INFECTION AND IMMUNITY, Aug. 1982, p. 622-6310019-9567/82/080622-10$02.00/0

Vol. 37, No. 2

12I-Peptide Mapping of Protein III Isolated from Four Strainsof Neisseria gonorrhoeae

RALPH C. JUDDLaboratory of Microbial Structure and Function, Rocky Mountain Laboratories, National Institute ofAllergy

and Infectious Diseases, Hamilton, Montana 59840

Received 22 December 1981/Accepted 9 April 1982

Gonococcal outer-membrane protein I (PI) and PIII were isolated by sodiumdodecyl sulfate-polyacrylamide gel electrophoresis from reduced and unreducedwhole-cell and outer-membrane lysates of four strains of nonpiliated (P-),transparent (O-) Neisseria gonorrhoeae. These proteins were radioiodinated anddigested with a-chymotrypsin. The resultant 125I-peptides were then resolved byhigh-voltage thin-layer electrophoresis, followed by ascending thin-layer chroma-tography, and visualized by autoradiography. Results corroborated previousobservations regarding the structural relationships of PIs having different appar-ent subunit molecular weights. All PIIIs had very similar apparent primarystructures, regardless of the strain from which they were isolated, the source (i.e.,whole cells or outer membranes), or the reduction state of the sodium dodecylsulfate lysates. By the techniques used, it appeared that PIII is structurally similarin all of the gonococcal strains studied, even though each strain had structurallyunique PIs.

The gonococcal outer membrane (OM) con-tains several proteins which are exposed on thecell surface. These proteins have been classifiedinto several groups (28) on the basis of theiroccurrences and their behaviors in sodium dode-cyl sulfate (SDS)-polyacrylamide gel electropho-resis (PAGE). The major, or principal, OMprotein (10), protein I (PI), is present in allgonococci studied (15, 16, 20, 22). The apparentsubunit molecular weight (aMW) of PI variesfrom strain to strain (11, 22, 23) but shows novariation among different phenotypes within asingle strain (15, 16, 22, 31). The aMW of PI isnot altered by varying the solubilization tem-perature (5, 18, 24) or by the presence (orabsence) of 2-mercaptoethanol (2ME) in thesolubilization mixture (18).

In a previous study, the technique of 125[_peptide mapping (23) was used to investigate thestructures of three PIs of different aMWs isolat-ed from 10 gonococcal strains. A seemingly highdegree of structural homology was found for notonly PIs having the same aMW but also for PIshaving aMWs of 34,000 (34K) and 33K; several32K PIs were closely related to each other butdistinct from the 34K and 33K PIs. Immunopre-cipitation experiments have shown that rabbitantisera raised against OM (18) or whole gono-cocci (27) react with homologous PI in OMvesicles or on intact cells, respectively, suggest-ing that PI could contribute to serotype-specificreactions (3, 9, 11)

PIIs are a heterogeneous family of OM pro-

teins which can vary widely in aMW, presence,and number within a single strain (9, 16, 20, 22,31, 32). They are distinguished by an increase inaMW when solubilized in SDS at 100 versus56°C (i.e., heat modifiable [5, 9, 24, 31]). Thepresence of some of these proteins, which allhave somewhat similar primary structures asshown by 125I-peptide mapping (24), correlateswith gonococcal colony opacity (22). The gain orloss of PITs occurs at a high rate, necessitatingsingle-colony transfer on an appropriate mediumto maintain the desired phenotype (21). Theseproteins have extensive surface exposure (1, 5,6, 26) and have been associated with gonococcalaggregation (20), susceptibility of gonococcus tokilling by serum (2), and interactions with eu-caryotic cells (8, 13, 25, 29). Immunoprecipita-tion with antisera raised against whole gono-cocci has shown that rabbit immunoglobulin Gcan combine with PIIs in situ on intact orga-nisms (27), demonstrating that these proteins,like PIs, are immunogenic and antigenic andmay, therefore, contribute to serotype-specificreactions.

PIIIs (5, 18) are characterized by an increasein aMW when solubilized in the presence of 5 to8% 2ME as compared with solubilization with-out 2ME (i.e., 2ME modifiable) [18]. The aMWsof the unreduced (30K PIII) and the reduced(31K PIII*) forms of PIII are the same in allstrains and intrastrain phenotypes studied todate (27). PIIIs, unlike some PIs and all PITs, arevery resistant to exogenous proteolytic cleavage

622

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

GONOCOCCAL Plll PEPTIDE MAPPING 623

(1, 26) and are weakly iodinated by lactoperoxi-dase (5). lodogen (17), a smaller surface-reactivecatalyst, more efficiently radiolabels PIlIs (un-published data), suggesting that PIlIs may haveless surface exposure and perhaps less accessi-bility to surface-reactive agents than do PIs andPlls.

PIII has been shown to exist in heteropoly-mers with PI in whole cells (WCs) and OMvesicles by both cross-linkage studies (18, 19)and radioimmunoprecipitation (18, 27). Since PIand PIII seem to coimmunoprecipitate, it isdifficult to establish whether antibodies reactdirectly with PIII or whether PIII is merely an"innocent bystander" immunoprecipitated byimmunoglobulin G-PI complexes. Since PIlIsappear to be found in all of the strains andphenotypes studied to date, seem to be exposedon the gonococcal surface, and are intimatelyassociated with PI, they may contribute eithercross-reactive serotypic antigens, if alike, orstrain-specific antigens, if different, to the vari-ous immunological reactions described for gono-cocci and gonococcal OM vesicles (17).

In this study, the structures of Pllls isolatedfrom four gonococcal strains, chosen because thePI of each strain has a different aMW, wereinvestigated by the technique of 125I-peptide map-ping. PI and PIII were obtained from each strainby SDS-PAGE of unreduced or reduced lysatesfrom WC and OM preparations. The proteinbands were excised, radioiodinated, and digestedwith a-chymotrypsin. Resultant 125I-peptideswere resolved by high-voltage electrophoresis,followed by ascending thin-layer chromatography(TLC). The migration of the 1251-peptides wasthen visualized by autoradiography, yielding char-acteristic "fingerprints" for the proteins underinvestigation. Results confirmed that there aretwo PI structural homology groups, one contain-ing the higher aMW PIs and the other containingthe lower aMW PIs; all Pllls appeared to bestructurally similar in all four strains studied.

MATERIA AND METHODS

Bacteria. Neisseria gonorrhoeae strains JS1 (origi-nal designation, F62), JS2 (original designation, 10677-3), JS3 (original designation, 120176-3), and JS4 (origi-nal designation, MSL-7040, 1972) were grown on cleartyping medium (21) as previously described (21). Orga-nisms were grown at 36°C in 5% CO2 for 18 h.Nonpiliated (P-), transparent (O-) organisms wereused throughout this study.OM preparations. OM vesicles were prepared by

shaking whole N. gonorrhoeae in 0.1 M Tris-1 MNaCl-0.02% sodium azide buffer (pH 8.0) at 430C with0.3-mm-diameter glass beads. Vesicles were then iso-lated by differential centrifugation and column chro-matography (Sepharose 6B; Pharmacia Fine Chemi-cals, Inc., Piscataway, N.J.) (23) and stored at -200Cuntil used.

SDS-PAGE. The proteins used in 125I-peptide map-ping were separated on 15% acrylamide (acrylamide/N,N'-methylenebisacrylamide ratio, 30:0.8) slab gelsby the Tris-glycine system of Laemmli (14). WCs orOMs were solubilized at 100°C in 10%o (wt/vol) SDS-10%o (vol/vol) glycerol-0.1 M Tris (pH 6.8) solubilizingsolution, either with or without 8% 2ME. Sampleswere electrophoresed at 45 mA until the tracking dye(bromophenol blue) had migrated approximately 75mm. The gels were then fixed in 7% acetic acid-25% 2-propanol, stained with 0.2% Coomassie brilliant blue,and destained until the background cleared. The low-molecular-weight markers Phosphorylase B (94K), bo-vine serum albumin (68K), ovalbumin (43K), carbonicanhydrase (30K), soybean trypsin inhibitor (21K), andlysozyme (14.3K) (Bio-Rad Laboratories, Richmond,Calif.) were included in each gel.2-D SDS-PAGE. Two-dimensional (2-D) SDS-PAGE

was used to visualize the change in aMW of PIIIs inthe presence of 2ME. WCs of each strain were solubi-lized in 10o (wt/vol) SDS-109o glycerol-0.1 M Tris(pH 6.8) without 2ME and subjected to SDS-PAGE asdescribed above. One of two duplicate lanes of eachstrain was excised from the gel and soaked in 0.2%(wt/vol) SDS-8% (vol/vol) 2ME-0.1 M Tris (pH 6.8)for 1 h at 56°C. The gel lanes of each strain, whichwere cut so as to include only the region containing PIand PIII, were then inserted, at a 900 angle to the firstdimension electrophoresis, into a slab gel apparatuswhich had a prepolymerized 15% running gel in place.A stacking gel consisting of4% agarose-0.2% (wt/vol)SDS-0.1 M Tris (pH 6.8) was held at 56°C untilneeded. Immediately before use, 2ME was added to1%. This solution was then pipetted around the re-duced gel lane and allowed to solidify. This was thenelectrophoresed, and the gel was fixed and stained asdescribed above.

Radioiodination and protease treatment of proteinbands. The radioiodination and a-chymotrypsin diges-tion of proteins in bands excised from acrylamide gelswere done by the procedures described by Elder et al.(4) and Swanson (23).

Eletrophoresis and chromatography. The 125I-peptidefiagments resulting from a-chymotryptic proteolysiswere suspended in a solution of L-leucine-L-arginine-L-tyrosine (0.2 mg/ml each) to yield approximately 50,000cpm/pl. A 2->tl amount of this mixture was spotted ontoa Polygram R Cel 300 (Brinkmann Instruments, Inc.,Westbury, N.Y.) thin-layer cellulose sheet (two samplesper sheet). High voltage thin-layer electrophoresis(TLE) was carried out at a constant 1,200 V for 45 minon a TLE 20 apparatus (Savant Instruments, Inc.,Hicksville, N.Y.) under Varsol cooled to 10°C by acirculating cooling bath. Electrophoresis buffer was asolution (pH 3.7) of water-acetic acid-pyridine(200:10:1). After electrophoresis, the sheet was air driedand cut down the center line. Each half was then turned900 and subjected to TLC, which was allowed to proceeduntil the solvent front (n-butanol-pyridine-water-aceticacid, 13:10:8:2) was within 2 to 3 mm of the sheet top.The sheets were dried, sprayed with 0.25% ninhydrin inacetone to locate the amino acid markers, and applied toX-ray film (XAR-5; Eastman Kodak Co., Rochester,N.Y.). The 16-h exposure was enhanced by using aCronex (Du Pont Co., Wilmington, Del.) intensifyingscreen at -76°C.

VOL. 37, 1982

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

INFECT. IMMUN.

A UNREDUCED

JSI JS2 JS3 JS4REDUCED

MW JSI JS2 JS3 JS4

43K

P.1-C A _B _

31KP.I11- 1- j- K- -3OKPIII- E F G H 30 I K;_;

30K,. - -G. -HC

B UNREDUCED

J51 J12 JS3 JS4

43K

P .I:'a_ K___30K P.111- e f 9 --. i -- 30K-

WITH?,'2AME

REDUCED

MW JS 1 JS2 JS3 1S4

- -

WIIUT

15J1 .12 J153 J154

P1w * P1w P1-wPill*- Pil*

FIG. 1. Coomassie brilliant blue-stained SDS-PAGE gels of WC and OM lysates of P- 0- N. gonorrhoeaestrains JS1, JS2, JS3, and JS4. (A) WC lysates solubilized in the absence or presence of 8% 2ME. Bands excisedfor peptide mapping: A, JS1 36.5K PI; B, JS2 36K PI; C, JS3 35.4K PI; D, JS4 35K PI; E-H, Unreduced PIlls(30K PIII) from each strain; I-L, unreduced Pllls (31K PIll) from each strain corresponding in aMW withreduced PIIls (31K PIII*); E'-H', Reduced PIlIs (30K PIII*) from each strain corresponding in aMW with 30KPllls; I'-L', reduced 31K PIII*s from each strain. (See Fig. 2A and 3 for 1251I-peptide maps of these proteins.) (B)OM lysates solubilized in the absence or presence of 8% 2ME. Bands excised for peptide mapping: a, JS1 36.5KPI; b, JS2 36K PI; c, JS3 35.4K PI; d, JS4 35K PI; e-h, unreduced 30K Pllls from each strain; e'-h', reduced 31KPIII*s. (See Fig. 2B and 4 for 125I-peptide maps of these proteins.) (C) 2-D SDS-PAGE of WC lysates. 1-D,lysates electrophoresed in the absence of 2ME; 2-D, central portion of lines shown in Fig. 1A were excised froma 1-D gel, soaked in 8% 2ME, and then electrophoresed in the presence of 2ME. Note that the position of thereduced PIII* is above the diagonal in each strain, indicating an increase in aMW. The PIs showed nomodification in aMW as a result of reduction.

Mixed 25I-peptide mapping. To compare the 1251_peptides of different proteins, 1 ,ul of each of theappropriate protein digests was mixed and applied tothe cellulose sheets (see Results for those proteins thatwere mixed). TLE and TLC were then performed as

described above.

RESULTS

SDS-PAGE of N. gonorrhoeae WC and OMlysates. Whole N. gonorrhoeae and OM prepara-tions of strains JS1, JS2, JS3, and JS4 weresolubilized in SDS either with or without 8%2ME and subjected to SDS-PAGE. The Coo-massie brilliant blue-stained PIs of each strainwere clearly visible in these preparations (Fig.1A and 2B). The PI aMW of each strain wasestimated to be as follows: JS1, 36.5K; JS2,36K; JS3, 35.4K; and JS4, 35K. These aMWswere higher than previously published values(23) and more in agreement with estimates of

other investigators for PIs from these strainsowing to the current use of different molecular-weight markers. There was no alteration ofaMW of any of the PIs when solubilized eitherwith or without 2ME. Each PI band (Fig. 1,bands A-D and 1B, bands a-d) was excised andexamined by 125I-peptide mapping.The PIIIs of each strain can be clearly seen in

the SDS-PAGE gel of the OM lysates (Fig. 1B).Comparison of the unreduced and reduced prep-arations shows the increase in aMW in thepresence of 2ME which characterizes these pro-teins. The aMW of the unreduced PIlIs wasestimated to be 30.4K (30K PIII), whereas thatof the reduced form had an aMW of 31.4K (31KPIII*).The PIIIs were much less apparent in gels of

WC lysates than they were in OM lysates.Nevertheless, 31K bands could be identified inthe reduced WC preparations, which showed a

-%>pl;.I I1*11II*

C

624 JUDD

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

GONOCOCCAL Plll PEPTIDE MAPPING 625

A JS1 36.5K

:

94

J3 354K

.i0~~~~~~~~~~~~.

ix. _

J54 35K

.. _I.. .

L.:1:

B JS1 365K JS 236K

A

JS335.4K

bi~

N.;K v .

1S4 35K

_I

l_ ...E4.

lT LC

FIG. 2. a-Chymotryptic 1251I-peptide maps of PIs from P- 0- organisms of N. gonorrhoeae strains JS1, JS2,JS3, and JS4. (A) PIs from WC lysates (see Fig. 1A); (B) PIs from isolated OM lysates (see Fig. 1B). Arrows andstars represent unique 125I-peptides as discussed in the text. TLC, Ascending thin-layer chromatography.

slightly increased staining intensity (Fig. 1A,bands I'-L') when compared with the unreducedpreparations. These bands had the same aMWas that of 2ME-modified PIIIs seen in OMlysates (Fig. 1B) and correlated in aMW with thePIII* seen above the diagonal in Fig. 1C andwith surface-iodinated PIIIs (12). In addition,the PIII* spot above the diagonal in Fig. 1C wasconfirmed to be surface iodinatable by 2-D SDS-PAGE of surface-iodinated WC lysates (data notshown) and of surface-labeled OM lysates (datanot shown). Based on these correlations, thesebands were designated as reduced 31K PIIIs. Aprominent 30K band could be seen in all strainsin both the unreduced and reduced WC lysates(Fig. 1A, bands E-H and E'-H'). In addition,bands corresponding in aMW to reduced PIIIs,not present in unreduced OM lysates, were seenin all unreduced WC lysates (Fig. 1A, bands I-J). These bands showed slight differences inaMW (<500) from strain to strain which did notoccur in any of the 30K PIIIs or the 31K PIII*s.The fate of 31K PIII bands in the reducedpreparations is not clear. A faint band directly

above the 31K PIII* of the JS2, 2ME-reducedWC lysate may correspond to the JS2 31K PIII;however, it seems likely that the 31K PIII*bands overlay the 31K PIII bands in reducedWC lysates. All of the above PIII bands wereexcised and examined by 125I-peptide mapping.2-D SDS-PAGE ofWC lysates. To confirm that

2ME modification of PIII occurred in WC ly-sates, 2-D SDS-PAGE was performed on WClysates of each N. gonorrhoeae strain. The 2-DSDS-PAGE of this experiment (Fig. 1C) showsthat 2ME modification of PIII, as evidenced bythe single spot falling well above the diagonal,occurs in the WC lysates of all four strains.Similar results were observed in 2-D gels ofOMvesicles (data not shown).

125I-peptide mapping. Those bands indicatedin Fig. 1 were excised and subjected to 1251_peptide mapping. The 125I-peptide maps of PIsfrom lysates of WCs and OMs from all fourstrains are shown in Fig. 2A and B. 125I-peptidemaps in Fig. 2, 3, and 4 are shown in groups offour, with JS1 and JS2 preparations on the upperleft and right, respectively, and JS3 and JS4

I

VOL. 37, 1982

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

is JS2E ~~~~~~~~~F

6;1

w ~~~~~~~~~~~~~~~. . ..

JS3 * 4G f H

I

A

..4s--...1%.-I..r:. t

:*..,: .0

JS2 IF,

AAMx* io!

aa

4 NL :...

..

JS..*-3

',

.. *1

JS 4

L'If

4.

..

4

FIG. 3. a-Chymotryptic 125I-peptide maps of Pllls from P- 0 WCs of N. gonorrhoeae strains JS1, JS2, JS3,and JS4. (A) Unreduced Pllls (30K Pllls); (B) unreduced proteins (31K PIII) corresponding in aMW to reducedPIlls (31K PIII*); (C) reduced PIII (30K PIII*) corresponding in aMW to 30K Pll; (D) 31K PIII*s. Letters oneach map correspond with bands shown in Fig. 1. Arrow indicates a variable peptide as discussed in the text.TLC, Ascending thin-layer chromatography.

Js' 452

IAt

S.S

:t

JS3K

JS4

I4*4&v

.#

Ike'

isI

I'

JIs 12

.4, :.:.fi&.:,; .Lit:

I.1TAs. #.V.-

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

GONOCOCCAL Plll PEPTIDE MAPPING 627

:s,*

*::.

*:':i.

J53 -- e..

g e..*

A

i, i?JS2el fIJS2

BFIG. 4. a-Chymotryptic 125I-peptide maps of Pllls from OM lysates of P- 0- organisms from N.

gonorrhoeae strains JS1, JS2, JS3, and JS4. (A) Unreduced Pllls (30K Plll); (B) reduced Pllls (31K PIII*).Letters on each map correspond with bands shown in Fig. 1B. Arrow indicates a variable peptide as discussed inthe text. TLC, Ascending thin-layer chromatography.

preparations on the lower left and right, respec-tively. The direction of high-voltage TLE isalways towards the top of each figure, and thedirection of ascending TLC is toward the left(Fig. 2A). The double 125I-peptides shown in theupper right corner of each map are felt to beartifactual, since they occur in maps of unrelat-ed, nongonococcal proteins (data not shown).

Virtually identical 125I-peptide maps of PI foreach N. gonorrhoeae strain were obtained fromSDS-PAGE gels ofWC lysates and OM prepara-tions (Fig. 2A and B). Minor variations in inten-sities and positions of 125I-peptides from OMand WC PIs were due to slight technical varia-tions in the radioiodination, TLE, and TLCprocedures. These variations were reflected inthe migration of amino acid markers (data notshown).The 125I-peptide maps of JS1 and JS2 PIs were

very similar, with two relatively acidic, hydro-philic 125I-peptides (as described by Tsai et al.[30]) occurring only in the JS2 PI (arrows, Fig.2A and B). The maps of JS3 and JS4 PIs werequite different from those of JS1 and JS2 PIs butwere very similar to one another; however, the

JS3 PI had one acidic, hydrophilic and twoapproximately neutral, hydrophilic 125I-peptides(30) not seen in the JS4 PI (double arrows, Fig.2A and B). In addition, three hydrophilic pep-tides (30; stars, Fig. 2A and B) were moreintensely labeled in the JS3 PI. Despite thedifferent overall 125I-peptide maps of the JS1 andJS2 PIs as compared with the JS3 and JS4 PIs,several peptides appeared to be shared by allPIs, suggesting that some structural homologyexists among all PIs. These results corroboratedearlier observations (23) regarding the structuralrelationships of PIs.

125I-peptide mapping of PIlls. The 125I-peptidemaps of PIls isolated by SDS-PAGE of bothWC (Fig. 3A-D) and OM (Fig. 4A and B) lysateswere quite distinct from any of the PI 125I1peptide maps. All of the PIIIs showed a remark-able degree of similarity both within a givenstrain and among different strains. Each group offour 125I-peptide maps represents the sameaMW PIII type, either reduced or unreduced,isolated from each of the four strains. Thepreparations are designated as follows: 30KPIll, 30K bands from WC and OM lysates

VOL. 37, 1982

14 I

ill .:,& ...6.4WIll,f:

Amil-

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

INFECT. IMMUN.

C

;#: Dr..v

X ov Q -:

FIG. 5. a-Chymotryptic 125I-peptide maps of mixtures of Pllls isolated from WC lysates of P- 0- N.gonorrhoeae strains JS1, JS2, JS3, and JS4. (A) Mixture of 30K PIIIs from each strain; (B) mixture of 31K PIllsfrom each strain; (C) mixture of 30K PIII*s from each strain; (D) mixture of 31K P11l*s from each strain; (E)mixture of JS1 30K PIII, 31K Plll, 30K PlIl*, and 31K PII*; (F) composite a-chymotryptic '25I-peptide map ofPIII. 0, Common to PIls; 0, varies; f), weak or variably resolved; 0, absent from 31K PIlls; 0, artifact; 0,varies as a group in TLE migration. Arrow indicates a variable peptide as discussed in the text. TLC, Ascendingthin-layer chromatography.

subjected to SDS-PAGE in the absence of 2ME(Fig. 3A and 4A); 31K Pll, -31K bands fromWC lysates subjected to SDS-PAGE in the ab-sence of 2ME (Fig. 3B); 30K PIII*, 30K bandsfrom WC lysates subjected to SDS-PAGE in thepresence of 2ME (Fig. 3C); 31K PIII*, 31Kbands from WC and OM lysates subjected toSDS-PAGE in the presence of 2ME (Fig. 3D and4B).Comparison of PIIIs within each strain. The

structural relationships of the various PIllsfound in each strain could be ascertained bycomparing 125I-peptide maps in the same relativepositions within the different PIII groupings(e.g., to compare JS1 Pllls, see Fig. 3A, upper

left, and Fig. 3B, upper left). The overall 1251_peptide map patterns of all WC PIII forms withineach strain suggest that all of these proteins havevery similar primary structures. The 30K PIII,the 30K PIII*, and the 31K PIII* were found tobe identical by 125I-peptide mapping for onestrain by mixing experiments (Fig. SE). The 31KPIIIs, though identical among strains (Fig. 3Band 5B), were different from 30K PIII, 30KPIII*, and 31K PIII*. Several peptides wereabsent from 31K PIII type as shown in Fig. 5F, acomposite PIII 125I-peptide map.The apparent identities of I25i-peptides from

30K Pll and 31K PIII* of OMs were alsoconfirmed by carrying out the 2-D separation on

628 JUDD

A |B

. .l. 1.~~~~~~~~~~~~~~~~~~.

..

,.4I

i

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

GONOCOCCAL Plll PEPTIDE MAPPING 629

mixtures of the PIII types of all four strains (datanot shown). The slightly different positioning ofsome of the peptides in the 125I-peptide maps ofthe OM PIIIs as compared with the WC PIIImaps (see Fig. 5F) is believed to be owing totechnical variations inherent in such complexprocedures.There are some intensity differences between

the WC and OM PIII peptide maps, with severalof the OM PIII peptides being more weaklylabeled than those in the WC PIII preparations.However, increased exposure times of the OMPIII maps confirmed that all of the peptides seenin WC PIII maps were present in the OM Plllpreparations as well. Mixing experiments of JS1PIlls from WCs and OMs showed that all butthree of the peptides migrated to the samelocation, with only the three peptides in thelower right of each PIII map being blurred (datanot shown). Therefore, it appears that the OMPIlIs are not significantly different from thesame PIII types found in WCs.

Interstrain comparison of PIIIs. Interstraincomparisons of each PIII type are made byobserving the four 125I-peptide maps within eachgroup (e.g., 30K PIII, Fig. 3A and 31K PIll, Fig.3B). It appears that each PIII type is structurallyidentical in all strains (confirmed by mixingexperiments) in both WC PIIIs (Fig. 3A-D and5A-D) and OM PIlIs (Fig. 4). One peptide,marked by an arrow in all Plll maps, varied inoccurrence (or possibly intensity), especially inthe WC PIII maps. It was generally more appar-ent in the 30K PIIIs than in the 31K PIIIs. Asnoted above for JS1, the 125I-peptide maps of30K PIII, 30K PIII*, and 31K PIII* were verysimilar in both WCs and OMs of strains JS2,JS3, and JS4. The 31K PIlIs, which were thesame in all strains, differed from the other PIIIsin that four peptides seen in all other PIIIs wereabsent from the 31K PIII preparations. Fourpeptides, denoted by open triangles in the PIIIcomposite map (Fig. SF), showed some varia-tion as a group in TLE migration. However,since the migration of these peptides varied as agroup in individual PIII preparations while be-having identically in mixed 125I-peptide maps,the observed variation in migration of thesepeptides is felt to reflect technical variationsrather than unique peptides (see Discussion).The above results support the following con-

clusions. (i) 125I-peptide maps obtained for PIand PIII from SDS-PAGE of WC lysates corre-spond to 125I-peptide maps of the same proteinsin OM preparations. (ii) PIII appears to besimilar or identical in all strains studied, regard-less of the PI type. (iii) 2ME modification of PIIIoccurs in both WC and OM lysates. This modifi-cation seems less complete in WC lysates than inOM preparations. (iv) The primary structure of

PIII as revealed by 125I-peptide mapping is notsignificantly altered as a result of 2ME modifica-tion. (v) A unique protein of -31K occurs in WClysates in the absence of 2ME. It exhibits slightaMW differences among strains and is similar instructure to PIII. (vi) The four PIs in this studycan be categorized into two 125I-peptide maphomology groups, one containing the two higheraMW PIs and the other containing the two loweraMW PIs. Despite these similarities, all PIs dohave differences in their primary structures.

DISCUSSIONInterest in PIII has increased since previous

studies have shown that PI and PIII exist as aheteropolymeric unit in situ in both intact N.gonorrhoeae (19) and OM vesicles (18). Immu-noprecipitation experiments have shown that PIand PIII coprecipitate (18, 27), prompting specu-lation about the role of PIII in immunologicalreactions on the surface of N. gonorrhoeae.

In this study, the technique of 125I-peptidemapping has been used to study the structuralrelationships of PIIIs from four strains of N.gonorrhoeae. PI and PIII bands from SDS-PAGE were obtained from both WC lysates andOM vesicles of each strain and were subjected to125I-peptide mapping. Several modifications ofthe procedure resulted in enhanced resolution ofthe 125I-peptides generated by a-chymotrypsindigestion. The use of the Savant TLE 20 appara-tus allowed more efficient cooling than did a flat-bed apparatus and resulted in more consistentmigration and greater resolution of 125I-peptidesin the electrophoresis step. Precise temperaturecontrol during TLE is very important. Tempera-ture variation is accompanied by a change in thepH of the running buffer, and this change altersthe charge on all of the peptides. If a peptideundergoes a charge change (i.e., deprotonates)at a pH that is close to that of the TLE runningbuffer, small pH fluctuations can have a pro-found effect on the electrophoretic mobility ofthe peptide. The four peptides in the PIII 125I1peptide maps, which show some variation inTLE migration, illustrate this point. Mixing ex-periments (Fig. 5) clearly demonstrated thatthese peptides are identical in the different prep-arations; yet even with the improved coolingprovided by the Savant apparatus, they showedsome variation from run to run.

Peptide mapping of PIs of the four strainsextended previous studies; although all PIsshared some structural similarity, two homologygroups, one containing the 36.5K and 36K PIsand the other containing the 35.4K and 35K PIs,were apparent. It is interesting that the peptideswhich were unique to the individual PIs were allhydrophilic. If they represent surface-exposedportions of the PI molecules, they may contrib-

VOL. 37, 1982

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

INFECT. IMMUN.

ute to antigenic differences observed among thePI types.The 125I-peptide mapping patterns obtained

for the various PIIIs indicated that, with theexception of the 31K PIII, all PIIIs seemedvirtually identical. There were no observed pep-tides which were unique to one strain, as wereseen in the 125I-peptide maps of PIs. One peptidein PIlIs isolated from WCs did seem to vary inoccurrence (or possibly intensity). This peptidewas not observed in PIIIs isolated from OMvesicles. With this exception, the Pllls fromOMs appeared to be very similar to those de-rived from WCs. Similarly, reduction did notsignificantly alter 125I-peptide patterns of PIII,indicating that no major change in primary struc-ture accompanies the increase in aMW of 2ME-modified PIlIs and supporting the suggestionthat PIII has an internal disulfide linkage which,when cleaved in reducing environments, allowsthe molecule to more fully unfold, thereby in-creasing the Stoke's radius and the aMW (18).Conversion of 30K PIII to 31K PIII* by 2ME

was clearly evident in OM vesicles (Fig. 1B) butwas less easily visualized in WC lysates. One-dimensional electrophoresis (Fig. 1A) showedthat only a small amount of Plll was convertedto the higher aMW form. However, 2ME treat-ment of the one-dimensional SDS-PAGE, fol-lowed by 2-D electrophoresis (Fig. 1C), showedthat the majority of the PIII material was con-verted to the 31K PIII* form, suggesting thatonly a partial conversion of PIII from 30K to31K occurs in WC lysates in the presence of2ME. This could indicate that WCs have manymore sites competing for the available reducingagent or that PIII is less accessible to reductionin WC preparations owing to decreased penetra-tion of the 2ME or limited exposure of PIII tothe reducing environment.The 31K PIII, which was present in unre-

duced WC lysates, was similar to the other PIIItypes but lacked several peptides seen in theother PIIIs. This protein was not readily radio-iodinated by surface-reactive agents (unpub-lished data), nor was it found in OM lysates,suggesting that the 31K PIII may not be aconstituent of blebbed OM vesicles. It mayrepresent a precursor of PIII or a modified formof PIII which functions in a different capacitythan does PIII. The fate of the 31K PIII inreduced WC lysates is not clear. It seems likelythat the 31K PIII* band overlays the 31K PIIIband and that the 125I-peptide map of 31K PIII*is actually a composite of the two proteins.Certainly, further study is necessary to elucidatethe role of the 31K PIII.There were a few 1251I-peptides which ap-

peared to migrate to the same location in the PIand PIII maps. Similar observations have been

made regarding structural relationships betweenPlIs and PIs (24), indicating that a certainamount of structural homology exists among allof the N. gonorrhoeae OM proteins studied by125I-peptide mapping. However, the differencesbetween the PI and PIII maps demonstrate thatPIII is not a cleavage product of PI. PIII alsoappears to be quite different than the Plls so farstudied (24; unpublished data).On the basis solely of structural data, it is

difficult to predict what the immunological rela-tionships of PIlls from different strains may be.It has been shown that PIls having similar 125i-peptide maps have different peptides exposed onthe bacterial surface (7). Furthermore, PIlswhich had similar structures had different anti-genicity as assessed by radioimmunoprecipita-tion of WCs; the observed differences in antige-nicity correlated with the surface exposure ofdifferent peptides (J. Swanson, 0. Barrera, andR. C. Judd, manuscript in preparation). Where-as PIIIs may have different peptides exposed onthe surface when associated with different PItypes, the virtually identical 125I-peptide maps ofthese proteins suggest that all PIlIs will have thesame surface exposure.The nature of PIII surface exposure in both

WCs and OM vesicles can be investigated byextrinsic 125I labeling, followed by peptide map-ping procedures similar to those discussed inthis study. This information may help elucidatefurther the contribution of PIII to the structureand immunobiology of the gonococcus.

ACKNOWLEDGMENTSI thank Susan Smaus for her expert assistance in preparing

this manuscript and Chuck Taylor and Bob Evans for theirfine photographic work. I also am especially grateful to JohnSwanson for his support and guidance and to the staff of theLaboratory of Microbial Structure and Function for theircritical evaluation and assistance in this work.

LITERATURE CITED1. Blake, M. S., E. C. GotschlHch, and J. Swanson. 1981.

Effects of proteolytic enzymes on the outer membraneproteins of Neisseria gonorrhoeae. Infect. Immun.33:212-222.

2. Brooks, G. F., C. J. Lammel, E. Z. Burns, and J. F.James. 1980. Confounding factors affecting normal serumkilling of N. gonorrhoeae colony phenotype variants, p.251-253. In D. Danielsson and S. Normark (ed.), Geneticsand immunobiology of pathogenic neisseria. University ofUmei, Umea, Sweden.

3. Buchanan, T. M., and W. A. Pearce. 1979. Pathogenicaspects of outer membrane components of gram-negativebacteria, p. 475-514. In M. Inouye (ed.), Bacterial outermembranes. John Wiley & Sons, Inc., New York.

4. Elder, J. H., R. A. Pickett II, J. Hampton, and R. A.Lerner. 1977. Radioiodination of proteins in single poly-acrylamide gel slices. J. Biol. Chem. 252:6510-6515.

5. Heckels, J. E. 1977. The surface properties of Neisseriagonorrhoeae: isolation of the major components of theouter membrane. J. Gen. Microbiol. 99:333-341.

6. Heckels, J. E. 1978. The surface properties of Neisseriagonorrhoeae: topographical distribution of outer mem-brane protein antigens. J. Gen. Microbiol. 108:213-219.

630 JUDD

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

GONOCOCCAL PIII PEPTIDE MAPPING 631

7. Heckels, J. E. 1981. Structural comparison of Neisseriagonorrhoeae outer membrane proteins. J. Bacteriol.145:736-742.

8. James, J. F., C. J. Lammel, D. L. Draper, and G. F.Brooks. 1980. Attachment of N. gonorrhoeae colonyphenotype varients to eukaryotic cells and tissues, p. 213-216. In D. Danielsson and S. Normark (ed.), Genetics andimmunobiology of pathogenic neisseria. University ofUmea, Umel, Sweden.

9. Johnston, K. H. 1978. Antigenic profile of an outer mem-brane complex of Neisseria gonorrhoeae responsible forserotypic specificity, p. 121-129. In G. F. Brooks, E. C.Gotschlich, K. K. Holmes, W. D. Sawyer, and F. E.Young (ed.), Immunobiology of Neisseria gonorrhoeae.American Society for Microbiology, Washington, D.C.

10. Johnston, K. H., and E. C. GotschlUch. 1974. Isolation andcharacterization of the outer membrane of Neisseriagonorrhoeae. J. Bacteriol. 119:250-257.

11. Johnston, K. H., K. K. Holmes, and E. C. Gotschlich.1976. The serological classification of Neisseria gonor-rhoeae. I. Isolation of the outer membrane complexresponsible for serotypic specificity. J. Exp. Med.143:741-758.

12. Judd, R. C. 1982. Surface peptide mapping of protein Iand protein III of four strains of Neisseria gonorrhoeae.Infect. Immun. 37:632-641.

13. King, G. J., and J. Swanson. 1978. Studies on gonococcusinfection. XV. Identification of surface proteins of Neis-seria gonorrhoeae correlated with leukocyte association.Infect. Immun. 21:575-584.

14. Laemmli, U. K. 1970. Clevage of structural proteins dur-ing the assembly of the head of bacteriophage T4. Nature(London) 227:680-85.

15. Lambden, P. R., and J. E. Heckels. 1979. Outer mem-brane protein composition and colonial morphology ofNeisseria gonorrhoeae strain P9. FEMS Microbiol. Lett.5:263-265.

16. Lambden, P. R., J. E. Heckels, L. T. James, and P. J.Watt. 1979. Variations in surface protein compositionassociated with virulence properties in opacity types ofNeisseria gonorrhoeae. J. Gen. Microbiol. 114:305-312.

17. MarkweU, M. A. K., and C. F. Fox. 1978. Surface-specif-ic iodination of membrane proteins of viruses and eucary-otic cells using 1,3,4,6-tetrachloro-3a,6ac-diphenylglycol-uril. Biochemistry 17:4807-4817.

18. McDade, R. L., Jr., and K. H. Johnston. 1980. Character-ization of serologically dominant outer membrane pro-teins of Neisseria gonorrhoeae. J. Bacteriol. 141:1183-1191.

19. Newhall, W. J., W. D. Sawyer, and R. A. Haak. 1980.

Cross-linking analysis of the outer membrane proteins ofNeisseria gonorrhoeae. Infect. Immun. 28:785-791.

20. Swanson, J. 1977. Surface components associated withgonococcal-cell interactions, p. 370-401. In R. B. Roberts(ed.), The gonococcus. John Wiley & Sons, Inc., NewYork.

21. Swanson, J. 1978. Studies on gonococcus infection. XII.Colony color and opacity variants of gonococci. Infect.Immun. 19:320-331.

22. Swanson, J. 1978. Studies on gonococcus infection. XIV.Cell wall protein differences among color/opacity colonyvariants of Neisseria gonorrhoeae. Infect. Immun.21:292-302.

23. Swanson, J. 1979. Studies on gonococcus infection.XVIII. 125I-labeled peptide mapping of the major proteinof the gonococcal cell wall outer membrane. Infect.Immun. 23:799-810.

24. Swanson, J. 1980. 125I-labeled peptide mapping of someheat-modifiable proteins of the gonococcal outer mem-brane. Infect. Immun. 28:54-64.

25. Swanson, J. 1980. Adhesion and entry of bacteria intocells: a model of the pathogenesis of gonorrhea, p. 17-40.In H. Smith, J. J. Skehel, and M. J. Turner (ed.), Themolecular basis of microbial pathogenicity. Verlag Che-mie, GmbH, Weinheim, Germany.

26. Swanson, J. 1980. Gonococcal growth inhibition and hy-drolysis of outer membrane proteins by alpha-chymotryp-sin, p. 19-24. In D. Danielsson and S. Normark (ed.),Genetics and immunobiology of pathogenic neisseria.University of Umei, Umea, Sweden.

27. Swanson, J. 1981. Surface-exposed protein antigens of thegonococcal outer membrane. Infect. Immun. 34:804-816.

28. Swanson, J., and J. E. Heckels. 1980. Proposal: nomencla-ture of gonococcal outer membrane proteins, p. xxi-xxvi.In D. Danielsson and S. Normark (ed.), Genetics andimmunobiology of pathogenic neisseria. University ofUmei, UmeA, Sweden.

29. Swanson, J., G. King, and B. Zel4. 1975. Studies ongonococcus infection. VIII. 125Iodine labeling of gono-cocci and studies on their in vitro interactions witheukaryotic cells. Infect. Immun. 11:453-459.

30. Tsai, C., C. E. Frasch, and L. F. Mocca. 1981. Fivestructural classes of major outer membrane proteins inNeisseria meningitidis. J. Bacteriol. 146:69-78.

31. VIri, M., and J. S. Everson. 1981. Comparative virulenceof opacity varients of Neisseria gonorrhoeae strain P9.Infect. Immun. 31:%5-970.

32. Walstad, D. L., L. F. Guyman, and P. F. Sparling. 1977.Altered outer membrane protein in different colonial typesof Neisseria gonorrhoeae. J. Bacteriol. 129:1623-1627.

VOL. 37, 1982

on July 4, 2020 by guesthttp://iai.asm

.org/D

ownloaded from