Intramolecular heterogeneity of IgM antibodies

MISSING IMAGE

Material Information

Title:
Intramolecular heterogeneity of IgM antibodies
Physical Description:
x, 96 leaves : ill. ; 29 cm.
Language:
English
Creator:
Giles, Robert Clay, 1952-
Publication Date:

Subjects

Subjects / Keywords:
Hybridomas   ( mesh )
Immunoglobulin M   ( mesh )
Immunoglobulins, Heavy-Chain   ( mesh )
Immunology and Medical Microbiology thesis Ph.D   ( mesh )
Dissertations, Academic -- Immunology and Medical Microbiology -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida.
Bibliography:
Bibliography: leaves 90-95.
Statement of Responsibility:
by Robert Clay Giles.
General Note:
Photocopy of typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000342610
oclc - 08412126
notis - ABX8759
System ID:
AA00009111:00001


This item is only available as the following downloads:


Full Text













INTRAMOLECULAR HETEROGENEITY OF IgM ANTIBODIES


BY

ROBERT CLAY GILES








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


1982





























"Take instruction, and riot silver,
And knowledge rather than choicest gold.
For wisdom is better than jewels;
And all desirable things can not compare
with her."

Proverbs 8:10-11 (NAS)
The Bible















ACKNOWLEDGEMENTS

First, thanks to Dr. Bruce Glick for his encouragement

to me that I pursue this undertaking and for his assistance

in allowing it to become possible. I would like to acknow-

ledge the continued support and encouragement of-my'advisor

Dr. William Clem. His experience and perspective in solving

scientific problems have been a major contribution to my

learning process. I am appreciative of. both his tutorage

and his friendship. I would also like to extend my appre-

ciation to Dr. Kenneth Berns as Chairman of the Department

of Immunology and Medical Microbiology and the other mem-

bers of my committee, Dr. George Gifford, Dr. Bill Holloman,

and Dr. Paul Klein, for their time and advice. I would also

like to thank Dr. David Klapper for serving as an advisor

for this study and for his collaboration.

Thanks go out to my fellow graduate students, Terry

Van Dyke, Jim Rusche, and Tom Rowe, for helping to make my

short stay in Gainesville a memorable one. Special thanks

to Erv Faulmann and Marie Hoo'ver for being my friends

through it all.

Thanks to my father and my mother for always giving

of themselves unselfishly and for their continued support

and encouragement during this endeavor of my life. I

would like to thank my wife and closest friend, Jansen,

iii









for her love and sacrifice during this time of our lives.

Without her understanding and support this work would not

have been possible. I would like to thank my son Clay

for the added joy he has brought to our lives and for

not keeping us up too many nights. I look forward to

growing and learning together as a family.

















TABLE OF CONTENTS


ACKNOWLEDGEMENTS .

LIST OF TABLES .

LIST OF FIGURES .

ABSTRACT . .


CHAPTER I

CHAPTER II


CHAPTER III


BACKGROUND

INTRASUBUNIT HOMOGENEITY IN HETERO-
GENEOUS IgM ANTIBODIES TO THE DNP
MOIETY DERIVED FROM A MURINE HYBRIDOMA
CELL LINE

Introduction . .
Materials and Methods .
Results . .
Discussion . .

INTRAMOLECULAR HETEROGENEITY OF TWO
IgM ANTIBODIES TO THE DNP MOIETY
DERIVED FROM MURINE HYBRIDOMA CELL
LINES


Introduction .
Materials and Methods
Results .
Discussion .


CHAPTER IV

REFERENCES .

BIOGRAPHICAL


SUMMARY AND CONCLUDING REMARKS .


SKETCH


Page
iii


vii









LIST OF TABLES

TABLE PAGE

1. Amino Terminal Sequences of Light Chains
from Unfractionated 19S and Active and
Inactive 7S Reductive Subunits from 14PAF 26

2. Amino Terminal Sequences of Heavy and Light
Chains of IgM Antibody to DNP from a Hybridoma
Line NP3-17 C1-20 . 55

3. Comparison of Hapten Elution Profiles of Mildly
Reduced SP2/0 1-64 Cl-12 IgM Antibody from Tm-
and Non-Tm-treated Culture Supernatants 75

4. Immunoprecipitation of Tunicamycin (Tm)-
treated and Non-Tm-treated Culture Supernatants
and Cell Lysates . 76









LIST OF FIGURES


FIGURE PAGE

1. Equilibrium dialysis of mouse hybridoma
protein 14PAF with DNP-E-aminocaproate 17

2. Equilibrium dialysis of DNP-E-aminocaproate
of 7S subunits from 14PAF . 20

3. Step-wise elutions of protein 14PAF from
DNP-lys-sepharose affinity column using
four concentratjo c .of hapten . 22

4. Alkaline urea polyacrylamide gel electro-
phoresis of isolated light chains from
unfractionated (A), inactive (B), and active
(C) 7S subunits . 25

5. Equilibrium dialysis with DNP-c-aminocaproate
of 14PAF recombinants using either heavy chains
from active subunits (AH) or inactive subunits
(IH) and light chains from active subunits
(AL ) . . 30

6. SDS-PAGE of recombinants formed using heavy
chains from either active or inactive subunits
combined with an equimolar amount of light
chains consisting of 60% from active subunits
and 40% from inactive subunits . 33

7. Mild reduction profile of 19S IgM from
hybridoma SP2/0 1-64 Cl-12 using increasing
amounts of 2-ME from left to right 44

8. Equilibrium dialysis of 19S IgM from
hybridoma SP2/O 1-64 Cl-12 with DNP-e-amino-
caproate . . 52

9. Alkaline urea polyacrylamide gel electro-
phoresis of isolated light chains from (A)
SP2/0 1-64 Cl-12 and (B) NP3-17 CI-20 54

10. Equilibrium dialysis using DNP-c-aminocaproate
of NP3-17 CI-20 "7S" fractions resulting from
mild reduction at dilute protein concentra-
tions (< 1 mg/ml) . 60

11. 10% polyacrylamide gel electrophoresis (SDS)
of NP3-17 CI-20 "7S" . 62


vii









FIGURE PAGE

12. 10% polyacrylamide gel electrophoresis of
timed trypsin digestions of SP2/0 1-64 CI-12
7S reductive subunits . 64

13. Sephadex G-200 elution profile of SP2/0 1-64
Cl-12 Fab's obtained after trypsinization
for five hours ... . 67

14. 10% polyacrylamide gel electrophoresis (SDS)
of unfractionated (A) and active (C) Fab's
obtained from five hour trypsinization of
SP2/0 1-64 C1-12; (B) contains inactive, non-
Fab material . . 69

15. Equilibrium dialysis of Fab fragments
obtained from a five hour trypsinization
of SP2/0 1-64 Cl-12 with DNP-6-aminocaproate 71

16. Autoradiogram of specifically purified
3 S-methionine molecules from supernatants
or cell lysates of Tm- or non-Tm-treated
SP2/0 1-64 Cl-12 cells . 74


viii
















Abstract of Dissertation Presented to the Graduate
Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree of
Doctor of Philosophy

INTRAMOLECULAR HETEROGENEITY OF IgM ANTIBODIES

BY

Robert Clay Giles

May, 1982

Chairman: Dr. L. William Clem
Major Department: Immunology and Medical Microbiology

A hybridoma line (14PAF) secreting 19S IgM antibodies

reactive with the DNP moiety was derived from a fusion of

DNP-primed murine spleen cells with the myeloma line P3-

X63-Ag-8. The failure to demonstrate ten binding sites per

pentameric molecule (six were measured) was attributable to

the presence of two different light (L) chains. Analysis

of the reductive subunits indicated the presence of two

types. One type (designated active) contained an average

of two homogeneous ligand binding sites per subunit and did

not contain L chains normally found in MOPC-21, the myeloma

protein secreted by the P3 myeloma line. The other type of

subunit (designated inactive) did not contain any DNP

binding sites and contained only MOPC-21 L chains. Recom-

bination studies with u and L chains from each type of

reductive subunit indicated that the noncovalent assembly

of the 2p-2L chain subunits was restricted by the L chains

ix









in such a way that the recombinant subunits were homogen-

eous in terms of L chains. Equilibrium dialysis studies

with active homogeneous recombinant subunits indicated the

presence of but one ligand binding site per 2p-2L chain

subunit.

Two additional murine anti-DNP IgM hybridoma anti-

bodies were derived using the non-immunoglobulin producing

myeloma lines P3-X63-Ag8.653 and SP2/O-Ag-14. These IgM

antibodies, NP3-17 CI-20 and SP2/0 1-64 CI-12, displayed

an average of five high affinity binding sites for the DNP

moiety. Reductive subunits of each of these proteins

absorbed to and were hapten eluted from an affinity column.

These subunits, when examined by equilibrium dialysis, each

contained an average of one high affinity binding site.

Structural analysis of these molecules indicated each to be

homogeneous. Results obtained from trypsin hydrolysis of

each molecule as well as studies performed with reductive

halfmers (H-L) indicated that differences in conformation

of the binding sites may account for the observed binding

heterogeneity.

The results of the chain recombination studies with

14PAF closely mimic the natural situation that occurs with

NP3-17 Cl-20 and SP2/0 1-64 Cl-12. These findings strongly

implicate the mechanism of assembly as a possible factor in

the generation of conformational differences that could ac-

count for the intramolecular heterogeneity of ligand binding

seen with certain IgM molecules.
x














CHAPTER I
BACKGROUND

The structure and physiochemical properties of mammal-

ian IgM immunoglobulins have been elaborated in considerable

detail largely through the study of homogeneous IgM myeloma

proteins. These studies demonstrate that secreted mammalian

IgM antibodies are pentameric molecules containing ten

heavy-light chain pairs and thus, by analogy with IgG

antibodies, should contain ten equivalent ligand binding

sites per molecule (reviewed 1). Although several studies

employing structurally homogeneous IgM monoclonal proteins

have demonstrated the presence of ten homogeneous, rela-

tively low affinity binding sites per molecule (2,3),

numerous other studies with conventional IgM antibodies

have indicated considerable heterogeneity of ligand binding

with valences of less than ten. In several instances

valences of less than ten were attributable to steric

factors related to antigen size, For example, in one

study the measured valences of IgM antibodies to dextran

progressively decreased as the size of the poly-glucosan

ligand was increased (4). Similarly in another study a

human IgM myeloma protein which bound human IgG was observed

to have an effective valence of five per intact pentamer

(one per subunit) whereas each of the ten Fabp fragments

contained an active binding site (5). In other cases,


-1-









haptens which are not likely to impose steric limitations

on the antibody combining sites were employed. In this

latter context it is important to point out that an average

of five high and five low affinity binding siteshas fre-

quently been observed in a variety of species (6,7,8,9).

One study utilizing an unusual IgM subunit containing only

a single heavy-light chain (10) demonstrated that 50% of

these subunits were retained by an immunoadsorbant column

previously used to isolate this heterogeneous antibody from

serum. Since it was also shown that greater than 90% of

the 7S subunits of this IgM antibody were retained by an

immunoadsorbant, the hypothesis was suggested that hetero-

geneity must exist within individual 7S subunits. However

due to the fact that the IgM was not homogeneous, critics

explained the data by the presence of different populations

of high and low affinity IgM molecules, half of which lack

sufficient affinity as heavy-light chain pairs to be

retained by an immunoadsorbant.

In terms of resolving this issue there would seem to

be two different, but not necessarily exclusive, possibil-

ities. The first is that ligand binding heterogeneity is

a consequence of differences between antibodies in a popu-

lation, i.e., intermolecular differences due presumably to

differences in antibody primary structure. The second

possibility is that ligand binding heterogeneity reflects

differences within individual antibody molecules, i.e.,

intramolecular differences. While the first possibility









is likely the case in some circumstances wherein obviously

structurally heterogeneous antibodies are employed, the

frequent finding of an average of half high and half low

(, 1% as high as those with high affinity) affinity sites

suggests consideration of the second possibility, Further-

more, one attempt at separating the two populations of

binding sites suggested that both were on the same IgM

molecule (11), hence adding credence to the possibility

of intramolecular heterogeneity.

The solution to this question has, to a large part, been

elusive due to difficulties in obtaining sufficient amounts

of structurally homogeneous IgM antibodies. A potential

approach to this question was offered by the observations

that the physicochemical properties of immunoglobulins from

lower vertebrates indicate that such animals may be

restricted to but one immunoglobulin isotype analogous to

IgM (reviewed, 12,13). Furthermore, studies using sharks

indicated that certain antigens, especially A-variant

streptococci, can elicit the production of very large

amounts of relatively homogeneous 19S antibodies (up to

10 mg/ml serum) (14,15). Although it has not been possible

to isolate a suitable hapten for studying affinities and

valences with this antigen (16), the finding in sharks of

reasonably good 19S antibody responses to the capsular

polysaccharide of pneumococcal cells and to the DNP hapten

covalently coupled to streptococcal cells provided sufficient

material for limited studies, The results of these studies

with shark IgM antibodies are summarized below.









Shark 19S antibodies to the Type III (S3) pneumococcal

capsular polysaccharide were isolated from immune sera by

affinity chromatography and subjected to equilibrium

dialysis using the S3 hexasaccharide hapten. The results

demonstrated several important points: (1) The IgM anti-

bodies contained an average of ten combining sites per

molecule, (2) in all cases, the antibodies showed marked

heterogeneity of affinities, (3) the antibodies were all

of low average affinities and (4) there was no increase in

the average affinity of the antibodies isolated from any

single animal for periods up to twelve months after initial

immunization. In order to determine if a single IgM mole-

cule contained ten equivalent combining sites, the anti-

bodies isolated from several animals were fractionated by

liquid isoelectric focusing. Equilibrium dialysis experi-

ments using focused fractions showed the presence of ten

functionally identical combining sites per 19S molecule.

As a proof of the structural homogeneity of focused

fractions, antibodies were mildly reduced, separated into

H and L chains, recombined to 7S subunits (2H-2L chains),

and tested for combining sites by equilibrium dialysis,

The results indicated that these 7S recombinants of focused

antibody fractions each contained two binding sites iden-

tical to those of the intact antibody whereas heterogeneous

(unfocused) recombinants or isolated H and L chains failed

to show any binding activity. The conclusion from this

study is that the heterogeneity of ligand binding exhibited




5



by nurse shark 19S antibodies to the capsular polysac-

charide of the Type III pneumococcus can be attributed

to intermolecular heterogeneity most likely at the primary

structural level (17).

In contrast to the results obtained in sharks with the

pneumococcal antigen, those obtained with antibodies to the

DNP moiety dictate a quite different conclusion (18),

Equilibrium dialysis studies using the hapten DNP-E-amino-

caproate with affinity purified nurse shark 19S antibodies

to DNP demonstrated several important points: (1) The 19S

antibodies exhibited heterogeneity of ligand binding with

an average of five high and five low affinity sites per

molecule, (2) the affinities of the low affinity sites were

approximately 1% of those of the high affinity sites, and

(3) no evidence for increased affinities was seen for up to

twenty-one months of immunization. To study the basis for

the ligand binding heterogeneity of the shark 19S anti-

bodies to DNP, several different approaches were employed.

The results of studies with subunits and proteolytic frag-

ments showed that neither steric hindrance nor allosteric

effects could account for the observed heterogeneity, i.e.,

the two forms of binding sites were on separable Fabp frag-

ments. In fact these results strongly suggested that both

types of combining sites were present within individual

reductive subunits. Preparative liquid isoelectric focus-

ing was used in an attempt to separate structurally homo-

geneous 19S antibodies. The results from equilibrium









dialysis studies indicated that each of 16 different focused

preparations contained an average of five high and five low

affinity sites.. These results seemingly argue strongly

against the possibility that the observed heterogeneity was

due to intermolecular heterogeneity. It seems highly

improbable that each of sixteen different isoelectrically

focused anti-DNP preparations would fortuitously be composed

of equimolar mixtures of 19S molecules, half of which have

ten equivalent ligand binding sites approximately 100 times

higher in affinity than those of the remaining low affinity

population. It would seem more likely that the isoelectric

focusing technique actually separated structurally homo-

geneous antibodies. This interpretation .is supported by the

results of recombination studies. Heavy and light chains

from focused molecules yielded 7S recombinants with high and

low affinity sites, albeit in low yields, indistinguishable

from their putative counterparts on the intact molecule.

Since heterogeneous recombinants resulted in binding sites

of only low affinities, it seems likely the focused prepara-

tions were structurally homogeneous and hence the observed

ligand binding heterogeneity must be an intramolecular

phenomenon (19).

The presence of binding sites of two different affin-

ities on a single antibody molecule can be explained by

primary structural and/or conformational differences between

the two types of sites. Although the presence of amino acid

sequence heterogeneity in the heavy and/or light chains of a









single IgM molecule seems unlikely, the lack of amino acid

sequence data on the shark IgM antibodies makes it impossible

to a priori rule out this explanation, Despite this

unclarified point, evidence from another experimental

approach seems to favor the latter possible explanation

mentioned above. Shark 19S antibodies to the DNP moiety

(exhibiting five high and five low affinity sites) and to

the S3 polysaccharide (exhibiting ten low affinity sites)

were each treated with 5 M guanidine-HC1 and studied by

equilibrium dialysis with the appropriate hapten after

removal of the denaturing solvent. The results showed this

treatment had no effect on the number or affinity of sites

on the antibodies to the S3 polysaccharide but had a consid-

erable effect on the antibodies to DNP. In fact the data

obtained with these latter antibodies indicated linear

Scatchard plots that readily extrapolated to a valence of

ten low affinity sites. Thus it appears that the guanidine-

HCl treatment converted the five high affinity sites to low

affinity ones. This was presumably accomplished by causing

conformational changes in or near the high affinity sites,

Unfortunately sufficient amounts of shark antibodies were

not available to perform detailed studies of the conforma-

tional relationships between Fabp fragments containing high

or low affinity sites.

In light of the suggestion that the intramolecular

heterogeneity of ligand binding by some shark (and perhaps

other species) IgM antibodies to the DNP moiety may result









from intramolecular conformational differences, it seems

appropriate to speculate on possible mechanisms responsible

for these putative differences. One possible explanation

is that IgM molecules may be assembled intracellularly by

two different mechanisms depending upon the cell involved.

Since the data discussed above suggest these differences

are in fact intrasubunit differences, it is conceptually

sound to'suggest that those lymphocytes secreting IgM

antibodies with ten homogeneous sites may do so by assembling

the subunits (2H-2L) from halfmers (H-L) as is the case for

a relatively limited number of IgM myeloma proteins studied

(20,21). In this case it would be predicted that the sub-

units should exhibit symmetry of ligand binding. On the

other hand, those cells secreting IgM antibodies with half

high and half low affinity sites may do so by assembling

the subunits in a different manner, i.e., either H-H-H-H-L-*

L-H-H-L or H-LH-H-L+L-H-H-L. If these latter modes o.f

assembly exist, as is the case with certain'other immuno-

globulin isotypes (22), it is conceivable (although admit-

tedly antidogma) that the formation of the first-H-L chain

pair may somehow influence the conformation of the second

pair so as to result in a different affinity site. A priori,

one cannot tell which (high or low affinity) site would be

formed first.

Since the suggestion of intramolecular conformational

differences may seem to some to reflect too much "antidogma,"

it would be appropriate here to briefly consider this issue.









There are several reports in the literature which document

the existence of small but detectable conformational dif-

ferences between Fab fragments derived from specific anti-

bodies and immunoglobulins (23,24). Admittedly these

proteins were heterogeneous (intermolecularl in primary

structure and hence the existence of such conformational

differences may not be surprising, Furthermore, while the

relatively recent surge of X-ray crystallographic data has

indicated considerable 3-dimensional similarities between

several homogeneous Fab fragments, there are also indica-

tions of slight differences (reviewed 25), Perhaps the most

important observations that could be cited in support of the

possible existence of intramolecular conformational hetero-

geneity are those of Edmundson and colleagues (26) with the

myeloma protein and dimer Bence-Jones protein of patient Mcg.

These data indicate rather conclusively that different

conformations of a polypeptide chain (L chains in their

study) can be derived from the same amino acid sequence

depending upon the other chain to which the peptide is

paired. Similar precedents have been reported for the

crystalline structures of dimeric insulin (27) and chymo-

trypsin (28). Thus it does not seem absurd at this point to

suggest that such conformational differences could exist and

be functionally important within structurally homogeneous

IgM molecules. Finally, since part of our rationale for

suggesting the role of intramolecular conformational differ-

ences to explain the apparent intramolecular heterogeneity









of ligand binding of some IgM antibodies was based upon the

results obtained with guanidine-HCI treated antibodies, the

results of Richards and coworkers with mouse IgA myeloma

protein 460 become pertinent (29). Their data indicated

that this protein, possessing two ligand binding specifi-

cities, could be rendered unreactive with one ligand (DNP)

by the guanidine-HC1 treatment whereas the affinity for the

other menadionee) was unaffected. Presumably this partial

loss in function resulted from an intracombining site

conformational change. It is unknown if such a change

could be detected by the methodology available.














CHAPTER II
INTRASUBUNIT HOMOGENEITY IN HETEROGENEOUS
IgM ANTIBODIES TO THE DNP MOIETY DERIVED
FROM A MURINE HYBRIDOMA CELL LINE

Introduction

In light of the immunologic importance of IgM it

becomes imperative to develop approaches for obtaining

sufficient amounts of structurally homogeneous mammalian

IgM antibodies with specificities for defined ligands in

order to clearly resolve the question of intramolecular

heterogeneity. The approach that seemed most promising and

was undertaken involved cell fusions (hybridomas). Chapter

II describes the initial results obtained with one murine

hybridoma line secreting 19S IgM antibodies reactive with

the DNP moiety. Although the secreted IgM product was a

structurally heterogeneous molecule (due to the presence of

two different light chains) it appeared to be assembled in

a random fashion from two pools of homogeneous subunits.

Furthermore in vitro recombination studies with the

component polypeptide chains revealed the surprising

findings that a) the subunit homogeneity was L chain

directed and b) each recombinant subunit exhibited but

one active site.


-11-









Materials and Methods



Hybridomas

The fusion techniques described by Galfre' et al.

(30) and Gerhard et al. (31) were employed with some modi-

fications. Briefly 1 x 108 BALB/c spleen cells (mice

injected 3 days previously with 50 ug 2,4-dinitrophenyl-

ficoll) and 1 x 107 myeloma cells (P3-X63-Ag8; derived by

Cotton and Milstein) (32) were mixed and centrifuged at

300g for 8 minutes. Cells were resuspended for fusion in

1 ml of a 50% solution (v/v) of polyethylene glycol (PEG-

1000, J.T. Baker Chemical Co.) in Dul.becco's Minimal

Essential Medium (DMEM, Gibco). .This suspension was

further diluted at a rate of 6 ml/min with DMEM over a

period of 5 minutes. The cells were washed and resuspended

in 35 ml of the selective medium whic-h contains hypoxan-

thine, aminopterin, and thymidine (33). The cells were

distributed in 100 ul aliquots into 96 well microtiter

plates (Falcon Microtest II) and incubated at 370C in a

humidified 5% C02-95% air mixture. After substantial

growth (, 10-12 days) supernatants were screened for IgM

antibodies to the DNP moiety using a radioimmunoassay (34).

Cultures positive for anti-DNP antibody were cloned by

limiting dilution in soft agarose (35). Individual posi-

tive clones (- 5 x 10 cells) were injected into BALB/c
mice primed with pristane to obtain ascitic tumors. One

such clone, designated 14PAF, was selected for the studies

reported here.









Immunochemical Procedures

Mouse antibodies to the DNP moiety were purified from

ascitic fluid by affinity chromatography, freed of hapten,
3
and examined by equilibrium dialysis against H-DNP-c-

aminocaproate as described previously (18). For calculating

protein concentrations of the pentameric IgM antibodies; a

molecular weight of 900,000 daltons and an extinction

coefficient (E280nm cm
coefficient (E280nm 1cm ) of 11.9 were assumed.

In one experiment 14PAF was subjected to 5.0 M

guanidine-HC1 for one hour at room temperature. The

guanidine-HC1 was removed by dialysis against Tris buffered

saline (0.15 M NaCl, 0.01 M Tris-HC1, pH 7.4). These

guanidine-HC1 treated antibodies were subjected to equili-

brium dialysis as previously described.

Mildly reduced 7S subunits of protein 14PAF were pre-

pared by subjecting purified 19S material to reduction with

0.1 M 2-mercaptoethanol in 0.5 M Tris-HCl, pH 8.0, for one

hour at 220C followed by alkylation with 0.15 iodoacetamide

for one hour on ice. Gel filtration under nondenaturing

conditions (Sephadex G-200 equilibrated with 0.15 M NaCl,

0.01 M Tris-HCl, pH 7.4) indicated that > 95% of such re-

duced and alkylated 7S material eluted in a volume expected

to contain %180,000 dalton proteins; analysis under

denaturing conditions without additional reduction indi-

cated that >90% of the 7S subunits dissociated into equi-

molar H and L chains. A small percentage of halfmer (H-L)

molecules was also detected. Mildly reduced and alkylated









H and L chains and halfmers were obtained by gel filtering

the 7S subunits on Agarose A5M columns equilibrated with

5 M guanidine-HCI containing 0.01 M iodoacetamide.

Recombinant molecules were prepared from mildly reduced

and alkylated H and L chains as described previously (17).

Briefly, the desired amounts of separated H and L chains

were mixed in 5 M guanidine-HCl containing 0.01 M iodoace-

tamide and concentrated by positive pressure dialysis to

\ 5 mg/ml while dialysing against Tris buffer, pH 7.4.

These recombinants were gel filtered under nondenaturing

conditions and >90% of the UV absorbing material eluted in

a volume expected to contain 180,000 dalton proteins;

electrophoresis of these recombinants in SDS polyacryla-

mide gels indicated the--.presence of equimolar H and L

chains (see Figure 5 for example). Alkaline-urea gel

electrophoresis of extensively reduced L chains was per-

formed by the method of Reisfield and Small (36). SDS-

polyacrylamide gel electrophoresis was performed according

to the method of Laemmli (37). Amino terminal sequence

studies were performed by Edman degradation using an auto-

mated Beckman sequenator. PTH derivatives were identified

by high pressure liquid chromatography (38).

Results

The hybridoma line (14PAF) used here readily prolif-

erated as ascitic tumors in pristane-primed BALB/c mice and

yields of 10-20 ml ascitic fluid were obtained from indivi-

dual mice. Affinity chromatography of these ascitic fluids









on TNP-lys-sepharose yielded from 5-10 mg/ml antibody.

These isolated antibodies were considered to be exclusively

19S IgM(K) based upon sedimentation velocity measurements,

immunodiffusion analysis with commercial antisera and

SDS-gel electrophoresis studies (data not shown).

Equilibrium dialysis of the isolated 19S antibody

against the hapten 2,4-dinitrophenyl-e-aminocaproate

yielded the Scatchard plot depicted in Figure 1. These

data indicate the hybridoma derived IgM antibody contained

an average of about six binding sites per molecule with an

affinity of 2 x 106M Furthermore, Sips analysis indi-

cated a heterogeneity index of 0.98 which presumably

reflected a high degree of homogeneity in the binding

constants of the sites being detected. Equilibrium dialysis

of the 19S antibody treated with 5.0 M Gn-HCl demonstrated

an identical Scatchard plot to the untreated antibody

(see Figure 1).

In order to determine if the observation of about six

binding sites per molecule was due to some peculiar steric

effects, reductive 7S subunits of protein 14PAF were prepared

and studied by equilibrium dialysis. As can be seen in

Figure 2, these 7S subunits appeared to exhibit a ligand

binding pattern identical to that observed with the parent

pentameric molecule, i.e., 1l.2 sites with an affinity of

% 2 x 106M-1 per every two H-L chain pairs. These reductive

subunits were then subjected to affinity chromatography on

DNP-lysine-sepharose. Approximately 55-60% of the material,


































Figure 1. Equilibrium dialysis of mouse hybridoma protein
14 PAF with DNP-E-aminocaproate.








































10.0


5.0
r


E2.0
10


>1.0
L..









designated as active, absorbed and was hapten-eluted from

the affinity matrix; the other 40-45%, designated as

inactive, did not absorb. Each of these populations of

subunits was examined by equilibrium dialysis. As pre-

sented in Figure 2, the active subunits .exhibited two homo-

geneous binding sites identical with respect to affinity to

those seen in the unfractionated material. The inactive

subunits contained no demonstrable binding sites for DNP.

The finding that both the active and inactive subunits

could be derived from functional IgM molecules prompted the

hypothesis that the secreted pentameric IgM may be assembled

randomly from these two populations of subunits. Thus, as

an indirect test of this hypothesis, an experiment was

undertaken to attempt fractionating the secreted 19S

molecules into subpopulations with differing numbers of

hapten binding sites. Specifically purified protein 14PAF

was reabsorbed to an affinity column and sequentially eluted

with increasing (arbitrary) amounts of DNP-OH. The elated

components were freed of DNP-OH, quantified for protein

and studied by equilibrium dialysis. The results, depicted

in Figure 3, indicate that the initial pentameric IgM anti-

body population was composed of subpopulations with differ-

ent numbers of binding sites: While this experiment was not

"fine tuned" to the point where it is possible to-clearly

say that the pentamers were assembled in a random fashion

from the two forms of subunits, the results are certainly

suggestive.

































Figure 2. Equilibrium dialysis with DNP-e-aminocaproate
of unfractionated (*-*), and active (o--),
and inactive (&--0) 7S subunits from 14PAF.



















CD


2.0

1.5

1.0

0.5


1.0
r






























Figure 3. Step-wise elutions of protein 14PAF from DNP-lys-
sepharose affinity column using four concentra-
tions of hapten. 18% of the eluted protein was
recovered using 3 x 10-4M hapten; 39% using
1 x 10-3M; 30% using 3 x I0- M, and 13% using
1 x 10- M.






































5.0


s 4.0
CD

- 3.0


- 2.0


1.0


I I I
M E
CO C









In light of the above finding of two apparently distinct

subpopulations of subunits, the possibility of intramole-

cular heterogeneity at the structural level was considered.

The first approach to this question involved subjecting the

mildly reduced and alkylated L chains from each subunit to

SDS polyacrylamide gel electrophoresis without additional

reduction. The banding patterns of these mildly reduced

chains were clearly different in that the L chains from

the active subunits exhibited a somewhat slower mobility

than those from the inactive subunits (see Figure 6 for

example); extensive reduction of each of these chains

resulted in mobilities expected of 22,000 dalton peptides.

The second approach involved subjecting extensively reduced

and alkylated light chains isolated from the unfractionated

subunits and the two fractionated subunit subpopulations

to alkaline-urea gel electrophoresis. The gel patterns of

the L chains from the two subpopulations were clearly

distinct (Figure 4). These suggestions of L chain differ-

ences between the two forms of subunits prompted limited

amino acid sequence studies. The results of these studies

(Table 1) indicated several important points. Firstly, the

L chains from the unfractionated parent molecules exhibited

considerable primary structural heterogeneity with two

different amino acids being demonstrable at 5 (and possibly

7) different positions in the first 12 residues. Secondly,

the L chains from the active and inactive reductive subunits

were each homogeneous (through 12 residues) but quite































Figure 4. Alkaline urea polyacrylamide gel electrophoresis
of isolated light chains from unfractionated
(A), inactive (B), and active (C) 7S subunits.




25








A B C








TABLE 1

Amino Terminal Sequences of Light Chains from
Unfractionated 19S and Active and Inactive
7S Reductive Subunits from 14 PAF

7S subunits
Position 19S unfractionated Active Inactive

1 Asp Asn Asp Asn

2 Val Ile Val lle

3 Val Val Val

4 Met Met Met

5 Thr Thr Thr

6 Gln Gln Gin

7 Thr* Thr* Ser

8 Thr Pro Thr Pro

9 Leu Leu Lys

10 Ser Ser Ser

11 Leu Met Leu Met

12 Thr* Thr* Ser


The amino terminal sequence of the light

(Svasti and Milstein, 1971) is shown for


chains from MOPC-21

comparison.


*Ser could be masked


MOPC 21

Asn

lle

Val

Met

Thr

Gin

Ser

Pro

Lys

Ser

Met

Ser








different from each other, A comparison of the amino termi-

nal sequences of the L chains from these two subpopulations

of subunits with th.e L chains from MOPC-21 (the L chain

secreted by the myeloma cell line used tn the original

fusion) indicated that those from the inactive subunits were

identical (through 12 residues) to those from MOPC-21, The

L chains from the active subunits were quite different.

Hence it seems quite likely that the latter L chains were

encoded for by the genome of the plasma cell used in the

fusion. Therefore the secreted pentameric IgM product of

the hybrid cell contained both MOPC-21 and "anti-DNP" L

chains.

The p chains from both the active and inactive subunits

from protein 14PAF were blocked at the amino terminus

(presumably due to a pyrollidone carboxylic acid) and thus

were not amenable to automated sequence analysis. In an

attempt to ascertain if these p chains were functionally

identical, recombinant non-covalent subunits were produced

between all possible equimolar combinations of separated

mildly reduced and alkylated v and L chains derived from both

types of subunits. Each of these recombinants was gel,

filtered and the 180,000 dalton material (representing

>90% of the total) subjected to affinity chromatography

on DNP-lys-sepharose. Greater than 95% of the recombinants

formed with L chains from active subunits and v chains from

either active or inactive subunits were observed to absorb

to (and to subsequently hapten elute from) the affinity









columns; less than 5% of either of the recombinants formed

with L chains from inactive subunits absorbed to the columns.

Equilibrium dialysis with these latter inactive recombinants

indicated no detectable combining sites for the DNP group;

similar studies had indicated that the isolated v or L

chains bound no detectable hapten (data not shown). On the

other hand, the active recombinants recovered by hapten

elution from immunoadsorbant columns were each observed,

in duplicate experiments, to have an average of about one

combining site with an affinity identical to that of the

parent molecule (Figure 5). Similarly a single study was

performed utilizing halfmers (covalently linked H-L's) from

active subunits obtained from the leading edge of the H

chain peak on the 5 M guanidine Agarose A5M column. After

removal of the guanidine by dialysis against Tris buffer,

these subunits (n 180,000 dalton material by gel filtration)

displayed an average of one combining site (identical to

Figure 5). To further prove the functional identity of the

v chains present in the original 19S molecules, mildly

reduced and alkylated p chains from either the active or

inactive reductive subunits were recombined with an equtmolar

amount of mildly reduced and alkylated L chains derived from

the original molecule (i.e., %60% "anti-DNP" and r 40%

MOPC-21 L chains). Each of these recombinants was then

gel filtered on Sephadex G-200 and the 2H-2L chain recombi-

nants (representing >90% of the material) were subjected to

affinity chromatography on DNP-lys-sepharose columns;






























Figure 5. Equilibrium dialysis with DNP-E-aminocaproate of
14PAF recombinants using either heavy chains from
active subunits (AH) or inactive subunits (IH)
and light chains from active subunits (AL).
Recombinants made using either heavy chain with
light chains from inactive subunits showed no
detectable binding.























IH-AL


1.5 2.0


(0


x
C.3









approximately 40% of the applied material in each case

failed to absorb to the column whereas the remainder ab-

sorbed and was recovered by hapten elution. The findings

from this experiment were a) the recombinant subunits which

absorbed to the affinity column contained an average of one

binding site identical in affinity to that seen in the

previous active recombinants (Figure 5) and the original

molecules, and b) the active recombinants contained only

L chains with an SDS polyacrylamide gel electrophoretic

mobility characteristic of the "anti-DNP" L chains whereas

the inactive recombinants appeared to contain only MOPC-21

L chains (Figure 6). Soft laser scans of these stained gels

clearly indicated each of the recombinants to be composed

of equimolar v and L chains. In a further attempt to

ascertain if 14PAF v chains preferentially form recombi-

nants with either of the two forms of L chains, mildly

reduced and alkylated chains from active reductive subunits

were mixed with a two-fold molar excess of 14PAF L chains

(consisting of a 60:40 ratio of "anti-DNP" and MOPC-21 L

chains). The resultant mixture was, without subsequent gel

filtration, then subjected to affinity chromatography on a

DNP-lys-sepharose column; approximately 52% of the material

failed to absorb to the column whereas the remaining 48%

absorbed and was recovered by hapten elution. Again, as in

the above case where equimolar v and L chains were mixed,

the active recombinants contained equimolar p chains and

"anti-DNP" L chains when examined by SDS gel electrophoresis.





























Figure 6. SDS-PAGE of recombinants formed using heavy
chains from either active or inactive subunits
combined with an equimolar amount of .light
chains consisting of 60% from active subunits
and 40% from inactive subunits. Lane A repre-
sents unfractionated recombinants; Lane B repre-
sents recombinants which absorbed to and were
subsequently hapten eluted from a DNP-lys-
sepharose affinity matrix; and Lane C represents
recombinants which did not absorb to the affinity
column.











A B C
H bmP









The inactive fraction from this experiment contained p

chains and a greater than three-fold molar excess of L

chains; these L chains were present in a 55:45 ratio of

MOPC-21 to "anti-DNP" L chains. The implications of these

results are considered below.

Discussion

The original purpose of the work undertaken here was

to utilize murine hybridomas as sources of homogeneous IgM

antibodies to test the hypothesis that certain IgM anti-

bodies can exhibit intramolecular heterogeneity of ligand

binding that is not attributable to primary structural

differences. Although the presence of two different L

chains in protein 14PAF precludes its utility in this regard,

the studies reported here did reveal several novel and

potentially important aspects of IgM structure and function.

It was not surprising, based upon the observation of "mixed"

molecules secreted by other hybridoma cell lines derived

from Ig secreting myeloma cells (39), that protein 14PAF

contained two different L chains. It was surprising, how-

ever, that protein 14PAF was seemingly randomly assembled

from 2p-2L chain subunits that were homogeneous in terms of

L chains. This finding suggests that each type of subunit

may be assembled from p-L chain (or perhaps u-u-L chain)

(40) intermediates which, can only associate with homologous

p-L (or L) chain intermediates to form 2u-2L chain subunits.

The possibility that this putative L chain restriction at

this level is due to the presence of two different p chains









is seemingly ruled out by the results of the in vitro

recombination studies. The finding that p chains from

either type of reductive subunit (active or inactive)

could combine with "anti-DNP" L chains to form active

recombinant subunits certainly indicates a high level of

functional similarity in these chains. Similarly, the

observation of homogeneous 2y-2L chain recombinants formed

with a mixture of L chains and V chains from either type

of subunit argues strongly that the intrasubunit L chain

restriction is attributable to the L chains themselves.

In this regard the results obtained from the recombination

experiment using a two-fold molar excess of L chains to v

chains are particularly important in that they show rather

conclusively that either type of L chain can combine equally

well with the p chains. For example if the V chains had a

preference for recombining with the "anti-DNP" L chains,

no V chains would be expected in the inactive recombinant

fraction as they would all be found in the active fraction.

Likewise, if the p chains had shown a preference for the

MOPC-21 chains, > 85% of the recombination mixture should

have been found in the inactive recombinant fraction. As

pointed out in the results, 52% of the recombinant mixture

was found in the inactive fraction (hypothetically, one

would predict 53% if assembly was random for each type L

chain). Furthermore, the presence of "anti-DNP" L chains

along with i chains in the inactive mixture demonstrates

that the "anti-DNP" L chains were not a limiting factor in









forming recombinants but that the p chains must have an

equal capacity to form inactive recombinants with MOPC-21

L chains or active recombinants with "anti-DNP" L chains,

In addition the presence of 55% MOPC-21 L chains to n 45%

"anti-DNP" L chains in the inactive fraction is almost

precisely what would have been expected if assembly of the

recombinants was random for both L chains (hypothetically,

one would predict 57% MOPC-21 L chains and 43% "anti-DNP"

L chains). It thus appears that assembly of 14PAF subunits

in vitro mimics assembly of subunits in vivo both with

respect to ratios of active to inactive subunits and the

homogeneity in terms of L chains. Hence, the conclusion

from this study is that the combination of one of the L

chains with the v chain produced by this cell line results

in an assembly intermediate that will only associate with

a homologous p-L chain pair to form a subunit. Such

restriction may result from different L chain imposed v

chain conformational differences; future work is required

to determine if such putative V chain conformational differ-

ences are caused by the seemingly large conformational

differences (as manifested by the differences in SDS gel

electrophoretic mobilities) between mildly reduced and

alkylated MOPC-21 and protein 14PAF "anti-DNP" L chains.

A final point of interest, and potential importance,

from these initial studies with protein 14PAF was the

failure to achieve the anticipated number of combining

sites in the active recombinants. The 2p-2L chain in vitro








recombinants formed with U and L chains (or halfmers)

from active 2p-2L chain reductive subunits should have

exhibited two ligand binding sites if the in vitro recom-

bination process worked perfectly, i,e., such as seemingly

was the case during intracellular assembly (see Figure 2).

On the other hand, due to the technical intricacies of such

experimental manipulations, a loss of (or failure to

recover) some of the active sit-es'wcol d not be unexpected.

Thus, for the sake of discussion, if 50% of the sites were

lost (as suggested by the data in Figure 5) it would be

expected that such losses might be random and hence about

25% of the recombinants should have two sites, 50% have one

site and 25% have no active sites. In such a case, only

75% of the 2V-2L chain recombinants would be expected to

absorb to an affinity matrix; these absorbed and hapten

eluted recombinants should exhibit an average of 1.3 sites

per molecule. The results obtained with 2p-2L chain recom-

binants formed with protein 14PAF "anti-DNP" L chains and

either of the two v chains (from active or inactive reduc-

tive subunits) clearly showed that >95% of the recombinants

absorbed to a DNP affinity column and therefore had at least

one active site. Furthermore equilibrium dialysis with

these active recombinants indicated an average valence of

only one. Hence the conclusion from these results must be

that each 2u-2L chain recombinant had but one active site

for ligand binding. The reason for this rather striking

result is unknown but it would seem appropriate to speculate









that it may be attributable to intrasubunit conformational

differences arising during the in vitro assembly; perhaps

one mildly reduced and alkylated p-L chain somehow

influenced the conformation of the other pair to result in

a 2p-2L chain recombinant with but one active site for the

DNP moiety. Such conformational differences, if demon-

strable in the in vitro recombinants of protein 14PAF,

would certainly justify future studies regarding this

possible explanation for heterogeneity of ligand binding

by other IgM antibodies.















CHAPTER III
INTRAMOLECULAR HETEROGENEITY OF TWO IgM ANTIBODIES
TO THE DNP MOIETY DERIVED FROM MURINE
HYBRIDOMA CELL LINES

Introduction

As previously discussed the results obtained by

several different laboratories using IgM antibodies from a

wide variety of sources indicated that some,. but not all,

of these antibodies appeared to exhibit intramolecular

heterogeneity of ligand binding. As mentioned, these

earlier studies were limited by the relatively small amounts

and the structural heterogeneity of the IgM molecules avail-

able. In order to clearly resolve this question of intra-

molecular heterogeneity it has become imperative to

develop approaches for obtaining sufficient amounts of

structurally homogeneous mammalian IgM antibodies with

specificities for defined ligands. Chapter II describes

the initial attempt at obtaining a homogeneous anti-DNP

IgM antibody using hybridoma technology. The study of the

molecule obtained, designated 14PAF, resulted in some rather

interesting and novel observations which might be pertinent

to the question of intramolecular heterogeneity. However,

it was felt that this molecule would not, in reality,

permit an adequate assessment of the hypothesis regarding

intramolecular heterogeneity due to the presence of two

different L chains in the secreted pentamer.


-39-









This present chapter describes additional cell fusions

which were undertaken using two myeloma lines, NP3 (41)

and SP2/0 (42), which do not produce immunoglobulin. Two

of the cell lines obtained from these fusions have been

designated NP3-17 CI-20 and SP2/0 1-64 Cl-12. Each of

these cell lines grows as ascitic tumors in mice and yields

moderate amounts (% 1 mg/ml) of 19S IgM antibodies to the

DNP moiety. Although apparently structurally homogeneous

(as manifested by alkaline urea gel patterns of isolated

L chains from both proteins and by limited amino acid

sequence analysis of H and L chains from one, NP3-17

C1-20), the hapten binding data obtained for each of these

proteins indicate an average of only five high affinity

binding sites for the DNP group per 19S molecule. Greater

than 95% of the reductive 7S subunits absorbed to and were

hapten eluted from a DNP-lys-sepharose affinity column.

When examined by equilibrium dialysis, each subunit contained

an average of one high affinity binding site. One experi-

ment aimed at defining the molecular basis of this observed

binding heterogeneity attempted to determine if asymmetrical

carbohydrate attachment to the molecule is in any way

involved. Tunicamycin, an antibiotic that prevents glyco-

sylation of glycoproteins, was used in an effort to isolate

carbohydrate-free'IgM antibodies and to study their binding

properties.

Other studies attempting to define the molecular basis

of the observed binding heterogeneity involved physically








separating the two types of sites, Mild reduction of each

of these IgM molecules yielded predominantly halfmers (H-L)

in dissociating buffers. Experiments designed to disrupt

the non-covalent associations between opposing halfmers

of a 7S subunit, as well as trypsin digestion of both 7S

molecules, indicated that conformational differences may

exist between the two binding site regions of an individual

7S subunit. A proposed mechanism for this heterogeneity

will be discussed.

Materials and Methods

Hybridomas

Cell fusions were performed as described in Chapter II

using the myeloma cell lines P3-X63-Ag8.653 (41) and

SP270-Ag-14 (42). After substantial growth, culture

supernatants were screened for IgM antibodies to the DNP

moiety using a radioimmunoassay (34) or passive hemagglutin-

ation with DNP conjugated sheep red blood cells. Cultures

were cloned in soft agarose (35) and individual positive

clones were subsequently injected into BALB]c mice primed

with pristane to obtain ascitic tumors. Two such clones,

designated NP3-17 Cl-20 and SP2/0 1-64 C1-12, were selected

for the studies reported here.

Immunochemical Procedures

IgM antibodies to the DNP moiety were purified from

ascitic fluid by affinity chromatography, freed of hapten

by Dowex 1X-8 ion exchange chromatography, and examined by

equilibrium dialysis against 3H-DNP-c-aminocaproate as









described previously (18). For calculating protein concen-

trations of the pentameric IgM antibodies, a molecular

weight of 900,000 daltons and an extinction coefficient

(E28 In 1 ) of 11.0 were assumed.
280nm 1cm
Initially, reduced 7S subunits of both proteins were

prepared by subjecting purified 19S material (concentra-

tions ranged from 3-10 mg/ml) to reduction with .01 M

and .005 M 2-mercaptoethanol for IgM antibodies from

NP3-17 C1-20 and SP2/0 1-64 C1-12, respectively. These

levels of 2-ME were determined empirically for each molecule

as that level of reducing agent which yielded the greatest

percentage of subunits containing covalently linked H and

L chains or halfmers (covalently linked 2H-2L chain 7S

subunits could not be obtained; Figure 7). These reduc-

tions were performed in 0.5 M Tris-HC1, pH 8.0 for one

hour at 220C followed by alkylation with 0.15 M iodoace-

tamide for one hour on ice. Gel filtration under non-

denaturing conditions (Sephadex G-200 equilibrated with

0.15 M NaCl, .01 M Tris-HC1, pH 7.4) indicated that \ 95%

of such reduced and alkylated 7S material eluted in a

volume expected to contain a 180,000 dalton proteins;

analysis under denaturing conditions without additional

reduction indicated the 7S subunits dissociated predomi-

nantly into halfmers (H-L molecules). Subsequently it was

found that halfmers could be obtained in non-denaturing

buffers by mild reduction of dilute protein solutions

(.5-1.0 mg/ml) in .5 M Tris-HCl pH 8.5.































Figure 7. Mild reduction profile of 19S IgM from hybridoma
SP2/0 1-64 Cl-12 using increasing amounts of
2 ME from left to right. Lane 1, .001 M;
Lane 2, .005 M; Lane 3, .01 M; Lane 4, .05 M;
Lane 5, .1 M. Profiles of IgM from NP3-17 C1-20
were similar with .01 M 2 ME being optimum for
the production of halfmers (H-L).















2 345


19S


7S

H-L


'"e *: '* ..
" *" -'* "'l ^ '








Alkaline-urea gel electrophoresis of extensively

reduced L chains was performed as in the preceding chapter

by the method of Reisfield and Small (36). SDS-polyacryla-

mide gel electrophoresis was performed according to the

method of Laemmli (37),

Amino terminal sequence studies were performed by

Edman degradation using an automated Beckman sequenator.

PTH derivatives were identified by high pressure liquid

chromatography (38).

Tryptic hydrolysis of 7S reductive subunits (5-10

mg/ml) was performed for various time periods following

the protocol of Klapper et al. (43). Briefly a 1:100

enzyme (trypsin, 3-X crystallized, Worthington Biochemical

Corp.) to substrate ratio was established and the tempera-

ture was maintained at 370C throughout the digestion.

Trypsinization was done in .25 M Tris-HC1, pH 8.3, and

.01 M CaCI2 was added to the reaction mixture prior to

adding trypsin (44). Tryptic digestions were halted by the

addition of equimolar amounts of soybean trypsin inhibitor

(45). Fabu fragments were isolated by gel filtration (G-200

Sephadex equilibrated with 0.15 M NaCl, .01 M Tris-HC1,

pH 7.4), Active Fab's were then selected using a DNP-

lysine-sepharose immunoadsorbant. Equilibrium dialysis was

performed on these Fab's as previously described in this

text. A molecular weight of 50,000 and an extinction coef-

ficient of 13.0 were used to calculate Fab concentrations

(44).









Attempts were made to dissociate putative active

halfmers by disturbing non-covalent interactions between

opposing u chains. Mildly reduced antibody was allowed

to absorb to a DNP-lys-sepharose affinity column followed

by exposure of the column to chaotropic or ionic dissociating

agents. Various concentrations of the following agents

were used: guanidine HC1, sodium thiocyanate, acetate

buffer (pH 5.0), isopropyl alcohol, NP-40, and Tris-HCl

(pH 8.5).

Affinity labeling of reductive subunits from SP2/0

1-64 Cl-12 was attempted to irreversibly block the active

anti-DNP site on each 7S monomer without affecting the

inactive site. A two-fold excess of dinitrofluorobenzene

(DNFB) and dinitrobenzenesulfonate (DNBS) to antibody was

used. Two milliliters of a 7.2 x 10-6 M concentration of

antibody in .1 M'borate buffer, pH 8.3 was allowed to react

for two hours at room temperature with either DNFB or DNBS
"_5
at concentrations of 1.44 x 105 M. After two hours an

equal volume of .1 M Tris-HC1, pH 8.0, was added to each

tube to stop the labeling reaction. The antibody was

dialyzed against Tris-buffered saline (.15 M NaCl, .01 M

Tris-HC1, pH 7.4) overnight. The treated antibody was

passed over a DNP-lys-sepharose affinity column and any non-

absorbing antibody was measured spectrophotometrically at

280nm. It was hoped that when such treated subunits were

placed in 2.0 M guanidine-HCl and allowed to refoldd" upon









dialysis against Tris-buffered saline that 50% of the

original inactive sites would become active.

Attempts were made to obtain an anti-idiotypic anti-

serum to IgM from SP2/0 1-64 Cl-12. It was hoped that such

antiserum might detect idiotypic differences between mole-

cules which could bind hapten and those which could not.

Each of three BALB/c mice was initially immunized intra-

peritoneally with 100 pg of 19S antibody from SP/20 1-64

Cl-12 emulsified in complete Freund's adjuvant. Each

mouse received three additional intraperitoneal injections

of protein without adjuvant over the next two months. One

month into the immunization schedule each mouse received

1 x 106 SP2/0.14 myeloma cells in order to produce an

ascitic tumor. Three days after the end of the last

immunization ascitic fluid was removed and assayed for

anti-idiotypic antibodies.

Tunicamycin Experiment

Cells. Six x 106 SP2/0 1-64 Cl-12 cells were indubated

in 10 ml of methionine-free Dulbecco's MEM (with L-gluta-

mine, penicillin and streptomycin) with 15% fetal calf

serum. The experiments were run in a humidified atmosphere

of 5% C02:95% air at 370C.

Cell labeling. Cells were labeled with 35S-methionine

(American Searle, Arlington Heights, IL, 900 Ci/mmole) by

the addition of 100 uCi to 10 ml cultures. When tunicamycin

(Tm) (Eli Lilly, Co., Indianapolis, IN) was employed, cells









were suspended in the above medium containing 10 pg/ml

Tm for one hour before addition of the label. Cells were

labeled for eight hours before supernatants were removed

for analysis.

Purification of antigen-specific counts. Cells were

removed from supernatants by centrifugation at 400g for

15 minutes. Cell-free supernatants were passed over a

DNP-lys-sepharose affinity column to purify antigen-

specific radioactive molecules. This purified antibody

was hapten eluted (.1 M DNP-OH) and subjected to Dowex

1X8 ion exchange chromatography to remove the DNP-OH.

It was necessary at this point to spike the radioactive

IgM with unlabeled IgM in order to avoid significant loss

of radioactivity on the Dowex column. Radioactivity was

assessed by adding 10 pl of sample to 2 ml of scintillation

cocktail containing 50% (v/v) Triton-X-100.

Gel analysis. Tm-treated and non-treated purified

antibodies were subjected to SDS-DATD-acrylamide gel chroma-

tography (46). Gels were dried, overlaid with Kodak X-OMat

AR film and stored between intensifying screens for desired

exposure times.

Sequential hapten elutions. Purified Tm- or non-Tm-

treated antibody was mildly reduced with .1 M 2ME for one

hour and alkylated (.15 M iodoacetamide) for one hour on

ice. These reductive subunits were allowed to absorb to a

small DNP-lys-sepharose column and subjected to step-wise

gradient hapten elutions (beginning with 1 X 10-6 M DNP-

OH) until all radioactivity was eluted.









Analysis of nonspecific radioactivity from supernatants.

Radioactivity from Tm- or non-Tm-treated supernatants was

examined for nonspecific IgM by immunoprecipitating super-

natant fluid which had been freed of specific antibody by

affinity chromatography. Twenty microliters of a rabbit

anti-mouse V chain antiserum was added'to .1 ml of anti-DNP

free supernatant and incubated for one hour on ice. This

was followed by addition of .2 ml of a 50% suspension of

Staphylococcus A (Cowan I strain) and incubated for 15

minutes on ice. This mixture was centrifuged for three

minutes at 15,000 rpm. Pellets were resuspended in 1%

SDS-Tris buffered saline and samples were removed for

scintillation counting.

Analysis of cell lysates. Pelleted cells from Tm-

or non-Tm-treated cultures were resuspended in 4 ml of

.1% NP40 in .04 M Tris-buffered saline with .01 M EDTA,

pH 8.3. Cell debris was removed by centrifugation at

500g for twenty minutes. Supernatants were then freed

of DNP specific antibodies by affinity chromatography.

Both purified antibody and antibody-free supernatants were

assayed by the methods described above.

Results

Functional and Structural Characterization of the IgM
Antibodies

The hybridoma lines utilized here (NP3-17 Cl-20 and

SP2/0 1-64 Cl-12) grow as ascitic tumors in BALB/c mice

and yield moderate amounts (b 1 mg/ml) of 19S IgM antibodies

to the DNP moiety. Equilibrium dialysis studies using the








hapten DNP-eaminocaproate indicate each to contain an
average of five high affinity (Kal07 M- ) binding sites

(Figure 8). Furthermore, % 95% of the reductive 7S

subunits of each of these proteins absorb to and subse-

quently can be hapten eluted from a DNP-lys-sepharose

affinity column. These subunits when examined by equili-

brium dialysis again using DNP-e-aminocaproate indicate

each to contain an average of one high affinity binding

site. Structural features of these two proteins indicate

a considerable degree of homogeneity as might be antici-

pated using non-producing myeloma cell lines as the parent

line for the fusion process. Alkaline-urea gel electro-

phoresis of the isolated L chains from each protein revealed

banding patterns indicating considerably restricted hetero-

geneity (Figure 9). In addition, limited amino terminal

sequence analysis of the H and L chains from the NP3-17

C1-20 protein indicate each to be homogeneous (Table 2).

The H and L chains from SP2/0 1-64 C1-12 IgM were not

amenable to sequence analysis (presumably due to blocked

amino terminii; pyrollidone carboxylic acid).

In experiments designed to measure the binding ability

of either of these 19S molecules after treatment with 5.0 M

guanidine-HC1 for one hour followed by dialysis against

Tris buffer, it was found by equilibrium dialysis that the

high affinity binding sites were totally destroyed; no low

affinity sites could be detected (data not shown).
































Figure 8. Equilibrium dialysis of 19S IgM from hybridoma
SP2/0 1-64 Cl-12 with DNP-E-aminocaproate.
Results for NP3-17 Cl-20 were similar.













20.0 x
18.0
16.0
(? 14.0
12.0 X
>< 10.0
< 8.0
6.0
4.0 -
2.0 x

5.0 10.0
































Figure 9. Alkaline urea polyacrylamide gel electrophoresis
of isolated light chains from (A) SP2/0 1-64
Cl-12 and (B) NP3-17 Cl-20.



















































. .









TABLE 2

Amino Terminal Sequences of Heavy and Light
Chains of IgM Antibody to DNP from a Hybri-
dome Line NP3-17 CL-20


RESIDUE H CHAIN L CHAIN

1 Glu Glu

2 Val Asn

3 Gin Val

4 Leu Leu

5 Gin Thr

6 Gin Gin

7 Ser Ser

8 Gin Pro

9 Pro Ala

10 Glu Ile

11 Leu Met

12 Val Ser









Separation of Halfmers (H-L)

Mild reduction of either of these IgM molecules

resulted in the formation of noncovalently associated

halfmers (H-L) in nondissociating buffers. These mole-

cules gel filtered on G-200 Sephadex as would be expected

of a 7S, 180,000 dalton protein. However on SDS-PAGE

(10% acrylamide) without further reduction these molecules

migrated as ,90,000 dalton proteins, i.e., halfmers. Upon

subsequent extensive reduction in SDS the v90,000 dalton

material dissociated into equimolar %70,000 dalton H chains

and .22,000 dalton L chains. Since these halfmers non-

covalently associated in nondissociating buffers it was felt

that appropriate denaturing or chaotropic agents might be

able to disturb the noncovalent interactions without

destroying the binding site of the "good" halfmers, there-

fore allowing elution and recovery of the "bad" halfmers.

Using a "grab bag" approach, various concentrations of the

following agents were used: guanidine-HC1, sodium thio-

cyanate, acetate buffer (pH 5.0), isopropyl alcohol, NP-40,

and Tris-HCl (pH 8.5). Initially it was observed that two

column volumes of 2.0 M guanidine-HC1 would elute 50% of

the "7S" antibody from the DNP-lys-sepharose column with

the remaining 50% being eluted with .1 M DNP-OH. After

dialysis against Tris buffer the guanidine-eluted fraction

would subsequently bind to the DNP affinity matrix.

Furthermore both the guanidine-eluted and the DNP-eluted

material exhibited one binding site per 2H-2L chain









subunit. The major problem in interpreting this experi-

ment was the observation that all of the bound antibody

could be eluted if more than two column volumes of 2.0 M

guanidine-HC1 was used. Therefore there are two distinctly

different interpretations of the data: 1) It is possible

that the two column volumes of guanidine-HC1 were eluting

preferentially the "bad" halfmers from the affinity column

and that any more than two volumes began to denature and

elute "good" halfrers. This liberal interpretation of the

data implies that if "bad" halfmers were eluted by the

smaller amount of guanidine-HC1, half of them must have

subsequently become "good" haTfmers when dialyzed into Tris

buffer. Proof of this interpretation would substantiate the

theory of conformational differences between opposing

binding sites of a "7S" molecule. 2) The more conservative

interpretation (and more likely) of the data would be that

the guanidine-HC1 elution was not preferential but rather

nonspecifically eluted 50% of the antibody. All but one of

the other dissociating agents tried either eluted all of

the antibody from the column or was not able to elute any

of the antibody. Only recently it was observed that .5 M

Tris-HC1 (pH 8.5) could elute 50% of the "7S" antibody

absorbed to the DNP affinity column. The use of .5 M

Tris-HCl was prompted by an observation which warrants some

discussion at this point. Rather serendipitously, it was

observed that mild reduction in .5 M Tris-HCl of either of

the two 19S molecules at dilute protein concentrations









(<1 mg/ml) yielded subunits of which 50% would bind to a

DNP affinity matrix (active) and 50% would not (inactive)

providing the protein was left undiluted as applied to the

affinity column. If the reduced mixture was first concen-

trated %90% of the antibody would bind to the column.

When examined by equilibrium dialysis the unconcentrated

inactive fraction displayed no binding sites, whereas the

active fraction displayed an average of one site per pre-

sumed 2H-2L chain 7S subunit (Figure 10). Interestingly,

if the inactive fraction was first concentrated to > 1 mg/ml

a small percentage (.3 sites/2H-2L chain subunit) of high

affinity sites could be generated. SDS-acrylamide gel

electrophoresis of each fraction demonstrated that these

molecules were predominantly halfmers (H-L) in dissociating

buffer (Figure 11).

Hapten Binding by Tryptic Fragments

The finding of only one active hapten binding site in

each reductive subunit from the two hybridoma-derived IgM

antibodies prompted studies aimed at separating "good"

(hapten binding) from "bad" (no hapten binding) Fab frag-

ments. The strategy employed for this purpose involved

attempts at preparing such Fab's by trypsinizing reductive

subunits. Originally each reductive molecule was subjected

to trypsinization for various times and aliquots were

analyzed by SDS-PAGE (.Figure 12). It was observed that

although relatively little digestion occurred within the

first two hours, digestion to Fabp's (defined as %50,000





























Figure 10.


Equilibrium dialysis using DNP-e-aminocaproate
of NP3-17 Cl-20 "7S" active (e--4) and inactive
(0--0) fractions resulting from mild reduction
at dilute protein concentrations (< 1 mg/ml).
The unfractionated starting material was iden-
tical to the active fraction. Concentration of
the inactive fraction (0--) resulted in the
generation of some sites.




60











2.0


(D I



.0 1..0







1.0 2.0































Figure 11. 10% polyacrylamide gel electrophoresis (SDS)
of NP3-17 Cl-20 "7S". Lane A, unfractionated;
Lane B, active fraction, and Lane C, inactive
fraction. D, E, and F depict the SDS-PAGE
patterns of samples A, B, and C after extensive
reduction.





62













A C O E F






























Figure 12. 10% polyacrylamide gel electrophoresis of timed
trypsin digestions of SP2/0 1-64 Cl-12 7S
reductive subunits. Length of exposure to
trypsin is designated in hours above each
sample. Panel A depicts unreduced samples;
Panel B depicts the same samples upon reduction
with .1 M 2 ME in 1% SDS in a boiling H20 bath
for three minutes.















A

0 I 2 3 4 5 24















B
0 I 2 3 4 5 24.


H -


A


H-L
FL


qm-









dalton material containing L chains and part of the u

chains) was fairly rapid between 3-5 hours; very little

change was seen over the course of the-subsequent 19 hours.

In addition, no F(ab)2 's were obtained since the inter-H

chain disulfides were cleaved by the mild reduction step

used in preparing the subunits. Large scale digestions

were performed for periods longer than three hours (five

hours for SP2/0 1-64 Cl-12 and 18 hours for NP3-17 Cl-20).

Very little difference, if any, could be detected in their

digestion profiles (G-200 and SDS-PAGE). The results of

these large scale digestions are represented in Figures

13 and 14. Surprisingly, only 30-35% of the total protein

(n50% of the expected yield) was recovered as Fab's when

the digested material was chromatographed on G-200 Sephadex

equilibrated in Tris buffer. The remainder of the material

was seen as small peptides. However, as demonstrated in

Figure 15, the Fab's which were recovered displayed one

high affinity binding site as measured by equilibrium dialy-

sis. An important finding was that those Fab's containing

one site accounted for .90% of the binding sites present

in the untrypsinized starting material. Thus the digested

Fab's seemingly were exclusively of the type that contained

no active sites for binding the DNP ligand.

Asymmetry of Carbohydrate

One possible explanation for heterogeneity of binding

within the pentamer and reductive subunits of these molecules

could be the presence of an asymmetrical attachment of
































Figure 13. Sephadex G-200 elution profile of SP2/0 1-64
Cl-12 Fab's obtained after trypsinization for
five hours. An eighteen hour trypsinization
of NP3-17 C1-20 produced a similar profile.

















7S
IgM
I


.5 4' 4'


.4 -

0
co .3-


.2


.1 -



50

FRACTION NUMBER


(2 ml fractions)


100


FE6
I































Figure 14.


10% polyacrylamide gel electrophoresis (SDS)
of unfractionated (A), and active (C) Fab's
obtained from five hour trypsinization of
SP2/0 1-64 Cl-12; (B) contains inactive, non-
Fab material. D, E, and F depict the SDS-PAGE
patterns of samples A, B, and C after extensive
reduction. SDS-PAGE patterns of NP3-17 C1-20
Fab's were indistinguishable.


. ^ v -











A


D


F































Figure 15. Equilibrium dialysis of Fab fragments obtained
from a five hour trypsinization of SP2/0 1-64
Cl-12 with DNP-E-aminocaproate. Results from
equilibrium dialysis with Fab's from NP3-17
C1-20 were identical.


















1.0


(0



C-,
Is-


1.0

r


2.0









carbohydrate (see 47). This possibility prompted an exper-

iment designed to obtain carbohydrate-free IgM antibodies.

Tunicamycin (Tm), an antibiotic that prevents glycosylation

of glycoproteins by blocking the formation of M-acetyl-

glucosamine-lipid intermediates (48), was used to block

glycosylation of the IgM produced by SP2/0 1-64 Cl-12. It
35
was found that culture supernatants from S-methionine

labeled cells treated for eight hours with Tm contained only

1% (15,000 cpm) of the antibody level (1.5 x 106 cpm) found

in supernatants from non-Tm treated 35S-methionine labeled

cells. Furthermore, the rgM that was purified from the

supernatants of Tm-treated cells appeared to be no differ-

ent from IgM purified from the supernatants of non-Tm

treated cells. This was evident by an autoradiogram of

purified antibody run on an SDS-DATD acrylamide gel (Figure

16) and by sequential hapten elutions of reductive 7S

subunits (Table 3). Radioactive material which was not

specifically purified by an affinity column was shown by

immunoprecipitation to be free of IgM (Table 4). Attempts

were made to isolate carbohydrate-free antibody from cell

lysates of Tm-treated- cells, but unfortunately, none was

detected, It is not known if lower levels of Tm or other

incubation times might be advantageous for the production

of unglycosylated IgM by this cell line.

Anti-idiotypic Antiserum

In terms of assessing the variability of antibody sites,

the utilization of anti-idiotypic antibodies has become a




























Figure 16.


Autoradiogram of specifically purified 35S-
methionine molecules from supernatants or
cell lysates of Tm- or non-Tm-treated SP2/0
1-64 Cl-12 cells.
Lane A and H Purified antibody from non-Tm-
treated cultures;
Lane B and G Purified antibody from Tm-
treated cultures;
Lane C and F Purified antibody from non-Tm-
treated cell lysates;
Lane D and E Purified antibody from Tm-treated-
cell lysates.
Samples E H were reduced with .1 M 2 ME and
boiled in 1% SDS for three minutes.




74








75








U
4*4-


C\j C 00 4O M C 0)0
u r- r- -- oo o C Co
SL- LaU r_ ( 0J r- O r- r- C
0 4 ,-- .C C. C-
0O.= (--- 0O 0 O l/-



> 4- 0C I-- .
C"0 oo Cm4 in o 0 oi (4 -'M in
0 cu c 4- r 0 C 3
-3 4.) 0
*r- to SC CCC C) c I O

an ,: o 4 Q E
4- I 3 0 0 0 0 0 0 00 0 Q C. OC,
>. C ER u C Co LO o CO o u oo
-0 m I- LU C) o o ko 0o m Lnm C) c-
4-o 0- -0m iL r,- 0 .- 4
o.n to E O O U
*.- C 0.. C 0
,n-- 4J tS- O C 0-0
i- <; Q.S-

4- Z V) 4J J "V> 0
0 01 *,- .*- 0 4-
S- f US S^
5. -4 0i 3 .0

CNJ ('3 C C 4- O
Sr- 4- E E c- .-
0 1 3 m3 3 .S-
*.- ,--- r. r- Q.


('4 ('4 C ') ( mM mCn 1 0
Ui LL U >, >. -
- ,4-) C) C) 01 ,- 4.J 4J )4-.-
a -r- 64 tu 00 t o N r N ..
_| :: +- O. r-- 0 O0 r- r- 4 0 & 0 .
C C1-4 CO .- a*- *,- o 0.

< C 0 0 0 Co c- 4- 1-L
(.- 41 4o 4- I -
S0-C E e-c m m U C o n (a o t s-
v/)- 4 4- o- r- 00 CC 0 .
4- 1 S 0 ( 0 M
0C0 -D +n- c- >,
0 0 4- 0 0 0 C 0 0 0 0
co u = E o 4 1.- S- (A 0
o = O .n CI C- C, CC C C ,- 4C) 0 to .0
a "a LS.- ( 4 C CD C C C C C .C 0 ).-
*r-u c 4- C Lij "4 LO O Co Co C 0 Q. > 4-
S- W a a (o 4- r. r,- d- d <- Co ( a) *'- C
go > G) 4-) X I C tiY V4 to- o
QL>I 0 1 -'0 40 0. I I U
E-- E O to ac in in to t
0 O1- .0 S- >0 > 0 >
S- *r- ) 4- *-- *-
*r-5 e 4-- Q. C. I I &04-
S a CI = 0 O' 0- 0- tOU
S- 4- I- o 4- 0 m C I O U
I U C *r- -.
C 4. (0 -- 0O V S-
o to S- E E c co co
z t- S|l |l | .l .i m | U. C E
--- r- 4-
Ln lqt Irt m m 0 C cU
0C I I I I I > -0 T U4.'
I U Co C Co C *r- u r-C to r-
0.. C r- r- .- -- r- r- 4) -o -O 5 Ea in
Z 0 U 4U 'U C
L) ? x x x x< to 0 0 LO 0o u
C ...J -I 4 0
C In r- Ln r- I .- -- C m
cu
1-9

















> -
*,- W1


.,- 00 000 CM
0.- 000


S-

0


4J C


4->S
*r~ *r
.0 0
rC0
=
00
0 0)





cC 4J
C
a0


oE

o )

0 0

0 o
0


0000
N< O z0


- O Ln O-
O o 0 0 Oc
o 0 0 0 0


raE
E
S- -

z
)4-


0
o




o o


I -
E 0













S.-
4.-) U
'-0
H- C

















*M 4
0 1
E-
P- 00
OC

S0.

4|-


0 0 0
CD CD C
CD C C)


0)
r1--




4J
c0







U U


-J 4.


I m


o E
E
0 0
o Uo
o 0u
I 0
0 0


C C












I-I *-(


4-)
*U



4-)
0


0
C-








4-



0.













4->
4-
U)


0
*0
4-)
0






5'





C
0




<4-

*>



C

*0










-o
'4-








*r-
4-








1-


'-4 I-6
'-4


0)


o
r-- -







S- 4-)
0 o











0
-e U





0 H
0 5-



Ua 4-



- C
0. 0






E .-E




4-J
4-3 0
4 4 C4


02


U )
9- 9-4
9-4
0 >


4-)

0


0
s-
43
0O



o-
U)
4--
0



to
E



4-
0


*r-
4-
0
S*-









o
4-
0







4-




U
0











O -r-
0







4-) 4-3

C u
S.-





* r-
4-






00




N0 r



O *
U,
















o E
(-) 4-







0
0









S*- 0
o) C
0

CO
5-






5- U)
3).


0
0






4.-)
a-













4-
*-





0





















rJ
4-.
C



0












0











a
*-
9-











4-
0->
0








o
Uo
0.







0
5-











*0

(U









*4O


s-
.c5-

C-









powerful tool (49). The purpose of developing such an

antiserum in this study was hopefully to allow detection

of idiotypic differences between "good" and "bad" halfmers.

Unfortunately no anti-idiotypic antibodies could be detected

in the ascitic fluids of these animals by Ouchterlony

analysis against the intact 19S molecule. Perhaps a better

approach might be to immunize guinea pigs and absorb this

antiserum with the necessary mouse proteins until desired

idiotypic specificity is achieved.

Affinity Labeling

Affinity labeling of various anti-DNP antibodies has

been successfully utilized to either completely or partially

block the antigen combining site via covalently coupling

antigen in the site (50). This approach was .attebp.ted with

hopes of obtaining "7S" subunits with inactive sites. By

placing these inactive subunits in 2.0 M guanidine-HCI and

subsequent dialysis against Tris-buffered saline the gene-

ration of active sites might be possible. This result

would argue that inactive sites have the capability of

binding to antigen but when paired with an active site

assume the wrong conformation for antigen binding. Unfor-

tunately in this experiment conditions were not appropriate

for affinity labeling as no inactive antibody passed through

the affinity column. It is quite possible that the appro-

priate conditions for affinity labeling could be achieved

and should be pursued in future studies.









Discussion

The purpose of the work described in Chapter III was

to utilize murine hybridomas as sources of homogeneous IgM

antibodies to test the hypothesis that certain IgM anti-

bodies can exhibit intramolecular heterogeneity of ligand

binding that is not attributable to primary structural

differences. The results of polyethylene glycol induced

fusions using non-immunoglobulin producing cell lines

indicated that stable, antigen specific, IgM secreting

cell lines can be obtained. The two IgM antibodies

obtained from the hybridomas displayed an unusually high

affinity (, 107M-1) for the DNP hapten. It is not known

if cells producing lower affinity IgM antibodies (104-105M-)

were present as a result of the fusion process and were

missed by the screening procedure used or if, for one

reason or another, cells producing such antibodies simply

did not exist. Nevertheless, these antibodies did meet the

necessary criteria for testing the original hypothesis in

that 1) they were specifically induced pentameric antibodies

which displayed binding heterogeneity, i.e., only five DNP

binding sites. This heterogeneity appeared to be of an

intramolecular nature as opposed to intermolecular hetero-

geneity based upon the following premises. It seems rather

unlikely, due to the nature of hybridoma derived antibodies,

that the two hybridoma cell lines derived (NP3-17 Cl-20 and

SP2/0 1-64 Cl-12) from the two different myeloma cell lines

used (NP3 and SP2/0) could both manufacture and secrete









in equal amounts two nondistinct populations of IgM anti-

bodies, one of which had ten binding sites for DNP and one

of which bad no binding sites for DNP, Furthermore,

essentially all (> 90%) of the "7S" reductive. subunits

absorbed to (and subsequently Kapten eluted from) a DNP-

lys-sepharose column. These subunits when measured by

equilibrium dialysis displayed an average of one site per

7S subunit. These results clearly support the interpreta-

tion that each 7S subunit must contain only one site, i.e.,

there were none which contained two and likewise there

were none which displayed no DNP-binding sites. 2) These

IgM antibodies also met the criterion of being structurally

homogeneous, at least by the methods employed. Comparison

of alkaline urea gel banding patterns of myeloma protein

L chains with the banding patterns of these hybridoma L

chains demonstrated considerable structural homogeneity

(for example, compare with the banding pattern of MOPC-21

L chain, Figure 4, Chapter II). Likewise limited amino

acid sequence analysis certainly indicated homogeneity

with NP3-17 CI-20. Undoubtedly, complete amino acid

sequences of both proteins or sequence analysis at least

through the first hypervariable region would be invaluable

in this regard. However, the question becomes more perti-

nent when putative inactive sites are separated from active

sites. Perhaps at that stage additional sequence data, as

well as peptide mapping, isoelectric focusing, and/or idio-

typic analysis would be beneficial.









The results from mild reduction of both IgM antibodies

indicated that the 19S molecules proceeded directly from

the 19S covalent pentameric form to covalent halfmers (H-L).

For example Figure 7 demonstrates that a concentration of

.001 M 2ME does relatively little to change the covalent

structure of the molecule whereas a concentration of .005

M 2ME has converted all of the heavy molecular weight

molecules predominantly to halfmers. The obvious inter-

pretation of this finding is that for these molecules the

inter-heavy chain disulfides are as susceptible to reduction

as are the inter-subunit disulfides. This fact could

become important in uncovering a possible mechanism for

intramolecular heterogeneity. According to Askonas and

Parkhouse (51, p. 632) the intracellular assembly pathway

for IgM "mirrors the resistance of interchain disulphide

bonds to reduction." If this is true with these hybridoma-

derived molecules, it would appear that these molecules are

assembled through a H-L chain intermediate based upon their

reduction profiles. If assembly of IgM is involved in the

observed heterogeneity, comparisons of reduction profiles

of these IgM antibodies with reduction profiles of other

IgM antibodies which possess ten homogeneous binding sites

might bring to light an assembly intermediate (for example

H-H-L, see 40) responsible for the phenomenon.

Examination of the results from experiments designed to

separate noncovalently associated halfmers yielded few clear

interpretations. This was in part due to the fact that the









amount of dissociating agent used (two column volumes of

2.0 M guanidine-HC1) was critical in removing 50% of the

antibody absorbed on the affinity column. Any more than

this amount or subsequent elutions with the same amount

eluted more antibody from the column. As pointed out

previously the most likely interpretation of these results

would be that the guanidine-HC1 elution was not preferen-

tial but rather nonspecifically eluted 50% of the antibody.

Likewise, results obtained when .5 M Tris-HCl was used to

dissociate the halfmers were equally difficult to interpret

due to the problems associated with total recovery of hapten

binding sites. Neither the 50% fraction which was eluted

from the affinity column by .5 M Tris-HCl nor the 50%

fraction which passed through the DNP-immunoabsorbant when

the mild reduction was performed on dilute solutions of

antibody, displayed any active binding sites by equilibrium

dialysis. The hapten eluted fractions (the remaining 50%)

in both cases displayed only one active site per 2H-2L

chain subunit, thus accounting for only one half of the

original sites in the unfractionated subunit population.

Although equilibrium constants were not measured for the

noncovalent association of halfmers used in this study,

one published report of halfmer association using three

other IgM monoclonal antibodies gives an equilibrium constant

of 2.3 X 106 moles halfmers- for the reaction one halfmer

I two halfmers (52). It is quite possible that equili-

brium constants of this magnitude or higher may exist for








the hybridoma-derived IgM antibodies described in this

study. If so, separation of the putative "good and bad

halfmers" using chaotropic or dissociating agents may be

impossible, as these agents may equally disrupt hapten

binding and halfmer association. Unlike the previous

study mentioned on halfmer association Parkhouse .(53, p.640)

in describing halfmer association with IgM from the mouse

myeloma MOPC 104E states:

Since, therefore, there is no pronounced
tendency for non-covalent interaction between
HL subunits it is no surprise to find that
the equilibrium for the reaction 2HL-IgMs
(7S) is not completely in favour of IgMs.

Quite possibly, intramolecular heterogeneity may arise not

as a result of the mode of assembly of the IgM molecule

but rather by the conformation that must be assumed by one

halfmer as it associates with another halfmer for which

it possesses strong noncovalent attractions, Likewise,

those molecules lacking strong noncovalent attractions can

readily form two functional binding sites per 2H-2L chain

subunit. Future experimentation in this area could provide

some rewarding information about intramolecular hetero-

geneity.

The most exciting observation concerning the possibility

of conformational differences between opposing binding sites

of the same 7S subunit was seen when tryptic fragments were

obtained from both IgM antibodies, As demonstrated in

Figure 13 only 50% of the expected yield of Fabp's was

obtained by gel filtration on G-200 Sephadex. Approx-








imately 90% of these Faby's absorbed to and were subse-

quently hapten eluted from a DNP-lys-sepharose affinity

column. Each of these Fab's contained one site as

measured by equilibrium dialysis (Figure 15) and accounted

for 90% of the binding sites present in the untrypsinized

starting material. The obvious conclusion from these

results is that the Fab's which were digested and therefore

not recoverable were exclusively of the type that contained

no active sites for binding the DNP ligand, This finding

is best explained by the presence of two populations of

conformationally distinct Fab's. It is hoped through future

experimentation with trypsin, or one of the other available

proteolytic enzymes, that conditions which will allow for

near 100% recovery of the Fab's can be ascertained.

Accomplishment of such a task will allow for careful invest-

igation at both the structural and functional levels of each

population of Fab's.

The finding of trypsin-sensitive Fab's might have some

connection with an observation made while attempting to

separate the two putative (good and bad) halfmers. The

50% fraction which passed through- the affinity column after

reduction of the IgM in dilute solution was found to be

extremely labile. Immediate analysis of this fraction

after affinity chromatography demonstrated the fraction was

predominantly (>9.0%) halfmers (Figure 11). However,

storage at 40C overnight or freezing and thawing of this

fraction produced a protein solution of which only a small








percentage was able to be concentrated using negative

pressure dialysis. The remainder of material passed

through the dialysis bag (MW exclusion of 12,000-15,000)'

Similar treatment of the 50% hapten eluted fraction had

little effect. It is conceivable that this labile halfmer

fraction, like the trypsin-sensitive Fab's, had obtained

a conformation which when purified in dilute solution was

extremely sensitive to proteolysis.

Recent studies have shown that the structure of the

carbohydrate on the H chain of certain antibodies may

influence the strength of the interaction of those anti-

bodies with polymeric antigen (54). Studies suggesting

the importance of the protein structure in determining

glycosylation have also been reported (55,56). These

studies in conjunction with one report of asymmetrical

attachment of carbohydrate to rabbit IgG prompted the

experiments in this study with the antibiotic tunicamycin.

One possible explanation of two distinct conformations for

a single protein could be the presence or absence of carbo-

hydrate. Unfortunately, as predicted by previous studies

with Tm (57), secretion of IgM from these cell lines was

dramatically decreased (, 99%) in the presence of Tm.

Likewise under these experimental conditions no intra-

cellular IgM from Tm-treated cells could be isolated,

Perhaps future endeavors with this antibiotic may determine

a suitable concentration and incubation period which would





85



allow for secretion of unglycosylated IgM. One possible

alternative to this approach might be mechanical stripping

of purified IgM.














CHAPTER IV
SUMMARY AND CONCLUDING REMARKS

The objective of this research was to clearly resolve

the question of intramolecular heterogeneity of ligand

binding by homogeneous IgM antibodies. This has been an

area of controversy for over a decade with relatively few

advancements being made in the last five years. By

utilizing hybridoma-derived IgM antibodies many of the

problems associated with this unresolved question have

been alleviated. This work also attempted to determine if

intramolecular ligand binding differences are due to

primary structural differences or conformational differences.

Chapter II described a somewhat unusual IgM molecule,

14PAF. This 19S IgM antibody exhibited an average of six

homogeneous relatively high affinity binding sites per

molecule for the DNP moiety. The failure to demonstrate

ten binding sites per pentameric molecule was attributable

to the presence of two different light (L) chains in the

secreted molecules. One of the L chains appeared, based

upon limited amino acid sequence studies, to be the L

chains normally found in MOPC-21, the myeloma protein

secreted by the P3 myeloma line. The other L chains,

presumably derived from the murine spleen cell used in the

fusion, were essential for anti-DNP combining sites.

Analysis of the reductive subunits of the secreted IgM


-86-









anti-DNP molecules indicated the presence of two types.

One type (designated active) absorbed to DNP affinity

matrices, contained an average of two homogeneous ligand

binding sites per subunit and did not contain MOPC-21

L chains. The other type of subunit (designated inactive)

did not contain any DNP binding sites and contained only

MOPC-21 L chains.

Polypeptide chain recombination studies with mildly

reduced and alkylated u and L chains from each type of

reductive subunit indicated that the u chains were function-

ally identical. Furthermore these in vitro recombination

studies indicated that the noncovalent assembly of the

2p-2L chain subunits was restricted by the L chains in

such a way that the recombinant subunits were homogeneous

in terms of L chains. Equilibrium dialysis studies with

active homogeneous recombinant subunits indicated the

presence of but one ligand binding site per,2V-2L chain

subunit.

Chapter III described the structural and functional

analysis of two additional murine anti-DNP IgM hybridoma

antibodies, NP3-17 Cl-20 and SP2/0 1-64 Cl-12. Each of

these 19S IgM molecules displayed an average of five high

affinity binding sites for the DNP moiety. Furthermore,

> 95% of the reductive subunits of each of these proteins

absorbed to and were hapten eluted from a DNP-lys-sepharose

affinity column. These subunits, when examined by equili-

brium dialysis using DNP-E-aminocaproate, indicate each to








contain an average of one high affinity binding site,

Structural analysis of each molecule indicated each to be

homogeneous. The most surprising finding with these

molecules was the ability to recover only 50% of Fab's

following trypsin hydrolysis. Interestingly, the 50%

that were recovered accounted for all of the binding sites

present in the starting 7S material. This finding, as

well as limited success with the separation of inactive

reductive halfmers (H-L) which subsequently became active,

indicate that differences in conformation of the binding

sites may account for the observed binding heterogeneity.

Although the study with 14PAF (Chapter II) did not directly

answer any questions regarding intramolecular heterogeneity,

the results of the chain recombination studies (one ligand

binding site per 2H-2L chain recombinant subunit) closely

mimic the natural situation that occurs with the two mole-

cules discussed in Chapter III, i.e. one binding site per

2H-2L chain subunit. These findings strongly implicate

the mechanism of assembly as a possible factor in the

generation of conformational differences that could account

for the intramolecular heterogeneity of ligand binding seen

with certain IgM molecules.

It is hoped that these studies have laid a foundation

from which important questions regarding the structure,

function, and intracellular assembly of IgM antibodies may

be answered, as well as uncovering other potentially reward-

ing areas of research. Optimistically, extensions of these





89



studies may result in a wide variety of new approaches

to probing and understanding B cell activation and

regulation.









REFERENCES


1. Metzger, H. (1970) Structure and function of M macro-
globulins. Adv. Immunol. 12:57-116.

2. Ashman, R., and Metzger, H. (1969) A Waldenstrom macro-
globulin which binds nitrophenyl ligands. J. Biol.
Chem. 244:3405-3414.

3. Young, N., Jacius, I.,and Leon, M. (1971) Binding
properties of a mouse immunoglobulin M myeloma
protein with carbohydrate specificity. Biochem.
10:3457-3460.

4. Edberg, S., Bronson, P., and Van Oss, C. (1972) The
valency of IgM and IgG rabbit anti-dextran anti-
body as a function of the size of the dextran
molecule. Immunochem. 9:273-288.

5. Stone, M.,and Metzger, H. (1968) Binding properties of
a Waldenstrom macroglobulin antibody. J. Biol.
Chem. 243:5977-5984.

6. Onoue, K., Grossberg, A., Yagi, Y.,and Pressman, D.
(1968) Immunoglobulin M antibodies with ten
combining sites. Science 162:574-576.

7. Oriol, R., Binaghi, R.,and Coltorti, E. (1971) Valence
and association constant of rat macroglobulin
antibody. J. Immunol. 104 (4):932-937.

8. Voss, E. and Sigel, M. (1972.) Valence and temporal
change in affinity of purified 7S and 18S nurse
shark anti-2,4 dinitrophenyl antibodies. J.
Immunol. 109 (4):665-673.

9. Oriol, R.,and Rosset, M. (1974) The IgM antibody site.
II. Comparison of the two populations of IgM
antibody sites and analysis of the behavior of
the low affinity population at equilibrium with
a dinitrophenylated hapten. J. Immunol. 112 (6):
2235-2240.

10. Frank, M.M.,and Humphrey, J.H. (1968) The subunits in
rabbit anti-Forssman IgM antibody. J. Exptl. Med.
127:967-982.

11. Kishimoto, T.,and Onoue, K. (1971) Properties of the
combining sites of rabbit IgM antibodies. J.
Immunol. 106:341-348.


-90-




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E1VMA2EI3_GJBNKW INGEST_TIME 2012-02-29T16:35:22Z PACKAGE AA00009111_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES