This item is only available as the following downloads:
THE PHYSICOCHEMICAL PROPERTIES
AND BIOLOGIC ACTIVITIES OF
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
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
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
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
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
13. Imaunoelectrophoretic and radioim3unoelectro-
phoretic analyses of anti-SI serum and yM-
antibody . 61
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
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
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,
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
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,
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
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-
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
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.
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.
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
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
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
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
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-
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.
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 H HQ
4 -N-C---- HO-CH2-CH2-SH -N-(---
REDUCTION WITH 2-MERCAPTOETHANOL
HH9 H H F)
HS-CH2CHOH)2CH2-SH -N-~-C- .-N
-N- H--2- IHH
REDUCTION WITH DITHIOTHREITOL
FIGURE 3: Reactions leading to reduction by 2-mercaptoethanol and
(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
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
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
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.
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-
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
60 100 140 180 220
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.~ ______ ~-
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
' YM yG
FIGURE 6: The antigenic analysis of horse anti-SI inmunoglobulins
and their isolated 'light' polypeptide chains. See text for expla-
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
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
S ai M-Ab ISll
.a / 0 0IM JE
I I--, .'. H N i
DIALYSIS vs 000525M H S-BORATE BUFFER, pH 85 I
7.1 DIALYSIS v 0OO. 25M
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
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
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
Titrate Free Sulhydryls.
to pH 2.5 with
drop at HCI.
Apply to 1.8 X 35 cm
Col at G-25 Sephadex
0.15M NaCI, 0.I001M EDTA,
Monitor fractions at 280 mI.
Use peak tube.
Titrate SH Protein Coni
30 mg yM-antibody in 6 ml.
Add stock, titrate. DTT to make
.0025M. Incubate at 370 for
Dial. vs. .00525M
phos-borate buffer, pH 8.5.
to 0.05M. Stand at
Room temp. I Hr
Free Light Chains Extent of Redn. yM-antibody.
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.
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.
12 6 0 24
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
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
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;
FIGURE 11: A series of ultracentrifuge frames from various reoxi-
dation experiments showing the variable results obtained.
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
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
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
Sl .. ---- -
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.
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
2000- i 20
1000 00 200
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
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
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
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
Present lins Present Counts/10 minb ODZgg80 CountslO min
yM-antibodies 5038 0.375 3057
yM-antibodies 5289 3.94 3619
yM-subunit 5098 4.58 2918
yG-globulin (normal) NDe 1389
yG-globulin (immune)d 1843 3.76 ND
6.4S immunoglobulins -- 4036
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
yG-globulin isolated from serum of the immunized horse by the same method used
to isolate yG globulin from the unimmunized horse.
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-
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
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.
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).
Ability of yM-subunits to Bind SI
as Detected by the Farr Method a
Vol 1251 -SI
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
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-
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
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 retention of antigenic integrity by the 125 SI is shown
in Figure 17. These data were obtained by adding increasing amounts
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
S" r -0.3
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
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
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
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.
Protection Against Type I Pneumococcus Infection in the
Mouse by Equine Anti-SI Immunoglobulins
10 to 0.01 pg/mla
10 to 0.1 pg/mla
10 to 0.001 pg/ml
1000 to 100 pg/ml
10 to 0.1 pg/ml
1000 and 100 pg/ml
100 and 10 Ag/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.
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
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-
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
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 --- -
yG-antibody precipitating and 6.4S --- no effect
yA-antibody binding precipitating? 10.0S --- ---
Properties of Immunoglobulins from Various Species
tivity Sedimentation Intact Heavy Light
onrre Coefficient Molecule Chain Chain
or Serunm S
--- --- Miller and Metzger, 1965
Lamm and Small, 1966
Lamm and Small, 1966
Small and Lamm, 1966
-- --- Kabat and Pedersen, 1938
--- --- Kabat and Pedersen, 1938
--- Pappenheimer, et al., 1940
or Serum Source Coefficient Molecule Chain Chain
Altmeier, W., Robbins, J. B., and Smith, R. T. (1965), In
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
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.
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.
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,
Draff, M. A., and Shulman, N. R. (1965), J. Exptl. Med. 121.
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.
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.
Franklin, E. C., and Ovary, Z. (1963), Immunol. 434.
Fudenberg, H. H., and Kunkel, H. G. (1957), J. Exptl. Med.
Gitlin, D., Rosen, F. S., and Michael, J. G. (1963), Pediatrics
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