INFECrION AND IMMUNITY, JUlY 1993, p. 3026-3031 Vol. 61, No. 70019-9567/93/073026-06$02.00/0Copyright © 1993, American Society for Microbiology

Protein D of Haemophilus influenzae Is Not a UniversalImmunoglobulin D-Binding ProteinKEN SASAKI' 2'3 AND ROBERT S. MUNSON, JR.2,3.4*

Connaught Centre for Biotechnology Research, Toronto, Ontario, Canada, 1 and Edward MallinckrodtDepartment ofPediatrics2 and Department ofMolecular Microbiology,4 Washington University School ofMedicine, and Division of Infectious Diseases, St. Louis Children's Hospital,3 St. Louis, Missour 63110

Received 4 January 1993/Accepted 7 April 1993

Haemophilus influenzae type b and nontypeable H. influenzae have been reported to bind humanimmunoglobulin D (IgD). IgD myeloma sera from five patients were tested for the ability of IgD to bind to H.influenzae. Serotype b strains bound human IgD in four of the five sera tested. IgD in the fifth serum boundstrongly to type b strain MinnA but poorly to other type b strains. Additionally, IgD binding was not observedwhen nontypeable strains were tested. The gene for protein D, the putative IgD-binding protein, was clonedfrom the IgD-binding H. influenzae type b strain MinnA and expressed in Escherichia coli. IgD binding to E.coli expressing protein D was not demonstrable. Recombinant protein D was purified, and antisera weregenerated in rabbits. Using these rabbit sera, we detected protein D in nontypeable as well as serotype b strainsby Western blotting (immunoblotting). In contrast, IgD myeloma protein 4490, which was previously reportedto bind to protein D by Ruan and coworkers (M. Ruan, M. Akkoyunlu, A. Grubb, and A. Forsgren, J.Immunol. 145:3379-3384), bound strongly to both type b and nontypeable H. influenzae as well as to E. coliexpressing protein D. Thus, IgD binding is a general property ofH. influenzae type b strains but not a generalproperty of nontypeable strains, although both type b and nontypeable strains produce protein D. With theexception of IgD myeloma protein 4490 binding, we have no evidence for a role of protein D in IgD binding toH. influenzae.

Haemophilus influenzae causes invasive and noninvasivediseases in children and adults. There are six encapsulatedserotypes, designated a through f, and nontypeable strains,which have no demonstrable capsule. More than 95% ofinvasive H. influenzae disease is caused by serotype bstrains. Nontypeable strains of H. influenzae are a majorcause of otitis media in children but do not usually causelife-threatening diseases in normal children in the developedworld (4, 20). These organisms cause pneumonia in elderlyand immunocompromised patients and have been implicatedin exacerbations of chronic bronchitis (12). In developingcountries, nontypeable strains cause lower respiratory infec-tions that often are fatal (11).

Vaccines designed to prevent diseases caused by serotypeb strains have recently become available and appear to havea high efficacy (13). These vaccines are composed of theserotype b capsular polysaccharide conjugated to protein;therefore, they are not efficacious against diseases caused bynontypeable strains. Outer membrane proteins have beenextensively studied as alternative vaccine candidates. Re-cently, a 42-kDa protein designated protein D, which isconserved among type b and nontypeable strains, was de-scribed (1, 5, 15). The protein was reported to be responsiblefor the binding of human immunoglobulin D (IgD) to Hae-mophilus strains. IgD-binding activity was detected in 127 of127 strains of H. influenzae examined. The gene was clonedand sequenced (5).We were interested in further characterizing the IgD-

binding protein present on the surface of H. influenzae.Although the published data and our results support the

* Corresponding author.

conclusion that protein D binds certain IgD myeloma pro-teins, our data question the generality of these findings.

MATERIALS AND METHODSBacterial strains and plasmids. H. influenzae type b

MinnA, Eagan, Durst, and 1613 were previously described(10). They have outer membrane protein subtypes 1H, 1L,2L, and 3L, respectively, and are representative of the typeb strains responsible for the majority of invasive diseases inthe United States. Nontypeable strains 12, 15, and 17 wereisolated from the ears of patients with otitis media and wereobtained from the collection of Stephen Barenkamp (2).Additional nontypeable strains were obtained from the col-lections of Stephen Barenkamp and Dan Granoff (14, 21). Acapsule-deficient mutant (RM804) of H. influenzae type bEagan was the gift of E. Richard Moxon (7). Escherichia coliJM101 and bacteriophages M13mpl8 and M13mpl9 wereobtained from New England Biolabs, Inc. (Beverly, Mass.).E. coli BL21(DE3) and BL21(DE3)/pLysS were obtainedfrom F. William Studier (18). Bacteriophage T7 expressionvector pT7-7 and mGPl-2, an M13 derivative containing thephage T7 RNA polymerase gene, were obtained from StanTabor (19).

IgD-binding assay. Human IgD myeloma sera were ob-tained from Moon Nahm (KN) and Jeanette Thorbecke (S1to S4). Purified myeloma protein 4490 was obtained fromArne Forsgren. Human agammaglobulinemic serum wasobtained from Dan Granoff. IgD-binding activity was as-sayed by a sandwich method. Whole-cell suspensions or cellsonicates were blotted onto nitrocellulose, and the mem-brane was blocked for 45 min in a 2% gelatin solution inTris-buffered saline. The membrane was washed for 10 minin Tris-buffered saline containing 0.05% Tween 20 (T-TBS)

3026

on July 21, 2019 by guesthttp://iai.asm

.org/D

ownloaded from

H. INFLUENZAE PROTEIN D 3027

B N S E

FIG. 1. Partial restriction map of pRSM1021. The)amplified by PCR and cloned into the NdeI and BapT7-7. Vector sequences are shown as a line; the genoa bar. The arrow represents the bacteriophage T7 pr(vant restriction sites are BglII (B), NdeI (N), Sau3A (SSspI (P), and BamHI (H). nt, nucleotides.

and then incubated for 90 min in IgD myeloma sewith T-TBS containing 1% gelatin. The menwashed three times with T-TBS and then tremin with goat anti-human IgD antibodies coalkaline phosphatase (Tago Inc., Burlingame, Cato 1/1,000 with 1% gelatin-T-TBS). The menwashed three times with T-TBS and thenwith nitroblue tetrazolium (NBT)-5-bromo-4-dolylphosphate (BCIP) solution as described pre

Cloning and expression of the protein D gene. TI(hpd) gene was amplified from chromosomal ]

influenzae serotype b MinnA by the polymerasetion (PCR) with a PCR kit from Perkin-Elmer 4walk, Conn.) in accordance with the manufacturtions. Oligonucleotide primers were prepared onthe published nucleotide sequence of protein rtypeable H. influenzae 772 (5). The primers werethat the complete gene would be amplified or, also that the signal sequence was deleted. In thestruct, the N-terminal cysteine residue of the mawas substituted for by a methionine. Additionallers were constructed so that they contained an j.the 5' end of the gene and a BamHl site at the 3gene. The sequences of the primers were as foll(GCAGCACATAflAAAC1TLAAAA(1UI&1GCC204 to 225), 5'TCAGCAGCACATATGAGC&GC(AAATAT.GGiCG~(positions 261 to 284), and 5'CGCCC (complement

1340 to 1361); the underlined nucleotides corresindicated positions in the sequence published by J(5), and the boldfaced nucleotides are the NdeI sittwo oligonucleotides and the BamHI site in the ticleotide.The PCR products were digested with NdeI

and ligated into pT7-7 that had been digested wienzymes. The ligation mixtures were transforrcoli JM101. Plasmids with the correct restricti(Fig. 1) were saved as pRSM1021 (hpd gene witpeptide) and pRSM1022 (hpd gene without thetide). The BglII-BamHI fragment of each pcloned into M13, and the nucleotide sequences ogenes were determined by the dideoxy metiSequenase kit (US Biochemicals, Cleveland, Ohdance with the manufacturer's instructions. 0thnant DNA manipulations were performed by staiods (16, 17).

Protein D was produced in strain JM101/pRSIabsence of transcription from the bacteriophamoter. For induction of higher-level expressioJM101/pRSM1021 and JM101/pRSM1022, tran:the hpd gene from the bacteriophage T7 pr(induced by the addition of isopropyl-3-D-thiogalside (IPTG) and infection with mGP1-2 as desc

P E H ously (9). In the BL21(DE3)/pLysS background, transcrip-tion of the hpd gene from the T7 promoter was induced bythe addition of 2 mM IPTG to mid-log-phase cultures.

Purification of recombinant protein D. At 2 h after theinduction of recombinant protein D synthesis in E. coli

hpd gene was BL21(DE3)/pLysS/pRSM1022, the cells were harvested bye is shown as centrifugation at 3,000 x g for 15 min at 4°C, frozen, andmoter. Rele- thawed for partial lysis. All subsequent operations were

i), EcoRI (E), performed at 4°C. Cells were broken by sonication in 10 mMpotassium phosphate buffer (pH 7.5) containing 0.1 mMphenylmethylsulfonyl fluoride. The sonicate was centrifugedat 12,000 x g for 20 min. Saturated (NH4)2SO4 was added tothe supematant to 50% saturation, and the mixture was

.rum diluted allowed to stand for 30 min. After centrifugation at 12,000 xnbrane was g for 20 min, the supernatant solution was retained. Satu-ated for 90 rated (NH4)2SO4 was added to 70% saturation; the suspen-njugated to sion was allowed to stand for 30 min and then centrifuged attlif.) (diluted 12,000 x g for 20 min. The supernatant solution was mixednbrane was with saturated (NH4)2SO4 to 80% saturation, allowed todeveloped stand for 30 min, and then centrifuged for 20 min. All

-chloro-3-in- precipitates were dissolved in 10 mM potassium phosphatewviously (9). buffer and dialyzed against the same buffer. The majority ofhe protein D the recombinant protein D was recovered in the 70%DNA of H. (NH4)2SO4 pellet. This material was loaded on a column ofchain reac- DEAE-Sephacel that had been equilibrated with 10 mMCetus (Nor- potassium phosphate buffer. The column was washed wither's instruc- 10 mM potassium phosphate buffer and then with potassiumthe basis of phosphate buffer containing a 0 to 0.5 M NaCl gradient. The) from non- protein concentration in each fraction was determined with adesigned so bicinchoninic acid protein assay kit from Pierce ChemicalIternatively, Co. (Rockford, Ill.), and the protein peaks were assayed fora latter con- protein D content by sodium dodecyl sulfate-polyacrylamidetture protein gel electrophoresis (SDS-PAGE). Two peaks (I and II)Iy, the prim- eluted with 10 mM potassium phosphate buffer; both con-NdeI site at tained protein D. An additional protein peak was observed' end of the during gradient elution. The protein D-containing peaksows: 5'TCA were combined, concentrated with a Centricon microcon-:C (positions centrator (10,000-molecular-weight cutoff; Amicon, Beverly,&ATICATC Mass.), and loaded on a column of CM-Sepharose that hadXGGATCCT been equilibrated with 10 mM potassium phosphate buffer.of positions The column was washed with 10 mM potassium phosphate-pond to the buffer and then with a 0 to 0.5 M NaCl gradient in potassiumJanson et al. phosphate buffer. The protein D-containing fractions werete in the first identified as described above, concentrated with a Centriconhrd oligonu- microconcentrator, and dialyzed against phosphate-buffered

saline (PBS).and BamHI Immunization of rabbits with recombinant protein D. Theith the same CM-Sepharose-purified recombinant protein D (100 ,ug) inned into E. 0.5 ml of PBS was mixed with 0.5 ml of Freund's completeion patterns adjuvant and subcutaneously injected into a rabbit. After 1h the leader month, 100 ,ug of protein D in PBS was mixed with Freund'sleader pep- incomplete adjuvant and injected. Serum was collected 10ilasmid was days after the second immunization.of the cloned SDS-PAGE and Western blot (immunoblot) analysis. SDS-hod with a PAGE was performed on 11% modified Laemmli gels asio) in accor- described by Lugtenberg et al. (8). Western blot conditionsLer recombi- were as previously described (9). Protein D was detectedndard meth- with anti-recombinant protein D rabbit serum and goat

anti-rabbit IgG antibodies conjugated to alkaline phos-41021 in the phatase (Tago).ige T7 pro-bn in strainsscription of RESULTS)moter wasactopyrano- IgD binding to H. influenzae. When a sandwich assay with:ribed previ- IgD myeloma serum KN and goat anti-human IgD-alkaline

I now------- ------ -------

VOL. 61, 1993

on July 21, 2019 by guesthttp://iai.asm

.org/D

ownloaded from

3028 SASAKI AND MUNSON

1:1,000 *

p1-s

1:2,000

1:3,000

1:5,000

1:10,000*

1:20,0000

1 2 3FIG. 2. IgD binding to H. influenzae. Cell sonicates (10 pLg of

protein) of H. influenzae serotype b MinnA were blotted ontonitrocellulose. After blocking was done, the membrane was incu-bated for 90 min in serum diluted 1/1,000 to 1/20,000, as indicated,with T-TBS containing 1% gelatin. The membrane was washed,treated with goat anti-human IgD-alkaline phosphatase diluted to1/1,000, and then developed with NBT-BCIP solution. Lanes: 1, IgDmyeloma serum KN; 2, normal human serum; 3, human agamma-globulinemic serum.

phosphatase was used, IgD binding to sonicates of serotypeb strain MinnA was observed (Fig. 2). IgD binding wasobserved with dilutions of serum KN of <1/20,000. Fordemonstration of specificity for immunoglobulins, dilutionsof serum from an agammaglobulinemic patient were substi-tuted for dilutions of serum KN. No reaction was observedwhen a 1/1,000 dilution of the former serum was used,indicating that the reaction was dependent on serum immu-noglobulins (Fig. 2). When normal human serum was used, aweak reaction was observed at a 1/1,000 dilution and noreaction was observed at a 1/10,000 dilution, consistent withthe low level of IgD found in normal human serum.

Cloning and expression of the protein D gene from strainMinnA. E. coli JM101/pRSM1021 and JM101/pRSM1022were constructed. Plasmids pRSM1021 and pRSM1022 con-tained the MinnA hpd genes with and without the signalsequence, respectively. Protein D expression was observedin strain JM101/pRSM1021 in the absence of induction of theT7 system. Transcription from the T7 promoter was inducedin both strains by infection with mGP1-2 and the addition ofIPTG. In the BL21(DE3)/pLysS background, transcriptionfrom the T7 promoter was induced by the addition of IPTG.Although a protein band with an apparent molecular weightof 42,000 was apparent in Coomassie brilliant blue-stainedSDS-polyacrylamide gels of sonicates of induced cultures,no IgD binding was detectable when these preparations wereprobed with serum KN (Fig. 3).

Sequence analysis of the cloned protein D gene. One poten-tial explanation for our inability to detect IgD binding toextracts ofE. coli strains expressing protein D is a cloning orPCR error. The inserts from pRSM1021 and pRSM1022 weretherefore cloned into M13, and the nucleotide sequenceswere determined. The nucleotide sequences of both clonedhpd genes were identical, with the exception of the modified5' end of the gene in pRSM1022. The nucleotide and derivedamino acid sequences were compared with the publishedsequence of the hpd gene from nontypeable strain 772. Therewere seven base-pair differences between the genes fromstrains MinnA and 772. Four silent base-pair differences

190--blnding to Soniento8

1 2 3 4

-IgD

.+lgD I * a

5 6 1 8 9

tgD-Biniding to Whole Cel.ls

1 2 3 4 5 6 1 8 9

lgD

.IgD *

FIG. 3. IgD binding to H. influenzae serotype b and nontypeablestrains and to E. coli clones expressing the protein D gene. Whole-cell suspensions or cell sonicates were blotted onto a nitrocellulosemembrane. IgD binding was assayed as described in the legend toFig. 2 with a 1/500 dilution of human IgD myeloma serum KN(+IgD). A control strip was treated identically except for theincubation in myeloma serum (-IgD). Lanes: 1 to 4, serotype bstrains MinnA, Durst, 1613, and Eagan, respectively; 5 to 7,nontypeable strains 12, 15, and 17, respectively; 8, JM101/pT7-7; 9,JM101/pRSM1021 (induced).

were observed in the 5' end of the genes, and three addi-tional changes were localized near the 3' end of the genes.Protein D of strain MinnA has alanine at position 327 andvaline at position 338 in place of the glutamic acid andalanine observed at these respective positions in strain 772protein D. The sequence can be retreived from GenBankunder accession number L15200. The sequence of protein Dfrom strain MinnA has also been reported by Janson et al.(6).IgD binding to various H. influenzae strains. Our inability

to observe IgD binding to recombinant protein D promptedus to perform additional experiments to further characterizeIgD binding to H. influenzae. When serum KN was used,IgD-binding activity was observed with 10 of 10 serotype bstrains tested. In contrast to the published data (1, 15), nobinding was observed when 25 nontypeable strains weretested. The nontypeable strains included 5 invasive strainscollected in Pakistan (21), 10 invasive strains collected in theUnited States (14), and 10 noninvasive strains collected inthe United States (2, 14). Data for representative strains areshown in Fig. 3. Interestingly, we found different IgD-binding intensities among the 10 serotype b strains. MinnAwas 1 of the 2 strains that showed the strongest IgD-bindingactivity among the 10 serotype b strains.IgD binding to H. influenzae tested with different IgD

myeloma sera. Our data did not agree with the previouslypublished data indicating that recombinant protein D andnontypeable H. influenzae strains bind IgD (1, 15). Wetherefore obtained four additional IgD myeloma sera andtested a subset of strains for their ability to bind IgD fromthese sera. The results are shown in Fig. 4. Sonicates of fourserotype b strains (MinnA, Eagan, Durst, and 1613) stronglybound IgD from four (KN, S2, S3, and S4) of the five seratested. Strain MinnA bound IgD from serum S1, but thebinding of IgD from serum S1 was negligible with the otherserotype b strains. IgD binding to nontypeable strains 12, 15,and 17 was negligible with all of the IgD myeloma proteins

INFECT. IMMUN.

on July 21, 2019 by guesthttp://iai.asm

.org/D

ownloaded from

H. INFLUENZAE PROTEIN D 3029

CKN 0

Si *

S2 0 ** S

S33 * * 0S4 * 0 0

2 3 4 5 6 7 8 9 10 11FIG. 4. IgD binding to H. influenzae tested with different IgD

myeloma sera. Cell sonicates (10 ,ug of protein) were blotted onto anitrocellulose membrane, and IgD binding was determined as de-scribed in the legend to Fig. 2 with IgD myeloma sera KN, Si, S2,S3, and S4. KN was diluted to 1/500. Sl through S4 were diluted toan IgD concentration of 6 j±g/ml. Row C contains samples thatwere not treated with serum. Lanes: 1 to 4, H. influenzae serotypeb MinnA, Durst, 1613, and Eagan, respectively; 5 to 7, nontypeablestrains 12, 15, and 17, respectively; 8, JM101/pT7-7; 9, JM101/pRSM1021 (uninduced); 10, BL21(DE3)/pLysS/pT7-7; 11, BL21(DE3)/pLysS/pRSM1021 (induced).

tested. Additionally, all of our E. coli clones expressingprotein D failed to bind the IgD myeloma proteins tested. Wethen obtained a sample of IgD myeloma protein 4490 fromArne Forsgren. This myeloma protein bound strongly toserotype b strains, nontypeable strains 15 and 17, and E. coliclones expressing protein D (data not shown).

Role of the capsular polysaccharide in IgD binding. Anobvious difference between the serotype b and nontypeablestrains is the presence of the capsular polysaccharide in theserotype b strains. To determine whether the capsular poly-saccharide was important in the binding of IgD to theserotype b strains under our conditions, we preincubatedIgD myeloma serum KN with capsular polysaccharide andcompared the activity of this serum mixture with that ofserum KN. Polysaccharide-treated serum KN and non-treated serum KN were both reactive in the dot blot assay.Additionally, we tested the binding of IgD from serum KNand serum S4 to a capsule-deficient mutant of serotype bstrain Eagan. This strain bound IgD as well as the parentalstrain (Fig. 5). Therefore, we concluded that the capsularpolysaccharide was not responsible for IgD binding to sero-type b Haemophilus strains under our assay conditions.

Presence of protein D in serotype b and nontypeable strainsof H. influenzae. When the protein D gene was expressedunder the control of the bacteriophage T7 promoter, largequantities of a protein with an apparent molecular weight of42,000 accumulated; the protein was presumed to be proteinD. Recombinant protein D synthesized by BL21(DE3)/

C

KN O

S4 *

2 3 4 5

FIG. 5. Binding of IgD to serotype b strains is not due to theserotype b capsular polysaccharide. Whole cells were dried ontonitrocellulose as described in the legend to Fig. 2 and probed withserum KN or serum S4. Row C contains samples that were nottreated with IgD myeloma serum. Lanes: 1, H. influenzae serotypeb MinnA; 2 and 3, nontypeable strains 15 and 17, respectively; 4, H.influenzae serotype b Eagan; 5, H. influenzae RM804, an isogeniccapsule-deficient mutant of strain Eagan.

94K-67K- -

43K- _

3OK-

20.1 K -

i4.4K-

2 3 4 5 6 7 8 9

FIG. 6. SDS-PAGE analysis of the purification steps for recom-

binant protein D. Proteins (30 ,ug) were separated by SDS-PAGEand stained with Coomassie brilliant blue. Lanes: 1, molecularweight standards (in thousands [K]); 2, sonicate; 3 to 5, insolublefractions at 50, 70, and 80% (NH4)2SO4 saturation, respectively; 6and 7, peaks I and II, respectively, eluted from the DEAE columnwith 10 mM phosphate; 8, protein peak eluted from the DEAEcolumn with the salt gradient; 9, CM-Sepharose-purified protein D.

pLysS/pRSM1021 was recovered in the 100,000 x g pellet ofsonicates, suggesting that it was present in the membranefraction. In contrast, protein D synthesized without theleader peptide and subsequent lipid modification was ob-served in the 100,000 x g supernatant. This protein waspurified to homogeneity by a combination of ammoniumsulfate precipitation, DEAE-Sephacel anion-exchange chro-matography, and CM-Sepharose cation-exchange chroma-tography as described in Materials and Methods. SDS-PAGE analysis of the purification is shown in Fig. 6. Rabbitantiserum against the purified protein was prepared and usedto determine whether all Haemophilus strains tested pro-duced protein D.When Western blotting was used, protein D was observed

in the four serotype b strains and the three nontypeablestrains tested (Fig. 7). A dot blot assay was used to charac-terize protein D expression in an additional 6 serotype b and22 nontypeable strains. Extracts from all H. influenzaestrains tested reacted strongly with the anti-recombinantprotein D rabbit serum.

DISCUSSION

Forsgren and coworkers (1, 3, 15) observed that strains ofH. influenzae and Moraxella catarrhalis bound human IgD.They identified in H. influenzae 772 a protein that has anapparent molecular weight of approximately 42,000 and thatbinds human IgD. The structural gene for this protein wascloned and sequenced (5). The recombinant protein boundhuman IgD. As the protein appeared to be highly conservedamong both serotype b and nontypeable strains, we soughtto further characterize this protein and evaluate the potentialfor including it in vaccines designed to prevent Haemophilusdiseases.When five myeloma sera were used, IgD-binding activity

was readily observed by immunoblotting of whole cells andsonicates of serotype b strains. Four of five IgD myelomaproteins tested were strongly reactive in this assay with all of

VOL. 61, 1993

on July 21, 2019 by guesthttp://iai.asm

.org/D

ownloaded from

3030 SASAKI AND MUNSON

observed binding to type b H. influenzae strains via a protein66K- D-independent mechanism and failed to observe binding tonontypeable H. influenzae strains.

45KK

31 K-

21.5K-

14.5K-

2 3 4 5 6 7 8 9 10 fl 12FIG. 7. Detection of protein D in H. influenzae and E. coli

strains by Western blotting. Proteins in sonicates of H. influenzaeand E. coli strains were separated by SDS-PAGE and electrotrans-ferred to nitrocellulose. Protein D was detected with rabbit anti-recombinant protein D serum diluted to 1/3,000. Lanes: 1 to 4, H.influenzae serotype b MinnA, Durst, 1613, and Eagan, respectively;5 to 7, nontypeable strains 12, 15, and 17, respectively; 8, JM101/pT7-7; 9, BL21(DE3)/pLysS/pT7-7; 10, blank; 11, JM101/pRSM1021(uninduced); 12, BL21(DE3)/pLysS/pRSM1021 (induced). Thirtymicrograms of protein was applied to lanes 1 through 9. Fivemicrograms and 1 pLg of protein were applied to lanes 11 and 12,respectively. Numbers at left are molecular weights in thousands(K).

the type b strains, while the fifth IgD myeloma proteinreacted strongly with strain MinnA and poorly with the otherstrains. These myeloma proteins did not bind to a panel ofnontypeable strains. Ruan and coworkers (15) reported thata myeloma protein designated 4490 bound to protein D. Incontrast to the results obtained with our panel of IgDmyeloma proteins, we observed that IgD myeloma protein4490 bound strongly to both type b and nontypeable strains.We cloned the hpd gene from strain MinnA and expressed

the protein in E. coli under the control of the bacteriophageT7 promoter. Upon induction, our E. coli clones carrying thehpd gene produced a large quantity of a protein with anapparent molecular weight of approximately 42,000. Ourmyeloma proteins did not bind to extracts of E. coli express-ing recombinant protein D. Because H. influenzae type bMinnA bound IgD and the strains expressing the recombi-nant protein failed to bind IgD, we conclude that protein D isnot sufficient for IgD binding to type b strains. In contrast,IgD myeloma protein 4490 bound strongly to E. coli express-ing protein D.Thus, IgD myeloma protein 4490 reacts with protein D.

Ruan and coworkers (15) prepared Fab and Fc fragmentsfrom this protein and demonstrated that both types offragments bound to H. influenzae, although only at concen-trations substantially higher than the concentrations neces-sary for IgD binding. They speculated that protein D mightrecognize heavy-chain structures located near the hingeregion. However, it is also possible that the reported Fcbinding was due to undigested IgD that was not detected intheir purified fractions and that protein 4490 is immunospe-cific for protein D. Alternatively, it is possible that allotypeor carbohydrate differences among the different IgD my-eloma proteins are responsible for the different bindingspecificities observed. Regardless of the mechanism, my-eloma protein 4490-protein D binding is not typical of thebinding of the panel of IgD myeloma proteins that weexamined. With the exception of myeloma protein 4490, we

ACKNOWLEDGMENTS

We thank Moon Nahm and Dan Granoff (Washington UniversitySchool of Medicine) and Jeanette Thorbecke (NYU Medical School)for the kind gifts of human serum and Arne Forsgren for myelomaprotein 4490. We thank Stephen Barenkamp and Dan Granoff for thenontypeable strains of H. influenzae and E. Richard Moxon for thenonencapsulated mutant of strain Eagan. We thank Hope WuellnerBrooks for technical assistance.These studies were supported by the Industrial Research Assis-

tance Program of the National Research Council of Canada.

REFERENCES

1. Akkoyunlu, M., M. Ruan, and A. Forsgren. 1991. Distribution ofprotein D, an immunoglobulin D-binding protein, in Haemo-philus strains. Infect. Immun. 59:1231-1238.

2. Barenkamp, S. J., and E. Leininger. 1992. Cloning, expression,and DNA sequence analysis of genes encoding nontypeableHaemophilus influenzae high-molecular-weight surface-ex-posed proteins related to filamentous hemagglutinin of Borde-tella pernussis. Infect. Immun. 60:1302-1313.

3. Forsgren, A., and A. Grubb. 1979. Many bacterial species bindhuman IgD. J. Immunol. 122:1468-1472.

4. Giebink, G. S. 1989. The microbiology of otitis media. Pediatr.Infect. Dis. J. 8:S18-S20.

5. Janson, H., L. 0. Heden, A. Grubb, M. Ruan, and A. Forsgren.1991. Protein D, an immunoglobulin D-binding protein of Hae-mophilus influenzae: cloning, nucleotide sequence, and expres-sion in Escherichia coli. Infect. Immun. 59:119-125.

6. Janson, H., M. Ruan, and A. Forsgren. 1992. High degree ofconservation of protein D genes from nontypable and type bstrains of Haemophilus influenzae, abstr. D-244. Abstr. 92ndGen. Meet. Am. Soc. Microbiol. 1992.

7. Kroll, J. S., and E. R. Moxon. 1988. Capsulation and gene copynumber at the cap locus of Haemophilus influenzae type b. J.Bacteriol. 170:859-864.

8. Lugtenberg, B., J. Meiiers, R Peters, P. van der Hoek, and L.van Alphen. 1975. Electrophoretic resolution of the major outermembrane protein of Escherchia coli K12 into four bands.FEBS Lett. 58:254-258.

9. Munson, R., Jr., and R. W. Tolan, Jr. 1989. Molecular cloning,expression, and primary sequence of outer membrane proteinP2 of Haemophilus influenzae type b. Infect. Immun. 57:88-94.

10. Munson, R S., Jr., B. Brodeur, P. Chong, S. Grass, D. Martin,and C. Proulx. 1992. Outer membrane proteins P1 and P2 ofHaemophilws influenzae type b: structure and identification ofsurface-exposed epitopes. J. Infect. Dis. 165(Suppl.):S86-S89.

11. Munson, R. S., Jr., M. H. Kabeer, A. A. Lenoir, and D. M.Granoff. 1989. Epidemiology and prospects for prevention ofdisease due to Haemophilus influenzae in developing countries.Rev. Infect. Dis. 11(Suppl.):S588-S597.

12. Murphy, T. F., and M. A. Apicella. 1987. Nontypeable Haemo-philus influenzae: a review of clinical aspects, surface antigens,and the human immune response to infection. Rev. Infect. Dis.9:1-15.

13. Murphy, T. V., K E. White, P. Pastor, L. Gabriel, F. Medley,D. M. Granoff, and M. T. Osterholm. 1993. Declining incidenceof Haemophilus influenzae type b disease since introduction ofvaccination. JAMA 269:246-248.

14. Musser, J. M., S. J. Barenkamp, D. M. Granoff, and R. K.Selander. 1986. Genetic relationships of serologically non-typable and serotype b strains of Haemophilus influenzae.Infect. Immun. 52:183-191.

15. Ruan, M., M. Akkoyunlu, A. Gmbb, and A. Forsgren. 1990.Protein D of H. influenzae; a novel bacterial surface proteinwith affinity for human IgD. J. Immunol. 145:3379-3384.

INFECT. IMMUN.

on July 21, 2019 by guesthttp://iai.asm

.org/D

ownloaded from

H. INFLUENZAE PROTEIN D 3031

16. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

17. Silhavy, T. J., M. L Berman, and L. W. Enquist. 1984.Experiments with gene fusions. Cold Spring Harbor Labora-tory, Cold Spring Harbor, N.Y.

18. Studier, F. W., and B. A. Moffatt. 1986. Use of bacteriophage T7RNA polymerase to direct selective high-level expression ofcloned genes. J. Mol. Biol. 189:113-130.

19. Tabor, S., and C. C. Richardson. 1985. A bacteriophage T7RNA polymerase/promoter system for controlled exclusive ex-

pression of specific genes. Proc. Natl. Acad. Sci. USA 82:1074-1078.

20. Wald, E. R. 1989. Haemophilus influenzae as a cause of acuteotitis media. Pediatr. Infect. Dis. J. 8:S28-S30.

21. Weinberg, G. A., A. Ghafoor, Z. Ishaq, N. K. Nomani, M.Kabeer, F. Anwar, M. I. Burney, A. W. Qureshi, J. M. Musser,R K Selander, and D. M. Granoff. 1989. Clonal analysis ofHaemophilus influenzae isolated from children from Pakistanwith lower respiratory tract infections. J. Infect. Dis. 160:634-643.

VOL. 61, 1993

on July 21, 2019 by guesthttp://iai.asm

.org/D

ownloaded from