, 11, 13, and 14LPS (or other (Fig. 10). Western blotting confirmed these observations andLittle or no extended them to include all extracts with the possible

PS or to any of exception of extract 15. In addition, the major antigensl. It is worth resolved for PEV by CIE displayed properties of heat andat at least four protease resistance anticipated for LPS (Fig. 7). Finally, theix monovalent major antigens resolvable by CIE for the individual extractsCIE immuno- showed immunological interrelationships comparable to'Ss or polysac- those established for the corresponding IATS 0-serotypesrecipitates in (7, 34).

Although LPS was clearly the dominant antigen for mostall 16 mono- of the monovalent extracts, other antigens were present.

tbbit anti-PEV Notable among these was a 16.2-kilodalton protein whichthat obtained was common to all extracts and which generated CIE

ohydrate (Fig. immunoprecipitate no. 41. The identity of this protein isinant reactions unclear at the present time. However, from the following,PS. Few other considerations, a relationship with pili (17, 44) appearsly labeled LPS unlikely. First, cells were grown under conditions (36) whichant to clearly usually suppress the production of pili. Moreover, neithernonas antigens formalinized nor unformalinized extracts showed evidence

INFECT. IMMUN.

on Novem


.org/D

ownloaded from

http://iai.asm.org/


2

5

1t

13

A

7 14

- aI

10 15

urnLI

it w

+-- SDS-PAGEFIG. 10. Analysis of monovalent extracts by SDS-polyacrylamide gel CIE. Monovalent extracts were analyzed as described in Materials

and Methods. Control lanes from the first-dimension polyacrylamide gel were stained for carbohydrate and have been aligned in the correctposition with respect to the second-dimension CIE gel. The anode is to the left and the top of all gels. In each case, numbers refer to theextract under investigation.

of pilus structures when viewed in the electron microscope.Second, the 16.2-kilodalton protein had a similar size in all16 extracts (Fig. 1B) but was clearly distinguishable fromPAO pilin by SDS-polyacrylamide gel electrophoresis (Fig.2). Third, as judged by immunoblotting, anti-PEV serumrecognized the 16.2-kilodalton proteins but not PAO pilin(Fig. 11). Conversely, antiserum raised to PAO pili reactedwith PAO pilin but not with the 16.2-kilodalton proteins. Inaddition, the 16.2-kilodalton proteins were not recognized byantiserum raised to PAK pili.The properties of this 16.2-kilodalton protein and also the

21.4-kilodalton polypeptide, viz., their presence in all ormany of the serotypes, their ready removal from the cellsurface by mild extraction or by homogenization, theirapparent absence from purified outer membranes, theirretention in supernatant fractions after centrifugation at105,000 x g, and the apparent absence of sugar residues,may suggest a relationship with (the proteinaceous moietyof) the slime glyco(lipo) proteins which have been describedfor this organism (2, 14, 40, 54). An alternative candidate isthe so-called original endotoxin protein described byTanamoto et al. (56). This protein has a molecular weight ofabout 22,000 and is reported to be a common surface antigenfor most serotypes of P. aeruginosa (9, 56).The antigenic and molecular composition of all 16

monovalent extracts appeared to remarkably simple, a phe-nomenon which can be attributed, at least in part, to the

selective removal of a limited number of cell-surface anti-gens during extraction with EDTA-glycine buffer. This wasnot unexpected. Hedstrom et al. (25) found release of LPSand outer membrane proteins D, E, G, and Hi but not F, H2,or I on exposure of sucrose-stabilized P. aeruginosa toTris-EDTA, and Hancock et al. (21) have documented therelease of LPS and protein E under similar conditions. Theproclivity of the organism to shed parts of its outer mem-brane is also underscored in a recent report documenting thespontaneous release of LPS during growth in defined me-dium (6). However, the simplicity of the polypeptide profilesobserved for this vaccine belie to some extent the complex-ity of the original monovalent extracts. Analyses of freshlyprepared unformalinized extracts by SDS-polyacrylamidegel electrophoresis, CIE, and Western blotting clearly sug-gest the presence of outer membrane proteins and flagellawithin the original extracts. Although these do not seem tobe major constituents of the vaccine, some, e.g., flagella andproteins F and H1/H2, appear capable of eliciting significantantibody responses (18, 24, 27). As a consequence, thepossible contribution of these components to the protectionafforded by PEV should not be ignored.

Formalinization appeared to impair analysis of the ex-tracts by causing aggregation of some of the protein compo-nents. Aggregation is slow, but it becomes a significantproblem after a few months of storage at 4°C (unpublishedobservations). The effect can probably be attributed to

.. I." 11111-

I t I F

4 I .4,

VOL. 51, 1986

..-.ANMIMIML

J,dl"- wmmwk-._.

..7..

I "A'.

on Novem


.org/D

ownloaded from

http://iai.asm.org/


1 2 3 4 5 6 7

:~~~ ~*-

m.

FIG. 11. Western blot analysis, showing t6 and of various subcellular fractions deriv4IATS serotype 6 with LPS6-adsorbed anti-ILane 1, Purified LPS from serotype 6 (0.9 lijextract 6 (2 jig of protein); lane 3, unformalirprotein); lane 4, purified outer membranes (2purified inner membranes (2 ,ug of protein); laprotein); lane 7, crude pili preparation from P(1 ,ug of protein); lane 8, supernatant fiSilverson blending of cells and centrifugatior(1.8 Kg of protein); lane 9, supernatant fr100,000 x g) from French-pressed cells (1.6 ,uno reaction can be detected for pilin, theserved for that fraction (lane 7) being attribiweak reaction observed for the 16.2-kilodaltatypical and probably reflects decreased leveparticular extract. The approximate sizes inthe prominent reacting bands are indicateimmunoblot.

8 9 LPS can form immunoprecipitates of a morphology verysimilar to that observed here for antigen 130/140 (26, 33). Inthe present instance, we suspect that antigen 130/140 repre-sents an unresolved complex of either LPS species or of LPSand protein molecules derived from the six extracts inquestion. Formaldehyde-induced cross-linking of LPS andprotein molecules may play a role here, as it might in the

-' 4-53 formation of the protease-sensitive anodal leg of some of theLPS antigens when resolved by CIE.As severe adverse reactions (including malaise, fever, and

localized induration and pain) were observed after adminis-tration of another LPS-based Pseudomonas vaccine (1, 45),it may appear remarkable that PEV-01 contains LPS as themajor antigen but yet elicits few toxic side effects (52). Thereason for this apparent discrepancy most likely relates to

1.09 differences in the quantities of LPS admninistered. Assuming,as we have found, that KDO represents 2.0 to 4.8% of the

.4-16 2 dry weight of P. aeruginosa LPS (see also references 12 and31), then one human dose of PEV contains at most about 30

*10 jig of LPS. In contrast, approximately 1.75 mg of LPS wasadministered to 70-kg patients with each dose of thehe reactions of extract heptavalent LPS vaccine Pseudogen (1). In additions, as theed from P. aeruginosa high-molecular-weight polysaccharide described by Pier andREV immunoglobulins. co-workers (46, 49) shares immunological identity with theg); lane 2, formalinized 0-antigenic side chains of LPS but contains no lipid A, it isnized extract 6 (2 jig of capable of eliciting an anti-LPS antibody response with noVtg of protein); lane 5, toxic side effects. Although this surface polysaccharide isme 6, flagella (0.3 ,g of approximately 1,000-fold less immunogenic than intact LPS. aeruginosa PAO/DB2 (11), the possibility that it may have a role in PEV-inducedraction obtained after immunity cannot be excluded and is now under investiga-raction (centrifugation tion. Certainly, a major antibody response elicited in humangofprotein) Noterthat volunteers immunized with PEV seemed to be directeddominant reaction oh- against the LPS components of this vaccine (Fig. 6). How-,utable to flagellin. The ever, it remains to be established whether LPS can accountton protein in lane 3 is for some or all of the protection afforded by PEV and whatls of the antigen in this role, if any, is played by other constituents of the vaccine.kilodaltons of some of Also, in the light of results presented herein, a carefuled at the side of the comparison of the relative efficacies and toxicities of two

vaccine products, PEV and Pseudogen, would appear to bewarranted.

reaction of the hydrated methylene glycol form of formalde-hyde [CH2(OH)21 with free primary amino groups in theproteins. The resultant aminomethylol derivatives can thenundergo further reaction with an amide, hydroxy, sulfhy-dryl, guanidinium, or imidazolyl function to generate pro-teins cross-linked by methylene bridges (53). An importantpractical consequence of these observations and the fact thatpolymerization is accelerated by heating at 100°C is thenecessity of removing excess Formalin by dialysis beforeanalysis on SDS-polyacrylamide gels; otherwise, additionalcross-linkage is effected.Another factor which may contribute to the time-

dependent reduction in complexity of stored vaccine ex-tracts is proteolytic degradation caused by trace amounts ofextracellular proteases. However, it should be stressed thatthe protective capacity of the (formalinized) vaccine isvirtually unaffected by prolonged storage (2 to 3 years) at 4or 37°C (unpublished observations).Why the major antigens present in extracts 4, 9, 11, 12, 14,

and 15 resolved as an ill-defined immunoprecipitate (no.130/140) during CIE of the polyvalent vaccine is not clear.The results of SDS-polyacrylamide gel CIE clearly showedthat, for four of these extracts (extracts 4, 11, 14, and 15),LPS-polysaccharide was a major antigen. Furthermore,other CIE studies ofPseudomonas preparations confirm that

ACKNOWLEDGMENTS

We thank W. Paranchych for supplying us with anti-PAO andanti-PAK pili serum, D. E. Bradley for providing the hyperpiliatedvariant DB2 of P. aeruginosa PAO, R. Cervi for preparing the pili,and Gillian Johnston for providing excellent secretarial service.Finally, we would like to express our gratitude to scientists atWellcome Biotechnology Ltd., notably J. F. Spilsbury and J. J.Miler, who were involved in the preparation of cells and of vaccineextracts, and to A. J. Beale for catalyzing our interest in this area andfor organizing the company's full financial support for the project.

LITERATURE CITED1. Alexander, J. W., and M. W. Fisher. 1974. Immunization against

Pseudomonas infection after thermal injury. J. Infect. Dis.130(Suppl.):152-158.

2. Bartell, P. F. 1983. Determinants of the biologic activity ofsurface slime in experimental Pseudornonas aeruginosa infec-tions. Rev. Infect. Dis. 5(Suppl.):971-978.

3. Bodey, G. P., R. Bolivar, V. Fainstein, and L. Jadeja. 1983.Infections caused by Pseudomonas aeruginosa. Rev. Infect.Dis. 5:279-313.

4. Bonner, W. M., and Laskey, R. A. 1974. A film detectionmethod for tritium-labelled proteins and nucleic acids in poly-acrylamide gels. Eur. J. Biochem. 46:83-88.

5. Burnette, W. N. 1981. "Western blotting": electrophoretic

INFECT. IMMUN.

on Novem


.org/D

ownloaded from

http://iai.asm.org/


transfer of proteins from sodium dodecyl sulphate-polyacryl-amide gels to unmodified nitrocellulose and radiographic detec-tion with antibody and radioiodinated protein A. Anal.Biochem. 112:195-203.

6. Cadieux, J. E., J. Kuzio, F. H. Milazzo, and A. M. Kropinski.1983. Spontaneous release of lipopolysaccharide by Pseudomo-nas aeruginosa. J. Bacteriol. 155:817-825.

7. Chester, I. R., P. M. Meadow, and T. L. Pitt. 1973. Therelationship between the 0-antigenic lipopolysaccharides andserological specificity in strains of Pseudomonas aeruginosa ofdifferent 0-serotypes. J. Gen. Microbiol. 78:305-318.

8. Chua, N.-H., and F. Blomberg. 1979. Immunochemical studiesof thylakoid membrane polypeptides from spinach and Chlam-ydomonas reinhardtii. A modified procedure for crossed-immunoelectrophoresis of dodecyl sulfate-protein complexes. J.Biol. Chem. 254:215-223.

9. Cryz, S. J., Jr. 1984. Pseudomonas aeruginosa infections, p.317-351. In R. Germanier (ed.), Bacterial vaccines. AcademicPress, Ltd. London.

10. Cryz, S. J., Jr., E. Furer, and R. Germanier. 1983. Protectionagainst Pseudomonas aeruginosa infection in a murine burnwound sepsis model by passive transfer of antitoxin A, anti-elastase, and antilipopolysaccharide. Infect. Immun.39:1072-1079.

11. Cryz, S. J., Jr., E. Furer, and R. Germanier. 1984. Protectionagainst fatal Pseudomonas aeruginosa burn wound sepsis byimmunization with lipopolysaccharide and high-molecular-weight polysaccharide. Infect. Immun. 43:795-799.

12. Darveau, R. P., and R. E. W. Hancock. 1983. Procedure forisolation of bacterial lipopolysaccharides from both smooth andrough Pseudomonas aeruginosa and Salmonella typhimuriumstrains. J. Bacteriol. 155:831-838.

13. Dawson, R. M. C., D. C. Elliot, W. H. Elliot, and K. M. Jones.1969. Data for biochemical research, p. 625-626. Oxford Uni-versity Press, Oxford.

14. Dimitracopoulos, G., and P. F. Bartell. 1980. Slime gly-colipoproteins and the pathogenicity of various strains of Pseu-domonas aeruginosa in experimental infection. Infect. Immun.30:402-408.

15. Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F.Smith. 1956. Colorimetric method for determination of sugarsand related substances. Anal. Chem. 28:350-356.

16. Dubray, G., and G. Bezard. 1982. A highly sensitive periodicacid-silver stain for 1,2-diol groups of glycoproteins andpolysaccharides in polyacrylamide gels. Anal. Biochem.119:325-329.

17. Frost, L. S., and W. Paranchych. 1977. Composition and mo-lecular weight of pili purified from Pseudomonas aeruginosa K.J. Bacteriol. 131:259-269.

18. Gilleland, H. E., Jr., M. G. Parker, J. M. Matthews, and R. D.Berg. 1984. Use of a purified outer membrane protein F (porin)preparation of Pseudomonas aeruginosa as a protective vaccinein mice. Infect. Immun. 44:49-54.

19. Goldman, R. C., and L. Leive. 1980. Heterogeneity of antigenic-side-chain length in lipopolysaccharide from Escherichia coli0111 and Salmonella typhimurium LT2. Eur. J. Biochem.107:145-153.

20. Hancock, R. E. W., and A. M. Carey. 1979. Outer membrane ofPseudomonas aeruginosa: heat- and 2-mercaptoethanol-modifiable proteins. J. Bacteriol. 140:902-910.

21. Hancock, R. E. W., R. T. Irvin, J. W. Costerton, and A. M.Carey. 1981. Pseudomonas aeruginosa outer membrane: pepti-doglycan-associated proteins. J. Bacteriol. 145:628-631.

22. Hancock, R. E. W., and H. Nikaido. 1978. Outer membranes ofgram-negative bacteria. XIX. Isolation from Pseudomonasaeruginosa PAO1 and use in reconstitution and definition of thepermeability barrier. J. Bacteriol. 136:381-390.

23. Hanessian, S., W. Regan, D. Watson, and T. H. Haskell. 1971.Isolation and characterization of antigenic components of a newheptavalent Pseudomonas vaccine. Nature New Biol.229:209-210.

24. Hedstrom, R. C., 0. R. Pavlovskis, and D. R. Galloway. 1984.Antibody response of infected mice to outer membrane proteins

of Pseudomonas aeruiginosa. Infect. Immun. 43:49-53.25. Hedstrom, R. C., R. K. Schockley, and R. G. Eagon. 1981.

Ethylenediaminetetraacetate-extractable protein-lipopolysac-charide complex of Pseudomonas aeruginosa: characterizationof protein components. J. Bacteriol. 148:995-997.

26. H$iby, N. 1975. The serology of Pseudomonas aeruginosaanalysed by means of quantitative immunoelectrophoreticmethods. I. Comparison of thirteen 0 groups of Ps. aeruginosawith a polyvalent Ps. aeruginosa antigen-antibody referencesystem. Acta Pathol. Microbiol. Scand. Sect. B 83:321-327.

27. Holder, I. A., R. Wheeler, and T. C. Montie. 1982. Flagellarpreparations from Pseudomonas aeruginosa: animal protectionstudies. Infect. Immun. 35:276-280.

28. Jones, R. J. 1979. Antibody responses in burned patientsimmunized with a polyvalent pseudomonas vaccine. J. Hyg.(Lond.) 82:453-462.

29. Jones, R. J., E. A. Roe, E. J. L. Lowbury, J. J. Miler, and J. F.Spilsbury. 1976. A new pseudomonas vaccine: preliminary trialon human volunteers. J. Hyg. 76:429-439.

30. Karkhaais, Y. D., J. Y. Zeltner, J. J. Jackson, and D. J. Carlo.1978. A new and improved microassay to determine 2-keto-3-deoxyoctonate in lipopolysaccharide of gram-negative bacteria.Anal. Biochem. 85:595-601.

31. Kropinski, A. M., L. C. Chan, and F. H. Milazzo. 1979. Theextraction and analysis of lipopolysaccharides from Pseudomo-nas aeruginosa strain PAO and three rough mutants. Can. J.Microbiol. 25:390-398.

32. Laemmli, U. K. 1970. Cleavage of structural proteins duringassembly of the head of bacteriophage T4. Nature (London)227:680-685.

33. Lam, J. S., L. M. Mutharia, R. E. W. Hancock, N. Hoiby, K.Lam, L. Baek, and J. W. Costerton. 1983. Immunogenicity ofPseudomonas aeruginosa outer membrane antigens examinedby crossed immunoelectrophoresis. Infect. Immun. 42:88-98.

34. Liu, P. V., H. Matsumoto, H. Kusama, and T. Bergan. 1983.Survey of heat-stable, major somatic antigens of Pseudomonasaeruginosa. Int. J. Syst. Bacteriol. 33:256-264.

35. Markwell, M. A. K., S. M. Haas, L. L. Bieber, and N. E.Tolbert. 1978. A modification of the Lowry procedure to sim-plify protein determination in membrane and lipoprotein sam-ples. Anal. Biochem. 87:206-210.

36. Miler, J. J., J. F. Spilsbury, R. J. Jones, E. A. Roe, and E. J. L.Lowbury. 1977. A new polyvalent Pseudomonas vaccine. J.Med. Microbiol. 10:19-27.

37. Montie, T. C., D. Doyle-Huntzinger, R. C. Craven, and I. A.Holder. 1982. Loss of virulence associated with absence offlagellum in an isogenic mutant of Pseudomonas aeruginosa inthe burned-mouse model. Infect. Immun. 38:1296-1298.

38. Mutharia, L. M., and R. E. W. Hancock. 1983. Surface local-ization of Pseudomonas aeruginosa outer membrane porinprotein F by using monoclonal antibodies. Infect. Immun.42:1027-1033.

39. Mutharia, L. M., T. I. Nicas, and R. E. W. Hancock. 1982. Outermembrane proteins of Pseudomonas aeruginosa serotypestrains. J. Infect. Dis. 146:770-779.

40. Nadaud, M. 1983. Effects of glycopolypeptide from the slime ofP. aeruginosa on phagocytosis by mouse macrophage.Zentralbl. Bakteriol. Mikrobiol. Hyg. Abt. 1 Orig. A 254:489-499.

41. Oakley, B. R., D. R. Kirsch, and N. R. Morris. 1980. Asimplified ultrasensitive silver stain for detecting proteins inpolyacrylamide gels. Anal. Biochem. 105:361-363.

42. Owen, P. 1983. Antigens of the Escherichia coli cell envelope,p.347-373. In 0. J. Bjerrum (ed.), Electroimmunochemicalanalysis of membrane proteins. Elsevier Science Publishers,Amsterdam.

42a.Owen, P. 1985. Crossed immunoelectrophoresis in the study ofouter membrane antigens, p. 207-212. In T. K. K. Korhonen,E. A. Dawes, and P. H. Makela (ed.), Enterobacterial surfaceantigens: methods for molecular characterization. Elsevier Pub-lishing Co., Amsterdam.

43. Owen, P., and C. Condon. 1982. The succinate dehydrogenaseof Escherichia coli: subunit composition of the Triton X-100-

VOL. 51, 1986

on Novem


.org/D

ownloaded from

http://iai.asm.org/


solubilized antigen. FEMS Microbiol. Lett. 14:223-227.44. Paranchych, P. A. S., L. S. Frost, M. Carpenter, G. D.

Armstrong, and T. H. Watts. 1979. Biochemical studies on piliisolated from Pseudomonas aeruginosa PAO. Can. J. Micro-biol. 25:1175-1180.

45. Pennington, J. E. 1974. Preliminary investigations ofPseudomo-nas aeruginosa vaccine in patients with leukemia and cysticfibrosis. J. Infect. Dis. 130(Suppl.):159-162.

46. Pier, G. B. 1983. Immunochemistry of Pseudomonas aerugi-nosa lipopolysaccharides and high-molecular-weight polysac-charides. Rev. Infect. Dis. 5(Suppl.):950-956.

47. Pier, G. B., M. Cohen, and H. Jennings. 1983. Further purifica-tion and characterization of high-molecular-weight polysaccha-ride from Pseudomonas aeruginosa. Infect. Immun. 42:936-941.

48. Pier, G. B., M. Pollack, and M. Cohen. 1984. Immunochemicalcharacterization of high-molecular-weight polysaccharide fromFisher immunotype 3 Pseudomonas aeruginosa. Infect. Immun.45:309-313.

49. Pier, G. B., and D. M. Thomas. 1983. Characterization of thehuman immune response to a polysaccharide vaccine fromPseudomonas aeruginosa. J. Infect. Dis. 148:206-213.

50. Pollack, M., and L. S. Young. 1979. Protective activity ofantibodies to exotoxin A and lipopolysaccharide at the onset ofPseudomonas aeruginosa septicemia in man. J. Clin. Invest.63:276-286.

51. Pollack, M., G. B. Pier, and R. K. Prescott. 1984. Immunizationwith Pseudomonas aeruginosa high-molecular-weight polysac-charides prevents death from Pseudomonas burn infections inmice. Infect. Immun. 43:759-760.

52. Roe, E. A., and R. J. Jones. 1983. Immunization of burnedpatients against Pseudomonas aeruginosa infection atSafdarjang Hospital, New Delhi. Rev. Infect. Dis. 5(Suppl.):922-930.

53. Schauenstein, E., H. Esterbauer, and H. Zollner. 1977. Al-dehydes in biological systems: their natural occurrence andbiological activities. Advanced biochemistry series, p. 1-24.Pion Ltd., London.

54. Sensakovic, J. W., and P. F. Bartell. 1977. Glycolipoproteinfrom Pseudomonas aeruginosa as a protective antigen againstP. aeruginosa infection in mice. Infect. Immun. 18:304-309.

55. Smyth, C. J., J. Siegel, M. R. J. Salton, and P. Owen. 1978.Immunochemical analysis of inner and outer membranes ofEscherichia coli by crossed immunoelectrophoresis. J. Bacte-riol. 133:306-319.

56. Tanamoto, K., C. Abe, Y. Homma, K. Kuretani, A. Hoshi, andY. Kojima. 1978. A compound possessing antitumour andinterferon-inducing activities derived from the common antigen(OEP) of Pseudomonas aeruginosa. J. Biochem. 83:711-718.

57. Westphal, O., and K. Jann. 1965. Bacterial lipopolyaccharides.Extraction with phenol-water and further applications of theprocedure. Methods Carbohydr. Chem. 5:83-91.

INFECT. IMMUN.

on Novem


.org/D

ownloaded from

http://iai.asm.org/

immunochemical and biochemical analysis of the polyvalent ... · polyvalent vaccine pevwas derived...

Documents