INFECTION AND IMMUNITY, Feb. 1973, p. 130-136Copyright 0 1973 American Society for Microbiology

Vol. 7, NQ. 2Printed in U.S.A.

Characteristics of Human Antibody toDiphtheria Toxin

M. BAZARAL, P. J. GOSCIENSKI,1 AND R. N. HAMBURGER

Department of Pediatrics, Immunology and Allergy Division, University of California, San Diego,La Jolla, California 92037

Received for publication 18 September 1972

Human antibody to diphtheria toxin was analyzed with an ammonium sulfateradioimmunoassay. In 15 immune sera with titers of between 0.03 and 25 units/ml, the avidity was similar to that of U.S. standard antitoxin. The proportion ofthe post-immunization anti-diphtheria antibody specific for fragment A was

variable, but in the 10 sera tested constituted a substantial fraction of the anti-body. Approximately 2% of healthy adult males tested failed to make an anti-diphtheria toxin response after two booster doses of 1.5 Lf adsorbed diphtheria-tetanus toxoid, although these nonresponders made precipitating antibody totetanus toxoid and had normal immunoglobulin levels.

The single polypeptide chain of intact diph-theria toxin (62,000 daltons) contains a specificprotease-sensitive site (1, 5). The nicked toxinproduced upon limited proteolysis consists oftwo peptides which are held together by a di-sulfide bond. The toxin and nicked toxin havethe same in vivo toxicity. After incubation ofnicked toxin with a thiol-reducing agent, thetwo polypeptides, fragment A (24,000 daltons)and fragment B (38,000 daltons), may be sep-arated. Fragment A catalyzes the adenosinediphosphate ribosylation of transferase H,whereas intact or nicked toxin, and fragmentB, have no enzymatic activity. Fragment B,however, is apparently required for the passageof fragment A into cells, since only toxin andnicked toxin, but not fragment A, are toxic invivo (2, 6).Both diphtheria toxin and fragment A are

soluble in 45% saturated ammonium sulfate,which permits the primary binding test of Farr(4) to be used to measure antibody produced tothese antigens. We have applied this methodto the study of some of the parameters of thehuman immune response to diphtheria toxoid.

MATERIALS AND METHODSDiphtheria toxin. Partially purified diphtheria

toxin (2,000 Lf/ml, lot D-249) was purchased fromConnaught Medical Research Laboratories, Toronto,Canada. Approximately 85% of the toxin in this prep-aration was nicked toxin. Pure diphtheria toxin andfragment A of diphtheria toxin were prepared from

I Present address: Dept. of Pediatrics, U.S. Naval Hos-pital, San Diego.

130

this material by using diethylaminoethyl (DEAE)-cellulose and Sephadex G-100 column chromatog-raphy as described by Collier and Kandel (1). Nocontamination could be detected in the purified toxinby electrophoresis on polyacrylamide gels in the pres-ence of sodium dodecyl sulfate (SDS gels) underconditions capable of detecting 1% of the total sam-ple applied. Fragment A contained no contaminantsdetectable by immunodiffusion (Fig. 1) or SDS gels(Fig. 2). Protein content of pure diphtheria toxin andof fragment A was estimated by the optical density at280 nm, assuming an extinction coefficient of 10.4(El %).(1 C%m)

Diphtheria toxin and fragment A were labeled byan adaptation of the chloramine-T method of Hunterand Greenwood (7). To 5 mCi of carrier-free Na 1251Iwere added 0.7 ml of diphtheria toxin in PBS (0.1 MNaCl, 0.05 M sodium phosphate, pH 7.35) and 0.1ml of chloramine-T (1 mg/ml in water). After a 10-min incubation period at room temperature, the re-action mixture was applied to a 20-ml bed volumeSephadex G-25 column equilibrated in PBS, andeluted at 1 ml/min to separate the radioactively la-beled protein from the other reactants. The fractionof effluent counts eluting with the protein was 99%when 0.35 mg of diphtheria toxin was labeled and21% when 0.7 mg of fragment A was labeled. On SDSgels the 1251 diphtheria toxin counts appeared in asingle band coincident with the carrier diphtheriatoxin, and the 1251 fragment A counts appeared astwo discrete bands; 55% of the counts were coin-cident with the carrier fragment A and 45% ran ata mobility appropriate for fragment A dimers. Afterreduction, 12% of the toxin counts ran as intact toxin,38% as fragment A, and 50% as fragment B. Thecounts in labeled reduced fragment A appeared intwo bands; 96% of the counts were coincident withthe carrier fragment A, and 4% ran as fragment Adimers. For assays of anti-fragment A activity, frag-

on June 2, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

HUMAN ANTIBODY TO DIPHTHERIA TOXIN

ment A (1.0 ug/ml) was incubated for 10 min atroom temperature in 0.1 M dithiothreitol before dilu-tion to the 5 ng/ml assay concentration.

The 125I label in both fragment A and diphtheriatoxin was completely precipitable with excess anti-body in the ammonium sulfate test system, but overa 3-month period of storage in the PBS containing0.1% NaN3 and 0.5% human serum albumin the frac-tion precipitable decreased to 92%. In the absence ofspecific antibody, 6% of the label precipitated in theammonium sulfate test system.

Diphtheria toxoid. A solution of purified toxin(8.4 mg/ml in 0.15 M sodium borate buffer, pH 7.9)in 0.07% formaldehyde (8) was incubated for 0, 8, 20,and 40 hr at 37 C. At the end of the incubation period,samples of treated toxin were diluted 1/200 into dilu-ent (0.25% human serum albumin in PBS) for toxicitytesting, and into SDS gel buffer containing 7% mer-captoethanol. In intradermal skin tests on rabbits,0.05 ml of a 42-ng/ml dilution of the 0-hr sample pro-duced a 3.0-cm diameter zone of erythema after 48hr, the 8-hr sample produced trace erythema, andthe 20- and 40-hr samples gave no reaction. On SDSgels, mercaptoethanol-treated samples gave a distri-bution of intact toxin, fragment A, and fragment Bsimilar to the original preparation, whereas the 8-hrsample dissociated only slightly, and the 40-hr sam-ple apparently contained no molecules which couldbe dissociated.

Immunodiffusion. Plates for Ouchterlony immu-nodiffusion (10) consisted of a 0.8-mm thick layer of1% agar in buffer (0.14 M NaCl, 0.005 M sodiumphosphate, pH 7.4), covered with a plastic template;the wells contained 50 gliters, and were 0.85 cmapart. Horse anti-diphtheria toxin (National DrugCo., Philadelphia, Pa.), diluted to 100 units/ml wasused for the analysis of diphtheria toxin and frag-ments.

Anti-tetanus antibodies were detected in humansera by reaction with tetanus toxoid (fluid toxoid,approximately 15 Lf/ml, Lederle Laboratories, PearlRiver, N.Y.). Serum samples which did not producea precipitin line after 72 hr of incubation at 23 C wereconsidered to be negative.

Electrophoresis. Electrophoresis in 10% acryla-mide gels containing 0.1% SDS was done as de-scribed by Weber and Osborn (13). Samples to be re-duced were incubated for 30 min at 23 C in buffercontaining 7% mercaptoethanol and 0.1% SDS be-fore electrophoresis.

Immunizations. Immunizations consisted of in-tramuscular injections of 0.5 ml of adsorbed com-bined diphtheria and tetanus toxoids, containingapproximately 1.5 Lf of diphtheria toxoid per dose(Merrell-National Laboratories, Swiftwater, Pa.).

Antibody assays. Test sera were diluted 1/20initially in PBS. Subsequent dilutions were made in1/11 normal rabbit serum (NRS), which contained anamount of carrier protein equivalent to 1/20 humanserum, as determined by the optical density at 280nm (OD2J0) of the redissolved precipitate formed in45% saturated ammonium sulfate (SAS). All dilutionsof U.S. standard diphtheria antitoxin (lot A-27, 6units/ml) were made in 1/11 normal rabbit serum.

6

5o

I0

030

FIG. 1. Immunodiffusion of diphtheria toxin andfragment A. The center well contains anti-diphtheriatoxin, well 1 contains purified diphtheria toxin (0.9mg/mO, wells 2 and 3 contain fragment A (0.8 mg/ml), and wells 4 and 6 contain partially purified diph-theria toxin (1.0 mg/mI) as supplied by ConnaughtLaboratories.

Assays were done in polystyrene test tubes (12 by 75mm). Labeled antigen (0.5 ml in 0.5% human serumalbumin-PBS) was added to 0.5 ml of dilutions of thetest sera with an automatic dispensor (Labindus-tries). After a 1-hr incubation period, 1.0 ml of 90%SAS in water was added, mixed thoroughly, and heldfor an additional 30 min. Samples were then centri-fuged in a swinging-bucket multiple-place head for45 min at 1,200 x g. The supernatant fluids were sub-sequently poured off, and the tubes were inverted todrain. The 125I in the precipitate was counted in anautomatic well counter (Nuclear-Chicago model 4330)using polystyrene test tubes (16 by 150 mm) as adap-tors for the assay tubes. All operations were carriedout at room temperature.

Percent of radioactivity precipitated was com-puted by the formula: counts in sample - NRScounts/counts precipitated by hyperimmune serum -NRS counts. No correction was made for reducednonspecific precipitation in the presence of increasedspecific precipitation. All measurements were theaverage of duplicate determinations. Samples werecounted until at least 25,000 decays were countedin the 100% samples; a 1-min counting period wasused for I25I diphtheria toxin, and a 10-min periodwas used for 1251 fragment A. Unless otherwise noted,antitoxin titers were measured by using 2.5 ng of125I toxin per ml as the antigen.

131VOiL 7, 1973

on June 2, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

BAZARAL GOSCIENSKI, AND HAMBURGER

MM_

Ii.,w ......

s. ....

i,.....

....

*:

\ .. /

45 6 7

FIG. 2. SDS gels of diphtheria toxin, toxoid, fragment A. Gel I contains 10 Mg of purified toxin; gel 2, 10 Mgof reduced purified toxin; gel 3, 10 Mg of fragment A. Gels 4 to 7 contain 2.1 Mg of reduced samples taken atvarious times during the conversion of toxin to toxoid. Gel 4, 0 hr; gel 5, 8 hr; gel 6, 20 hr; gel 7, 40 hr. Thedirection of migration (anode) is toward the top of the photograph.

Because of the differences in gamma globulin con-centration among normal human sera, it was neces-

sary to ascertain that these differences do not greatlyaffecb the apparent titer for a given concentration ofantibody. Assays of 'Aloo dilution of an antiserumdiluted in several different concentrations of carrierdemonstrated that 'A, normal rabbit serum was suf-ficient to produce maximal specific precipitation ofantigen. Doubling the carrier concentration resultedin the precipitation of an additional 2% of the totalamount of antigen, resulting from an increase in thenonspecific precipitation. Decreasing the carrierconcentration by 1/3 caused a decrease of 5% in thetotal antigen precipitated. Since in a series of 23 un-

selected normal human sera the total protein precipi-tated in the assay conditions differed by no more

than +62% and -29% from the mean, normal varia-tion in serum immunoglobulin concentrations is un-

likely to produce significant error. Substantial error

could result, however, if the test serum had abnor-mally low immunoglobulin concentrations. To pre-clude this possibility, the OD25O of the redissolvedprecipitate was measured in samples with a low ap-parent titer at a !Ao dilution.

RESULT

Assay of standard diphtheria antitoxin.Assays of U.S. standard diphtheria antitoxinwere done by using diphtheria toxin and frag-ment A (Fig. 3). The antigen binding capacity

(ABC 33%) per unit is defined as the amountof antigen precipitated by one unit of antitoxinat the concentration of antitoxin required toprecipitate 33% of a specified amount of anti-gen in the ammonium sulfate assay. The ABC33% at 2.5 ng of diphtheria toxin per ml is 0.87Mg/unit, the ABC 33% at 25 ng of diphteriatoxin per ml in 4.3 ,g/unit, and the ABC 33%at 2.5 ng of fragment A per ml is 0.35 ,g/unit.The ratio of the antigen binding capacity fordiphtheria toxin to the antigen binding capac-ity for fragment A is 2.48 and is similar to theratio of the molecular weights of diphtheriatoxin and fragment A, which is 2.67. This re-sult implies that the antigenic sites recognizedby the standard antitoxin are present on bothparts of the molecule at approximately thesame density (sites/dalton) on each of the poly-peptide chains. The ration of the ABC 33% attwo different antigen concentrations is a mea-sure of the avidity of the antiserum; in thiscase the ratio ABC 33% at 2.5 ng/ml: ABC 33%at 25 ng/ml is 0.20 and may be used as a refer-ence figure for comparison with experimentalsera.Accuracy of the technique was estimated by

assaying 10 individually diluted samples at dif-ferent concentrations. At 6.0 x 10-3 units/ml,

132 INFECT. IMmNf

on June 2, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

HUMAN ANTIBODY TO DIPHTHERIA TOXIN

O .0%0%f 0o--o 2.5 ng/mi 1"R1 Fragment //6 ~~~~~~~~A[0ax

X 200_/z/

0 100 /

50 -

z

20

20

5-

I

0 2 5 10 20 50 100

% OF RADIOACTIVITY PRECIPITATEDFIG. 3. Binding curves for U.S. standard diphthe-

ria antitoxin lot A-27.

the standard deviation was 3.0%; at 1.2 x 10-sunits/ml, 6.3%; at 2.4 x 10-i units/ml, 10.3%;and at 4.8 x 10-5 units/ml, 25.4%. The lowerlimit of the useful range of the assay, if only a

single determination is done, is taken to be2.5 x 10-' units/ml; since 0.025 ml of serumis used in the assay, this corresponds to 1.0 x10- 2 units/ml in the test serum. Additionalaccuracy may be achieved by the use of mul-tiple determinations. Serum anti-diphtheriatoxin levels as low as 2 x 10- 4units/ml are de-tectable but are subject to large errors result-ing from variations in the protein concentrationof the test sera.

Blocking with toxin and toxoid. To verifythe specificity of the assay, blocking curveswere run with diphtheria toxin and a 40-hr tox-oid preparation, using a human antiserum todiphtheria toxoid at a dilution which precipi-tated 87% of the 125I toxin in the absence ofadditional unlabeled toxin (Fig. 4). The reactionmay be completely blocked with both toxin andtoxoid. Over most of the curve, the toxin andthe toxoid block the reaction to the same ex-

tent, which suggests that the antigenic deter-minants common to diphtheria toxin and tox-oid are stable during the conversion of toxin totoxoid.

Characteristics of human immune sera.Selected human sera were studied with respectto their anti-diphtheria toxin level, avidity, andanti-fragment A level. Antibody levels in the

sera were determined relative to standardcurves, similar to those in Fig. 3, generated inthe same run by serial twofold dilutions of U.S.standard diphtheria antitoxin lot A-27. Dupli-cate assays were done on each sample at dilu-tions in which 25 to 55% of the antigen precipi-tated, except when the antibody level at 12odilution precipitated less than 25% of the anti-gen (Table 1). The sera 1 to 10 are drawn fromnormal adults with an unknown previoushistory of diphtheria immunizations; postim-munization sera were drawn 2 weeks afterinoculation with adult-type combined diphthe-ria-tetanus toxoid. Sera 11 to 15 were from chil-dren who had had a diphtheria-tetanus immu-nization series and were drawn before a boosterimmunization and 5 months afterwards. Over awide range of antibody levels, no large differ-ences in apparent avidity, as indicated by theratio of units at 2.5 ng/ml to units at 25 ng/ml,were observed, although generally the humanantisera were slightly more affected by dilution(less avid) than the hyperimmune horse stand-ard serum. There is, however, substantial vari-ation among human immune sera in the propor-tion of the antitoxin which is directed againstfragment A, but after a booster there is anti-fragment A in all the responder sera studied.

I_

w

4

C

C')

z

0C.)

25

20

15

10

5

0 10 2 0.4 .08 .016 .0064

CONCENTRATION OF BLOCKER (ug/mI)FIG. 4. Blocking of the reaction of 125I diphtheria

toxin with human anti-toxoid by toxin and toxoid.The ordinate point marked "NRS" indicates thequantity of antigen precipitated in the absence ofspecific antibody.

I I I I I

KTOTAL COUNTS

UNBLOCKED

//I*-*TOXINs--oTOXOID

//.

NRS __ 0 /_I

I I I ~~~I I I

133VoL. 7, 1973

on June 2, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

BAZARAL, GOSCIENSKI, AND HAMBURGER

TABLI 1. Characteristics of human antisera

Anti-diphtheria toxin (units/ml) Units at 25 ng | Anti-fragment A Units anti-frag-Boost

A2n/ A2n/ |a2n/permi)Aunits ment A/unitsAt 2.5 ng/mi At 25 ng/ml at 2.5 ng/ml) unt/ianti.toxina

1. First

2. First

3. First

4. First

5. First

6. FirstSecond

7. First

8. PreFirst

9. PreFirst

10. PreFirst

11. PreFirst

12. PreFirst

13. PreFirst

14. PreFirst

15. PreFirst

a'At 2.5 ng/ml.

2.48

0.044

6.10

1.13

1.23

0.031>0.6

0.119

0.0751.58

0.1253.36

0.8327.8

0.972.84

2.563.20

0.574.7

2.309.88

0.1104.13

1.94

0.063

5.70

0.87

0.88

0.031

0.113

0.0551.06

0.1093.27

0.5816.3

0.93

2.18

0.58

1.83

0.103

0.78

1.43

0.93

0.77

0.88

1.0

0.95

0.730.67

0.870.97

0.700.59

0.96

0.85

1.0

0.80

0.94

2.74

0.077

5.74

1.30

0.40

0.014

0.126

<0.0040.220

0.0951.32

0.0256.72

0.60

2.18

0.12

2.36

0.049

1.10

1.75

0.94

0.58

0.33

0.45

1.06

<0.050.14

0.760.37

0.030.24

0.62

0.85

0.02

1.03

0.45

Unresponsiveness to diphtheria immu-nization. Two individuals who have remainedSchick positive after repeated diphtheria tox-oid immunizations were studied. One is a 7-year-old girl who received a primary diphthe-ria-pertussis-tetanus immunization series anda booster injection of toxoid, after which shewas Schick positive. Her anti-diphtheria toxintiter measured before and after a secondbooster was less than 0.0002 unit/ml. Immuno-logical evaluation was initiated because shedeveloped meningitis and pyarthrosis due toHemophilus influenzae type b at 6 years of age.Her immunoglobulin levels are normal, andshe has no apparent immunological deficiency.A second nonresponder, a 37-year-old man,

had had a primary immunization in childhoodin addition to at least three subsequent boosterdoses of toxoid and has remained Schick posi-tive. His anti-diphtheria toxin titer both be-fore and after an additional booster was lessthan 0.0002 unit/ml. There is no suggestionof a generalized immune deficiency, and hehas normal immunoglobulin levels (entry 8,Table 2). His wife has a normal level (0.02units/ml, no recent booster), and the anti-diph-theria toxin response of his children is entirelynormal (entries 11-15, Table 1).An investigation of the frequency of unre-

sponsiveness was made using sera drawn 2weeks after booster immunization ofyoung men(U.S. Marine Corps inductees). Of 214 men

134 INFECr. IMMUNrr

on June 2, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

HUMAN ANTIBODY TO DIPHTHERIA TOXIN

tested, 7 had less than 0.0002 unit/ml antitoxin(entries 1-7 in Table 2). Six of seven men weregiven a second booster 30 days after the first;four of these men made no immune response

detectable in sera drawn 2 weeks afterwards.Thus, the frequency of unresponsiveness toone booster dose of toxoid in this population isabout 3%, and the frequency of unresponsive-ness to a second booster is about 2%. Absenceof an immune response to diphtheria appears

to be an isolated deficiency, unrelated to a

generalized immune deficiency, since the im-munoglobulin levels are normal (Table 2) andsince all but one of these men made precipi-tating antibody to the tetanus toxoid withwhich they were concurrently immunized.

DISCUSSIONAnti-diphtheria toxin assays recently re-

viewed by Van Ramshorst (11) include bio-assays and passive hemagglutination. A spe-cialized radioimmunoassay has also been usedand has demonstrated that in the sera testedthe primary binding titer is proportional to theneutralization titer (9). The ammonium sulfateprimary binding assay described here has sub-stantial advantages in accuracy and ease ofperformance and is as sensitive as other tech-niques. The finding that human immune sera

have an avidity similar to the standard serum

independent of antibody levels is not unex-pected since in all probability the individualstested have had multiple immunizations.Both fragment A and fragment B are immu-

nogenic, and fragment A retains its ability toreact with antibody when separated from thediphtheria toxin molecule. It has been sug-gested (5) that the immunological reactions ofspontaneously dissociated fragments of diph-theria toxin may account for the presence ofsome of the multiple antigenic components pre-viously observed in purified toxin preparations.Free fragment A is, however, unlikely to bepresent in the immunizing toxoid, since cross-

linking resistant to SDS and mercaptoethanoloccurs between fragment A and fragment Bduring formaldehyde treatment. Indeed, sinceintracellular dissociation of the fragments isapparently necessary for the toxic activity, theobserved cross-linking is sufficient to accountfor the loss of toxicity during formaldehydetreatment. It seems likely then that the anti-fragment A resulting from immunizations withtoxoid is produced to the fragment A in theconfiguration in which it occurs in associationwith fragment B, and that many of the antigenicdeterminants are retained when fragment A is

TABLE 2. Immunoglobulina levels in men unrespon-sive to a single booster dose of diphtheria toxoid

Antitoxin Antitoxin Antiafter first after sec- teta IgG IgM IgA

boost ond boost e n (mg%) (mg%) (mg %)(units/ml) (units/ml)

1. <0.0002 0.068 + 1,500 300 1652. <0.0002 <0.0002 + 840 200 3653. <0.0002 0.112 _ 840 99 4404. <0.0002 <0.0002 + 1,500 220 3805. <0.0002 <0.0002 + 1,500 150 2706. <0.0002 + 929 210 2607. <0.0002 <0.0002 + 1,000 220 1908. <0.0002 _ + 1,550 204 210

aIgG, IgM, IgA; immunoglobulins G, M, and A, respec-tively.

rendered enzymatically active by dissociationfrom fragment B.The specific immunological unresponsive-

ness to diphtheria toxoid which occurs in asmall proportion of otherwise normal individu-als has been previously noted (3). In at leasttwo individuals that we have studied, no anti-body is detectable despite a documented his-tory of a primary immunization series and sub-sequent boosters. The percentage of nonre-sponders after two booster immunizationsfound in this study was 2%. Although specificnonresponsiveness to diphtheria toxoid is arelatively rare characteristic, it may be of sub-stantial interest should it prove to be a geneticeffect similar to the genetically determinednonresponsiveness of rodents to low doses ofspecific protein antigens (12).

ACKNOWLEDGMENTSSupport of this work has been provided by a Public Health

Service grant from the National Institute of Child Health andHuman Development.We appreciate the technical assistance of Eva Lydick.

LITERATURE CITED

1. Collier, R. J., and J. Kandel. 1971. Structure and activ-ity of diphtheria toxin. J. Biol. Chem. 246:1496-1503.

2. Drazin, R., J. Kandel, and R. J. Collier. 1971. Structureand activity of diphtheria toxin H. J. Biol. Chem. 246:1504-1510.

3. Edsall, G., J. S. Altman, and A. J. Gaspar. 1954. Com-bined tetanus-diphtheria immunization of adults: useof small doses of diphtheria toxoid. Amer. J. Pub.Health 44:1537-1545.

4. Farr, R. S. 1958. A quantitative immunochemical meas-ure of the pritnary interaction between I* BSA andantibody. J. Infect. Dis. 103:239-262.

5. Gill, M. D., and L. L. Dinius. 1971. Observations on thestructure of diphtheria toxin I. J. Biol. Chem. 246:1485-1491.

6. Gill, M. D., and A. M. Pappenheimer, Jr. 1971. Struc-ture-activity relationships in diphtheria toxin. J.Biol. Chem. 246:1492-1495.

7. Hunter, W. M., and C. C. Greenwood, 1962. Prepara-

135VoL 7, 1973

on June 2, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

136 BAZARAL GOSCIENSKI, AND HAMBURGER

tion of iodine-131 labelled human growth hormone ofhigh specific activity. Nature (London) 194:495-496.

8. Lorence, H. S., and A. M. Pappenheimer, Jr. 1948. Im-munization of adults with diphtheria toxoid. Amer. J.Hygiene 47:226-232.

9. Newcomb, R. W., K. Ishizaka, and B. L. DeVald. 1969.Human IgG and IgA diphtheria antitoxins in serum,nasal fluids, and saliva. J. Immunol. 103:215-224.

10. Ouchterlony, 0. 1968. Handbook of immunodiffusion andimmunoelectrophoresis. Ann Arbor Publishers, Inc.,Ann Arbor.

INFECT. IMMUNITY

11. Van Ramshorst, J. D. 1971. Titration of diphtheria andtetanus antitoxins in sera of low titer. Bull. W.H.O.45:213-218.

12. Vaz, N. M., E. M. Vaz, and B. B. Levine. 1970. Relation-ship between histocompatability (H-2) genotype andimmune responsiveness to low doses of ovalbumin inthe mouse. J. Immunol. 104:1572-1574.

13. Weber, K., and M. Osborn. 1969. The reliability of mo-lecular weight determinations by dodecyl sulfate-poly-acrylamide gel electrophoresis. J. Biol. Chem. 244:4406-4412.

on June 2, 2018 by guesthttp://iai.asm

.org/D

ownloaded from