607th Meeting
Held at the Polytechnic of Central London on 19-2 1 December 1983
Effector Functions of the Constant Region of
ImmunoglobuIin Molecules
Biochemical immunology Group Colloquium Organized and edited by
C. R. Stokes (Bristol)
Isotype switching in myeloma and hybridoma cells compared with LPS blasts
A. RADBRUCH, F. SABLITZKY and
K. RAJEWSKY
Insritutfcr Genet&, Uniuersitiit zu Kcln, Wyertal 121, 5000
Koln 41. Federal Republic of’Germany
We have shown that spontaneous isotype switch variants
occur at low frequency in myeloma and hybridoma cell
lines. These cells have deleted the wild-type C H gene from
the active chromosome. The recombination mostly happens
outside the ‘switch-regions’. LPS-stimulated B cells, however, switch isotype frequently and delete C p from both
chromosomes.
Antigenicity and effector functions of human immunoglobulin G1 proteins having Cy2 or
Cy3 deletions
ROYSTON JEFFERIS, MOHD I. NIK JAAFAR,
RUTH McADAM and MATTHEW WALKER
Department of Immunology, The Medical School, Vincent
Drive, Birmingham BI5 2TJ. U . K .
Equivocal and often contradictory results have been
obtained using proteolytic fragments and sub-fragments of
human IgG to localize sites controlling effector functions.
In contrast hinge-deleted paraproteins have made a significant contribution to our understanding of structure-function relationships and allow the C l q binding site to be
localized to the N-terminal region of the Cy2 domain (Klein
et al., 1981). Similar studies employing a series of variants
resulting from genetic or biosynthetic abnormalities may
allow sites controlling other effector functions to be precisely mapped.
Recently we have isolated two paraproteins, TIM and
SIZ, having extensive or complete deletions of the Cy2 and
Cy3 domains respectively (Nik Jaafar et al., 1983, 1984).
Thus protein TIM reacted with McAb-reactive pFc’ but not
with anti-Fc McAb that were unreactive with pFc’. In contrast protein SIZ reacted with some anti-Fc McAb that
were not reactive with pFc’. A third group of anti-Fc McAb
were identified that did not react with either TIM or SIZ
proteins and are assumed to be directed against epitopes
dependent on the quaternary structure of the intact Fc
region. Proteins TIM and SIZ have been used together with
the McAb panel to probe effector sites on IgG molecules.
Proteins TIM and SIZ are of the IgG1 subclass as
demonstrated by reactivity with monoclonal and polyclonal
antibodies. However, neither protein bound SPA. This is
Abbreviations used : Ig, immunoglobulin; McAb, monoclonal
antibodies.
VOl. 12
consistent with X-ray crystallographic studies which
demonstrate that contact residues within both Cy2 and Cy3
domains are necessary for SpA binding (Marquart &
Deisenhofer, 1982). Histidine-435 located within the interdomain region is essential to SpA binding and IgG3
proteins having arginine at this position d o not bind SPA. It
is interesting to note that ‘general’ rheumatoid factor recognizes the G a antigen which is also expressed in IgG 1 , 2 and
4 but not IgG3 proteins. Similarly, a number of murine
McAbs were produced that had a non-IgG3 specificity. The
possibility that these McAbs overlap in molecular specificity with human rheumatoid factor and the SpA binding
site was investigated in parallel studies with a monoclon.tl
human IgM rheumatoid factor (RF-AN) secreted by an
Epstein-Barr-virus-transformed lymphocyte cell line (Jefferis et al., 1984). Neither protein TIM or SIZ expressed the
epitope recognized by RF-AN and the non-IgG3 murine
McAb suggesting that the quaternary structure of the intact
Fc is essential for reactivity with these antibodies. Studies
employing an IgG3( 15,16) protein of non-Caucasian origin
allowed further structural definition of the epitope recognized. Caucasian IgG3 proteins are characterized by the
presence of arginine at position 435. However, a common
mongoloid allotype IgG3m( 15,16) has been shown to have
histidine at residue 435. These proteins bind to SpA (Van
Loghem et al., 1982). The rheumatoid factor RF-AN and
the murine non-IgG3 McAb all were reactive to high titre
with IgG3m(15,16) protein GOE. Thus histidine-435 is
critical to the expression of a conformational epitope
localized to the interdomain region of IgG molecules.
Neither protein TIM or SIZ inhibited polyclonal IgG
binding to monocytes (Woof et al., 1984), suggesting that
the effector site recognized by Fc receptor is contributed to
by both Cy2 and Cy3 domains either directly or through the
137
738
BIOCHEMICAL SOCIETY TRANSACTIONS
influence of one domain on the conformation of the other. It
may be anticipated that the conformationally ‘soft’ Cy2 may
be dependent on the influence of the Cy3 in maintaining
native conformation. The Cy3 domain is known to be
conformationally more stable. We have demonstrated that
fully reduced and alkylated Fc can be renatured with the
expression of epitopes localized to the Cy3 domain whilst
Cy2 epitopes were not expressed.
Jefferis, R., Nik Jaafar, M. 1. & Steinitz, M. (1984) Immunol Lett.
in the press
Klein, M., Haeffner-Cavaillon, N., Isenman, D. E., Rivat, C.,
Navia, M. A,, Davies, D. R. & Dorrington, K. J. (1981) Proc.
Natl. Acad. Sci. U.S.A. 18, 524-528
Marquart, M. C Deisenhofer, J. (1982) Immunol. Toduy3, 160-166
Nik Jaafar, M. I., Lowe, J. A., Ling, N. R. & Jefferis, R. (1983)
Mol. Immunol. 20, 679-686
Nik Jaafar, M. I., Lowe, J. A., Ling, N. R. & Jefferis, R. (1984)
Mol. Immunol.21, 137-143
Van Loghem, E., Frangione, B., Becht, B. & Franklin, E.C. (1982)
Scand. J . Immunol. 15, 275-278
Woof, J. M., Nik Jaafar, M.I., Jefferis, R. & Burton, D. (1984)
Mol. Immunol. 21, 523-527
C1 activation by immunoglobulin and immunoglobulin antibodies
NEVIN C. HUGHES-JONES,
ARNOLD FEINSTEIN, NEIL E. RICHARDSON,
BARBARA D. GORICK and
JONATHAN C. HOWARD
MRC Mechanisms in Tumour Immunity Unit,MRC Centre,
Cambridge CB2 4AT. U.K., and ARC Institute of Animal
Physiology, Babraham. Cambridge CB2 4AT, U.K.
C1, IgG and IgM molecules all co-exist within the blood
plasma and yet no activation of C1 takes place. What then
are the changes that take place in the IgG and IgM molecules following their combination with antigen that allows
the C1 molecule to be bound and why does this binding result in activation? Although C1 can be bound and activated
by surfaces without the participation of antibody, it seems
probable that C1 and Ig have evolved in a co-operative
manner so that the requirements for C1 activation are met
by the structural organization of Ig. Thus, in order to discuss
effector mechanisms of Ig molecules, it is necessary to delineate the requirements of the C1 molecule.
It is now well recognized that the binding of C1 to
immune complexes and the subsequent activation of the
molecule are two separate processes and activation does not
inevitably follow binding. The C l molecule is a composite
molecule consisting of two subcomponents: (1) a linear
tetramer containing two C l r and two Cl s molecules, which
are serine esterases in the zymogen form, and (2) the C l q
moiety, which has no enzymic activity and is essentially a
mechanical tool for binding the zymogens to the immune
complexes. The C l q molecule has six binding sites for complexes, one site on each of its ‘heads’, and the essential
requirement for stable binding is that at least two of the
heads must be involved. Binding by a single bond is so weak
that it lasts for less than a second (approx. K =
5 x l O 4 ~ - ’ ) ,whereas divalent binding lasts for several
minutes on average (approx. K = 5 x ~O’M-’) and the
dissociation rate for three-headed binding is measured in
hours (approx. K = 1 x ~O’OM-I)(Hughes-Jones, 1977). C l r
and C 1s exist as single polypeptides of about 85000 M , and
on activation are split into two unequal chains, the smaller
chain acquiring the active enzyme site. The C l r molecule is
split first and then acts on CIS. What is unknown at the
present time is the mechanism by which C l r is split following Clq binding. Two suggestions have been made. First,
that there is an allosteric change in Clq which leads to an
allosteric change in a C 1r molecule and that this results in
the appearance of an active enzyme site. Secondly, it is suggested that a separate binding site on the immune complex
for the ClrCls tetramer is essential and that the binding of
the tetramer to both C l q and the complex allows for autoAbbreviation used: Ig, immunoglobulin.
activation of the enzyme (Goers et al., 1977; Hughes-Jones
& Gorick, 1982).
Turning now to IgM, we do know that this is a molecule
that can exist in two forms. Pentameric IgM in solution has
a central (Fc), disc from which F(ab)z arms extend radially,
the disc and the arms being in the same plane. The (Fc),
disc contains the Cp2 domain (which is the equivalent of
the hinge region in IgG) together with the Cp3 and Cp4
domains (Feinstein & Beale, 1977). When forming complexes in antigen excess, most of the IgM molecules retain
this planar form, whereas complexes in antibody excess
have IgM mainly in a ‘staple’ form with the F(ab)z domains
bent at right angles to the central ( F c ) ~disc. We have looked
at C1 binding and activation by complexes in both antigen
and antibody excess. The antigen used was dextran particles
and the antibody, the mouse myeloma MOPC 104E. C l q
binds well to antibody present in the staple form, almost all
the IgM molecules being able to form stable binding sites
with a K of about 5 x IO’M-’, indicating binding by two
heads. Moreover, activation of C1 is rapid, being complete
in about 15min.
On the other hand, in antigen excess, with IgM mainly in
the planar form, Clq binding occurs but only about 10% of
the IgM molecules have binding sites. The binding constant
for Clq-IgM interaction is also of the order of 5 x lo7M - I ,
indicating binding by two heads, but it is not clear whether
the binding that does take place is to a few molecules which
may be in the staple form or whether binding can take place
with some IgM molecules when still in the planar form.
Although C1 binds under these circumstances, it does not
activate, showing that whatever the conditions are that lead
to activation, they are not present in antigen excess. These
findings indicate that sites for C 1q binding become available when the IgM molecule acquires the staple shape and
also that the correct conditions are then present for C1
activation.
The next question is, what is the mechanism for the
appearance of Clq binding sites when IgM assumes the
staple form? In order to answer this, we have looked at the
binding of Clq to the intact IgM pentamer and to various
fragments of IgM. The binding constant for the Clqpentamer IgM reaction is of the order of 5 x lo5M - I , a relatively weak reaction indicating that only one binding site is
available. A similar value is obtained for C 1q binding to the
pentameric (Fc), disc (Cp3 and Cp4 domains). The
similarity between these K values indicates that no additional C 1q binding sites become exposed by removal of the
F(ab), arms. The F(ab’)z fragments have no sites for C l q
binding. On the other hand, 7s subunits containing all the
variable and constant domains are able to bind Clq. These
findings indicate (1) that the 7s IgM subunits each have a
pre-existing binding site, (2) that in the assembly of the
1984