The physicochemical properties and biologic activities of gamma M-globulins

MISSING IMAGE

Material Information

Title:
The physicochemical properties and biologic activities of gamma M-globulins
Physical Description:
vi, 98 leaves : ill. ; 29 cm.
Language:
English
Creator:
Hill, William Carl, 1933-
Publication Date:

Subjects

Subjects / Keywords:
Antibodies   ( mesh )
Antigens   ( mesh )
Gamma-Globulins   ( mesh )
Horses   ( mesh )
Macroglobulins   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1966.
Bibliography:
Includes bibliographical references (leaves 93-98).
Statement of Responsibility:
by William Carl Hill, Jr.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 25883955
ocm25883955
System ID:
AA00012913:00001


This item is only available as the following downloads:


Full Text









THE PHYSICOCHEMICAL PROPERTIES

AND BIOLOGIC ACTIVITIES OF

KM-GLOBULINS


















By
WILLIAM CARL HILL, JR.










A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY









UNIVERSITY OF FLORIDA

APRIL, 1966
















ACKNOWLEDGEMENTS

The author wishes to express his sincere thanks to Dr. John J.

Cebra for his continuing encouragement, suggestions, and assistance

during the course of the work described in this dissertation and

bringing it to its final form.

Other members of the supervisory committee, Dr. Emanuel Suter,

Dr. Mendel Herzberg, Dr. Melvin Fried, and Dr. Paul Elliott have

rendered invaluable assistance which was greatly appreciated.

Members of the Department of Veterinary Science, especially

Dr. Wayne W. Kirkham, assisted with the numerous immunization injec-

tions and bleedings of the horse during the course of this study.

Without the untiring efforts of this group much of this work would

not have been accomplished.

And lastly, I must thank Laurie, my wife, for her patience,

support, and many hours she spent typing the manuscript.


















TABLE OF CONTENTS

Page
ACKNOWLEDGEMENTS ................... ...... ii

LIST OF TABLES .................... ..... iv

LIST OF FIGURES ..........................v

INTRODUCTION .. .. .. .. .. .. ... .. .. ... 1

REVIEW OF LITERATURE. ...................... 4

MATERIALS AND METHODS ...................... .24

RESULTS . .40

DISCUSSION. . .. .. .77

BIBLIOGRAPHY ... .. .. ... .. .. .. .. 93

BIOGRAPHICAL SKETCH ........................99
















LIST OF TABLES

Table Page

1. Quantitative Coprecipitation of 1251 Labelled SI
by Anti-horse Immunoglobulins and 7M-antibodies,
yM-subunits, or Other Horse Immunoglobulins. ... 65

2. Ability of yM-subunit to Bind SI as Detected by
the Farr Method .. .. .. .. ... ... .68

3. Protection Against Type I Pneumococcus Infection
in the Mouse by Equine Anti-SI Immunoglobulins. .76

4. A Comparison of the Properties Found for Horse
Immunoglobulins in this Study. .... .. 91

5. Properties of Immunoglobulins from Various Species. .92

















LIST OF FIGURES


Figure Page

1. Diplococcus pneumoniae. Type I, with its poly-
saccharide capsule stained with fluoroscein
conjugated yM-antibody. ... 25

2. Quantitative precipitin curves obtained for
sera taken at various times after initiation
of immunization with Type I pneumococci. ... 27

3. Reactions leading to reduction by 2-mercapto-
ethano. and dithiothreitol. 34

4. Resolution of yM-antibody and '6.4S immuno-
globulins' on Sephadex G-200. 43

5. Immunoelectrophoretic analysis of the horse anti-
SI components and their reaction with SI. .. .44

6. The antigenic analysis of horse anti-SI immuno-
globulins and their isolated 'light' polypeptide
chains. . ... ....... .45

7. Depolymerization of 7M-antibody by reduction with
two different thiol compounds used at varying con-
centrations and reaggregation of the subunits by
reoxidation. ... ..... .49

8. Antigenic relationship and antibody activity of
yM-antibody, its subunit and products of reoxidation. .51

9. A scheme for the study of the effect of 0.0025 M
Dithiothreitol on yM-antibody. ... 52

10. Release of free light chains by DTT. ... 54

11. A series of ultracentrifuge frames from various
reoxidation experiments showing the variable
results obtained. . .. .57

12. Separation by gel filtration on Sephadex G-200 of
the 7.2S (leading part) and 4.9S (trailing peak)
subunits of yM-antibody. ... 58










Figure Page

13. Imaunoelectrophoretic and radioim3unoelectro-
phoretic analyses of anti-SI serum and yM-
antibody . 61
125
14. 125 SI precipitated and coprecipitated by
anti-SI. . ... .... 63

15. Gel filtration of 1251 Dextran on Bio-Gel
P-30. . ... .... .. .70

16. Bio-Gel P-300 chromatography of 125I SI. ... 72

17. Integrity of the antigenic determinants of SI
following labelling with 1251 ... .73
















INTRODUCTION

This dissertation describes a series of investigations into

the nature of the reductive depolymerization of horse anti-pneu-

mococcal polysaccharide (Type I) yM-globulin, and reaggregation

of the subunits by reoxidation of disulfide bonds. Several pre-

liminary observations on the activity of horse anti-pneumococcal

antibodies have appeared in a previous publication by the author

(Hill, 1964). These observations may be summarized as follows:

1) The yM-antibody at a concentration of 0.5% had an s20,w of

17.3S and the reductively produced subunit at the same concentra-

tion had an s20,w of 6.1S.

2) Concentrations of 2-mercaptoethanol of at least 0.075 M

were required to depolymerize completely the yM-antibody.

3) Depolymerization under these conditions followed by treat-

ment with 1 M propionic acid resulted in the liberation of 32% of

the light chains of the molecule.

4) yM-globulin contains some determinants that are antigenically

different from some determinants on yG-globulin.

5) A portion of the anti-pneumococcal antibodies could be spe-

cifically precipitated with a polygalacturonic acid pecticc acid).

This is a reflection of the high galacturonic acid content of the

Type I polysaccharide.










Many of the physicochemical properties of the specific yM-anti-

body paralleled those found by others for various pathologic macro-

globulins. Since relatively large amounts of specific yM-antibody

were available for physicochemical studies, a detailed study of its

molecular properties was undertaken in conjunction with measurements

of its various biologic activities.

Previously, with a few exceptions, it had been assumed that

reduction with mercaptoethanol somehow altered the molecule in such

a way that it no longer reacted with its antigen. A detailed study

of the effect of reduction on the biologic activity of yM-antibody

would require a knowledge of the mildest conditions necessary to

achieve depolymerization, the extent of depolymerization at reducing

agent concentrations lower than necessary to depolymerize all the

molecules, and the number of disulfide bonds that are broken during

the course of depolymerization of the molecule.

Once the questions regarding the physicochemical manipulations --

of the yM-antibody were answered, it became necessary to develop or

adapt techniques to detect and quantitate the antibody activity of

the intact 7M-antibody as well as any activity retained by the sub-

unit. Visualization of any antibody binding activity by the sub-

unit would depend on indirect precipitation of the antigen which had

been labelled in a manner permitting itsveasy detection in a milieu

of proteins. The conditions of labelling had to be such that the

antigenic determinants of the antigen, the SI polysaccharide in our

case, were not destroyed. To this end, it was concluded that the

most desirable label would be a radioisotope which could be used for










qualitative detection of SI precipitation, by the technique of radio-

immunoelectrophoresis, and which could also be used for quantitative

determinations of its binding to yM-antibody and its subunits.

Since it has been established that depolymerization of yM-anti-

body abolishes or diminishes secondary effects of its antibody activ-

ity, the question that presents itself is: Can at least some of

these activities of macroglobulin be restored if one uses techniques

of oxidative reassembly that are known to restore activity to cer-

tain enzymes which are also inactivated reductively?

The experiments presented in the Results section of this dis-

sertation in part help to answer these questions. As the author

delved further into the nature of 7M-antibody, more questions were

raised, and some of these unanswered questions are treated in the

Discussion section.
















REVIEW OF LITERATURE

Immunology had its beginning in 1798 when a country doctor,

Edward Jenner, observed that dairy workers did not contract small-

pox if they had previously had a rather innocuous disease called

cowpox. Jenner deliberately infected persons with cowpox thereby

successfully protecting them from smallpox. From this momentous

beginning immunology has grown in two directions. The first is a

continuation of the work Jenner began, that is, protecting people

from the ravages of disease. The second direction involves a con-

tinuing attempt to understand the basic mechanisms of the immune

response. The attempt to quantitate antigen-antibody reactions and

to characterize the molecules involved in immune reactions has led

to the development of a new discipline, immunochemistry. Immuno-

chemistry employs the systems and nomenclature of immunology and

the tools and methods of analytical biochemistry. The concerns of

immunochemists have expanded beyond the field of immunology and

they have obtained information about the structure and properties

of certain proteins and other macromolecules that has contributed

to biochemical knowledge.

By 1936 it had been clearly established that antibodies were

actually modified serum proteins. Techniques had been developed

that allowed one to obtain solutions in which 98% or more of the

protein was specific antibody. Heidelberger and Pedersen (1937)









were able to study the sedimentation characteristics of such puri-

fied antibodies derived from different species and reactive with

different immunogens. The antibodies studied by these authors were

isolated from rabbits immunized with either egg albumin or Type III

pneumococci and horses immunized with Type I pneumococci. The anti-

bodies from the rabbit reactive with either immunogen were found to

have a sedimentation coefficient (s20) of 7.0S to 7.4S while anti-

pneumococcus antibodies from the horse had an s20 of 18.2S. In an

extension of this work, Kabat and Pedersen (1938) determined the

molecular weight of antibodies from several animal species reactive

with the same pneumococcal polysaccharide. Their results may be

summarized as follows: 1) Antibodies from rabbit and monkey were

found to have an s20 = 7S and a molecular weight of about 157,000

and 2) antibodies from pig, cow, and horse were found to have an

s20 = 18S and a molecular weight of 930,000. From these data the

authors concluded that mammals fell into two groups with respect to

the size of antibody molecules which they synthesized. One group

was composed of those animals that produced antibodies of a size

similar to that of the globulins found in largest amount in human

serum, i.e. having an s20 = 7S while the other group consisted of

animals producing antibodies of a much larger size. It was later

shown that while the response of the horse to pneumococcal poly-

saccharides was consistent with this grouping, the reaction of the

horse to diptheria toxoi;, as well as many other immunogens, was

inconsistent. Immunization of the horse with many immunogens could

and did lead to the production of antibodies of the smaller (820 =

7S) variety (Pappenheime:, t al.. 1940).









The molecules in the group of serum proteins known as anti-

bodies are formed in response to a stimulation by a 'foreign' sub-

stance, called an antigen or an immunogen, and each antibody mole-

cule reacts specifically with that immunogen, or a portion thereof,

which initiated its formation. Antibodies have the physicochemical

characteristics that place them in the group of serum proteins that

have been called 7-globulins. The 7-globulins were operationally

defined, in part, as that group of serum proteins that migrated

slowest in an electrical field when in an alkaline solution, i.e.

pH 8.6. Closer scrutiny has indicated that antibody activity can

be found in almost all of the electrophoretic groups of serum pro-

teins. It has been suggested that the term 'immunoglobulin' be

adopted which more accurately defines the groups than does '7-glob-

ulin'. The serum immunoglobulins are grouped into one of three dis-

tinct classes, yG-, yA-, or yM-immunoglobulins (Kunkel, 1960), called

alternately IgG-, IgA-, and IgM-immunoglobulins respectively (Bull.

WHO, 1964). A fourth class of immunoglobulins, yD-immunoglobulin,

has been reported, but its function is as yet unknown (Rowe and

Fahey, 1965). The above system of nomenclature reflects the close

interrelationship of the immunoglobulins. Molecules of the three

principal classes differ markedly from each other in their physico-

chemical properties and in their antigenic constitution. However,

they also have in common certain peptide sequences and polypeptide

chains (Fahey, 1963; Migita and Putnam, 1963).

Those pathologic proteins produced by multiple myeloma patients

have served as a source of globulins that are more homogeneous than










normal y-globulin. The myeloma protein of a multiple myeloma

patient seems to be produced by a clone of cells arising from a

single stem cell and the pathologic protein may comprise up to 80%

of the patient's total serum proteins (Putnam, 1960).

A large fraction of the multiple myeloma patients excrete a

protein, called Bence-Jones protein, in their urine. It has been

conclusively demonstrated that these are monomers or dimers of light

chains of the same antigenic type as those found on the intact mye-

loma protein of the same patient (Heremans, 1960).

Of all the immunoglobulins, the 7G-globulins have been most

thoroughly studied. This is mainly because of their relative abun-

dance in normal serum and the ease of their isolation as a homoge-

neous preparation. A molecule of yG-globulin has been found to con-

sist of two pairs of polypeptide chains covalently joined together

through disulfide bonds. The pairs of polypeptide chains consist

of one light chain, which has a molecular weight of 22,000 to 23,000

(Pain, 1963; Small, et al., 1963; Marler, et al.. 1964; Small and

Lamm, 1966), and of one heavy chain, having a molecular weight of

about 50,000. The total molecular weight of yG-globulin is 140,000

to 150,000 (Pain, 1963; Small, et al., 1963; Marler, et al., 1964;

Small and Lamm, 1966), and the molecule has a sedimentation coeffi-

cient of 6.5S. The heavy chain carries the antigenic determinants

that are unique to the class of yG-immunoglobulins and it is referred

to as the 7-chain. The glycopeptide moiety, which accounts for 2% to

3% of the total weight of the 7G-molecule, is also associated with

the y-chains (Fleishman, et al., 1963; Press and Porter, 1964). The

light chains of the human species occur as two major antigenically









distinct types and are denoted as K- (Kappa) and X- (Lambda) chains.

Light chains are also found on yA- and yM-globulins and seem to be

essentially similar regardless of the class of globulins from which

they are isolated (Chaplin, et al., 1965; Wier, 1964). The yG-

immunoglobulins have been found to contain two antigen binding sites

per 140,000 molecular weight (Karush, 1956, 1957; Singer and Campbell,

1955).

The information available on yA-globulins is relatively incom-

plete. These immunoglobulins are present in human serum in lower

concentrations than yG-globulins. Estimates of yA-globulin con-

centration have ranged from about 150 mg% (Heremans, 1960; Chodirker

and Tomasi, 1963) to 400 mg% (Fahey and Lawrence, 1963); however,

these proteins are found in relatively high concentration in various

human secretions, i.e. those of the lactating mammary gland, intes-

tinal secretions, and nasal secretions, saliva and tears (Hanson,

1961; Tomasi and Zigelbaum, 1963; Chodirker and Tomasi, 1963; Tomasi,'

et al., 1965). A yA-antibody has been isolated from precipitates,

formed with horse antiserum, along with other immunoglobulins; how-

ever, its ability to precipitate antigen is in doubt (Klinman, et

al., 1965). The yA-globulins appear to contain heavy chains of

higher molecular weight than that of yG-globulin (Cebra and Small,

1966). These heavy chains are immunogenically distinct and are

referred to as a chains. The yA-globulin molecules also contain

light chains, as do the yG- and yM-globulins (Wier, 1964). yA-glob-

ulins have an s20,w = 7S to 10S (Smith and Brown, 1950) but may

occur as multiples of the 7S molecule. The glycopeptide moiety









present in human and horse 7A-globulin comprises some 5% to 107 of

the total weight of these proteins (Heremans, 1959; Schultze, 1959).

The remaining class of immunoglobulins, the 7M-globulins, are

presently coming under intensive study and it is this class of

immunoglobulins that is the subject of this dissertation. Much of

the information about yM-globulin that has accumulated to date has

come from investigations of those yM-globulins produced in large

quantities by humans suffering from Waldenstr8m's macroglobulinemia

(Waldenstr8m, 1944; Kunkel, 1960). The macroglobulin from these

patients has been found to have an s020,w = 18S and a molecular

weight estimated to be about 1,000,000. The molecule has a some-

what higher carbohydrate content than does yG-globulin. Normal

human 7G-globulin was found to contain 2.49% to 2.6% carbohydrate

while yM-globulin contained 9.8% to 10.23% carbohydrate (MUller-

Eberhard, Kunkel and Franklin, 1956; Miller and Metzger, 1965).

The antibodies produced by the horse, cow, etc. in response to pneu-

mococcal polysaccharides, as mentioned above, are of the yM-globulin

class. This yM-antibody response was found not to be restricted to

the few animals noted above but to be a rather ubiquitous response.

Bellanti, et al. (1963) showed that the earliest detectable anti-

body produced by rabbits and humans upon immunization with various

Salmonella species was of the yM-globulin type. Other examples of

yM-globulin production include: Human isoagglutinins (Pedersen,

1945), rheumatoid factor of man (Edelman, et al., 1958; Heimer, et

al., 1958), and rabbit anti-sheep red blood cell antibodies (Fahey

and Humphrey, 1962). More recently, yM-antibodies have been iso-

lated which were specifically reactive with one of two haptens.










Rockey, et al. (1964) isolated yM-antibodies to the hapten p-azo-

phenyl-p-lactoside from immune horse serum. Onoue, et al. (1965)

isolated yM-antibodies from the rabbit reactive with the hapten p-

azobenzenearsonate. The titer of circulating 7M-antibodies is sig-

nificantly higher than the yG-antibody only early in an immune re-

sponse, a notable exception being the case of the pig, cow, or horse

immunized with the various pneumococci.

Testing the assumption that the macromolecular globulin from

macroglobulinemic patients was an aggregate of smaller molecules,

Deutsch and Morton (1957, 1958) attempted to dissociate the mole-

cules using various combinations of salt, pH and temperature, but

the native molecules retained their integrity during these manipu-

lations. However, addition of sulfhydryl compounds such as 2-

mercaptoethanol or cysteine resulted in the conversion of the macro-

globulins (18S) to subunits (6.5S). The conditions of reduction

used by Deutsch and Morton were 0.1 M mercaptan in phosphate buffer

(p = 0.2; pH = 7.4) and incubation at room temperature for 24 to

48 hours. If the mercaptan was subsequently removed by dialysis

the amount of 6.5S material was decreased and there was an appear-

ance of components which sedimented faster (13S and 18S). Removal

of the mercaptan by dialysis in the presence of 0.02 M iodoacetamide,

a sulfhydryl blocking agent, prevented the reassociation and all

the protein remained as the subunit. Dissociation of yM-antibody

under the conditions described above led to a concomitant loss of

antibody activity as measured by precipitation or agglutination.

Thus the decrease of antibody activity in serum following reduction










has been used as an indication of the proportion of specific activity

associated with the yM-globulin. Reduction of 7G-antibody under the

same conditions does not appear to affect its precipitating or agglu-

tinating activity or its physicochemical properties. Thus mild (0.1 M

mercaptan) reduction appears to destroy the ability of yM-antibody of

the appropriate specificity to agglutinate or immobilize bacteria

(Bellanti, et al. 1963; Nossal, et al.1964), to agglutinate eryth-

rocytes (Fudenberg and Kunkel, 1957; Chan and Deutsch, 1960; Bauer

and Stavitsky, 1961), to sensitize erythrocytes for complement lysis

(Stelos and Taliaferro, 1959), to neutralize phage (Uhr and Finkel-

stein, 1963; Bauer, et al., 1963) and specifically to precipitate

antigen (Josephson, et al., 1962; Benedict, et al., 1963). In only

one of these studies was an attempt made to isolate specific macro-

globulin antibody and to correlate the apparent loss of activity

with reductive depolymerization (Fudenberg and Kunkel, 1957). Re-

duction of the red blood cell cold agglutinins with sulfhydryl re-

agents produced 6.5S subunits from the antibody. These subunits were

found to be neither capable of agglutinating the red blood cells nor

of rendering the red blood cells susceptible to agglutination by

various sera in a Coomb's test. However, another of the above studies

(Chan and Deutsch, 1960) suggested that there was conservation of

antibody combining sites following reduction. 'Rh' saline agglutinin

titers fell to zero when dialyzed vs 0.1 M mercaptoethanol. Removal

of the mercaptan by dialysis vs buffer did not restore the ability

of the serum to agglutinate red blood cells, however, such cells

treated initially with the reduced serum could be agglutinated with

an antibody to the saline agglutinins. The proteins rendering the









cells Coomb's reagent positive were, most probably, subunits of the

agglutinin that had an intact antibody site. In subsequent exper-

iments, Deutsch and Fudenberg, using isolated saline agglutinins,

could not demonstrate any Coomb's reagent activity of the reduced

antibody (Chan and Deutsch, 1960). More recently Jacot-Guillarmod

and Isliker (1962, 1964) have shown that reduced human anti-A and

anti-B yM-globulins could inhibit hemagglutination by intact anti-

body and could also specifically render the appropriate erythrocytes

susceptible to agglutination by an anti-globulin serum. Onoue, et

al. (1964) used yet another technique, radioimmunoelectrophoresis

(Hochwald, et al. 1961), to demonstrate the binding of radiolabelled

antigen by yM-subunits. The yM-antibody was isolated from rabbits

immunized with p-azobenzenearsonate conjugates. The antibody was

totally reduced to 6S subunits and then electrophoresed in agar gel.

The slides were then developed with a mixture of sheep anti-rabbit

yM-globulin and p- I azobenzenarsonate-insulin conjugate. Radio-

autographs indicated that the labelled antigen was specifically bound

in the band of precipitation given by the subunits.

Although the reductively produced subunits of 7M-antibody are

incapable of eliciting many of the secondary effects of the antigen-

antibody interaction, i.e. those that are due to its cross-linking

properties, at least a portion of these activities can be restored

by oxidative recombination (Jacot-Guillarmod and Isliker, 1962;

Schrohenleher, et al., 1964). Jacot-Guillarmod and Isliker succeeded

in restoring agglutination activity of reduced human anti-A or anti-B

yM-isoantibodies by dialyzing the reduction mixture against Veronal.










buffer, pH 7.4, according to the procedure which was used by Deutsch

and Morton (1958) to recombine subunits of Waldenstrbm's macroglobulin.

Jacot-Guillarmod and Isliker could not demonstrate recombination of

the subunits or recovery of activity if the reduction was accomplished

using a sulfhydryl reducing agent such as sodium thioglycolate or

cysteamine. Reduction with sodium borohydride (0.01 M to 0.05 M) did

not completely abolish the activity of the antibody, however, recovery

of a majority of the initial activity by reoxidation was demonstrated.

Likewise, Schrohenloher, et al. (1964), reoxidized rheumatoid factor

using the same method and restored a significant level of its ability

to precipitate with aggregated human 7-globulin.

A feature that distinguishes an antibody from other proteins is

its ability to combine specifically with antigen. The antigen-anti-

body interaction is most easily visualized when the two components

are in the proper proportions so that a precipitate is formed. If

one attempts to explain this precipitation on a molecular basis,

it must be postulated that the antigen be capable of having more

than one antibody bound to it and conversely, tnat the antibody must

be capable of binding more than one antigenic site. Ample evidence

is now available that yG-antibody contains two and only two anti-

gen binding sites (Karush, 1956, 1957). The number of binding sites

of a 7M-antibody molecule has not been established with certainty.

Recently, Onoue, et al. (1965) presented evidence that indicates

there are six binding sites per 1,000,000 molecular weight for rabbit

yM-anti-p-azobenzenearsonate. If the recently determined molecular

weight for this molecule, of 850,000 to 900,000 (Lamm and Small, 1966)










is accepted then there are more likely five sites per molecule, or a

number equivalent to the number of 6S subunits that make up the whole

molecule.

The secondary biologic activities of antibodies differ. The

more important of these differences may be summarized as follows:

a) Agglutination and its inhibition. Draff and Shulman (1965)

found that, on a weight basis, yM-antibody had a higher agglutinating

activity than did yG-antibody to the same hapten. These authors

isolated rabbit yG- and yM-antibodies to the phenylarsonate hapten.

Then, using human erythrocytes coupled to arsanilate as the indicator,

they found that on a weight basis four times as much yG-antibody as

yM-antibody was required to give the same reproducible end point.

Calculations based on the currently accepted molecular weights of

7G- and yM-globulin indicates that, on a molecular basis, the 7M-

antibody is 26 times as effective as yG-antibody. A second study by

these authors determined the least amount of hapten required to inhib-:

it the agglutination of the conjugated erythrocytes by a particular

class of antibody. It was found that about 200 times as much hapten

was required for inhibition of the yM-antibody as was needed for

inhibition of a.. equal weight of 7G-antibody.

The results of Robbins, et al. (1965) are in close agreement

with the results cited above. These authors investigated the ability

of purified rabbit yG- and 7M-antibodies to Salmonella typhimurium

to agglutinate these bacteria. They found that on a molar basis, yM-

antibody was some 22 times more efficient in agglutinating the bacteria

than 7G-antibody. The other parameters which were also investigated

will be discussed below.










A more striking difference between the ability of yG- and yM-

antibody to cause agglutination was shown by Greenbury, et al.

(1963). yM- and yG-antibodies were isolated from rabbits immue to

human blood group A erythrocytes and their ability to cause hemag-

glutination was compared using a '50% agglutination' end point. They

found that while only 25 molecules of yM-antibody per cell were

required to produce this end point, 19,500 molecules of yG-antibody

per cell were required to achieve the same effect.

b) Opsonization. The in vivo clearance of S. typhimurium was

used by Robbins, et al. (1965) to quantitate the opsonic activity

of yG- and yM-antibodies to this organism. These authors found that

pretreatment of the bacteria with as little as 0.0006 mg/ml of yM-

antibody would result in phagocytosis of 50% of the organisms within

one hour after their injection into the tail vein of a mouse, while

concentrations of 0.131 mg of yG-antibody per ml were required to

produce the same degree of phagocytosis. It is interesting co note

that reduction of the yM-antibody with 2-mercaptoethanol almost abol-

ished the agglutination titer, however, the decrease in opsonic activ-

ity was not as dramatic. A concentration of 0.03 mg/ml of the re-

duced and blocked yM-antibody was required to give 50% phagocytosis

in one hour.

c) Placental passage. Transport of immunoglobulins across the

placental membrane is a highly selective process. The selection is

not due to size differences but to structural differences among the

immunoglobulins. Human yG-globulin is freely transported across the

membrane while neither yA- nor yM-globulin crosses this barrier,










(Franklin and Kunkel, 1958; Gitlin, et al., 1963). Brambell, et al.

(1960) found that there was a differential in the rates at which

papain fragments of rabbit 7G-globulins were transported across the

placenta. The Fc fragment (newer nomenclature for Porter's Fragment

III of rabbit 7-globulin, [Porter, 1959]) seems important for yG-

globulin transport since it crosses the placenta 11 times more rapidly

than does the Fab fragment (newer nomenclature for Porter's Fragments

I and II, [Porter, 1959]), but not quite as rapidly as intact yG-

globulin. Contrary to findings with humans, Hemmings and Jones (1962)

found that immunization of pregnant rabbits with human red blood cells

resulted in hemagglutinins of equal titer in both the maternal and

fetal circulation. In both cases, the hemagglutinins were restricted

to that region of a sucrose gradient after centrifugation in which

one would expect to find macroglobulin but not 7S 7-globulin.

d) Skin and smooth muscle sensitization. The 'skin fixation'

of an antibody depends on its ability to remain localized at an

intradermal injection site in guinea pig skin and subsequently cause

an anaphylactic reaction at the site when the antigen is administered

systemically. This technique is commonly called passive cutaneous

anaphylaxis (PCA). The ability of 7G-globulin of other species to

become 'fixed' in the guinea pig skin, while yA- and /M-globulins of

heterologous species do not possess this property, again seems to

reside in the Fc fragment portion of the active molecule. The Fc

fragment can block the fixation of intact yG-globulin while Fab frag-

ment cannot (Ovary and Karush, 1961). The inability of human yM-

antibody to fix in the guinea pig skin while yG-antibody to the same









antigen did become fixed was shown by Ovary, et al. (1960). It has

also been found that normal human yA-globulin as well as yA-myeloma

proteins would not sensitize guinea pig tissues (Franklin and Ovary,

1963).

Genetic Control and Cellular Synthesis. Two sets of alleles

controlling the light chains of human and rabbit yG-globulins respec-

tively have been useful as tools for study of genetic control of

immunoglobulin synthesis. The class specific antigen sites are

located on the heavy chain portion of the y-globulin molecule. Since

there are three such class-specific chains in humans, that is, 7, a,

and p chains, and two human light chain types, K and X, there are

six possible combinations of light and heavy chains for human immuno-

globulin molecules. The six possible combinations have been identified

in normal human serum. The various pairs are: yK, y7, OK, Xy, pK,

and p% (Fahey and Goodman, 1964). In recent studies, Pernis and

Chiappino (1964); Pernis, et al. (1965); and Bernier and Cebra (1964,

1965), found that individual cells of human spleen contained only

one class of heavy chain and one type of light chain.

In the rabbit spleen it was shown that a majority of the cells

(63% to 78%) that stained with a fluorescent anti-light chain also

were stained by an anti-y chain that had been coupled with a con-

trasting fluorochrome (Cebra and Goldstein, 1965). A large part of

the remaining fluorescing cells (14% to 21%) were stained by a fluor-

escent reagent specific for p chains (Cebra, Colberg, and Dray, 1966).

Also, it has been reported by Schoenberg, et al. (1965), that the

synthesis of yM-antibody in the rabbit was associated with a non-

phagocytic mononuclear cell while the yG-antibody to the same antigen










(diptheria toxoid) was synthesized in typical plasma cells. Thus,

yM-globulin appears to be synthesized in cells separately from the

other classes of immunoglobulins.

Sequential Appearance of Different Classes of Antbodyuring

an Antigenic Stimulation. The sequence of the synthesis of the

various classes of antibodies has been investigated using various

systems. A thorough study by Bauer, et al. (1963) used the bacte-

riophage T2 and key-hole limpet hemocyanin to stimulate rabbits

antigenically. The sequence of synthesis and secondary response

was similar with both antigens, therefore a description of only one

set of results is necessary. Antibodies were detected as early as

six days after a single injection of the antigen, 90% of which were

associated with yM-globulin as defined by its elution character-

istics from a DEAE-cellulose column. The proportion of antibody

eluted in the 7G-globulin region continued to increase with time

while that eluted in the yM-globulin region decreased so that by

day 16 the activity was about equally distributed between the two

classes. Some 60 days after injection of the antigen more than 99%

of the antibody activity was eluted in the yG-globulin region.

Restimulation of the rabbits at day 50 resulted in a marked increase

in both classes of antibodies, however, the absolute amount of yM-

antibody neither exceeded that produced by the initial stimulation

nor did it reach its maximum more rapidly. The level of yG-antibody

increased rapidly, reaching a maximum several fold greater than that

previously reached and in a shorter interval. The levels of yM-

antibody reached following the secondary injection accounted for

only a small fraction of the total activity of the serum.










Smith (1960) investigated the immune response of more than 150

premature and full term human infants. It was found that the infants'

anti-Salmonella tyhi agglutinin titers against the H antigen approx-

imated those of the immunized mothers, however, in only four cases

did the infants appear to passively receive anti-0 antibodies from

the mother. Subsequent immunization of infants having a H or 0 titer

less than 1:5 with commercial 'triple typhoid vaccine' produced an

increase of H agglutinin titers but no increase in 0 agglutinin

titers even after repeated multiple injections. The H agglutinin

was found to be sensitive to 2-mercapthoethanol and to have a sedi-

mentation coefficient equal to or greater than 19S as determined in

the Yphantis-Waugh partition cell (Waugh and Yphantis, 1953).

Evidence now available indicates that this sequential appear-

ance may be more apparent than real. The methods of assay of the

yM- and yG-antibodies are several fold more sensitive to detection

of yM-antibodies than yG-antibodies. Using techniques to minimize

the difference in detection limits, Altmeier, et al. (1965) have

reportedly reduced the time difference between detection of yM-anti-

body and detection of yG-antibody to as little as two days.

On a cellular level, Schoenberg, et al. (1965) found that the

appearance of cells of rabbit spleen making a particular class of

specific antibody paralleled the appearance of the corresponding

circulating antibody. In rabbits stimulated by a single injection

of diptheria toxoid there first appeared only cells producing yM-

antibodies. There was a slow increase of plasma cells making yG-

antibody which was evident around day 13 and which continued while

the relative proportion of cells producing yM-antibody decreased.










Phylogeny of the Immune Response. Investigations of the ability

of invertebrates to respond to an antigenic stimulus with production

of a protein analogous to mammalian antibody have yielded only nega-

tive findings. It would appear that phagocytosis together with intra-

cellular digestion is the main line of resistance by invertebrates to

infectious agents.

Thus, present evidence indicates that the antibody response is

restricted to the vertebrates, and that some of the lowest forms seem

to be capable of forming some type of antibody. Papermaster, et al.

(1962, 1963) has presented evidence that would suggest that develop-

ment of adaptive immunity was correlated with the development of

thymus and lymphoid tissue. Serum proteins having the electrophoretic

properties of mammalian 7-globulin have not been demonstrated in the

lowest vertebrates, however, there are proteins that migrate in the

P, or fast y region (Papermaster, et al., 1962). The presence or

absence of proteins migrating in the y-globulin region cannot itself

be used as prima facie evidence .of the ability, or lack of ability,

to produce antibodies. Elasmobranchs, frogs, and teleosts have been

shown to produce antibodies that have a sedimentation coefficient

of 19S and which are sensitive to mercaptoethanol, thereby resembling

mammalian macroglobulin. After an extended period, following stimu-

lation, cold blooded vertebrates begin to make antibodies that have'

a sedimentation coefficient of 7S (Uhr, et al., 1962; Gray, 1963).

The 7S antibody differs from yG-globulin of mammals in that it also

was mercaptoethanol sensitive.










Marchalonis and Edelman (1965) found that the elasmobranch,

Mustelus canis (smooth dogfish) could produce 17S, mercaptoethanol

sensitive antibodies when immunized with hemocyanin. The serum also

contained proteins with a 7S sedimentation coefficient that gave a

reaction of identity with the 17S antibody in gel-diffusion plates.

However, antibody activity in this 7S fraction could not be demon-

strated.

Demonstration of an antibody response in the lower vertebrates

is possible only if the correct antigen and reacting conditions are

chosen. Thus, an absolute statement on the ability or inability of

the invertebrates to produce antibodies may be unjustified at this

time as the experimental conditions imposed upon the animal and/or

the system used to detect its antibodies may not have been the correct

ones for successful stimulation.

Role in Specific Immunity. Antibodies play an important role

in a mammal's resistance to infectious agents. Pathogenic bacteria,

following reaction with antibody, are dealt with in two ways which

may or may not be interrelated. One of these, opsonization, has

been discussed above. Briefly, a bacterium that has been opsonized

with specific antisera is literally coated with antibody molecules,

Interposition of the antibody molecules between the bacterium and

the phagocyte facilitates engulfment of the bacteria by the phagocyte.

At least part of the complement system may be required for maximum

opsonization by a given amount of antibody (Boyden, et al., 1965).

Briefly, the complement system is composed of some 10 to 11 serum

proteins, some of which may act as enzymes (Becker, 1965).









The action of complement, in conjunction with antibody, is most

dramatically demonstrated in its ability to kill, by lysis, certain

bacteria. This bactericidal activity can be demonstrated in vitro

and forms the basis of a very sensitive assay for the presence of

small amounts of antibody to gram negative bacteria (Muschel and

Treffers, 1956). The gram negative bacteria are especially suscep-

tible to the bactericidal activity of specific antibody. Robbins,

et al. (1965) used this assay system to compare yG- and 7M-anti-

S. typhimurium. It was found that the yM-antibody more effectively

sensitized S. typhimurium for lysis by complement. As little as

0.05 jg 7M-antibody/ml caused detectable kill while 0.9 pg yG-anti-

body/ml (or 18 times more) was required to produce detectable kill.

Onoue, et al. (1965) compared the ability of rabbit 7G- and

yM-antibodies to the hapten p-azobenzenearsonate to sensitize sheep

red blood cells for lysis with guinea pig complement to which this

hapten had been coupled. These authors found that, on a weight basis,

30 to 50 times more 7G-antibody than yM-antibody was required to

achieve hemolysis of 50% of the cells in the test system.

Humphrey and Dourmashkin (1965) reported that sheep erythrocytes

had a mean maximum number of binding sites for yM-antibody of about

90,000 while there were about 600,000 sites to which yG-antibody could

bind. It was also reported that only two to three molecules of yM-

antibody were required to produce a hole in the erythrocyte membrane,

while 2,000 to 3,000 yG-antibody molecules were required. Calcula-

tions based on the above data and on the premise that one yM-molecule,

or two adjacent yG-molecules will produce a hole, indicated that the










probability of having two yG-molecules occupy adjacent sites would

require numbers of molecules of yG-antibody similar to those found

experimentally, i.e. 2,000 to 3,000. The interpretation of the data

of Onoue, et al. (1965) need not take into consideration different

numbers of antigenic sites. Both the yG- and yM-anti-azobenzenear-

sonate are directed against the same antigenic determinant, there-

fore, on any conjugated cell there should be an equal number of sites

available for both classes of immunoglobulins but there is still a

substantial difference in the abilities of the two classes of mole-

cules to sensitize for lysis.

Experimental evidence indicates that the complement-antibody

mediated lesions that lead to cytolysis are common to all cell types

(Green and Goldberg,1960). The susceptibility of a wide variety

of mammalian cells to destruction by complement in conjunction with

specific antisera was summarized by Osler (1961). One would feel

that the role of the various classes of immunoglobulins in mediating

complement lysis of erythrocytes may also apply to the various nucle-

ated mammalian cells.

The immunogen eliciting the antibodies used in this study was

the polysaccharide capsule of Diplococcus pneumoniae, Type I. The

polysaccharide, upon isolation from the pneumococci, was found to

have a molecular weight of 171,000 (Record and Stacey, 1948), and

to contain 60% uronic acid and 30% reducing sugars (Heidelberger,

et al., 1936). The uronic acid was found to be D-galacturonic acid

(Heidelberger, 1958).
















MATERIALS AND METHODS

Immunization of the Horse. Hyperimmune serum was obtained by

immunizing a horse with saline suspensions of heat-killed Diplococcus

pneumoniae, Type I. The bacteria were initially passed through

mice until they had become heavily encapsulated. Figure 1 shows

the pneumococci with their rather large capsule stained with fluoro-

scein labelled yM-antibody. Cultures were then grown in brain heart

infusion broth (BHI, Difco) enriched to contain 1% sheep blood.

The bacteria were harvested, washed free of growth medium, suspended

in 0.9% saline, and then heated to 600C for one hour. The horse

(500 kg) received a primary stimulation with three 30 mg portions

of the heat-killed pneumococci given intramuscularly (rump) on days

1, 20, and 29. Sera from trial bleedings obtained were found to

have a low agglutinating titer and showed only a trace of precipitins

up to 48 days after the initial injection. An injection of 120 mg

of bacteria was then made into the jugular vein on day 48. There-

after, intravenous injections of 120 mg of bacteria were made on

days 63, 97, 113, 118, and 158. A copious bleeding, 3 1, was taken

15 days after the first intravenous injection and the serum was

found to contain 1.63 mg of precipitating antibody per ml of serum.

Subsequent bleedings of 3 1 each were taken on days 97, 113, 118,

158, 165, 180, 219, 362, and 431. The serum from each bleeding was

titered for antibodies that precipitated with the Type I polysac-

charide before and after reduction. Each 3 1 bleeding yielded 1.2 1

to 1.5 1 of serum.

























































FIGURE 1: Diplococcus pneumoniae, Type I, with its polysaccharide
capsule stained with fluoroscein conjugated 7M-antibody.










The precipitating antibody level has remained at about 1.5

mg/ml of serum over this period of 431 days after having reached

a peak titer of 2.4 mg/ml at day 97. There has been only a slight

increase in the level of precipitins that are resistant to reduction

by 0.1 M mercaptoethanol, presumably 7G-antibody. At most, these

resistant antibodies account for 187 of the total specific antibody

precipitated by the polysaccharide. Figure 2 shows the precipitin

curves for selected representative bleedings obtained using untreated

and reduced serum.

Soluble Antigens

Preparation of Specific Polysaccharide. D. pneumoniae,

Type I were grown for 24 to 36 hours in 4 1 batches of BHI enriched

to contain 1% sheep blood. The bacteria were then harvested by

centrifugation and washed several times with saline to remove traces

of the medium. The bacteria were then suspended in 1 1 of 0.15 M,

phosphate buffered saline, pH 7.2, and deproteinized by several

cycles of mixing with chloroform in a Waring blender. The clear

aqueous layer containing the polysaccharide was concentrated ten to

twenty-fold by pressure dialysis against distilled water. The poly-

saccharide was then precipitated by making the aqueous solution 1.3 M

in sodium acetate and then by adding 1.5 volumes of 95% ethanol.

After two further precipitations, the polysaccharide was lyophilized

from a distilled water solution. The material appeared relatively

homogenous by sedimentation velocity analysis and had a sedimentation

coefficient of 3.OS. Hereafter, the polysaccharide from Type I

pneumococci will be referred to as SI.














































FIGURE 2: Quantitative precipitin curves obtained for sera taken
at various times after initiation of immunization with Type I
pneumococci. The sera used for the precipitin curves were collected
from bleedings taken on days 97 (0), 165 ( ), and 431 (0). Pre-
cipitin curves obtained following reduction of the sera with 0.1 M
mercaptoethanol are indicated by the broken lines. The amount of
SI indicated on the abscissa was added to 0.5 ml of untreated or
reduced antiserum. The washed precipitates were dissolved in 1.0
ml of 0.1 N NaOH for optical density measurements at 280 mp.









125 125
Trace abelling SI with Radioiodine I ( I SI). The SI

was extrinsically labelled with 125I by the following procedure:

100 mg of SI was mixed with 500 mg triphenylphosphorus (Peninsular
125
Chem. Research, Gainesville, Fla.), 9 mg KI and 2 mg 125 (New

England Nuclear, Boston, Mass.). All moisture was removed by gentle

heating. Anhydrous benzene (40 ml) was added and the reaction mix-

ture was heated and allowed to reflux for 72 hours. The benzene

phase was decanted and the solids were dissolved in 5 ml water and

washed into a suitable container. The polysaccharide was then pre-

cipitated by the addition of an equal volume of 95% ethanol. After

standing in the cold for several hours, the precipitated polysac-

charide was collected by centrifugation. After dissolution in 5

ml water, the polysaccharide was again precipitated with an equal

volume of 95% ethanol as before. This dissolution-precipitation

was continued until the supernatant fluid was found to be free of
125
radioactivity. Following the removal of non-bound 1I, the con-

jugate was dissolved in a small quantity of water, 5 ml to 10 ml,

and dried by lyophilization. This same procedure was followed in
125
the preparation of 125I labelled dextran, Type II (SII) pneumo-

coccal polysaccharide, and human A-substance. The characteristics

of these products are given in Results.

Isolation of Purified Antibodies

Horse Anti-SI Antibodies. A quantitative precipitin curve

was obtained for the serum from each bleeding and these data were

used to determine the amount of polysaccharide needed for precipi-

tation of the maximum amount of antibody. The optimal amount of










SI was then added to a large volume of the serum and the mixture

was incubated at 37C for one hour and then at 20C overnight. The

precipitate was collected by centrifugation and washed twice with

cold saline. The precipitate was then suspended in a quantity of

0.05 M acetate buffer, pH = 3.5, equal to one-half the original

serum volume and allowed to stand at room temperature for one hour.

The portion of the precipitate which did not dissolve was then

removed by centrifugation. The supernatant now contained about

1 mg of protein from each ml of serum used. The extinction coef-

ficient (E1% ) of 15.0, found for horse yM-globulin (McDuffie
280 mu
and Kabat, 1956), was used to determine protein concentrations.

The protein eluted from the antigen-antibody precipitate was about

82% 7M-globulin while the balance consisted of immunoglobulins which

sedimented with an S20,w = 6.4S. This slower sedimenting fraction

will be referred to as '6.4S immunoglobulins'.

Under the conditions used for eluting the antibodies from the

SI, the SI remained insoluble. The insolubility of the SI was most

probably due to the low pH of the eluting buffer and the antibody

which was not eluted under these conditions. Direct evidence indi-

cating an absence .of solubilization of the SI include: Absence of

immune precipitation upon restoration of physiological conditions

and no evidence of soluble complexes, of antigen and antibody, upon

ultracentrifugal analysis. If soluble SI was present in the anti-

body solution it would be removed by retardation on the Sephadex-

G-200 column.










Separation of yM-antibody from the '6.4S Immunoglobulins'.

The technique of molecular sieving in columns of Sephadex G-200

(Pharmacia), equilibrated in 0.1 M phosphate buffer, pH 6.8, was

used to separate the various immunoglobulins. The proteins that

had been eluted from the polysaccharide were equilibrated with the

same phosphate buffer by dialysis and 4 ml portions containing

about 10 mg of protein were applied to a column of Sephadex G-200

(2.5 cm x 60 cm). The proteins were eluted by passage of the

phosphate buffer through the column. The effluent was monitored

by measuring absorption at 280 mm. Fractions of approximately

4.7 ml were collected.

Fractionation of '6.4S Immunoglobulins'. The pool of '6.4S

immunoglobulins' from the above G-200 column were equilibrated with

0.01 M phosphate buffer, pH 7.5, and applied to a DEAE-cellulose

column, 0.7 cm, i.d. x 14 cm, equilibrated in the same buffer. A

homogeneous fraction of yG-antibody was not retained by the column

and was collected after one column volume of buffer had been passed.

Further protein fractions were obtained by stepwise elution with

this buffer which was made 0.05, 0.1, 0.2, and 0.3 M in NaCl. The

fraction eluting at 0.3 M NaCl was found to be nearly homogeneous

yA-globulin or the 'T-globulin' of van der Scheer (1940).

Preparation of Normal Horse yG-globulin. Serum was collected

from a non-immunized horse and the globulins were precipitated by

making the serum 37% saturated with ammonium sulfate. After standing

in the cold overnight, the precipitated proteins were collected by

centrifugation and washed several times with cold 37% saturated










ammonium sulfate. The precipitate was then dissolved in 0.00525 M

phosphate-borate buffer, pH 8.4 (Press, et al., 1960), then dialyzed

overnight against the same buffer. The soluble globulins were

applied to a triethylaminoethyl-cellulose (Biorad Lab., Richmond,

Calif.) column equilibrated in the same buffer. That protein frac-

tion which was not retained by the column was concentrated by pres-

sure dialysis and contained only 7G-globulin.

Reduction of Immunoglobulins

Depolymerization of yM-globulin to Subunits Using Monothiols.

The proteins of whole serum which were susceptible to reduction were

reduced by making the serum 0.085 M in mercaptoethanol (Morton and

Deutsch, 1957) followed by incubation at 370C for 30 minutes or,

alternatively, by making the serum 0.1 M in thioglycolate (Sigma

Chemical Co., St. Louis, Mo.) followed by incubation for one hour

at 370C. The free sulfhydryl groups were blocked by alkylation

with a 10% excess of monoiodoacetamide (Mann Research Labs, New York,

N. Y.) which was added following reduction. The excess reagents

were removed by dialysis against 0.1 M phosphate buffer, pH 6.8.

Reduction of yM-antibody was carried out by adding the desired

amount of a stock solution of 0.85 M mercaptoethanol or 1 M thio-

glycolate to the protein in a 0.1 M phosphate,buffer, pH 6.8, and

proceeding as for whole serum. Concentration of the stock solutions

of reducing agent were determined by titration with 5,5-dithiobis-

(2-nitrobenzoic acid) as described below.

Depolymerization of yM-globulin Using Dithiothreitol (TT).

The principal advantage of using. DTT is that much lower concentrations









of DTT are needed to depolymerize the 7M-globulin. The greater

effectiveness of DTT is a consequence of the structure of the

molecule as described by Cleland (1964). Reduction of a disulfide

bond by the previously described thiols required the participation

of two molecules of the thiol. The first step in the reduction

involves the formation of a mixed disulfide (between the protein

and the thiol) and one free -SH group on the protein. This step

is followed by a reaction of the mixed disulfide with a second

thiol molecule to yield a second free sulfhydryl group on the pro-

tein. DTT circumvents necessity for the participation of a second

thiol molecule since each PTT molecule contains two sulfhydryl groups

and the proper configuration to complete the reduction (Figure 3).

Depolymerization of yM-globulin with DTT was accomplished by adding

the appropriate volume of a solution of DTT that was approximately

ten times more concentrated than that finally required. Again the

concentration of this solution was determined by titration. The

reduction mixture was then incubated at 370C for one hour.

Re-aggregation of yM-subunits by Reoxidation.

a) The yM-antibody (2.5 mg/ml) was reduced in 0.1 M phos-

phate buffer, pH 6.8, at a final concentration of 0.1 M thioglycolate

(Sigma Chemical Co., St. Louis, Mo.) upon incubation for one hour at

370C. Complete reduction to subunits was confirmed by alkylation of

a small portion of the reaction mixture and subsequent sedimentation

analysis. The balance of the reaction mixture was passed through a

column of Dowex 1 x 8 acetate equilibrated in 0.1 M sodium acetate.

The subunits were then allowed to reassociate while being concentrated










to 5 to 7 mg/ml by pressure dialysis against 0.14 M phosphate-

borate buffer, pH 8.0.

b) The yM-globulin (6 mg/ml) was reduced with 0.0025 M

DTT (0.005 M in -SH) at 370C for one hour. Complete reduction to

subunits was confirmed by sedimentation analysis. Reassociation

was then allowed to proceed while the subunits were dialyzed against

0.00525 M phosphate-borate buffer, pH 8.4, containing 0.002 M Versene.

Separation of Polypeptide Chains. Either yG- or yM-globulin

was equilibrated with 0.55 M Tris (Tris hydroxymethyll] aminomethane)

buffer, pH 8.2. The solution was then made 0.75 M in mercaptoethanol

to reduce the globulin which was then alkylated, and fractionated on

a column of Sephadex G-100, equilibrated in 1 M propionic acid,

according to the method reported by Fleishman, et al., (1962).

Specific Precipitation and Coprecipitation of SI. Data for

quantitative precipitin curves were obtained by adding increasing

volumes of-a 1 mg/ml solution of SI to 0.5 ml of the test serum or

a saline solution containing 0.5 mg of purified yM-antibody. The

mixtures were incubated at 370C for one hour and then at 20C over-

night. The precipitates were collected by centrifugation, washed

twice with cold saline, and then dissolved in a convenient volume

of 0.1 N sodium hydroxide. The protein present in the precipitate

was determined by optical density measurements at 280 mi.
125
The amount of 125I SI required to give the maximum amount of

precipitate with a given amount of intact yM-globulin and the amounts

of rabbit antiserum (anti-horse immunoglobulins) required to precipi-

tate maximally a given amount of yM-antibody or normal yG-globulin













HHO
-N-_ -C---
rH2
H



HO-CH2-CH2


H H
-N_- -C--.
H2


-NH |
H &


HHO H HQ
4 -N-C---- HO-CH2-CH2-SH -N-(---
H-.-CH2-CH20H S-
H
+
HO-CH2-CH2-S-S-CH2-CH2-OH
-SH


REDUCTION WITH 2-MERCAPTOETHANOL


HH9 H H F)
HS-CH2CHOH)2CH2-SH -N-~-C- .-N
H2 {H2

JH2
SH


CHOH




HH8 H H /S \
-N- H--2- IHH

CHOH


REDUCTION WITH DITHIOTHREITOL


FIGURE 3: Reactions leading to reduction by 2-mercaptoethanol and
dithiothreitol.










(and '6.4S immunoglobulins') were determined separately in advance.

For quantitation of the amount of labelled polysaccharide that was

coprecipitated, 0.5 mg amounts of intact yM-antibody, reduced and

alkylated yM-antibody, normal yG-globulin, '6.4S immunoglobulin' or

other horse globulin preparation were mixed with the amount of 1251

- SI that was optimal with respect to the same weight of intact

yM-antibody. Each globulin solution was set up in triplicate with

SI. After incubation of the reaction mixtures for one hour at 3700C

precipitation could be seen only in the tubes containing intact yM-

antibody or '6.4S immunoglobulins'. At this time, that amount of

rabbit anti-globulin required to precipitate maximally the horse

immunoglobulin in a given tube was added to each tube with the

exception of one set of tubes which served as a control for amount

of polysaccharide which could be directly precipitated by intact yM-

antibody. After further incubation at 370C for one hour and at 4C

overnight, the precipitates derived from each immunoglobulin were

dissolved in 1.0 ml Hyamine and placed in vials with 5.0 ml of

scintillation fluid (Kinard, 1957) and counted for ten minutes in

a liquid scintillation counter (Packard Instrument Co., Series 314A).

The third precipitate from each set was dissolved in 0.1 N sodium

hydroxide and the optical density at 280 mi was determined.

Immunodiffusion, Immunoelectrophoresis, and Radioimmunoelectro-

phoresis. The various immunodiffusions and immunoelectrophoreses

were carried out on glass slides (either 1" x 3" or 2" x 3") covered

with a 2 mm layer of 0.8% lonagar made up in 0.05 M Veronal buffer,

pH 8.6. Electrophoresis was for 2.5 hours at 3 ma/cm (width).










For radioimmunoelectrophoresis, slides prepared as above and

the same conditions of electrophoresis were used. These slides

were developed by filling the troughs with a 1 mg/ml solution of

1251 SI plus the appropriate goat or rabbit anti-horse globulin.

The developed gel was washed in saline, then in water, and then

dried and stained with amido black dissolved in a solution of four

parts water, four parts methanol, and one part glacial acetic acid.

Radioautography was carried out by placing the slide in contact with

Kodak Industrial X-ray film (Type KK) for one to three weeks. The

technique of radioimmunoelectrophoresis has been previously described

by Hochwald, et al. (1961).

Preparation of Specific Reagent Sera

Goat Anti-horse yM-globulin. A goat was injected twice with

two 9.25 mg amounts of yM-antibody given four weeks apart. For the

first injection the antigen was incorporated in complete Freund's

adjuvant and for the second injection, it was precipitated with alum.

One-half of each antigen dose was injected intramuscularly into each

hind leg. The goat responded with precipitating antibodies, the bulk

of which reacted with the class specific portion of the yM-globulin

molecule. A small amount of the antibody cross-reacted with the

other immunoglobulins.

Rabbit Anti-horse Immunoglobulins. An anti-immunoglobulin

serum with broad specificity was prepared by injection of the unfrac-

tionated immunoglobulins that were eluted from the SI polysaccharide.

About 3 mg of the globulins were emulsified in complete Freund's adju-

vant and injected into multiple subcutaneous sites (Bernier and Cebra,










1965). Serum was collected two or three weeks after immunization.

This antisera reacted with common antigenic sites as well as with

the class specific sites of horse yM-, 7G-, and yA-globulins.

Rabbit Anti-Horse Liht Chain. Light chain from either yG-

or yM-globulin was emulsified with complete Freund's adjuvant and

was injected at multiple subcutaneous sites as described above. The

resulting antisera were specific for light chains.

Titration of Free Sulfhydryl Groups. The colorimetric method

for the determination of sulfhydryl groups has been described by

Ellman (1959). An appropriate dilution in 0.1 M Tris acetate, pH

8.0, 0.002 M EDTA of the material to be titrated was mixed with

0.02 ml of 0.01 M 5,5-dithiobis-(2-nitrobenzoic acid), (DTNB, Aldrich

Chemical Co., Milwaukee, Wise.). To titrate sulfhydryl groups on

proteins the solution was made 0.5% in sodium dodecylsulfate. The

final volume was adjusted to 3.02 ml and the reaction mixture allowed

to stand at room temperature for 30 minutes. The optical density

was then determined at 412 mu. The concentration of sulfhydryl groups

could be calculated from the molar extinction coefficient of 13,600

for the thiophenol, the reduced form of the reagent DTNB. The num-

ber of disulfides reduced in yM-antibody were determined by reducing

the molecule with the desired concentration of dithiothreitol. The

reaction was terminated by lowering the pH to 2.5 with 2 N HC1. Excess

reducing agent, as well as by-products of the reduction, were removed

by passing the reaction mixture through a column of Sephadex G-25

equilibrated with 0.15 M NaCl brought to pH 2.5 with HC1. The frac-

tion containing the highest protein concentration, as determined by










optical density at 280 mi, was then used to titrate the free sulf-

hydryls. The extent of polymerization was determined by chromato-

graphy on Sephadex G-200.

Preparation of Fluorescent yM-antibody. Two ml of buffer con-

taining 11.4 mg of yM-antibody were equilibrated with 0.01 M phos-

phate, 0.05 M NaCl, pH 7.5 by dialysis and the solution was then

placed in an ice bath. The pH was adjusted to 9.3 with sodium hydrox-

ide and maintained at this pH with a Radiometer Autotitrator (Radio-

meter Corp., Copenhagen, Denmark) during the first two hours of con-

jugation. A 140 pg amount of fluorescein isothiocyanate (Baltimore

Biological Labs., Baltimore, Md.) was dissolved in a very small

volume of acetone and was added to the protein solution. The mix-

ture was stirred constantly for the initial two hours of reaction

and placed at 20C overnight. The reaction mixture was then diluted

1:10 with the pH 7.5 buffer and applied to a DEAE-cellulose column

(2.5 x 40 cm) equilibrated with the 0.01 M phosphate buffer 0.05 M

in NaCl. Components were eluted using stepwise increases in the

NaCl content of this buffer. The fraction eluted with 0.2 M NaCl

in the 0.01 M phosphate buffer, pH 7.5, had optical densities at

280 mu and 495 mp in the ratio of 3.2:1 and was found to give very

good staining with little or no non-specific staining. The chromato-

graphic fractionation of the conjugate was adapted from Cebra and

Goldstein (1965).

Antibody Assay Using Agglutination of SI Coated Erythrocytes.

Washed sheep erythrocytes were sensitized for agglutination using

the method suggested by Askonas, et al., (1960). Sheep red blood










cells (0.2 ml) that had been previously washed three times with

saline buffered at pH 7.4, were suspended in 10 ml of buffered saline

and then 0.2 ml of a filtrate from an 18 to 24 hour culture of Type

I pneumococci was added. After the mixture had incubated at 370C

for 45 minutes, the cells were washed three times with buffered

saline containing 1% Bovine Serum Albumin (Armour, Chicago, Ill.)

and then were finally suspended in 20 ml of buffered saline.

Hemagglutination was then carried out by preparing two-fold

serial dilutions of the antibody preparation or antiserum in plastic

trays having multiple depressions. To 0.2 ml of each dilution was

added 0.1 ml of the cell suspension. The mixture was allowed to

stand at room temperature for two hours and the agglutination pat-

terns were recorded. It was found that readings taken at two hours

were not appreciably different from those taken at the end of 12 to

18 hours.

Mouse Protection. Young adult mice obtained from the Charles

River Farm, Massachusetts, were used for in vivo assay of various

antibody preparations. In each case, six hour cultures of highly

virulent D. pneumoniae, Type I, were diluted with 0.9% saline to a

concentration of 1,000 organisms/ml as determined in a particle

counter (Coulter Electronics, Hollywood, Fla.). Equal volumes of

this suspension and the antibody preparation to be tested were mixed

and 0.2 ml portions were injected into the mice intraperitoneally;

thus, each mouse received 100 organisms. The animals were observed

for a 36 hour period and the number of deaths were recorded.
















RESULTS

Immunoglobulin Response of the Horse. The initial antibody

response of the horse immunized with Type I pneumococci was described

in a Master's Thesis by the author (Hill, 1964). At the time the

thesis was completed, only data of bleedings taken through day 180

were available (Figure 2). Since then, the horse has been bled

numerous times with no additional antigenic stimulation since day

158. The level of circulating antibodies, as determined by quanti-

tative precipitin tests, has decreased slightly from 2.4 mg/ml

determined for the bleeding of day 97, to 1.5 mg/ml, where it has

remained for 31 months. Continued evaluation of the proportion of

antibody in the serum that retained its' ability to precipitate SI

following treatment with mercaptoethanol has not indicated more than

a slight increase of these resistant antibodies (yG- and yA-immuno-

globulins). Thus, it appears that this horse will probably not show

the conversion to predominately yG-antibody production that is char-

acteristic of many mammals' response to various immunogens.

Elution of the antibodies from an antigen-antibody precipitate

at a low pH, as described in Materials and Methods, led to the iso-

lation of a mixture of antibodies of different classes. The majority

of the antibodies were of the yM-globulin class as evidenced by

sedimentation analysis (Figure 7A,) and immunoelectrophoretic analysis

(Figure 5A). Of the total quantity of antibodies eluted from the










precipitate, 82% was determined to be yM-globulin by gel filtra-

tion on Sephadex G-200 (Figure 4). In Figure 5B one can see that

the mixture of 6.4S components were capable of precipitating SI.

A detailed examination of the components that comprised the smaller

immunoglobulin eluted along with the yM-antibody and resolved from

it on Sephadex G-200, indicated that this fraction was composed of

two populations of globulins (Figure 6A), differing in average net

charge as reflected by their different electrophoretic mobilities

(Figure 5A). Their difference in charge was subsequently utilized

to separate the components of the '6.4S immunoglobulins' by ion

exchange chromatography on a column of DEAE-cellulose. The mole-

cules that were not retained on the column under the starting con-

ditions, i.e. pH 7.5, 0.01 M phosphate buffer, were found to have

the electrophoretic properties ascribed to yG-globulin when examined

by immunoelectrophoresis (Figure 5C). A later fraction, eluted

from the column when the sodium chloride concentration of the buffer

was increased to 0.3 M had an electrophoretic mobility somewhat

faster than the 7G-globulin (Figure 5C). In fact, its mobility

approximated that of the previously described horse 'T'-globulin,

and, in accord with the newer nomenclature, it will be referred to

as yA-globulin (Rockey, et al.. 1964).

The Antigenic Composition of Horse Anti-SI Immunoglobulins.

An antigenic analysis of the various immunoglobulin classes

described above showed that certain antigenic determinants were

shared among all three classes. These common determinants are

expressed as reactions of partial identity in double gel diffusion










plates as shown in Figures 6A and B. In Figure 6A one can see the

relatively sharp, single precipitin band formed as the yM-antibody

(well labelled yM) reacted with its antiserum (labelled A-Ig). The

same antiserum (A-Ig) reacted with components in the well contain-

ing the fraction from the G-200 column that was eluted after the

yM-antibody, labelled 6.4S. The band closest to the A-Ig well was

formed by yG-globulin as evidenced by its complete fusion with the

band formed between A-Ig and chromatographically pure yG-globulin

(yG). Where this band meets the upper yM-antibody band there is

partial fusion, or a reaction of partial identity. However, there

is also a spur that appears on the opposite side of the yM-antibody

band, indicating the presence on the yG-antibody of antigenic sites

that do not occur on the yM-antibody. The precipitin band nearest

the '6.4S' well was formed by the 7A-globulin mentioned in the pre-

vious section. Here the band was formed by precipitation due to

antibodies directed against the class-specific sites on the mole-

cule, the antibodies reacting with the common sites having been

effectively removed by precipitation with yG-antibody. Hence, the

cross reaction with the yM-antibody band is less apparent, but the

reaction of its unique sites is more apparent. One also can notice

that the yM-antibody band spurs through both the yG- and 7A-anti-

body bands, indicating the existence of sites peculiar to yM-immuno-

globulin alone.

Figure 6B is an antigenic analysis of the fractions obtained

from the '6.4S immunoglobulins' using DEAE-cellulose chromatography.

It can be seen that the protein that was not retained on the column























O-



04



0-2




60 100 140 180 220

ELUTION VOLUME
T~i (ml)







FIGURE 4: Resolution of yM-antibody and '6.4S immunoglobulins'
on Sephadex G-200. The protein which was eluted after the passage
of one column volume of buffer was solely yM-antibody while that
eluted later constituted the '6.4S immunoglobulins'. The immuno-
globulin mixture was applied to a 2.5 cm x 100 cm reverse-flow
Sephadex Laboratory Column, filled with Sephadex G-200 equilibrated
in 0.1 M phosphate buffer, pH 6.8.

















~F" : ~II


k~~-;; *V.~ ______ ~-


A- Ig

SrG

A- Ig

2







FIGURE 5: Immunoelectrophoretic analysis of the horse anti-SI
components and their reaction with SI. Plate A shows that the
yM-antibody (yM) chromatographed on Sephadex G-200 was free of
other immunoglobulins recognized by the anti-immunoglobulin serum
(A-Ig). However, the '6.4S' component (6.4S) contained two
electrophoretically distinct populations of molecules, one of
which migrated as yG-globulin (yG). Plate B shows that the
isolated yM-antibody and the 6.4S components were capable of
forming precipitin bands with SI. Plate C shows the separation
of the two electrophoretic populations of the '6.45' component
on DEAE-cellulose. The fraction labelled 1 was eluted with 0.01
M phosphate buffer, pH 7.5, while a later fraction was eluted
with the same buffer which had been made 0.3 M in NaCl (labelled
2).












yG 2


A- g


YM

A-L


G


lm


' YM yG





LYG, M


FIGURE 6: The antigenic analysis of horse anti-SI inmunoglobulins
and their isolated 'light' polypeptide chains. See text for expla-
nation.











under the starting conditions (here labelled 1) shows a reaction of

identity with the yG-globulin (yG). The protein that was eluted

with the buffer made 0.2 M in NaCl (here labelled 2) shows a reaction

of partial identity with the yG-globulin. The proteins in '2' form

a spur over the yG-band, however, while the yG-band does not form

a spur through '2'. This indicates that '2' contained both yG- and

yA-immunoglobulins. The protein eluted at 0.3 M NaCl (3), forms a

precipitin line which passes through the line formed by yG-globulin

(double spurring). Thus, the protein in '3' has protein molecules

which possess antigenic determinants that are different from yG-

globulin and conversely, 7G-globulin has unique sites. These two

proteins also share sites and these are evidenced by the reaction

of partial identity that also occurs between these two bands. Thus,

the second component that made up the'6.4S immunoglobulins' along

with yG-antibody, was antigenically distinct from yG-antibody and

was yA-antibody, based on its electrophoretic properties.

The reaction of partial identity that has been described as

occurring between the different classes of immunoglobulins can be

in part interpreted as indicating that certain polypeptide chains

are present in all three immunoglobulins. Small polypeptide chains,

called light chains, were isolated from yG-globulin and yM-antibody

using the system devised for isolation of light chains from rabbit

yG-globulin. In Figure 6C one can see that light chains from 7G-

globulin (L7G) and from yM-antibody (LyM) form precipitin bands

which show complete reactions of identity when diffused against an

anti-light chain serum (A-L). The same reaction was present if the










A-L was anti-LyM, anti-LyG or a mixture of the two. Thus, as far

as this system can detect, the light chains from the two immuno-

globulins are identical. The isolated light chains show a reaction

of partial identity with either yG-globulin or yM-antibody and in

each case the intact immunoglobulin forms a band which shows a spur

beyond the light chain band (Figure 6D).

Physicochemical Properties of yM-antibody, Following Treatment with

Reducing Agents

Reduction with Monothiols. The immunoglobulins of the yM-

class are uniquely susceptible to depolymerization by reduction

under rather mild conditions. The.conditions generally employed

call for incubation times varying from 30 minutes to 48 hours in

the presence of 0.1 M 2-mercaptoethanol, thioglycolate, mercapto-

ethanolamine or one of several other thiols.

For a better understanding of the process of depolymerization

of the yM-antibody, it seemed desirable to carry out the reduction

under conditions as mild as possible, that is, using minimal con-

centrations of reducing agent for short time intervals (15 to 60

minutes) and at a pH around neutrality. To determine accurately

the lowest concentration of reducing agent that could be used to

effect depolymerization, a stock solution of mercaptoethanol (stan-

dardized by titration) was used to make a series of solutions of

7M-antibody 0.04 M, 0.06 M, 0.075 M, and 0.1 M in reducing agent.

The solutions were incubated at 370C for 30 minutes, the free

sulfhydryl groups were then blocked with iodoacetamide to stop

reduction, and the extent of depolymerization was estimated by










sedimentation analysis (Figures 7B and C). Under these conditions,

depolymerization of all of the molecules of yM-antibody was not

achieved in solutions having a concentration of mercaptoethanol

below 0.075 M.

Subsequently, concentrations of 0.085 M mercaptoethanol, and

incubation for 30 minutes at 370C has been used to give maximum

yield of subunits. yM-antibody depolymerized under the above con-

ditions, or any other set of conditions tried,,that resulted in

the formation of 6.1S subunits, was no longer capable of precipi-

tating SI (Figure 8A). Attempts to demonstrate that depolymeri-

zation of the yM-antibody by reduction had exposed or produced new

antigenic sites or separated groups of sites present in the intact

molecule were not successful. The subunits completely cross-reacted

with intact yM-antibody when tested in gel diffusion with either

goat or rabbit anti-horse yM-globulin (Figure 8B). Thus, it

appeared that the subunits, which no longer could precipitate

antigen, each still contained a set of antigenic sites representa-

tive of those which could be detected on the whole molecule. Con-

centrations of mercaptoethanol around 0.1 M exceed the level of

reducing agent found sufficient to free a significant portion of

the light chains from covalent bonding with heavy chains in horse

yG-globulin (Fleishman, et al., 1963). Thus, there seemed to be a

possibility that the subunits produced from yM-antibody contained

light chains which had been freed from covalent bonding with p-

chains during the process of depolymerization.


















REDUCTION WITH REDUCTION WITH
2-MERCAPTOETHANOL DITHIOTHRIEITOL

A
S ai M-Ab ISll



.a / 0 0IM JE









I I--, .'. H N i
DIALYSIS vs 000525M H S-BORATE BUFFER, pH 85 I



4SS
7.1 DIALYSIS v 0OO. 25M
PHOS-BORATE BUFFER.
PpH85









FIGURE 7: Depolymerization of yM-antibody by reduction with two
different thiol compounds used at varying concentrations and
reaggregation of the subunits by reoxidation. A. yM-antibody
purified on a column of Sephadex G-200; B. 7M-antibody reduced
with 0.04 M (upper) and 0.06 M (lower) mercaptoethanol; C. yM-
antibody totally depolymerized by 0.08 M thioglycolate; D. Reag-
gregation of 'C' by reoxidation; E. yM-antibody only partially
depolymerized by reducing 10.7 disulfide bonds with 0.0005 M
dithiothreitol; F and G. 7M-antibody almost completely depoly-
merized with 0.0025 M dithiothreitol which cleaved 22 disulfide
bonds per molecule, F at 16 minutes and G at 48 minutes after
reaching 56,100 rpm; H. Aggregates formed by reoxidation of
above.










Reduction with Dithiothreitol DTT). Reduction of protein

disulfides with DTT is a reaction that proceeds much more rapidly

and requires much lower concentrations of reducing agent than

reduction of the same number of disulfide bonds with the monothiols

described above. yM-antibody was reduced to its subunits after

incubating it for one hour at 370C in the presence of 0.005 M DTT

(0.01 M sulfhydryl group), as detected by ultracentrifugal analysis.

Again, the subunits prepared under these mild conditions could not

directly precipitate antigen.

Reduction of a 5 mg/ml yM-antibody solution with only 0.0025 M

DTT under the conditions used above, resulted in about 70% of the

yM-antibody being depolymerized to subunits. The scheme of a

relatively complete study of the reduction of the molecule under

these conditions is outlined in Figure 9.

One portion (7.5 mg) of 7M-antibody reduced with 0.0025 M DTT,

was brought to pH 2.5 immediately after incubation to prevent fur-

ther reduction and was then passed through a column of Sephadex

G-25 to remove the reducing agent from the protein. The peak pro-

tein fraction, as determined by optical density measurement at 280

mza, was used to titrate the free protein bound sulfhydryl groups.

The fraction was found to be 6.6 x 10-7 M in yM-antibody (using

940,000 as the m.w. of yM-globulin) and to be 2.89 x 10-5 M in free

sulfhydryl groups. On a molecular basis there were 44 free sulf-

hydryl groups per yM-molecule or 22 disulfide bonds had been broken

per molecule by the DTT. This was in direct agreement with a

similar study of human yM-globulin, which could be reduced to








































.95
YM








FIGURE 8: Antigenic relationship and antibody activity of yM-anti-
body, its subunit and products of reoxidation. In plate A it can
be seen that yM-antibody (yM) and the mixture of products of reoxi-
dation (yM ) form precipitin bands with SI, however, the reduced,
alkylated subunits (7Mr) do not precipitate SI. Plate B shows that
the subunit (yMr) remains antigenically identical to intact yM-
antibody. Plate C shows another reoxidation that resulted in com-
ponents capable of precipitating SI. Plate D shows that neither
the 7.3S nor the 4.9S subunits of yM-antibody could precipitate SI
in gel.


























I-
Titrate Free Sulhydryls.


1.5 ml
to pH 2.5 with
drop at HCI.

Apply to 1.8 X 35 cm
Col at G-25 Sephadex
equilibrated in
0.15M NaCI, 0.I001M EDTA,
pH 2.5.

Monitor fractions at 280 mI.
Use peak tube.



Detn.
Titrate SH Protein Coni


30 mg yM-antibody in 6 ml.
Add stock, titrate. DTT to make
.0025M. Incubate at 370 for
I hr.
I


I
Reoxidation

0.5 ml
Dial. vs. .00525M
phos-borate buffer, pH 8.5.


I
Add lodoacetamide
to 0.05M. Stand at
Room temp. I Hr


I
Separate subunlts
from intact
Free Light Chains Extent of Redn. yM-antibody.


1.5 ml
Dial vs.
0.5M HPr 0.5 ml
6M Urea Dial vs.pH 6.8 Apply to -20ol.
Apply to phosphate buffer Apply t G-20Col.
1.8 X 50 cm col. and analyze in Equilibrated in
of G-100 ultracentrifuge O.IM phos. pH 6.8


FIGURE 9: A scheme for the study of the effect of 0.0025 M
Dithiothreitol on yM-antibody. Parameters covered in each step
are indicated in block letters.


c.










subunits upon cleavage of 22 disulfide bonds with 0.01 M boro-

hydride (Jacot-Guillarmod and Isliker, 1962).

Another portion of yM-globulin (17.5 mg) which had been reduced

in 0.0025 M DTT, was blocked by making it 0.05 M in 0.1 M phos-

phate buffer, pH 6.8. Ultracentrifugal analysis was used to obtain

a rough approximation of the extent to which the yM-antibody had

been reduced to subunits (70%). A better estimation was obtained

by gel filtration through a column of Sephadex G-200. From the

elution diagram of the column, as monitored by optical density

measurements at 280 mp, it was determined that 57% of the yM-anti-

body had been completely depolymerized to subunits.

The remaining 7.5 mg of the reduced and blocked antibody was

equilibrated with a solution of 6 M Urea, 0.05 M in propionic acid

and chromatographed on Sephadex G-100 to obtain an estimate of the

amount of light chains that had been released from the molecule by

the DTT. A clean separation of light chains was not obtained on

the column as there was an unresolved component that chromatographed

between the subunit and the light chains which was, probably free

heavy (ui) chain (Figure 10).

A five fold reduction of the concentration of DTT, that is to

0.0005 M, led to depolymerization of only a very small part of the

total yM-globulin present (Figure 7E); however, titration of the

disulfide bonds broken indicated that an average of 10.7 disulfides

had been broken per yM-molecule. Thus, it appears that reductive de-

polymerization of macroglobulin requires cleavage of very resistant

or very many disulfide bonds.



































z 0.2



0.








12 6 0 24
FRACTION NUMBER






FIGURE 10: Release of free light chains by DTT. yM-antibody was
reduced with 0.0025 M DTI, alkylated with iodoacetamide, dialyzed
vs 0.5 M propionic acid made 6 M in Urea. Gel diffusion of the
various fractions against an anti-light chain serum showed precipi-
tin reactions only in the trailing shoulder.










Reassociation of yM-subunits by Reoxidation. Preliminary

experiments to demonstrate reassociation of the reductively pro-

duced subunits of yM-antibody were carried out by reducing the anti-

body with 0.1 M thioglycolate. The reducing agent was subsequently

removed by passage of the reaction mixture through Dowex-l acetate

resin, which retained the thioglycolate. Reaggregation was accom-

plished by dialyzing the protein against 0.14 M phosphate-borate

buffer, pH 8.0, while concentrating it two to three fold by pres-

sure dialysis. Sedimentation analysis, following the spontaneous

reoxidation, indicated boundaries having s20,w = 6.8S, 9.9S, 11.6S,

and 17.1S (Figure 7D). This mixture of subunits in various stages

of reaggregation was shown to be capable of precipitating with SI

in agar gel whereas a portion of the 7M-antibody which had been

acetylated following reduction did not precipitate with SI (Figure

8A).

At that ratio of subunit protein:SI giving maximum precipita-

tion, as determined by quantitative precipitin curves, 42% of the

reoxidized subunits could be precipitated or coprecipitated relative

to the same weight of intact, untreated 7M-antibody. Thus, in this

heterogeneous mixture, one was able to demonstrate the presence of

a significant part of the precipitin activity in the reoxidized

aggregates when compared to that activity present in the same weight

of intact yM-antibody. Solutions of yM-antibody at concentrations

of 5 to 6 mg/ml were reduced with 0.0025 M DTT as described in

Materials and Methods. The extent of reductive fragmentation as

estimated by sedimentation analysis, indicated that around 70% of










the protein was converted to subunits, the balance being intact

molecules. Sedimentation analysis after three to five hours

dialysis against the reoxidation buffer, 0.00525 M phosphate-

borate buffer, pH 8.4, indicated that virtually all of the yM-

antibody was subsequently in the form of subunits. This further

reduction was probably due to the continued presence of DTT and

its increased activity (about ten fold) in the range of pH 8 (Cle-

land, 1964). Reoxidation was allowed to proceed for some 70 hours

in the cold as the DTT was dialyzed away, following which the

remaining free sulfhydryls were blocked with iodoacetamide. A

series of 12 to 15 reduction-reoxidation experiments were per-

formed using DTT, and Figure 7H shows the sedimentation analysis

of the products formed in one instance. Figure 11 shows representa-

tive sedimentation analyses which indicate the range of subunit

aggregates formed by reoxidation. In instances when higher mole-

cular weight aggregates were not obtained by reoxidation of sub-

units, the subunitss' themselves showed a tendency to separate. It

was found that the 6.1S subunit could then be resolved into two

components following reoxidation. One of these sedimented with an

s20,w = 4.9S while the second component had an s20,w = 7.2S (Figure

7H).

These two components were separated by chromatography on

Sephadex G-200 and the elution pattern is shown in Figure 12.

Neither of these components, nor a mixture of them, were capable

of either precipitating SI, or of agglutinating Type I pneumococci;




























17.IS


B 4.9S


\ 9.8S





SD 4.9S
7.1IS


FIGURE 11: A series of ultracentrifuge frames from various reoxi-
dation experiments showing the variable results obtained.






























0.0
4 o.od


C /






15 I FRACTION NUBE 43 47 51






FIGURE 12: Separation by gel filtration on Sephadex G-200 of the
7.2S (leading part) and 4.9S (trailing peak) subunits of yM-anti-
body. The respective molecular weights (as determined by the s and
D method) are 180,000 and 100,000 suggesting the 4.9S subunit is
one-half of the 7.2S subunit.










however, both rendered the pneumococci agglutinable by a goat anti-

horse yM-globulin serum indicating that at least a portion of the

subunits retained antigen binding activity.

Molecular weight estimations based on sedimentation-diffusion

measurements, as described by Svedberg and Pedersen (1940), indi-

cated approximate molecular weights of 100,000 for the smaller

component (20,w = 4.9S; D20,w = 3.82 x 10-7 cm2/sec) and 180,000

-7 2
for the larger component (s20,w = 7.2S; D20,w = 4.64 x 10 cm /sec).

The values obtained are consistent with the larger component being

a complete subunit (two pairs of light and heavy chains) and the

smaller component being half of a subunit (one pair of light and

heavy chains).

An occasional reoxidation has produced a third component, which

sedimented with an s20,w = 10S (Figure 11). Attempts to isolate this

component have thus far been unsuccessful. However, in each case

where this 10S component could be demonstrated, the reoxidation mix-

ture could precipitate SI in gel diffusion (Figure 8). Thus, it

would appear that the minimum size aggregate of yM-subunits which

were capable of cross-linking antigen had an s20,w = 10S. Such

aggregates could consist of two subunits (four pairs of light and

heavy chains) but their precise characterization requires their iso-

lation. The smaller units, s20,w = 4.9S and 7.2S have been shown to

be incapable of cross-linking antigen molecules but apparently many

of them have at least one intact binding site as evidenced by the

experimental results described in the next section.










Retention of Antibody Activity b the Reductively Produced Sub-

units of yM-antibody. Several reports have appeared in the litera-

ture suggesting that at least a portion of the reductively pro-

duced subunits of yM-antibody bind specifically to antigenic sites,

(Chan and Deutsch, 1960; Jacot-Guillarmod and Isliker, 1962 and

1964; Onoue, et al., 1964), however, the extent of this binding has

not been quantitated.

Preliminary investigations using SI labelled with 125I to

detect binding of subunits to antigen were carried out using radio-

immunoelectrophoresis. This technique depends on the anti-SI sub-

units being precipitated in the gel by an anti-yM-globulin which in

turn will bind the radiolabelled SI. A normal photograph of the

developed and stained slide and a radioautograph of the same slide

are shown in Figures 13A and B. Radioautographs prepared with whole

anti-SI serum shows that radioactivity was precipitated only in the

region in which yM-antibody would be expected to be found. A con-

comitant experiment carried out using the same antisera after it

had been reduced with 0.1 M mercaptoethanol and then alkylated with

iodoacetamide (the well labelled WSr in Figures 13A and B) also

showed a band of radioactivity (Figure 13B) only in the position that

one finds yM-subunits precipitated by the anti-immunoglobulin serum.

No band of radioactivity could be detected in the absence of the

anti-immunoglobulin serum (not shown). The results obtained using

isolated yM-antibody were similar to those described above. In Fig-

ure 13C one can see precipitin bands were formed with the yM-anti-

body (yM) and anti-yM-antibody (A-7M) 1251 SI (SI*) mixture as




























A-yM*sr


Sl .. ---- -

yMr
*A-yMSI*-


rni w #4'


FIGURE 13: Immunoelectrophoretic and radioimmunoelectrophoretic
analyses of anti-SI serum and yM-antibody. The pictures in the
left hand column are normal photographs while those in the right
column are the corresponding radioautograph. The symbols used
indicate: A-Ig, rabbit anti-immunoglobulins; A-7M, goat anti-yM-
antibody; yM, yM-antibody; yMr, yM-antibody reduced with 0.085 M
mercaptoethanol and alkylated with iodoacetamide; WS, whole horse
antiserum; WSr, horse antiserum reduced with 0.1 M mercaptoethanol;
SI*, 1251 labelled Type I polysaccharide; +, anode.


I










well as with SI* alone. However, the yM-subunits (yMr) only formed

a band with the A-yM + SI* mixture (Figure 13, A and B top, C and D

bottom). The radioautograph shows that each of the bands had radio-

label associated with it and therefore SI, and that there were no

additional detectable labelled precipitin bands.

This qualitative demonstration of at least partial retention

of active antibody sites following treatment with 0.1 M mcercapto-

ethanol led to an attempt to quantitate this activity. The method

of choice was tube coprecipitation, a procedure in which the reactions

that occurred in the agar gel were essentially carried out in a

series of tubes. This was accomplished by allowing the antibody

preparation to react with a predetermined optimal amount of 125 -

SI for one hour at 370C. An amount of anti-horse immunoglobulin

serum that would maximally precipitate the anti-SI was added. The

mixture was incubated at 370C for one hour, then at 20C overnight.

The precipitate was collected, washed, and used for subsequent deter-

minations. The experimental results shown in Figure 14 indicate

that, in this system, the amount of non-specific coprecipitation is

not large relative to the amount of 125I SI specifically precipi-

tated. These data were obtained by adding increasing quantities of

1251 SI to a duplicate series of tubes containing intact yM-anti-

body. One of the series also received increasing amounts of rabbit.

anti-yM-globulin over a range that included amounts of the rabbit

antisera equal to or greater than that required to just precipitate

all of the yM-antibody. Following incubation at 370C for one hour

and at 20C for 18 hours, the precipitates were collected, washed


































/ o 0
S / /
3000 / / -30


S0 /
2000- i 20
'/1


10


1000
1000 00 200
g SI('5)









FIGURE 14: 125I SI precipitated and coprecipitated by anti-SI.
The optical density at 280 ml is shown by the solid line while the
counts are shown by the broken line. The corresponding density-
count lines use the same symbols. The closer circles ( 0 ) repre-
senting direct precipitation by anti-SI while the closed triangles
( ) represent coprecipitation.










several times with cold saline, then dissolved in 0.1 N NaOH. A

0.2 ml portion of the dissolved precipitate was counted in a liquid

scintillation counter and the balance was used to determine the

total amount of precipitate formed from optical density measurements.

As can be seen in Figure 14, although there was, in some cases, as

much as an 8.5 fold difference in the amount of protein precipitated

in the two series, the number of counts precipitated by antibody in
125
the presence of a given amount of antigen ( 125 SI) were essentially

identical whether anti-yM-globulin had been added or not.

Since the above experiment indicated that non-specific precipita-

tion of label was negligible in this system, tube coprecipitation

seemed a reliable method to use to quantitate the antibody activity

of the subunits.

The data presented in Table '1 are the results of two series of

experiments performed on different yM-antibody preparations. It was

obvious from these results, that the subunits almost quantitatively
125
retained their ability to bind antigenic determinants of the I -

SI. The yM-subunits coprecipitated 104% (Experiment 1) and 95.7%

(Experiment 2) of the antigen directly precipitated by intact 7M-

antibody. Under the conditions of the experiment, yG-globulin iso-

lated from a nonimmunized horse also precipitated some counts in

the presence of the anti-immunoglobulin. This precipitation of 125

- SI may be interpreted to mean either or both of two possibilities.

The first of these was that the counts found in precipitates formed

with normal horse 7G-globulin were truly the result of non-specific

coprecipitation, in which case the retention of active sites should


















TABLE I

Quantitative Co-precipitation of 1125 Labeled S I by Anti-Horse
Immunoglobulins andyM-antibodies, yM-subunits, or Other Horse Immunoglobulins
Horse Globulin Anti-Horse Experiment I Experiment 2
a Immunoglobu-
Present lins Present Counts/10 minb ODZgg80 CountslO min
yM-antibodies 5038 0.375 3057
4685 2959
yM-antibodies 5289 3.94 3619
4929 3306
yM-subunit 5098 4.58 2918
5069 2842
yG-globulin (normal) NDe 1389
ND 1393
yG-globulin (immune)d 1843 3.76 ND
2034 ND
6.4S immunoglobulins -- 4036
3551


aA 0.5 mg amount of TM-globulin or yM-subunits and a 0.8 mg amount of the other
immunoglobulins was used to yield the same final amount of precipitate.
bNet counts after correcting for background (457, exp. I; 449, exp. 2).

COptical density expressed as that of a solution of the precipitate in 1.0 ml of
O.IN NaOH.

yG-globulin isolated from serum of the immunized horse by the same method used
to isolate yG globulin from the unimmunized horse.
eNot Done.










be quantitated with respect to the counts precipitated by intact

yM-antibody in the presence of the anti-immunoglobulins. Expressed

in this manner, the yM-subunits coprecipitated 99.5% (Experiment 1)

and 82.2% (Experiment 2) of the antigen that was coprecipitated by

yM-antibody. The other explanation could be that there were cir-

culating antibodies in the normal serum that cross-reacted with the

SI. Polyuronic acids are prevalent in nature, for example pectin

and pectic acid, and some of these are known to cross-react with SI

(Hill, 1964). If this is the case, then the counts in the copre-

cipitates of the yG-globulin, normal and immune, may not be true co-

precipitation but may be due to the presence of pre-existing cross-

reacting antibody.

Finally, Table 1 shows that the pool containing the '6.4S

immunoglobulins' were, on a weight basis, slightly more effective

as coprecipitants with antiglobulins than either yM-antibody or its

subunits.

In another series of experiments, agglutination of SI sensitized

sheep erythrocytes was used as a sensitive indicator for cross-linking

in order to assay for multiple antibody sites, on a single subunit of

yM-antibody. Using intact yM-antibody, as little as 0.53 pg/ml of

antibody could be detected. No agglutination could be detected using

the 4.9S or 7.1S subunit preparations, even at concentrations of

about 3,000 ig/ml. However, the addition of a small amount of a

rabbit anti-horse immunoglobulin serum to the reaction mixture pro-

duced agglutination in those tubes in which the concentration-of:sub-

unit was relatively high, i.e. over 0.5 mg/ml.











The quantitative retention of antibody active sites by the 7.1S

subunit was demonstrated using the Farr technique. In the Farr tech-

nique, the soluble antigen-antibody complexes are precipitated at

50% saturation with ammonium sulfate (Farr, 1958). The results of

this experiment are shown in Table 2. At the maximum precipitation

of radioactivity by the subunit, obtained with 0.2 ml of 1251 SI,

the subunit bound 96% of the radioactivity which was bound by intact

yM-antibody indicating that the 7.1S subunits quantitatively retained

their ability to bind antigenic determinants.

To carry out the experiments described in the proceeding sec-

tion, it was necessary to prepare a radiolabelled polysaccharide.

Since the method used for this purpose was a novel one and has not

been described in detail, this section is concerned with the prep-

aration of extrinsically labelled polysaccharides.
125
Characterization of the 125I labelled SI. The need for a radio-

labelled SI could be satisfied by one of two general methods. The

SI could be labelled in vivo by feeding the growing bacteria a radio-

active precursor which would be incorporated, at least in part, into

the polysaccharide capsule. This method requires a relatively large

amount of the label as part of it would be incorporated into other

cell components or products. Once the radioactive polysaccharide

was laid down, it would be necessary to isolate the SI as previously

described. Another method would be to extrinsically label the SI.

Technically, this method was preferred as it involves a minimum num-

ber of manipulations of radioactive substances.

A method of extrinsically labelling the SI with radioiodine was

suggested in the introduction to a paper by Landauer and Rydon (1953).




















TABLE 2

Ability of yM-subunits to Bind SI
as Detected by the Farr Method a


Antibody


0.75 mg




7.1S subunit
0.75 mg


Vol 1251 -SI


Counts/10 minb
in nrecnitate


0.05 ml
0.10 ml
0.20 ml
0.30 ml


0 ml


0.05 ml
0.10 ml
0.20 ml
0.30 ml


a) In the Farr method (1958) the antigen is incubated with the-anti-
body preparation, then the antibody is precipitated with ammonium
sulfate. In this case the antibody was precipitated at 50%
saturation with the salt. The above counts have been corrcted
for a small amount of non-specific precipitation of the I SI.

b) Corrected for background counts.


in r p_-,- i- t











In this paper it was mentioned that alkyl halides could be prepared

by the reaction of triphenyl phosphate methiodide with an alcohol

according to the following reaction:

(PhO)3PMeI + ROH ----> RI + PhOH + Me PO(OPh)2
Where: R = alkyl residue
Ph = phenyl residue
Me = methyl group

This reaction was found to be adaptable to iodinating carbo-

hydrates using the procedure described in the Materials and Methods

section.

In an early attempt to assure ourselves that the 125I was

covalently bound to the carbohydrate, dextran, of average molecular

weight of 153,000, was reacted according to the described procedure

and then was passed through a 2.5 x 40 cm column of Bio-Gel P-30

(Bio-Rad Labs., Richmond, Calif.). Figure 15 shows that, by far,

the majority of the radioactivity (bar graph) was associated with

dextran (line graph). There remained a small amount of free radio-

iodine which was retarded on the column and eluted after the dex-

tran.

In all preparations of SI there was a small amount of contami-

nant that adsorbed at 280 mp. To exclude the possibility of the

radiolodine being associated solely with this material, the 125 -

SI was subjected, in turn, to digestion with RNase, DNase, and

pronase, at the optimal pH for each enzyme, with concomitant dialysis

against buffer at the appropriate pH. Following each treatment there

was a small reduction in specific activity which could be accounted

for by the appearance of radiolabel in the dialysate. Most of the

































40- -8,000


S0- 3-6,000


320- -6,000


10- -2,000



0 4 8 12 16 20 24 28 32 36 40













FIGURE 15: Gel filtration of 125 Dextran on Bio-Gel P-30. The
concomitant elution of the majority of the 1251 (bar graph) with
the dextran (line graph) indicates that the 125I was bound to the
dextran. Carbohydrate was determined in the effluent by the orcinol
method and is expressed as 10 x optical density at 540 mu. Counts
were determined for 1 ml samples of each fraction.











280 mi absorption remaining was that expected from the concentration

of pronase used for digestion. This enzyme treated 1251 SI was

then passed through a 2.5 x 90 cm column of Bio-Gel P-30.. In Fig-

ure 16 it can be seen that the uronic acid curve and the radio-

activity curve are almost superimposable while the 280 mg absorbing

material eluted after a larger volume of water was passed through

the column.
125
The retention of antigenic integrity by the 125 SI is shown

in Figure 17. These data were obtained by adding increasing amounts
125
of yM-antibody to a constant amount of the 125I SI. The precipi-

tate was washed with cold 0.9% saline and dissolved in 1 ml of hya-

mine. The precipitate and the first supernatant, normalized to 1

ml, was counted in the scintillation counter. It can be seen in this

figure, that at maximum precipitation, at least 70% of the total

counts were precipitated by the anti-SI (shown by open squares in

Figure 17). The radioactivity that remained in the supernatant after

precipitation with the anti-SI is shown by the closed circles and it

amounts to only 207. of the total radioactivity added to the mixture.

The open triangles are the sum of the counts precipitated plus those

remaining in the supernate at each point. The decrease noted here

between the sum and the total activity added is most probably due to

quenching by the protein present in the antibody precipitated label.

The labelling reaction requires rather long exposures of the poly-

saccharide to temperatures above ambient and during this time there

may be some loss of antigenic reactivity which could account for the

non-precipitated counts.































S" r -0.3

S, -5000
I20- -0
a s*-0-2


10
-01

1000
0- 0 -0
100 200 300 400





FIGURE 16: Bio-Gel P-300 chromatography of 1251 SI. The SI
elution is shown by the closed circles ( 0 ); radioactivity by
open rectangles ( ) and optical density at 280 mi by open circles
( 0 ). The fractions represented by the closed triangles ( )
gave a brown color in the uronic acid determination due to the
protein present.






















16,000-



12,000



8,000



4,000-

2,000-


0 0.2 0.4 0.6 0.8 1.0

ml ANTI SI







FIGURE 17: Integrity of the antigenic determinants of SI following
labelling with 1251. The radioactivity precipitated by the indicated
amounts of anti-SI are indicated by the open rectangles ( ), that
remaining in the supernatant by the closed circles ( 0 ), and the sum
of the two at each point by the open triangles ( ).










Hence, from the data found, it would appear that a significant

portion of the radioiodine is covalently bound to the polysaccharide.

In the case of dextran, it was found that the technique resulted in

conjugation of 80 moles of iodine per mole of dextran (mw = 153,000)

or, one iodine atom per 10 glucose residues. Evidence seems to

indicate that the number of moles of iodine reacting with a mole of

SI is about the same. The molecular weight of SI is 171,000, or

slightly more than the dextran, while the number of counts per unit

time per mg are also slightly higher. Labelling with this number

of molecules of iodine was sufficient for the proposed use but did

not result in an extensive loss of antigenic specificity. This tech-

nique has been successfully employed in another laboratory to radio-

label human erythrocyte 'A' substance. Thus, it would seem that

the basic technique, with slight modifications, will be useful in

extrinsically labelling carbohydrates in general.

Biologic Activity of Anti-SI Antibodies. The exquisite sen-

sitivity of mice to fatal infections with pneumococci provides a

sensitive system for examining the in vivo activity of various anti-

body preparations. The protective activity of antisera is directly

related to its content of antibodies to the polysaccharide of the

capsule.

The anti-SI serum, diluted to a final concentration of 0.001

ig/ml of antibody as determined by quantitative precipitin curves,

protected 50% of the mice to a challenge dose of 100 pneumococci.

As little as 0.0001 ag/ml of purified yM-antibody would confer the

same degree of protection. Thus, the purified antibody was ten times






75




more effective in protecting the mice. This apparent increase in

the ability to prevent a lethal infection in mice upon isolation

of the 7M-antibody was probably due to removal of the other, less

efficient, antibodies from competition for antigenic sites. In

contrast, 10 pg/ml of yG-antibody or 0.1 pg/ml of yA-antibody were

required to protect 50% of the mice.

Reduced serum or reduced yM-antibody did not protect the mice

at concentrations up to 10 pg/ml. The 7.1S subunit ofyyM-antibody,

which was shown to render the bacteria agglutinable by an anti-

immunoglobulin serum, conferred no protection when tested at 100

or 1,000 pg/ml. The results of experiments yielding the above data

are shown in Table *3J.










TABLE 3

Protection Against Type I Pneumococcus Infection in the
Mouse by Equine Anti-SI Immunoglobulins


Whole Antiserum



Reduced Antiserumb

yM-Antibody



Reduced yM-Antibodyb



7G-Antibody





yA-Antibody


Concentration

10 to 0.01 pg/mla

0.001 pg/ml

10 to 0.1 pg/mla

10 to 0.001 pg/ml

0.0001 pg/ml

1000 to 100 pg/ml

10 to 0.1 pg/ml

1000 and 100 pg/ml

10 pg/ml

1.0 pg/ml

100 and 10 Ag/ml

1.0 pg/ml

0.1 pg/ml


a) Concentration of antibody based on quantitative precipitin curves
on whole serum.

b) Reduced at a final concentration of 0.1 M 2-mercaptoethanol.


Survival

26/26

5/9

0/15

39/40

2/5

10/10

0/25

10/10

6/9

0/9

15/15

9/10

5/9

















DISCUSSION

Investigations of the physicochemical properties of horse yM-

antibody undertaken in this study show that: The yM-antibody

possess antigenic sites that are unique to the 7M-globulins as well

as antigenic sites that are shared with the other class of immuno-

globulins. The intact molecule was depolymerized to subunits by

reduction of disulfide bonds with various thiol reagents, however

while certain secondary activities were lost, the subunits of

certain sizes retained the ability to specifically bind antigen.

Reoxidation of the disulfide bonds of yM-subunit gave circum-

stantial evidence that a combination of subunits having an S20,w

of about 10S was the minimum size that would cross-link antigen

molecules and lastly, yM-antibody was found to be 105 times more

effective than yG-antibody and 103 times more effective than yA-

antibody in protecting mice from a lethal infection when challenged

with an injection of 100 lethal doses of Type I pneumococci.

Purified horse macroglobulin antibody was employed in the initial

studies which defined yM-globuliin (71M-globulin) as an immunoglobulin

(Heidelberger and Pedersen, 1937; Kabat and Pedersen, 1938) and which

distinguished this protein from yG-globulin on the basis of its sed-

imentation properties and molecular weight (Kabat, 1939). In the

intervening years, most of the comparisons of the antigenic and

physicochemical properties of 7M-globulin with those of yG-globulin










have utilized human yM-immune globulin, normal yM-globulin,

Waldenstrim's macroglobulin, yM-globulin that was rich in rheuma-

toid factor, or whole serum. In this present study specifically

purified horse yM-antibody directed against the defined antigen,

SI, has been again employed. Many of the relationships between

yM- and yG-globulins and certain characteristics of yM-globulin

that were found for the horse immunoglobulins paralleled or were

similar to those previously found for the corresponding human glob-

ulins. Thus, purified horse 7M-antibody was found to contain

unique antigenic sites which permitted its differentiation from

7-globulins of other classes (yA, yG). The class specificity of

human 7M-globulin has been shown by Franklin and Kunkel (1958),

who used a crude normal yM-globulin preparation, and by Korngold

and van Leeuwen (1957), who used WaldenstrUm's macroglobulin. Also,

just as human or rabbit yM- and yG-globulins have been found to be

antigenically related (Fahey, 1963; Cebra and Goldstein, 1965) due

in part to the presence of similar light chains in both molecules,

the horse yM-antibody was found to be antigenically related to yG-

globulin. Light chains prepared from both horse yG- and yM-globulins

gave reactions of identity when tested with rabbit anti-horse yG-

globulin. Light chains prepared from both horse yG- and yM-globulins

gave reactions of identity when tested with rabbit anti-horse yG-

light chain sera. Chaplin, et al. (1965), on the other hand, have

found close agreement between the amino acid compositions of the

light chains derived from horse yG- and yA-globulin, so it appears

that all three horse immunoglobulins (yG-, yA-, and 7M-) will be

found to contain the same kinds of light chains, as is the case in

the human system.










Using a series of pathologic human macroglobulins, Deutsch and

Morton (1957, 1958), first showed that yM-globulins were susceptible

to reductive depolymerization with 0.1 M mercaptoethanol. A similar

reduction of horse yM-antibody to subunits was found to occur on

incubation for 30 minutes with concentrations of mercaptoethanol

0.075 M and higher. Incubation in mercaptoethanol at concentrations

of 0.04 M to 0.06 M resulted in depolymerization of only a portion of

the molecules. Another reducing agent, dithiothreitol, was found

to depolymerize reductively yM-antibody at much lower concentrations

than that required for mercaptoethanol. The increased efficiency

of dithiothreitol can be easily shown to be a result of the nature

of the molecule. Each molecule contains two sulfhydryl groups

separated by four carbon atoms. The two sulfhydryl groups are

necessary to reduce completely both members of the broken disulfide

and the bond angles of the DTT molecule are such that a stable,

five atom ring structure can be formed. Horse yM-antibody is com-

pletely reduced to subunits by as little as 0.005 M dithiothreitbl

at pH 6.8 or 0.002 M at pH 8.5.

The concentration of mercaptoethanol required to dissociate

the horse yM-antibody is consi higherr than that required to

halve the '5S product' of pepsin digestion of rabbit yG-globulin

(0.008 M mercaptoethanol, pH 5, one hour at 37C) (Mandy and Nisonoff,

1963). In fact, the concentration of a monothiol reducing agent

(0.075 to 0.1 M) required to depolymerize 7M-globulin was very close

to that (0.06 M) found by Fleishman, et al. (1963) to be sufficient

to reduce all of the interchain disulfide bonds in horse yG-globulin.











Thus, it appears that reductive depolymerization of macroglobulin

antibody requires cleavage of very resistant or very many disulfide

bonds. Indeed, Jacot-Guillarmod and Isliker (1962) found that about

22 disulfide bonds were cleaved on depolymerization of human 7M-

globulin with 0.01 M borohydride and, in fact, this is the same num-

ber as we found to be cleaved in horse yM-antibody following reduc-

tion with 0.0025 M dithiothreitol. One might anticipate that some

of the light chains of yM-globulin would no longer be covalently

bound to the rest of the molecule after its depolymerization.. Starch

gel electrophoresis of reduced horse yM-antibody under conditions

known to dissociate the chains of reduced 7G-globulin, pH 3.0, 8 M

urea, (Edelman and Poulik, 1961) did reveal the presence of a

second, rapidly migrating component in the subunit mixture, which

could have been from light chain (Hill, 1964). Free light chain

was then isolated by dissociating it from the subunit in 0.5 M

propionic acid, 6M in urea and subsequent gel filtration.

Reduction of the horse,yM-antibody to subunits resulted in loss

of its ability to precipitate antigen in solution or in-agar gel.

Through the years, the loss of a variety of activities indicative

of antigen-antibody interaction upon reduction of whole serum or

immune globulin has been taken to demonstrate the macroglobulin

nature of the antibody under study. However, that loss of precipi-

tating or agglutinating ability upon reduction need not indicate

inactivation of the antibody combining sites of yM-globulin was sug-

gested by the finding of Chan and Deutsch (1960). These authors











showed that reduced human Rh saline agglutinins of the yM-type still

specifically associated with the surface antigens of the appropriate

erythrocytes although they no longer agglutinated the cells. The

erythrocytes coated with 7M-subunits could then be agglutinated by

a rabbit anti-human macroglobulin serum. Recently, Schrohenloher,

et al. (1964) have found that subunits obtained by reduction of

human macroglobulin, which was rich in a pathologic anti-human yG-

globulin (rheumatoid factor) still associated with the human glob-

ulin antigen to form soluble complexes. Onoue, et al. (1964) used

yet another technique, radioimmunoelectrophoresis (Hochwald, et al..

1961), to demonstrate the binding of 131I labelled p-azobenzen-

earsonate-insulin conjugate by the subunits of rabbit yM-antibody

prepared by reduction of a purified antibody preparation containing

39% yM-globulin. A specific sheep anti-rabbit yM-globulin was used

to coprecipitate the labelled antigen in agar gel along with the

subunits. Using radioimmunoelectrophoresis we confirmed the reten-

tion of binding activity by the yM-subunits that were present in

both reduced anti-SI sera or that were derived from purified, homo-

geneous 7M-antibody by using 1251 SI and a goat anti-horse yM-

globulin for coprecipitation. To quantitate the amount of antigen

binding activity retained by the yM-subunits, relative to their

parent intact yM-antibody, quantitative tube coprecipitation was

used. There was no significant difference between the amount of

125I SI coprecipitated with anti-yM-globulin by the intact yM-

antibody or by the yM-subunit. Thus, our results indicate that

reductive depolymerization of purified horse yM-antibody does not

appreciably inactivate antibody combining sites.











The integrity of a particular tertiary structure of a protein

is usually a requirement for its activity. Proteins,which have some

functional activity which can be measured, often show a great alter-

ation in activity, either qualitatively or quantitatively, when

changes in their tertiary structure are induced. In addition,

certain complex proteins, composed of several polypeptide chains

or subunits, shown an alteration in their functional activity upon

suffering changes in their quaternary structure. For example,

enzymes whose quaternary structures are stabilized by electrostatic

as opposed to covalent bonds, can be dissociated and reassociated

freely. Glutamic dehydrogenase can be dissociated into several sub-

units that no longer have glutamic dehydrogenase activity but which

now have alanine dehydrogenase activity (Tomkins and Yielding, 1961).

This conversion is freely reversible and, in fact, seems to be

utilized as a control mechanism for a particular reaction.

Restoration of the original configuration and activity of a

protein is usually more difficult if the depolymerization involves

breaking covalent bonds. The covalent bond almost exclusively

involved in the stabilization of the quaternary structure of pro-

teins is the disulfide bond. Disulfide bonds may be formed bet-

ween cysteine residues on the same polypeptide chain (intrachain)

or between residues on different polypeptide chains (interchain).

With rare exceptions, one is unable to break selectively one par-

ticular bond within one of these groups and not the others. Pro-

tein molecules often contain several intrachain as well as several

interchain disulfide bonds. The selectivity one can achieve in the










reduction of these disulfide bonds leading to depolymerization of

a protein will be dependent on the folding of the chains and their

propensity, or lack of it, to remain associated by non-covalent

forces. The interchain disulfide bonds of yG-globulin are readily

accessible to reduction by thiols, complete reduction being achieved

with as little as 0.75 M thiol. However, there is little or no

detectable dissociation of the various polypeptide chains or com-

binations of polypeptide chains that are known to make up the mole-

cule. Thus, the quaternary integrity of yG-globulin remains intact

until one brings about unfolding of the molecule by interference

with its non-covalent bonds. Reagents commonly employed are 1 M

propionic acid, 6 M urea, or 6 M to 8 M guanidine. The propensity

of the polypeptide chains of yG-globulin to remain associated con-

trasts with an apparent lack of non-covalent association of the sub-

units of yM-globulin. One need only to reduce a requisite number

of interchain disulfide bonds to achieve dissociation of the mole-

cule into its subunits.

Reducing agents used at concentrations which just lead to

complete depolymerization of yM-globulin do not selectively reduce

only those interchain disulfide bonds holding the subunits together

but also reduce a significant number of disulfide bonds between the

p-polypeptide chain and the light polypeptide, which remain associated

as do the heavy and light chains of yG-globulin. Upon subsequent

treatment and chromatography in the presence of propionic acid and

urea, 32% of the total light chains can be shown to be no longer

covalently bound to p-chains. The location of the disulfide bonds










responsible for the linking of the subunits in the yM-globulin mole-

cule, which can be between light-light, -p, or light-u polypeptide

chains, has not been elucidated.

Reoxidation of reduced intrachain disulfide bonds under condi-

tions that favor disulfide interchange, will lead to reformation of

most of the original disulfide bonds. Of course, those disulfide

pairs which did not become grossly disorientated will be correctly

reoxidized early on while other pairs form at random. The final

disulfide bonding achieved will be that results in the most thermo-

dynamically stable configuration under that condition. The disul-

fide bonds in native protein represent the most stable configur-

ation under physiologic conditions (Givol, et al., 1965). When

more than one polypeptide chain is involved, reoxidation to an

active configuration depends on the proper interchain disulfide

bonds being reformed as well as the intrachain disulfide bonds of

the polypeptide chains. Again, if the chains remain associated by

non-covalent bonds, it is relatively easy to have the same sulfhydryl

residues reoxidize and form a disulfide bond. If the chains become

even partially dissociated or rearranged, the reformation of the

proper disulfide bond become less probable.

Reoxidation of the subunits of horse yM-globulin fit the latter

category. Reduction of the molecules leads to a complete dissocia-

tion of the 7.1S subunits and the production of a large number (21

to 44 depending on the concentration of thiol) of free sulfhydryl

residues. Assuming that the 44 free sulfhydryl groups are about

equally distributed among the five subunits, then there are about










nine sulfhydryl groups in each subunit. At least a portion of these

remain sterically close to each other due to non-covalent forces,

hence, are able to readily reform, however, this still allows for a

large number of possible intra-subunit combinations of disulfide

bonds if all are sterically possible in addition to the inter-sub-

unit disulfide bonds. Most likely the probability of formation of

an intra-subunit disulfide bond upon reoxidation is much higher than

the chance of formation of a link between two subunits and this

makes reaggregation of two subunits difficult. Formation of the

correct disulfide bond within the cell synthesizing yM-antibody is

facilitated by an innate propensity to non-covalently associate,

thereby aligning the sulfhydryl groups that will subsequently co-

valently link the subunits. In vitro the subunits prepared by the

described method do not appear to exhibit this propensity. Either

the concentrations of the subunits achieved in the reoxidation mix-

tures are too low to result in close packing of the molecules so

that the electrostatic forces can become effective, or disruption

of large numbers of disulfide bonds upon reductive depolymerization

results in modifications of the tertiary structure of the subunits

and a concomitant interference with complementary electrostatic

bonding.

The aggregates that are formed by reoxidation are able to

express at least some of their former biologic activity, for example,

the precipitating of SI or agglutinating of sensitized erythrocytes.

Restoration of the ability to produce some of the secondary effects

of yM-antibody that are due to its cross-linking properties by










oxidative recombination of its subunits has been reported by Jacot-

Guillarmod and Isliker (1962). These workers succeeded in restoring

the agglutination titer of reduced human anti-A or anti-B yM-iso-

antibodies by immediate dialysis of the reduction mixture against

Veronal buffer, pH 8.4, according to the procedures which were

used by Deutsch and Morton (1958) oxidatively to recombine subunits

of Waldenstrbm's macroglobulin. Likewise, Sqhrohenloher, et al..

(1964) have partially recombined the subunits of rheumatoid factor

in the same manner and observed that the reoxidized preparations

regained the ability to precipitate with human 7-globulin. Removal

of thioglycolate from depolymerized horse yM-antibody on Dowex-l-

acetate, followed by pressure dialysis against phosphate-borate buf-

fer, pH 8.0 or depolymerization with 0.0025 M DTT followed by dialy-

sis against phosphate-borate buffer, pH 8.4, leads to partial reas-

sociation of subunits and restoration of precipitating ability.

Thus, it seems that yM-subunits retain active antibody sites but that

they are unable to cross-link particles or antigen. Even partial

reassociation of the subunits restores significant precipitating

ability. The subunits presumably cannot cross-link antigen because

either they are univalent or because their sites are so oriented

that they cannot combine with more than one high molecular weight

antigen at a time.

Subunits of the above type that are reaggregated in a fashion

that permits them to express a secondary biologic activity may be

artifacts and may bear no resemblance to the native molecule. Inter-

mediate polymers of the 7.1S subunit, between one (the subunit) and










five (the intact yM-globulin) are not found by reductive depoly-

merization but only by reoxidation of the subunits. In connection

with this, one wonders why it is necessary to break so many disul-

fide bonds to depolymerize the yM-globulin. I feel that, in

reality, there are relatively few easily reduced disulfide bonds

holding the subunits together. Before the molecule can dissociate,

sufficient disulfide bonds, other than those holding the subunits

together, must be broken in order to modify the secondary and terti-

ary structure. There is circumstantial but reasonable evidence to

support the above observations. If a large number of disulfide

bonds do hold the subunits together and all that is necessary to

depolymerize the molecule is to break these, then using limiting

levels of reducing agents one would expect random assortments of

subunits combinations could be found. As mentioned earlier, this

is not the case.

Although the bulk of the antibodies (82% ) precipitated from

the horse anti-SI serum by the polysaccharide preparation were of

the yM-class, yG- and yA-antibodies were also identified. The for-

mation of antibodies of three classes (yG-, yM-, and yA) to a sin-

gle antigen has been described for the human (Fahey and Goodman,

1964) and the rabbit (Onoue, et al., 1964 ). Recently Rockey, t

al. (1964) have described the presence of a new 10S 71-antibody,

as well as yA-, yM-, and three distinct yG-antibodies to the p-

azophenyl-p-lactoside group in the horse. In our own anti-SI sys-

tem, the actual specificities of the yA-, and yG-, and yM-anti-

bodies have not been defined as narrowly. However, it appeared

that at least a part of the yG- and yA-antibodies reacted with the

same polysaccharide molecule that was precipitated by the yM-antibody.










Since relatively large amounts of these separate classes of

antibody are available, it is possible to compare certain of their

biologic properties such as their opsonizing and complement fixing

activities. Some preliminary attempts to show complement fixa-

tion by the anti-SI antibodies have not met with success. yM-

antibody from other animals are known to efficiently fix comple-

ment, for example, rabbit anti-sheep erythrocytes. Guinea pig serum

is the usual source of hemolytic complement and also served as the

source reported above. Thus, one can say that horse anti-pneumo-

coccal antibody does not fix guinea pig complement. In our case,

we also used guinea pig serum as a complement source. That the

participation of guinea pig complement in an antigen-antibody reac-

tion that takes place in horse antiserum may be blocked by compon-

ents of the horse's own complement system has been suggested by

Nelson (1965). Using purified horse yM-, yG-, and yA-anti-SI we

could not find any hemolytic complement activity with SI conju-

gated sheep erythrocytes. Thus, it appears that guinea pig comple-

ment is not compatible with the antibodies of the horse. The role

of horse complement has not been investigated.

In the horse anti-SI system, as well as in the S. typhimurium

system of Robbins, et al. (1965), yM-antibody was found to have a

much higher in vivo opsonic activity in the mouse than either yG-

or 7A-antibody. In vivo protection of 50% of the mice challenged

with 100 pneumococci, about 100 L.D. 50's, required 105 times more

yG-antibody than yM-antibody. The yA-antibody was also less effective









than 7M-antibody, 10 times more 7A-antibody than yM-antibody being

required to achieve the same degree of protection. The protection

conferred by antibodies to pneumococci is by virtue of the bacteria

becoming very susceptible to phagocytosis. The mere coating of

antibody onto the capsular surface does not seem to explain 'opsoni-

zation' in this case. The yG-antibodies have antibody activity

directed against the polysaccharide and solutions of equal concen-

tration would contain approximately five times as many molecules

of 7G- as 7M-antibody. Also, a given solution of yM-antibody, upon

reduction would again have five times as many units capable of

binding to the surface of the pneumococci, yet, in both of these

cases relatively large amounts of the smaller proteins are required.

Thus, the explanation for its greater effectiveness must lie in the

integrity of the 7M-molecule or its unique quaternary structure.

The physical structure of horse 7M-globulin has not been eluci-

dated. From what is known, many structures can be postulated. The

molecule is made up of five covalently linked subunits, each con-

sisting of two light chains and two heavy chains, analogous to yG-

globulins; yet, no more than five antibody sites have been detected

in the intact molecule and consequently only a single site on each

subunit is apparent. The 7M-globulin molecule is only slightly

wider than 7G-globulin, but much longer; 47Ao x 946Ao for horse yM-

globulin and 39AO x 286Ao for horse 7G-globulin (Campbell and Bullman,

1952). If these dimensions reflect a more or less stable configura-

tion for the molecule then the spatial arrangement of the subunits

in an intact 7M-globulin molecule is not easily visualized. The










subunits have a molecular weight somewhat higher then 7G-globulin,

hence, a slightly larger size. They are more asymetrical than yG-

globulin, but not as asymmetric as yM-globulin itself. The asym-

metry has been ascertained from the frictional ratios found for the

human to be 1.92, 1.69, and 1.50 for yM-, yM-subunit, and yG-glob-

ulin respectively (Miller and Metzger, 1965). The problem of

building a model of yM-globulin then depends on how one may co-

valently bind five subunits into a unit that is less than 1.25

times as wide and 3 times as long as its individual subunit.

Table 4 summarizes the findings of the author for horse yM-

antibody, yM-subunit, yG-antibody, and yA-antibody. Table 5, as

a comparison, shows the findings of others for various immunoglob-

ulins.














TABLE 4

A Comparison of the Properties Found for Horse Immunoglobulins
in this Study


Type of Antibody Sedimentation Molecular Effect
Antibody Activity Coefficient Weight of Thiols

yM-antibody precipitating and 17.3S 900,000 depolymerization
agglutinating

yM-subunit
4.9S some binding 4.9S 100,000 ---
6.5S quantitative binding 6.5S --- ----
7.1S quantitative binding 7.1S 180,000 ----
9.8S possible precipitating 9.8S --- -
and agglutinating

yG-antibody precipitating and 6.4S --- no effect
agglutinating

yA-antibody binding precipitating? 10.0S --- ---








TABLE 5

Properties of Immunoglobulins from Various Species

Molecular Weight
tivity Sedimentation Intact Heavy Light
onrre Coefficient Molecule Chain Chain


Antibody Ac
or Serunm S


Species
and Globulin

Human
WaldenstrSm's
Waldenstr8m's
subunit
yG-globulin

Rabbit
yM-globulin

7M-subunit

yG-globulin


Swine
7M-globulin


Equine
yM-globulin

yG-globulin


anti-pneumococcus



anti-pneumococcus

anti-diptheria
toxoid


17.90S
7.07S

6.71S


890,000
185,000

160,000


--- 900,000

--- 180,000

--- 140,000



18.00S 930,000


19.30S

7.20S


70,000



53,000


Reference


--- --- Miller and Metzger, 1965
-- ibid

ibid


22,000 to
23,000


22,000 to
23,000


Lamm and Small, 1966

Lamm and Small, 1966

Small and Lamm, 1966


-- --- Kabat and Pedersen, 1938
Kabat, 1939


--- --- Kabat and Pedersen, 1938
Kabat, 1939
--- Pappenheimer, et al., 1940


920,000

184,000


pathologic
pathologic

normal


anti-S.
enteritidis
anti-S.
enteritidis
normal


or Serum Source Coefficient Molecule Chain Chain
















BIBLIOGRAPHY

Altmeier, W., Robbins, J. B., and Smith, R. T. (1965), In
Preparation.

Askonas, B. A., Farthing, C. P., and Humphrey, J. H. (1960),
Immunology 3L 336.

Bauer, D. C., Mathies, M. J., and Stavitsky, A. B. (1963), J.
Exptl. Med. 117. 889.

Bauer, D. C., and Stavitsky, A. B. (1961), Proc. Natl. Acad.
Sci. U. S. 47. 1667.

Becker, E. L. (1965), in Ciba Foundation Symposium on Complement
(G. E. W. Wolstenholme and J. Knight, eds., ) Little, Brown,
and Co., Boston, Mass., p. 58ff.

Bellanti, J. A., Eitzman, D. V., Robbins, J. B., and Smith, R. T.
(1963), J. Exptl. Med. 117. 479.

Benedict, A. A., Larson, C., and Nik-Khab, H. (1963), Science
139. 1302.

Bernier, G. M., and Cebra, J. J. (1964), Science 144. 1590.

Bernier, G. M., and Cebra, J. J. (1965), J. Immunol. 95. 246.

Boyden, S. V., North, R. J., and Faulkner, S. M. (1965), in Ciba
Foundation Symposium on Complement (G. E. W. Wolstenholme and
J. Knighti,eds.,) Little, Brown, and Co., Boston, Mass., p. 190ff.

Brambell, F. W. R., Hemmings, W. A., Oakley, C. L., and Porter, R. R.
(1960), Proc. Roy. Soc. B151. 478.

Bulletin World Health Organization (1964), 30. 447.

Campbell, D. H., and Bullman, N. (1952), Prog. in the Chem. of Org.
Nat. Prod. 9 443.

Cebra, J. J., and Small, P. A. (1966), In Preparation.

Cebra, J. J., Colberg, J. E., and Dray, S. (1966), J. Exptl. Med.
123. 547.

Cebra, J. J., and Goldstein, G. (1965), J. Immunol. 95. 230.

Chan, P. C. Y., and Deutsch, H. F. (1960), J. Immunol. 85. 37.










Chaplin, H., Cohen, S., and Press, E. M. (1965), Biochem. J.
95 256.

Chodirker, W. B., and Tomasi, T. B. (1963), Science 142. 1080.

Cleland, W. W. (1964), Biochemistry 3L 480.

Deutsch, H. F., and Morton, J. I. (1957), Science 125, 600.

Deutsch, H. F., and Morton, J. I. (1958), J. Biol. Chem. 231,
1107.

Draff, M. A., and Shulman, N. R. (1965), J. Exptl. Med. 121.
631.

Edelman, G. M., Kunkel H. G., and Franklin, E. C. (1958), J.
Exptl. Med. 108, 105.

Edelman, G. M., and Poulik, M. D. (1961), J. Exptl. Med. 113.
861.

Ellman, G. L. (1959), Arch. Biochem. Biophys. 82. 70.

Fahey, J. L. (1963), J. Immunol. 91. 438.

Fahey, J. L., and Goodman, H. (1964), Science 143, 588.

Fahey, J. L., and Humphrey, J. H. (1962), Immunol. 5 104.

Fahey, J. L., and Lawrence, M. E. (1963), J. Immunol. 91, 597.

Farr, R. S. (1958), J. Inf. Dis. 103, 239.

Fleishman, J. B., Pain, R. H., and Porter, R. R. (1962), Arch.
Biochem. Biophys. Suppl. 1. 174.

Fleishman, J. B., Porter, R. R., and Press, E. M. (1963), Biochem.
J. 88. 220.

Franklin, E. C., and Kunkel, H. G. (1958), J. Lab. Clin..Med.
52. 724.

Franklin, E. C., and Ovary, Z. (1963), Immunol. 434.

Fudenberg, H. H., and Kunkel, H. G. (1957), J. Exptl. Med.
106 689.

Gitlin, D., Rosen, F. S., and Michael, J. G. (1963), Pediatrics
31 197.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EEHJ5WRFD_PG2U12 INGEST_TIME 2012-12-07T22:44:20Z PACKAGE AA00012913_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES