| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Molecular Biology and Pharmacology (A.M.J.,
M.P., I.B.) Washington University School of Medicine St
Louis, Missouri 63110
Institute for Biomedical Aging Research
(P.B.) Austrian Academy of Sciences A-6020 Innsbruck,
Austria
Division of Reproductive Biology (C.K., A.W.J.H.)
Department of Gynecology/Obstetrics Stanford University Medical
Center Stanford, California 94305
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
-subunit noncovalently bound to a hormone-specific
ß-subunit. The two subunits are highly intertwined and only the
heterodimer is functional, implying that the quaternary structure is
critical for biological activity. To assess the dependence of the
bioactivity of hCG on the heterodimeric interactions, - and
ß-subunits bearing mutations that prevent assembly were covalently
linked to form a single chain hCG. Receptor binding and signal
transduction of these analogs were tested and their structural
integrity analyzed using a panel of monoclonal antibodies (mAbs). These
included dimer-specific mAbs, which react with at least four different
epitope sites on the hormone, and some that react only with the free
ß-subunit. We showed that there was significant loss of quaternary
and tertiary structure in several regions of the molecule. This was
most pronounced in single chains that had one of the disulfide bonds of
the cystine knot disrupted in either the - or ß-subunit. Despite
these structural changes, the in vitro receptor binding and
signal transduction of the single chain analogs were comparable to
those of the nonmutated single chain, demonstrating that not all of the
quaternary configuration of the hormone is necessary for biological
activity. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
-subunit and a hormone-specific
ß-subunit. The crystal structure of hCG (2, 3) showed that the two
subunits are intimately associated with each other along much of their
surfaces, each subunit having remarkably similar folds with two hairpin
loops at one end and a single loop at the other. The two loops of one
subunit are adjacent to the single loop of the other. At the core of
the hCG - and ß-subunits is a cystine knot motif previously
observed in several growth factors (4). The carboxy-terminal region in
the ß-subunit forms a ‘seat belt’ around the -subunit that
stabilizes the native heterodimer. This structural configuration,
together with the heterodimeric requirement for receptor binding,
implies that the complex quaternary structure of the gonadotropins is
crucial for biological activity. Recently a number of observations
using single-chain glycoprotein hormone variants, where the - and
ß-subunits are covalently linked, have suggested that the quaternary
configuration is not a prerequisite for receptor binding. Single-chain
mutants of hFSH and hCG displayed normal receptor binding and signal
transduction, despite modeling that suggested that some of the -ß
subunit alignments seen in the crystal structure could not be
maintained (5). For example, molecular modeling of tethers lacking a
linker sequence reveal that the alignment of the /ß domains in the
single chains differ substantially from that seen in the heterodimer.
Moreover, CG single-chain mutants bearing deleted disulfide bonds in
the cystine knot of either the - or ß-subunits were biologically
active (6, 7). These data suggested that the extensive interactions
holding the subunits together in the wild type was not essential for
receptor recognition. However, the extent of the structural loss in
these single-chain variants was not known, and thus a clear
relationship between the heterodimeric configuration and biological
activity could not be assessed from the earlier work.
Here we sought to directly examine the importance of the integrity of
the quaternary structure for the biological function of hCG. We
introduced mutations into subunit regions that are known to be critical
for assembly and secretion of the heterodimers (8, 9, 10). Those mutations
that prevented subunit association were then generated by site-directed
mutagenesis in the - or ß- domains of single-chain hCG. This
approach permits analysis of receptor binding properties of residues
critical for assembly that were previously not possible due to the
inability of mutated subunits to form heterodimer. These analogs were
examined for structural changes and activity on the assumption that if
they are biologically active and the /ß domains not associated in
a native configuration, it would indicate that the precise wild-type
conformation is not essential for maximal hormone function. The
structure of these single-chain analogs, which include mutations in the
-helix of the - subunit and disulfide bond mutations in the -
and ß-subunits, was assessed using a panel of conformationally
sensitive monoclonal antibodies (mAbs). This panel included eight mAbs
with specificity for only the intact heterodimer and two that react
only with free ß-subunit (11, 12, 13). We show that the single-chain
analogs were disrupted in their heterodimeric configuration and yet
biologically active. These results are consistent with the hypothesis
that the quaternary configuration of the /ß heterodimer is
critical for secretion and intracellular stability but not for receptor
binding and signal transduction.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
-Helix of the -Subunit-helix on the -subunit (amino acids 38–42;
Fig. 1
) has been implicated in the
assembly (reaction) (8, 10, 14, 15, 16, 17, 18). Previously we showed that
substitution of Thr 39 with Phe (T39F) or Ala (T39A) eliminated
or reduced /ß heterodimer formation (8). To examine the role of
the adjacent residues, an additional triple mutant was constructed,
39TPL41 changed to 39AAA41. This mutant, together with T39F or
T39A variants, was cotransfected with the hCGß gene in Chinese
hamster ovary (CHO) cells, and several clones expressing either the
same amount of subunits or excess ß were isolated (Fig. 2). Assembly was determined by
immunoprecipitating labeled subunits in the media with either - and
ß-antisera that recognize noncombined subunits and the corresponding
heterodimers. Because the level of ß-subunit synthesized is the same
or in excess, the amount of -subunit precipitated by both antisera
should be the same if complete heterodimer formation has occurred (8).
In the case of cells expressing the wild-type heterodimer, both the
- and ß-antisera coprecipitated a comparable level of -subunit
(lanes 1 and 2). This implies that combination was quantitative. The
two bands on the gels corresponding to the ß-subunit contain one or
two asn-linked oligosaccharides (19). The assembly of all mutants was
reduced although the extent varied, e.g. for the T39A
mutant, the efficiency of dimer formation was 10–30% (lanes 3 and 4),
whereas for the T39F (lanes 5 and 6) and the 39AAA41 (lanes 7 and
8) mutants, less than 5% of total -subunit formed heterodimer. This
is reflected by a parallel increase in the proportion of the free form
of the -subunit, which is more heterogeneous and migrates slower on
SDS-PAGE than the dimer form (20) (compare lane 1 with lanes 3, 5, 7,
and 9). This heterogeneity is due to hyperglycosylation of the
asn-linked carbohydrates in the noncombined -subunit (20 21B ). The
effect of these mutations is apparently reversible since assembly of
39AAA41 is significantly increased in clones expressing more ß-
than -subunit (compare lanes 7 and 8 with lanes 9 and 10; see
Discussion). That the triple alanine substitution
dramatically reduces assembly is consistent with previous studies (10, 15), which found that determinants for assembly are not restricted to
the Thr 39 but also include neighboring residues.
|
|
39–41 sequence in
heterodimer recognition for the receptor (14, 18). Assessing the role
of the 39–41 sequence on ligand binding by site-directed
mutagenesis is difficult due to the inhibition of assembly when this
region is altered. To by-pass the assembly step the above mutations
were constructed in the -domain of the single chain. These mutants
were secreted efficiently, and their receptor binding affinity and
signaling were tested on CHO cells expressing the human LH/CG receptor
(Table 1). Both the ßT39A and
triple-alanine mutant bound the receptor with high affinity. However,
ßT39F showed a 42-fold reduction compared with the unmutated
single chain (ß) (Table 1). Similar results were obtained when the
mutants were tested with the human kidney 293 line expressing the LH/CG
receptor (data not shown). To assess whether there were changes in
signal transduction, cAMP production was determined. While signal
activation generally paralleled binding affinity (Table 1), ßT39F
exhibited a 9-fold reduction in signaling compared with the 42-fold
decrease in binding affinity. This implies some uncoupling of the two
events. The data show that threonine 39 in the -subunit is not only
associated with the assembly step but also involved in receptor
binding.
|
T39A and ß39AAA41 mutants are
biologically active, it is not known whether the mutated - and
unmodified ß-domains form a heterodimeric-like interaction. To
examine this point, the single-chain analogs were analyzed with a panel
of mAbs that recognize different epitopes sensitive to conformational
changes. The mAbs used include eight dimer-specific mAbs (epitopes
c1–c4; see Materials and Methods for definition of epitope
regions) and two that bind only free ß-subunit (epitopes ß6 and
ß7). If the /ß heterodimeric interaction is disrupted, the
dimer-specific mAbs will not recognize the single-chain mutants because
the dimer-specific epitope is lost. However, if the domains are in a
noncombined conformation, the free subunit- specific mAbs should detect
them. Nonreducing Western blotting in the absence of heat denaturation
showed that six of the eight dimer-specific mAbs, 10, 40, 53, 55, A407,
and B109, reacted with all of the single-chain mutants. Figure 3 shows representative blots for mAbs 40
(panel A), A407 (panel B), and 53 (panel C). In each blot, lanes 1 and
5 correspond to unmutated single-chain and hCG heterodimer,
respectively. These dimer-specific mAbs gave a positive reaction with
all the -single-chain mutants (molecular mass = 47–50 kDa).
These results show that while these -mutants cannot assemble
efficiently to form heterodimer, when they are incorporated into the
single chain, some heterodimer-like epitopes are generated. It is
curious that when the mutants are probed with mAbs 407 and 53,
heterogeneity between 47 and 60 kDa is observed. These forms may be due
to altered conformation and/or changes in oligosaccharide processing
caused by the mutations. The mAbs also detected high molecular mass
proteins, the identities of which are not clear, but aggregates of
misfolded hCG variants and complexes formed during purification of
urinary hCG have been observed (21A, 22). In previous studies of a
series of tethered gonadotropin variants, there was no correlation
between the proportion of aggregates formed by a particular mutant and
the in vitro biological activity (6, 7). Thus, it is
unlikely that the aggregates are contributing significantly to the
overall potency of the tethered hormones. Further resolution will
require purification and direct examination of their biological
activity.
|
) does not react with the single chain
bearing either the ßT39F (lane 2) or ß39AAA41 (lane 3)
mutation and only partially with ßT39A variant (lane 4). This
suggests that the mutations in the region 39–41 disrupted the mAb
26 epitope, paralleling the extent of the mutated subunit to form
heterodimer (see Fig. 2
). Another dimer-specific mAb, designated 45,
displayed weak reactivity with all variants including the unmutated
single chain (Fig. 4B). As described in Materials and
Methods, this mAb binds to residues in the cystine knot region.
The result implies that the single chains have lost some heterodimeric
contacts at the mAb 45 epitope (see below). That the low
immunoreactivity of the mutants to mAbs 26 and 45 was not due to
denaturation during blotting was demonstrated by enzyme-linked
immunosorbent assays (ELISAs) (Fig. 4C and Table 2A). The ßT39A mutant had 14%
cross-reactivity with mAb 26 compared with the unmutated single chain,
and the ßT39F and ß39AAA41 variants were not detected (Table 2A). No ELISA signal was detected with dimer-specific mAb 45 (Fig. 4D
and Table 2A). As a further control we also tested the reactivity of
mAb 26 in an ELISA containing no Tween 20 to assess whether the
detergent caused denaturation (Fig. 4C, inset). The absence
of Tween did not significantly alter the ratio of immunoreactivity of
the mutants compared with reaction mixtures containing detergent.
Further ELISA experiments performed in the absence of Tween with mAbs
55, 64, and 68 did not affect the ratio of analog immunoreactivity
(data not shown). As discussed in Materials and Methods, the
amount of each -mutein added to the ELISA reaction mixture was
determined by a ß-subunit-specific mAb. Although the ß-domain is
not mutated, such quantitation may be subject to overall conformational
changes induced by the -mutations. To address this point, the mutein
concentrations were also determined for the ELISA by an RIA protocol
containing polyclonal antiserum against the ß-subunit. We reasoned
that such a quantitation method would be less susceptible to major
conformational changes caused by the mutations. Using this assay the
immunoreactivity of the mutants was comparable to the mAb-based
quantitation (Table 2B). Taken together, these data support the
blotting experiments that the heterodimeric interaction at epitopes
corresponding to mAb 26 and 45 were disrupted in all of the
single-chain variants.
|
|
/ß heterodimeric interaction is
disrupted in the mutants, they should be more accessible to mAbs that
are specific for the free ß-subunit. To test this prediction, two
mAbs, designated 64 and 68, which recognize epitopes in different
regions of only the free ß-subunit (11, 23) were tested with the
single-chain variants (Fig. 5). As
expected, both mAbs did not detect hCG heterodimer (lane 5), whereas
the noncombined ß-subunit in the CR127 preparation was detected (lane
5, Mr = 30 kDa). When these mAbs were tested against the unmutated
single chain (lane 1), weak activity was observed, suggesting that some
of the - and ß-contact sites in the single chain do not form a
tight heterodimeric complex. This is consistent with data obtained with
dimer-specific mAb 45 above, which did not detect the unmutated single
chain. The mAbs 64 (panel A) and 68 (panel B) reacted strongly to all
three of the mutated single chains. Comparable data were obtained by
ELISA (Table 2A); the mAbs reacted with the single-chain mutants in the
order ßT39F, ß39AAA41 > ßT39A > unmutated
single chain, which is the inverse of reactivity seen with the
heterodimer-specific mAb 26 (Fig. 4). When the mutants were assayed
according to the polyclonal-based quantitation described above, their
reactivity to mAbs 64 and 68 was 2- to 6-fold greater compared with
CGß (Table 2B). The data show that loss of heterodimeric
specificity is associated with gain of reactivity to free subunit
determinants and demonstrate that the mutations T39F, T39A, and
39AAA41 disrupted the /ß domain interactions in the single
chain.
|
- subunit are at residues 10–60, 28–82, 32–84, 7–31, and 59–87
(2, 3); the six bonds of the ß-subunit are 9–57, 34–88, 38–90,
23–72, 93–100, and 26–110 (2, 3). In each subunit the first three
bonds comprise the cystine knot (2, 3). Monomeric - or
CGß-subunits containing certain disulfide mutations are unable to
fold appropriately and form heterodimers (24, 25, 26). However, when either
a mutated - or ß-domain was incorporated in the single chain, they
were biologically active (6, 7). Earlier studies showed that the extent
of heterodimer formation of the -subunit bearing either the cystine
mutations 10–60 or 32–84 was less than 5% (7). To assess the
heterodimeric-like behavior of these -disulfide bond mutants in the
single chain, a Western blot analysis was performed using the mAbs
described above; two representative dimer-specific mAbs 10 and 40 are
shown (Fig. 6; see also Table 3). Each reacted with the unmutated
single chain and the hCG heterodimer (lanes 1 and 4). In the case of
mAb 10 (panel A), CGß 10–60 was not detected (lane 2) but
CGß 32–84 was recognized (lane 3). No significant reactivity of
mAb 40 with these mutants was seen (panel B, lanes 2 and 3). Similar
results were obtained when blots were probed with the other
dimer-specific mAbs described in Table 3 (data not shown). The ELISA
profiles also show a comparable decrease in immunoreactivity of the
mutants to these dimer-specific mAbs (Table 3).
|
|
, C
and D, and Table 3). The absence of signal was not due to a
preferential loss of the mutants since they were detected with
polyclonal antiserum against the -subunit (panel E). Clearly, the
disulfide bond mutations in either the - or ß-subunit disrupted
the quaternary relationship between the subunit domains in the single
chain.
As shown above for the -helix mutants, these disulfide bond
variants, compared with the nonmutated single chain, might be expected
to manifest a greater signal with the ß-subunit-specific mAbs.
Consistent with this prediction (Fig. 7),
the -mutants have an increased immunoreactivity to mAbs 64 (panel A)
and 68 (panel B). These results were also supported by ELISA assays
(panels C and D and Table 3). However, in contrast to these data, there
was no detectable immunoreactivity of the ß-disulfide bond mutants to
mAbs 64 or 68 except for the reactivity of ß38–90 with mAb 64 (Fig. 8). Moreover, reactivity of the
ß-mutants was significantly reduced with five additional mAbs that
recognize the dimer or noncombined ß-subunit (data not shown). The
ß-mutants were not detected by ELISAs containing the
ß-subunit-specific mAbs 64 and 68 (Fig. 8, C and D, and Table 3).
Thus, in contrast to the -disulfide bond single-chain mutants that
are detected by mAbs 64 and 68, these mAbs fail to recognize epitopes
on the ß-subunit (23). As performed above for the 39–41 mutants,
we tested the ELISA immunoreactivity in the absence of Tween with
representative dimer-specific mAb, e.g. 55, and both the
ß-subunit-specific mAbs 64 and 68; in all cases there were no
significant differences in the ratio of immunoreactivities compared
with assays containing detergent (data not shown).
|
|
). In the case of
the immunoreactivities to the dimer-specific mAbs, those mutants that
scored weak or not detectable by the mAb quantitation gave similar
signals whether the muteins were quantitated by mAb or polyclonal
antisera. There are variations with some of the mAbs and disulfide
mutants when the two quantitation methods are compared (Table 4). The
largest difference (5-fold increase) is seen when the 10–60 and
ß9–57 mutants are probed with mAb 10. (When 10–60 is analyzed by
Western blot with mAb 10, a significant reduction in signal is observed
compared with the ELISA. The reason for this variation is not clear.)
Although the signal is increased for some of the other muteins, the
overall immunoreactivity is still 6–10 times less than the unmutated
single chain. Despite these differences, the data obtained from two
widely different quantitation protocols show that loss of heterodimeric
immunoreactivity is markedly reduced with a corresponding increase in
reactivity of the noncombined ß-subunit-specific epitopes. Taken
together the results further support the hypothesis that these
mutations disrupted the heterodimeric configuration of the CGß
single chain, but despite these modifications, the mutants nevertheless
display high-affinity receptor binding.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
- and ß-subunits of hCG are genetically linked as a single
polypeptide chain, but prevented from forming the intimately associated
dimer-like configuration, the biological activity is nevertheless
retained. We approached this issue with two distinct sets of mutations.
One set involved the amino acids in the -subunit that have direct
contact with the ß-subunit, and the other dealt with altering the
tertiary structure of the subunits by mutating the disulfide bonds. The
single-chain model was critical for these studies since it overcame the
limitation of assembly defects that would hinder analysis of these
regions as seen in earlier work because of low heterodimer
production.
The sequence 39–41 of the -subunit is contained in the long loop 2
which interfaces with loops 1 and 3 of the CGß subunit. Previous
studies have demonstrated that mutations in this region inhibit
assembly with the ß-subunit of CG (8, 10, 15) or FSH (16). Threonine
39 apparently occupies a pivotal role in this sequence; based on the
crystal structure, it resides on the external face of the dimer, and
there is a hydrogen bond between this residue and asparagine 52, which
bears an oligosaccharide that is involved in assembly and biological
activity (42). In addition, it has been shown that Thr 39 can be
O-glycosylated in vivo, which prevents formation of the
heterodimer (29, 30). Residues 40–50 in the -subunit contain the
only helical structure in the dimer, and Thr 39 occupies the
NH2-terminal cap (31) of a helix. The cap apparently
initiates helix formation (31). The hydrogen bond between Thr 39 and
Asn 52 contributes to the secondary structure of the helix and thus
mutation of this threonine residue presumably disrupts helical
stability. The introduction of the bulky phenylalanine would lead to a
greater perturbation of the helix. The single-chain data suggest that
Thr 39 has a role in receptor binding because the binding of ßT39F
was inhibited 42-fold compared with the nonmutated control. Because
there is a correlation between the inhibition of receptor binding of
ßT39F and the loss of immunoactivity to dimer mAb 26 with a
corresponding gain of activity to free subunit-specific mAbs 64 and 68,
it could be concluded that the quaternary structure is involved in the
biological activity (Fig. 3). However, although the 39AAA41 mutant
exhibits a similar structural change, as reflected by its decreased
reactivity to mAb 26 and an increase in signal to mAb 64/68, it still
binds to receptor with an affinity comparable to wild type. We suggest
that the determinant associated with the 39–41 sequence in receptor
binding can be uncoupled from its requirement for /ß assembly and
implies that Thr 39 interacts with the receptor.
Although T39 is important for /ß-subunit recognition, it is
curious that mutations that reduce heterodimer formation did not have
such an inhibitory effect on the association of /ß-domains in the
single chain. For example, the monomeric -subunit bearing the T39F
mutation does not assemble with the ß- subunit, but it forms a
heterodimeric-like association when the mutant is incorporated into the
single chain (Fig. 3). Because the /ß-domains are covalently
linked in the single chain, the local concentration of the domains at
the /ß-interface may be sufficient to promote the productive
interaction of the two subunits. This is consistent with our data that
indicate when the ratio of ß-subunit exceeds the -subunit more
heterodimer is observed. We suggest that the assembly reaction proceeds
in at least two phases: the first involves the recognition by the
subunits to initiate pairing, which occurs in the ER lumen through the
possible aid of one or more chaperones and the second involves the
final functional dimeric configuration of the subunits. This is
supported by studies that show that synthesis of an assembly-competent
hCGß subunit is associated with multiple intermediate forms (32).
Given the dramatic effects cysteine mutations have on the secretion and
assembly of the individual subunits (7, 24, 25, 26), it was unexpected that
disrupting the cystine knot in either the - or ß-subunit domain
would not significantly affect the receptor binding or signal
transduction. Many of the mAbs used here are known to react with
different regions of the hormone (11, 12, 23, 33, 39), and thus lack of
mutant immunoreactivity is presumably due to altered conformation of
the subunit(s) rather than abolishing recognition at a specific
cysteine residue. These observations further support the hypothesis
that all of the wild-type heterodimeric interactions are not essential
for receptor binding and signal transduction. Clearly, domains from
both subunits are necessary since free subunits are not active. It
appears that if the crucial domains for receptor binding/signal
transduction are present, hormone function is preserved; even some loss
of tertiary structure is tolerated as shown by the reduced
immunoreactivity of the ß-disulfide bond mutants to all mAbs that
recognize the ß-subunit. Binding of one domain to the receptor may
restore altered conformation of the other domain(s). However, although
the native quaternary structure may not be essential for receptor
activation, we cannot exclude the possibility that the function of the
heterodimeric configuration is required to maintain circulatory
half-life of the hormone.
The unmutated single-chain hCG itself has lost some quaternary
structure since it is not recognized by dimer-specific mAb 45. Epitope
mapping studies using mutagenesis (33) indirectly imply that this mAb
binds in the vicinity of the cystine knots at the contact sites in the
heterodimer (see Materials and Methods). Additionally, a
free subunit-specific mAb, which also binds near the ß-subunit
cystine knot (residue 89), reacts positively with the unmutated single
chain. Therefore, there is some aberration in this area created by
linking the C terminus of the ß-chain to the N terminus of the
-chain. That the unmutated single-chain hCG is as potent as the
heterodimer, despite having such an open structure at the core of the
molecule, is further evidence that neither all of the native
heterodimeric configuration nor an intact cystine knot is a
prerequisite for receptor binding. While we cannot exclude some native
quaternary interactions in the mutants, e.g. the sequences
responsible for receptor binding, our functional data are consistent
with the conclusions from crystal structure analysis that the hCG dimer
structure is open with relatively modest affinity between the subunits
(3). These data imply that gonadotropins with different conformations
can combine with the receptor and activate adenylate cyclase.
The data also support the hypothesis that the cystine knots are
scaffolds in the subunits that play a role in heterodimer formation
(6). It is interesting that the biologically active form of cystine
knot-containing proteins is either a homo- or a heterodimer (34). This
capacity to dimerize is critical for the glycoprotein hormones in which
the common -subunit is shared by four hormones. Arakawa et
al. (35) have shown that the subunits of brain-derived
neurotrophic factor and neurotrophin-3, members of the cystine knot
family which exist as homodimers, formed biologically active
heterodimers. It is intriguing to consider that the common -subunit
or ß-subunit may form dimers with subunits of other proteins that
have a cystine knot motif. This may be a mechanism by which the free
ß-subunit exhibits a growth factor effect observed on certain cancer
cells (36); it could form a dimer with monomeric growth factors and
then bind to a receptor on the tumor cells.
These studies have important biological implications for producing recombinant multisubunit proteins, since it may not be necessary to ensure that the product has the exact native conformation. It is known that, even in large protein-protein interfaces, only a few residues are actively involved in binding to receptor (37). Thus, the approach of covalently linking subunits from multimeric proteins, combined with mutagenesis, is a promising strategy for creating less complex molecules. The removal of nonessential residues could lead to determining critical structural requirements for signal transduction and thus facilitate the design of small peptide agonists or antagonists.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Antibodies
mAb B110 was obtained from Dr. W. Moyle (Robert Wood Johnson
Medical School, Piscataway, NJ); B109 and A407 were obtained from Drs.
S. Birken and R. Canfield (Columbia University, New York, NY). The
Innsbruck (INN-) series of mAbs were described previously (11, 12). All
mAbs are purified ascites except INN-hCG-10 and INN-hCG-40, which are
unfractionated ascites. All INN-hCG-x mAbs will be referred to as mAb
x. Rabbit polyclonal antiserum against hCGß- and -subunits were
prepared in this laboratory. The epitopes for these mAbs on hCG have
been extensively characterized by two-site antibody compatibility
assays, inter- and intraspecies cross-reactivities, peptide mapping,
and site-directed mutagenesis (11, 12, 23, 33, 38, 39). The majority of
epitopes are discontinuous, and the reactivity of the mAbs is
conformationally dependent. Nine distinct epitope regions have been
defined on the ß-subunit (ß1–ß9), six on the -subunit
(1–6), and four regions only present on the holohormone, termed
c1 through c4 (Fig. 1). The ß6- and ß7-epitope regions are specific
for the free ß-subunit. The epitope region ß1 involves
CGß-subunit amino acid residues 10 and 60 (33). Residue 89 in CGß
was also shown by mutagenesis to be involved in both ß1- and
ß7-epitope regions (A. M. Jackson, P. Berger, A. J.
Lapthorn, N. W. Isaacs, P. J. Delves, T. Lund, and I. Roitt,
unpublished data)and ß1 overlaps with region ß7 and c3. Therefore
it is likely that mAbs specific for epitope region c3 (mAb 45) and ß7
(mAb 68) bind to a similar area of the hormone (core of the molecule in
Fig. 1).
Construction of -Subunit Mutants and Single-Chain Analogs
The construction of pM2 plasmids containing
-subunit T39F and T39A was described earlier (8). The following
primers were used for PCR overlap mutagenesis to generate the triple
mutant TPL39–41AAA:
1. 5'-TAC TTT GTC GAC AAA TGA TAA TTC AGT GAT TGA-3' SalI
2. 5'-AGA TCC GGA TCC ACA GTC AAC CGC CCT-3' BamHI
3. 5'-AGA GCA TAT CCC GCG GCC GCT AGG TCC AAG AAG-3'
4. 5'-CTT CTT GGA CCT AGC GGC CGC GGG ATA TGC TCT-3'
Sense and antisense primers 3 and 4 correspond to amino acid residues
35–45 of the -subunit (bp 277–309), and primers 1 and 2 correspond
to bp 42–65 and 709–729 in the 5'- and 3'-nontranslated
sequence, respectively (43). The hCG wild-type gene, subcloned in
pM2 (18), was used with primers 1 and 3 to create the first
fragment. Primers 2 and 4 generated the corresponding overlapping
fragment. These fragments were then used as the templates with primers
1 and 2 to synthesize the complete -subunit mutant, which resulted
in a BamHI site at the 5'-end and a SalI site at
the 3'-end. These sites were used to subclone the fragment into
pM2. The single-chain analogs were engineered using the
naturally occurring XbaI site at amino acid 35 of the
-subunit gene (7). An XbaI digest of the -mutants
excised a 1900-bp fragment, which was subcloned into an XbaI
digest of pM2HA-CGß(5) to generate single-chain
variants. All constructs were sequenced to ensure no errors were
introduced during the PCR.
Transfection, Clone Selection, Labeling, and
Immunoprecipitation
Plasmids containing the mutant genes were transfected into CHO
cells and screened as previously described (19, 40). Clones were
immunoprecipitated as previously described (8, 24).
Receptor Binding and cAMP Assays
The single-chain variants were quantitated by an hCG RIA
(Diagnostic Products Corp, Los Angeles, CA) using hCGß
polyclonal antiserum. Conditioned media were incubated with CHO or
human fetal kidney 293 cells stably transfected with human LH/CG
receptor, and the binding affinity was determined as previously
reported (41). The cAMP accumulation (intra- and extracellular) was
determined using the NEN Life Science Products (Boston,
MA) flashplate assay as per manufacturer’s instructions. Briefly,
5 x 104 CHO cells stably transfected with the LH/CG
receptor were incubated for 2 h at room temperature with hCG dimer
or mutants, [125I]cAMP was added, and the cells were
incubated for 17 h at room temperature. The radioactivity in the
flashplate was determined with a Packard Top Counter (Packard
Instruments, Meriden, CT).
Western Blot and ELISA
Concentrated conditioned medium of single chains was quantitated
by polyclonal RIA (Diagnostics Products, Los Angeles, CA)
using either the hCGß antiserum in the kit or our -antiserum. We
elected to use this polyclonal- rather than a monoclonal-based assay to
minimize the potential error in quantitation due to structural changes
caused by a particular mutation. For mutants in the -subunit domain
of the single chain, antiserum against the ß-subunit was used whereas
anti--serum was employed for mutations in the ß-subunit. In
several studies using site-directed mutagenesis, we observed minimal
changes in recovery when such muteins were immunoprecipitated with
polyclonal antisera. Equivalent amounts of each protein were loaded
onto 12.5% SDS-polyacrylamide gels in the absence of heat and
ß-mercaptoethanol. Proteins were transferred onto nitrocellulose and
probed with mAbs and visualized on film by chemiluminescence (Tropix,
Bedford, MA). All experiments contained a blot probed with polyclonal
antisera to ensure that comparable amounts of material was loaded on
the gels.
To determine the immunoreactivity of the variants in solution assays,
conditioned media were examined by ELISA. All test mAbs were titrated
against purified hCG, except for the ß-subunit-specific mAbs, which
were titrated with recombinant CGß subunit. The mAb used to determine
the amount of each variant in the condition media was based on the
mutated subunit. In the case of changes in the -domain, mAb 2 was
the capture mAb. This mAb, which recognizes the free and dimer forms of
the CGß subunit, exhibited comparable binding to the -mutants and
nonmutated single chain. Likewise, for the mutations in the ß-subunit
domain, mAb 132 was used since its epitope includes amino acids 17–22
of the -subunit, which is minimally influenced by mutations on the
ß-subunit. The detection of the complex was achieved with polyclonal
ß-antiserum for mAb 2 and polyclonal -antiserum for mAb 132.
The muteins were also quantitated for the ELISA using a
polyclonal-based RIA; for mutations in the -subunit, antiserum
against the ß-subunit was used, whereas anti--serum was used for
muteins containing changes in the ß-subunit domain. This quantitation
protocol should be less sensitive to conformational changes induced by
the mutations. When the two quantitation methods are compared, there
was a 0- to 5-fold variation in the determinations for some of the
muteins. Despite these differences the trends are unchanged.
Equal amounts of each analog were used in all analyses. The ELISA
plates were prepared according to the following: 1) The wells
(Corning, Inc., Corning, NY) were coated with 100
µl/well of the test (capture) mAb in carbonate-bicarbonate buffer
[0.29% (wt/vol) NaHCO3; 0.17% (wt/vol)
Na2CO, pH 9.6) at 4 C overnight at a concentration that
yields a linear response to the appropriate standard; 2) After coating,
blocking was performed with 200 µl 0.05% Tween-PBS containing 0.5%
BSA; 3) Condition media containing variant were added; 4) Rabbit
anti-hCGß or sera (at a dilution titrated for each coated mAb)
were added; 5) Alkaline phosphatase-conjugated goat antirabbit IgG
(Sigma, St. Louis, MO) was added; and 6) The reactions
were developed with p-nitrophenyl phosphate (1
µg/ml). At each step the plates were incubated for 1 h at
37 C, and the wells were washed three times with Tween-PBS between each
step. After the addition of p-nitrophenyl phosphate, the
plates were incubated at room temperature for 1 h, and absorbance
at 405 nm was determined in an ELISA plate reader.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
A.M.J. is a Lalor Foundation Research Fellow, and P.B. is funded by the Austrian Science Fund (P13652-Gen). This work has been supported by a grant from Organon.
Received for publication March 30, 1999. Revision received August 27, 1999. Accepted for publication September 9, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
subunit and either CGß or the
FSHß subunits: role of a linker sequence. Mol Cell Endocrinol 125:71–77[CrossRef][Medline]
subunit
is critical for intracellular behavior but not for in vitro
biological activity. J Biol Chem 272:18098–18103-subunit required for assembly with the
chorionic gonadotropin ß subunit. Mol Endocrinol 6:261–271[Abstract]
subunit
of a human carcinoma choriogonadotropin. J Biol Chem 261:10475–10477-subunit important in holoprotein formation and receptor
binding. Endocrinology 134:1768–1770-subunit of lutropin. Biochemistry 30:1858–1864[CrossRef][Medline]
subunit contain structural components
for subunit heterodimerization and human follitropin-receptor binding.
Arch Biochem Biophys 329:127–135[CrossRef][Medline]
-subunit. J Biol
Chem 268:21613–21617
subunit of human choriotropin that are important for interaction with
the lutropin receptor. J Biol Chem 266:14251–14255 subunit.
Differential processing of free and combined forms in human trophoblast
and transfected mouse cells. J Biol Chem 262:14197–14203-subunit. J Biol Chem 269:25543–25548-like material from bovine
pituitaries. Removal of its O-linked oligosaccharide permits
combination with lutropin-ß. J Biol Chem 259:2662–2666 subunit. Biochem Biophys Res Commun 122:1260–1267[CrossRef][Medline]
subunit of human chorionic
gonadotrophin: molecular basis of immunologically and biologically
active domains. J Endocrinol 140:145–154
subunit of the four human glycoprotein hormones. J Mol Appl Genet 1:3–18[Medline]
This article has been cited by other articles:
 |
|
 |
|
  S.F. de Medeiros and R.J. Norman Human choriogonadotrophin protein core and sugar branches heterogeneity: basic and clinical insights Hum. Reprod. Update, January 1, 2009; 15(1): 69 - 95. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Jablonka-Shariff, T. R. Kumar, J. Eklund, A. Comstock, and I. Boime Single-Chain, Triple-Domain Gonadotropin Analogs with Disulfide Bond Mutations in the {alpha}-Subunit Elicit Dual Follitropin and Lutropin Activities in Vivo Mol. Endocrinol., June 1, 2006; 20(6): 1437 - 1446. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Garcia-Campayo, I. Boime, X. Ma, D. Daphna-Iken, and T. R. Kumar A Single-Chain Tetradomain Glycoprotein Hormone Analog Elicits Multiple Hormone Activities In Vivo Biol Reprod, February 1, 2005; 72(2): 301 - 308. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Garcia-Campayo, A. Jablonka-Shariff, and I. Boime A Single-chain Bifunctional Gonadotropin Analog Is Secreted from Chinese Hamster Ovary Cells as Two Distinct Bioactive Species J. Biol. Chem., October 22, 2004; 279(43): 44286 - 44293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Sohn, H. Youn, M. Jeoung, Y. Koo, C. Yi, I. Ji, and T. H. Ji Orientation of Follicle-stimulating Hormone (FSH) Subunits Complexed with the FSH Receptor: {beta} SUBUNIT TOWARD THE N TERMINUS OF EXODOMAIN AND {alpha} SUBUNIT TO EXOLOOP 3 J. Biol. Chem., November 28, 2003; 278(48): 47868 - 47876. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. B. Fralish, B. Dattilo, and D. Puett Structural Analysis of Yoked Chorionic Gonadotropin-Luteinizing Hormone Receptor Ectodomain Complexes by Circular Dichroic Spectroscopy Mol. Endocrinol., July 1, 2003; 17(7): 1192 - 1202. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. B. Fralish, P. Narayan, and D. Puett Consequences of Single-Chain Translation on the Structures of Two Chorionic Gonadotropin Yoked Analogs in {alpha}-{beta} and {beta}-{alpha} Configurations Mol. Endocrinol., April 1, 2003; 17(4): 757 - 767. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Narayan, J. Gray, and D. Puett Yoked Complexes of Human Choriogonadotropin and the Lutropin Receptor: Evidence that Monomeric Individual Subunits Are Inactive Mol. Endocrinol., December 1, 2002; 16(12): 2733 - 2745. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Garcia-Campayo, T. R. Kumar, and I. Boime Thyrotropin, Follitropin, and Chorionic Gonadotropin Expressed as a Single Multifunctional Unit Reveal Remarkable Permissiveness in Receptor-Ligand Interactions Endocrinology, October 1, 2002; 143(10): 3773 - 3778. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Xing, W. Lin, M. Jiang, R. V. Myers, D. Cao, M. P. Bernard, and W. R. Moyle Alternatively Folded Choriogonadotropin Analogs. IMPLICATIONS FOR HORMONE FOLDING AND BIOLOGICAL ACTIVITY J. Biol. Chem., December 7, 2001; 276(50): 46953 - 46960. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Garcia-Campayo and I. Boime Independent Activities of FSH and LH Structurally Confined in a Single Polypeptide: Selective Modification of the Relative Potencies of the Hormones Endocrinology, December 1, 2001; 142(12): 5203 - 5211. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Hiro'oka, D. Maassen, P. Berger, and I. Boime Disulfide Bond Mutations in Follicle-Stimulating Hormone Result in Uncoupling of Biological Activity from Intracellular Behavior Endocrinology, December 1, 2000; 141(12): 4751 - 4756. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Darling, J. A. Wilken, A. K. Miller-Lindholm, T. M. Urlacher, R. W. Ruddon, S. A. Sherman, and E. Bedows Functional Contributions of Noncysteine Residues within the Cystine Knots of Human Chorionic Gonadotropin Subunits J. Biol. Chem., March 30, 2001; 276(14): 10692 - 10699. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Ben-Menahem, A. Jablonka-Shariff, R. K. Hyde, M. R. Pixley, S. Srivastava, P. Berger, and I. Boime The Position of the alpha and beta Subunits in a Single Chain Variant of Human Chorionic Gonadotropin Affects the Heterodimeric Interaction of the Subunits and Receptor-binding Epitopes J. Biol. Chem., August 3, 2001; 276(32): 29871 - 29879. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
