Scand.J. Immunol., Vol. 5, 1976.
`
`Reassembly of Immunoglobulin M Heavy and
`Light Chains In Vitro
`K. E. SCHROHENLOHER & R. B. HESTER
`Division of Clinical Immunology and Rheumatology, Department of Medicine,
`and Division of Clinical Pathology, Department of Pathology,
`University of Alabama School of Medicine, Birmingham, Alabama, USA
`
`Schrohenloher, R.E. & Hester, R.B. Reassembly of Immunoglobulin M Heavy
`and Light Chains In Vitro. Scand.J. Immunol. 5, 637-646, 1976.
`
`Reduced and alkylated monoclonal IgM was fractionated into [i and light (L)
`chains by gel chromatography in IN acetic acid. Equimoiar mixtures of the
`chains formed a noncovalently bonded structure in 0.0IM sodium acetate
`buffer, pH 4.1, that had the properties of a half subunit. The latter reassociated
`into a subunit-like structure after transfer into 0.08M sodium phosphate buffer,
`pH 7.5. The similarity of the reconstituted IgM subunit (IgMs^ to that of the
`native molecule was established by its physicochemical and immunochemical
`properties. Comparable products were obtained on reassembly of the alkylated \j.
`and L chains from several other monoclonal IgM. The presence of active bind-
`ing sites for IgG on subunits reconstituted from the chains of proteins with
`anti-IgG activity further indicated correct assembly of the \JL and I^ chains.
`High yields of subunit-like products were also obtained by assembly of [J. chains
`from one protein and L chains from another. Evidence was obtained that L
`chains of appropriate specificity can substitute for the homologous chain in the
`formation of the active site. Heterogeneous mixtures of high molecular weight
`products were generated from [J. and L chains that were not alkylated. Reduc-
`tion and alkylation demonstrated that the products represented polymers of
`reconstituted IgMs. Significant levels of anti-IgG activity were detected in the
`polymeric IgM generated from the chains of active proteins by precipitation with
`aggregated IgG.
`
`R. E. Schrohenloher, Ph.D., Division of Clinical Immunology atui Rlieionarohgy,
`Department of Medicine, University of Alabama School of Medicine, Birmingham,
`AL 35294, USA
`
`Recent studies from this laboratory demon-
`strated that active products can be regenerated
`from the isolated polypeptide chains of a mono-
`clonal IgM (protein Po) that binds human IgG
`(27). Approximately 80% of the materials
`derived from equimoiar quantities of the [i and
`V. chains from
`this protein sedimented as a
`heterogeneous group of polymers that were
`completely precipitated with heat-aggregated
`IgG. A large quantity of anti-IgG activity
`was also detected by indirect hemagglutination.
`Evidence was obtained that both chains eon-
`tributed to the formation of the combining
`site in a specific matiner. Thus, this reactive
`
`IgM resembled IgG antibodies in this respect (5,
`10, 19—22). Recombination of y and light (L)
`chains has been shown to give products that
`closely resembled native IgG with respect to
`physicochemical and antigenie properties (1, 9,
`17, 29). Similarly, recombination of the ^i and v.
`chains from protein Po gave well-defined struc-
`tural units that closely resembled the subunits
`of the native protein (23).
`In contrast to the results obtained with pro-
`tein Po, the component polypeptide chains from
`several other monoclonal IgM proteins failed
`to give significant yields of active products or
`subunit-like structures when subjected to identi-
`
`Merck Ex. 1040, pg 1158
`
`

`
`638 R. E. Schrohenloher & R.B. Hester
`
`cal conditions for recombination. These con-
`ditions were similar to those used for the re-
`generation of IgG molecules from y and L
`chains (1, 9, 17, 29). Briefly summarized, \i and
`L chains were isolated from reduced protein
`by gel chromatography in IN acetic acid. Mix-
`tures of the chains were first dialyzed into
`0.08M sodium phosphate buffer, pH 7.5, con-
`taining O.OIM dithiothreitol (DTT) in order to
`re-establish noncoyalent interactions. Further
`reassociation through disulfide bond formation
`was then effected by prolonged incubation at
`2°G after the DTT had been removed by di-
`alysis.
`A general approach to the in vitro reassembly
`of isolated polypeptide chains into IgM sub-
`units and polymeric molecules is described in
`the present communication. The studies also
`provide additional information on the inter-
`action of n and L chains and on the roles of
`the component chains in the formation of the
`combining site.
`
`MATERIALS AND METHODS
`Isolation of IgM. The IgM proteins used in this
`study were obtained from patients with Wal-
`denstrom's niacroglobulincmia. The various pro-
`teins were precipitated from plasma by the
`addition of 15 volumes of cold distilled water.
`Proteins Da and Gr were resuspended in saline
`and the precipitation procedure repeated twice.
`Proteins Po, Go, and Le were further purified
`by gel chromatography on Sephadex G-200 in
`O.IM sodium acetate buffer, pH 4.1, as pre-
`viously described (26, 27). The resulting prep-
`arations were essentially free of materials sedi-
`menting in the ultracentrifuge slower than the
`major macroglobulin component and demon-
`strated not more than trace amounts of IgG and
`IgA by immunoelectrophoresis using specific
`antisera.
`
`Reductive cleavage. The interchain disulfide
`bonds of the IgM proteins were cleaved by in-
`cubation for 3 h with O.OIM DTT at room
`temperature in 0.08M sodium phosphate buf-
`fer, pH 7.5. The protein concentration was
`10 mg/ml. Free sulfhydryl groups were alkyl-
`ated by dialysis against 0.02M iodoacetamide
`
`ill the pH 7.5 buffer for 4 h at room tempera-
`ture with stirring (26).
`Isolation and recombination of polypeptide
`chains. Reduced IgM or reduced and alkylated
`IgM was fractionated into u and L chains by
`gel chromatography on Sephadex G-150 in IN
`acetic acid as previously described (27). Equi-
`moiar quantities of the isolated |i and L chains
`were recombined and diluted to a protein con-
`centration of 0.1 mg/ml by addition of O.OIM
`sodium acetate buffer, pH 4.1. The mixture was
`thoroughly dialyzed against the same pH 4.1
`buffer. Further reassociation was achieved by
`dialysis into O.OIM sodium phosphate buffer,
`pH 7.5, and, finally, 0.08M sodium phosphate
`buffer at
`the same pH. Mercaptoethano!
`(O.OIM) was added to the O.OIM sodium phos-
`phate buffer when polymeric products were
`prepared from polypeptide chains with free
`sulfhydryl groups. The products were brought
`to a protein concentration of 10 mg/ml by
`uitrafiltration
`through
`type PM-10 Diaflo
`membranes (Amicon Gorp., Cambridge, Mass.).
`
`Products were fractionated by gel chroma-
`tography on Sephadex G-200 in the 0.08M
`sodium phosphate buffer. Eluted fractions were
`concentrated to approximately 10 mg/ml by
`uitrafiltration as described above.
`Analytical procedures. Analytical ultracentrif-
`ugal studies were performed in a Beckman Mod-
`el E ultracentrifuge as previously described
`(25). Before analysis, samples were thorough-
`ly dialyzed against the 0.08M sodium phos-
`phate or other indicated buffer. Sedimentation
`rates were corrected
`to standard conditions
`(^21).It) ^^'^ refer to a sample concentration of
`approximately 10 mg/ml unless otherwise
`stated. Sedimentation rates at infinite dilution
`(s^'jo.ir) were estimated by linear extrapolation
`of the s.,fi „. values determined at five sample
`concentrations, using
`the method of
`least
`squares. The partial specific volumes of the
`native proteins and reconstituted products were
`assumed to be 0.725 ml/g (16).
`Polypeptide chain composition was de-
`termined after dialysis against IN acetic acid
`by gel chromatography on Sepbadex G-150 in
`IN acetic acid (8). Relative yields were calcu-
`lated from absorbance measured at 280 nm.
`
`Merck Ex. 1040, pg 1159
`
`

`
`Reassembly of IgM Heavy and Light Chains 639
`
`Immunoelectrophoresis was performed as pre-
`viously described (27). Antisera were obtained
`from Meloy Laboratories, Springfield, Va. Disc
`electrophoresis in 5% polyacrylamide was per-
`formed
`in sodium phosphate-sodium docecyl
`sulfate (SDS)-urea buffer (13).
`
`RESULTS
`Interaction of \i and L chains
`Application of the previously used recombi-
`nation conditions (23, 27) to alkylated [i and X
`chains isolated from Gr IgM gave only a small
`yield (18%) of a subunit-like component. Ex-
`amination of the other products recovered by
`chromatography on Sephadex G-200 indicated
`that the [i chains underwent extensive self-
`association instead of interacting with X chains
`to form the four-chain structure characteristic
`of
`immunoglobulins.
`Immunoelectrophoresis
`and polyacrylamide gel electrophoresis in urea-
`SDS buffer demonstrated that the remaining fi
`chains were present as aggregates in the fraction
`excluded from the column, whereas the remain-
`ing A, chains were eluted after the subunits. In
`contrast, the yield of subunit-like component
`obtained from the alkylated \i and y. chains of
`Po IgM exceeded 50%.
`The self-association of \i chains was further
`characterized by sedimentation-velocity studies
`on alkylated Gr \i chains in buffers of different
`pH and concentration. A single major compo-
`nent was observed in 5mM glycine-HGI buffer,
`pH 2.5, having an s'\,,, „, value of 3.OS (Fig. la).
`Asymmetry of the boundary indicated the pres-
`ence of a small quantity of larger components,
`which was more apparent in older preparations.
`When examined in O.OIM sodium acetate buffer,
`pH 4.1, the s^.,fl ,^ value of the major component
`was 3.6S (Fig. lb). A second component having
`an s^2fl,?/' of approximately 8S was also evident.
`Further increases of either pH or buffer con-
`centration resulted in greater aggregation. In
`O.IM sodium acetate buffer, pH 4.1, major
`components sedimenting at approximately 6S
`and 7S were apparent, in addition to larger
`aggregates and a small quantity of 5S material
`(Fig. lc). Similarly, most components sedi-
`mented at approximately 7,5S and faster in
`
`a b c d
`
`Fig. 1. LJUraucntnlugiil paiicrns of [i chains from
`protein Gr in {a) 5mM glycine-HCl buffer, pH 2.5,
`{b) O.OIM sodium acetate buffer, pH 4.1, (c) O.IM
`sodium acetate buffer, pH 4.1, and {d) 4 mM sodium
`acetate buffer, pH 5.4. The pattern in a was recorded
`after 64 min at 68,000 rpm and that in b after 64 min
`at 60,000 rpm; those in c and d were recorded after
`the same interval at 56,000 rpm. The concentrations
`of the samples were {a) 1 mg/ml, (fe) 9 mg/ml, (c) 10
`mg/ml, and {d) 6 mg/ml. The direction of sedimenta-
`tion in this and subsequent figiu-es is to the right.
`
`Reactivity with human IgG was determined
`by indirect hemagglutination of tanned sheep
`erythrocytes coated with human IgG and pre-
`cipitation of heat-aggregated human IgG (27).
`
`Merck Ex. 1040, pg 1160
`
`

`
`4mM sodium acetate buffer, pH 5.4 (Fig. Id).
`Analysis
`in 4mM sodium phosphate buffer,
`pH 7.5, demonstrated a broad boundary sedi-
`menting at approximately l lS and only a small
`quantity of 3S or 4S material.
`The above studies suggested that successful
`reassembly of the IgM subunit might be achiev-
`ed by conditions that would permit |i/L inter-
`actions to be re-established in the absence of
`extensive [t-chain self-association. Sodium ace-
`tate buffer, O.OIM and pH 4.1, was selected for
`this purpose. Equilibration of an equimoiar
`mixture of the \i and "k chains from reduced
`and alkylated Gr IgM gave nearly complete
`recombination. Ghromatography on Sephadex
`G-200 in the pH 4.1 buffer demonstrated that
`most of the sample eluted as a single component
`witb only a small quantity of lower molecular
`weight materials present (Fig. 2a). The elutioii
`volume indicated that the product was smaller
`than IgG but larger than cither of the com-
`ponent chains. Sedimentation analysis of the
`isolated product demonstrated a symmetrical
`boundary having an s*.*^ „. value of 4.3S (Fig.
`3a).
`Its diffusion
`coefficient
`{D".,,,,^) was
`4.7 XIO"'' cm^/sec. The latter was estimated by
`extrapolation of values obtained by the method
`of Ehrenberg (7) over the range of 2 to 10
`mg/ml. A molecular weight of 80,000 daltons
`was calculated from these values by the Sved-
`berg equation (31). Dissociation in IN acetic
`acid followed by gel chromatography on Sepha-
`dex G-150
`in the acetic acid indicated
`the
`presence of equimoiar quantities of the [) and X
`chains. These results suggested that the [.i and \
`chains interacted to form a relatively stable
`two-chain structure that represented half of the
`natural four-chain subunit.
`
`640 R. E. Schrohenloher & R. B. Hester
`
`a
`
`1.5
`
`1.0
`
`0.5
`
`1.5
`
`10
`
`05
`
`1.5
`
`10
`
`05
`
`c OC
`
`O
`OJ
`
`<
`CD
`
`a:o
`
`to
`CD
`
`10
`
`50
`40
`30
`20
`FRACTION NUMBER
`
`Further reassociation of the pH 4.1 product
`was achieved by dialysis sequentially
`into
`O.OIM and 0.08M sodium phosphate buffer,
`Fig. 2. Chromatography on Sephadex G-200 (2.5 X
`pH 7.5. Ghromatography on Sephadex G-200
`100 cm column) of (a) an equimoiar mixture of ji. and
`in the latter buffer revealed a large quantity
`X chains from protein Gr in O.OIM sodium acetate
`of a component that closely resembled the sub-
`buffer, pH 4.1, (ft) reconstituted Gr IgM subunits in
`units of the native IgM (IgMs) (Fig. 2b). Sedi-
`0.08M sodium phosphate buffer, pH 7.5, and (c)
`reconstituted Gr IgM in 0.08M sodium phosphate
`mentation analysis of this component gave a
`buffer, pH 7.5. The quantities of protein applied to
`single symmetrical boundary having an 5".,^ ,^
`the columns were (a) 50 mg, {b) 100 mg, and (c) 85 mg.
`value of 6.69S (Fig. 3b). Native Gr IgMs gave
`The horizontal bars indicate column tubes pooled
`a similar sedimentation pattern and an 5*.,^ „.
`for further studv.
`
`Merck Ex. 1040, pg 1161
`
`

`
`a
`
`Fig, 3. Ultracentritugal patterns of (a) reconstituted
`Gr IgM half subunics in O.OIM sodium acetate buffer,
`pH 4.1, (fc) reconstituted Gr IgM subunits in 0.08M
`sodium phosphate buffer, pH 7.5, (c) reconstituted
`Gr IgM in the pH 7.5 buffer, and (d) reduced and
`alkylated reconstituted Gr IgM in the pH 7.5 buffer.
`Patterns a and h were recorded after 80 min at
`56,000 rpm; those in c and d were recorded after 32
`min at the same speed. The protein concentration
`of each sample was 10 mg/ml.
`
`value of 6.81S. The distribution of components
`obtained by gel chromatography of the recon-
`stituted IgMs in IN acetic acid demonstrated
`26% L chain. The yield of L chain from the
`native IgM was 29%. The reconstituted and
`native subunits gave similar precipitin
`lines
`on immunoelectrophoresis using antiserum spe-
`cific for IgM or X-type Bence Jones protem.
`Each demonstrated a single component that re-
`acted with both antisera. The small quantity
`of aggregated materials recovered in the void
`volume (Fig. 2b) appeared to contain only |i
`chains by immunoelectrophoresis. Both |.i and X
`determinants were detected in the low molec-
`ular weight materials eluted after
`the sub-
`units.
`Comparable results were obtained on re-
`assembly of the (i and L chains of several other
`
`Reassembly of IgM Heavy and Light Chains 641
`
`monoclonal IgM. The reconstituted and native
`IgMs of protein Da gave comparable sedimen-
`tation patterns having s*^.,,, ,^ values of 6.30S
`and 6.57S, respectively. The yield of L chains
`by gel chromatography of the reconstituted
`IgMs in IN acetic acid was 22 % and that of
`the native IgMs was 26%. Well-characterized
`subunits were also obtained from the |.t and x
`chains of two monoclonal IgM (Co and Le)
`that reacted with IgG (Fig. 4a and 4b). The
`reconstituted IgMs from protein Co (Fig. 4a,
`upper pattern) sedimented at 6.IS when ex-
`amined at 8 mg/ml. The sedimentation rate of
`the subunit derived from protein Le (Fig. 4b,
`upper pattern) was 6.3S at 7 mg/ml. These
`values were in agreement with those obtained
`for the respective native IgMs at similar con-
`centrations. The product obtained from
`the
`u and X chains of protein Po (Fig. 4c, upper
`pattern) demonstrated greater heterogeneity
`than the other reconstituted IgMs. An estimated
`65% of the materials detected sedimented at
`approximately 6S; the rest sedimented at ap-
`proximately 5S. The subunit prepared from
`the native protein sedimented as a single com-
`ponent at 6.4S (23).
`
`Although the subunits produced from IgM
`anti-IgG factors by reduction and alkylation
`often fail to show activity by serologic pro-
`cedures, they frequently form soluble complexes
`with human 7S IgG that can be detected by
`ultracentrifugal analysis (23, 24, 26). The sub-
`units of proteins Co, Le, and Po each formed
`such complexes. Similarly the subunits recon-
`stituted from the chains of these proteins re-
`acted with IgG to form faster-sedimenting com-
`ponents not present in either the IgG or the
`subunit preparations (Fig. 4). The sedimentation
`characteristics of the complexes formed by re-
`constituted Co IgMs and reconstituted Le IgMs
`were similar to those of the native subunits.
`Those formed by reconstituted Co TgM sedi-
`mented at 8.2S and were poorly resolved from
`the unreacted components (Fig. 4a, lower pat-
`tern). As previously observed for subunits of
`several anti-IgG factors (23, 24, 26), the yield
`of complexes was dependent on the quantity
`of IgG present, and the molecular size of the
`complex, as indicated by its sedimentation rate.
`
`Merck Ex. 1040, pg 1162
`
`

`
`642 R. E. Schrohenloher & R.B. Hester
`
`Fig. 4. Ultracentrifugal ex-
`periments demonstrating tht-
`reaction of
`reconstituted
`IgM subunits with human
`IgG.
`a. Upper pattern: reconsti-
`tuted Co
`IgM
`subunits
`(8 mg/ml). Middle pattern:
`IgG control
`(2 mg/ml:.
`Lower pattern: mixture con-
`laining
`reconstituted Co
`IgM subunits (7 mg/ml) and
`[gG (2 mg/ml).
`/'. Upper pattern: reconsti-
`uited 1-e IgM subunits (7
`mg/ml). Middle pattern:
`IgG control (3 mg/ml). Lower pattern: mixture containing reconstituted Le IgM subunits (6 mg/ml) and
`IgG (3 mg/ml).
`c. Upper pattern: reconstituted Po IgM subunits (9 mg/ml). Middle pattern: IgG control (1.5 mg/ml). Lower
`pattern: mixture containing reconstituted Po IgM subunits (7 mg/ml) and IgG (1.5 mg/ml).
`Experiments were performed at 20 C in 0.08M sodium phosphate buffet, pH 7.5 The patterns in a and b
`were recorded after 64 min at 56,000 rpm; those in c were recorded after 80 min at the same speed.
`
`did not vary. The complexes formed by re-
`constituted Le IgMs sedimented at 7.7S (Fig. 4b,
`lower pattern). Even though the reconstituted
`Po IgMs differed ultracentrifugally from the
`native IgMs, complexes sedimenting faster than
`the IgG control were indicated after the addi-
`tion of IgG (Fig. 4c, lower pattern). Reaction
`of native Po IgMs with IgG gave complexes
`sedimenting at 8S (23).
`
`Formation of hybrid IgMs molecules
`Well-characterized products were obtained in
`high yields (60%-90%) on assembly of the
`fi chains from one protein and the L chains
`from another. The hybrid subunits generated
`from Co n chains and Po K chains, shown in
`Fig. 5a, sedimented as a single relatively sym-
`metrical boundary at 6.4S. Those generated by
`the interaction of Po \.\ chains and Co x chains,
`shown in Fig. 5b, demonstrated the presence
`of a small quantity of lower molecular weight
`materials and sedimented somewhat slower
`(5.7S). Each hybrid IgMs gave a single line on
`immunoelectrophoresis with antisera specific for
`human IgM and kappa-typc Bence Jones pro-
`tein, similar in appearance to those given by
`native IgMs. Comparable hybrid products were
`obtained by recombination of the ft and v.
`chains from protein Po with the complementary
`chains from protein Da.
`
`The capacity of the hybrid IgMs prepa-
`rations to bind IgG was examined ultracentrif-
`ugally at two concentrations of IgG (1 mg'm!
`and 2 mg/ml). The concentration of the hybrid
`IgMs was approximately 8 mg/ml. The Po
`^/Da K, Da [.t/Po x,, and Co ^/Po x hybrids
`failed to interact with the IgG at either con-
`centration. The experiment with the Co |.i/Po x
`hybrid shown in Fig. 5a is representative of
`this group. In contrast, large quantities of
`complexes sedimenting at approximately 9S
`were observed after the addition of IgG to the
`Po [.i/Co X hybrid (Fig. 5b). The quantity of
`complex, but not its sedimentation rate, was
`dependent on the quantity of IgG present.
`
`Regeneration of polymeric IgM from isolated
`polypeptide chains
`IgM subunits can be reassociated into poly-
`meric molecules by reestablishing intersubunit
`disulfide bonding (26). Since this bonding re-
`quires the presence of free sulfhydryl groups,
`the polypeptide chains used for regeneration of
`polymeric IgM were not alkylated. Noncova-
`lent (i/L interactions were reestablished by di-
`alysis into O.OIM sodium acetate buffer, pH4.1,
`as described for the alkylated chains. Oxidation
`of the free sulfhydryl groups was prevented by
`the acid medium during this step as well as
`during isolation (3). Further reassociation of the
`
`Merck Ex. 1040, pg 1163
`
`

`
`Reassembly of IgM Heavy and Light Chains 643
`
`.ilkylation dissociated 60% of fraction A into
`•iubunlts that sedimented at 6.2S (Fig. 3d). The
`remainder sedimented as a heterogeneous mix-
`liire of aggregated materials at approximately
`25S. Ultracentrifugal analysis of fraction B
`demonstrated a single component that sedi-
`mented at 6.4S. The major component in frac-
`tion G sedimented at 5S. Fraction D was not
`characterized. Fractions A, B, and G each gave
`a single precipitin line of gamma mobility on
`immunoelectrophoresis with antisera specific
`lor IgM and lambda-type Bence Jones protein.
`Noncovalently bound immunoglobulin chains
`were not detected in fraction A, B, or G by
`polyacrylamide gel electrophoresis in urea-SDS
`buffer. Each fraction demonstrated multiple
`bands near the top of the gels. However, heavy
`and ligbt chains were readily detected in frac-
`tions A and B after reduction and alkylation
`(fraction C was not examined). It was evident
`from
`these results that extensive interchain
`disulfide formation had occurred.
`Comparable yields of macromolecular prod-
`ucts were obtained with the component poly-
`peptide chains isolated from reduced Go IgM
`and from reduced Po IgM. Both demonstrated
`large amounts of anti-IgG activity by precipi-
`tin curve analysis with heat-aggregated human
`IgG (Fig. 6). Precipitin curves given by the
`native proteins are shown for comparison.
`
`DISGUSSION
`Results obtained in this study indicate that [i
`chains undergo self-association that is depend-
`ent on pH and buffer concentration. A single
`component having a sedimentation rate of 3S
`was dominant in dilute pH 2.5 buffer. The
`major component at pH 4.1 in O.OIM sodium
`acetate buffer sedimented at a similar rate.
`Gomparison to the
`j% ,^ values of y-chain
`dimers (5.2S) (29) suggested that the \i chain
`was present as a monomer under these con-
`ditions. However, the size and number of com-
`ponents became greater as the pH or buffer
`concentration was increased. The formation of
`multiple components having sedimentation rates
`of approximately 7.5S and faster (not corrected
`for concentration effect) by the |.i chains in
`
`Fig. 5. Ultracentrifugal studies on ihu iiuurai.ih)ii of
`hybrid IgM subunits with human IgG.
`a. Upper pattern: Co [A/PO X hybrid subunits (9 mg/
`ml). Middle pattern: IgG control (2 mg/ml). Lower
`pattern: mixture containing the Co jx/Po x hybrid
`subunits (7 mg/ml) and IgG (2 mg/ml).
`b. Upper pattern: Po ['/Co x hybrid subunits (11
`mg/ml). Middle pattern: IgG control (1 mg/ml).
`Lower pattern: mixture containing the Po ji/Co x
`hybrid subunits (9 mg/ml) and IgG (1 mg/ml).
`Experiments were performed at 20 C in 0.08M
`sodium phosphate buffer, pH 7.5. The patterns were
`recorded after 64 min at 56,000 rpm.
`
`mixture into subunits was facilitated by equi-
`libration with O.OIM sodium phosphate buffer,
`pH 7.5, containing O.OIM tiiercaptoethanol.
`This concentration of mercaptoethanol was suf-
`ficient to prevent premature oxidation of the
`sulfhydryl groups. Retnoval of the mercapto-
`ethanol from the reaction mixture by dialysis
`into 0.08M sodium phosphate buffer, pH 7.5,
`permitted spontaneous disulfide bond formation
`and polymerization of the regenerated subunits.
`Gel chromatography of the products obtained
`in this way from the polypeptide chains of Gr
`IgM (Fig. 2c) revealed a large quantity of mac-
`romolecular material (fraction A) and smaller
`quantities of lower molecular weight com-
`ponents (fractions B, C, and D). Ultracentrif-
`ugal analysis of fraction A showed that it was
`a heterogeneous mixture of two prominent com-
`ponents, sedimenting at 14S and 16S, and lesser
`amounts of a spectrum of slower- and faster-
`sedimenting materials (Fig. 3c). Reduction and
`
`Merck Ex. 1040, pg 1164
`
`

`
`644 R. E. Schrohenloher & R. B. Hester
`
`NATIVE Co IgM
`
`RECONST, Co
`
`NATtVE Po
`
`RECOMST Po IgM
`
`0.4
`
`0 6
`
`O.B
`
`1.0
`
`AGGREGATED IqGfmg)
`Fig. 6. Precipitation of heat-aggregated human IgG by
`(a) native and reconstituted Co IgM and {h) native
`and reconstituted Po IgM. Analyses were performed
`on 0.2-mg samples.
`
`4mM sodium acetate buffer, pH 5.4, was in
`marked contrast to human y chain. The latter
`was shown to exist primarily as dimers in this
`buffer (29). Analysis of chain recombination
`products obtained under conditions that failed
`to give significant yields of immunoglobulin-
`likc structures suggested that aggregation of the
`[I chain prevented its reassociation with L
`chains. The differences in the behavior of the
`polypeptide chains from different IgM when
`subjected to the recombination conditions used
`in the previous studies (23, 27) may reflect dif-
`ferences in the self-associative properties of the
`(1 chains.
`Mu chains were shown to interact with L
`chains in O.OIM sodium acetate buffer, pH 4.1,
`to form a noncovalently bonded structure with
`the properties of a half subunit. With equimolar
`quantities of the chains, the reaction was quan-
`titative and the product
`relatively homo-
`geneous. Its elution volume from Sephadex
`G-200, sedimentation rate, and apparent molec-
`
`ular weight were consistent with its identity
`as a half subunit. The molecular weight, as
`indicated by sedimentation and diffusion data,
`was somewhat lower than expected from the
`accepted value for IgMs of 185,000 daltons
`(16). This may have been due to partial dis-
`sociation into free ji and X chains. The presence
`of equimolar quantities of \i and X chains was
`confirmed after dissociation in IN acetic acid.
`The pH-dependent
`interaction of |.l and L
`chains therefore appears to be reversible. Suzuki
`& Deutsch (30) observed partial dissociation of
`reduced and alkylated IgM into \y and L chains
`in a dilute sodium formate buffer, pH 4. Thus,
`the stability of the noncovalent bonds between
`[I and L chains may vary for different proteins
`or may be altered by different buffers.
`The reconstituted half subunits were reassoci-
`ated into subunit-like structures by increasing
`the pH to 7.5 and the buffer concentration to
`0.08M. Larger polymers were not formed in
`experiments with alkylated chains. The small
`quantities of aggregated materials recovered in
`most experiments consisted principally of it
`chains. The similarity of the reconstituted sub-
`units to those from the native molecule was
`established by sedimentation properties, poly-
`peptide chain composition, and behavior on
`immunoelectrophoresis. The capacity of the
`subunits reconstituted from the polypeptide
`chains of proteins with anti-IgG activity to
`form soluble complexes with 7S IgG further
`indicated correct assembly of the [i and L
`chains.
`The polymeric products generated from IgM
`polypeptide chains with free sulfhydryl groups
`resembled
`those obtained by reassembly of
`IgM suhunits (26). The reconstituted polymeric
`IgM was shown to be composed of a large
`quantity of subunits that closely resembled
`those of the native molecule ultracentrifugally.
`Release of subunits by reduction further estab-
`lished that polymerization was due to inter-
`subunit disulfide bond formation. Comparison
`of polyacrylamide gel electrophoretic patterns.
`before and after reduction also indicated exten-
`sive interchain disulfide bond formation. Re-
`generation of active products from the isolated
`chains of IgM proteins with anti-IgG activity
`
`Merck Ex. 1040, pg 1165
`
`

`
`hy the present recombination scheme is con-
`sistent with the results previously obtained with
`Po IgM. Since J chain is recovered from re-
`duced IgM primarily in the L-chain fraction
`(15, 27), it was present in the recombination
`mixtures. Its fate during the process was not
`established in the present study. Evidence has
`been obtained that polymeric reconstituted Po
`IgM contained covalently bonded J chain (27).
`Previous experiments have shown that sig-
`nificant levels of anti-IgG activity were not
`detected when chains from an inactive Walden-
`strom's IgM were substituted for either chain
`of Po IgM (27). Although it was not determined
`whether actual combination of the chains oc-
`curred, the observations suggested that both fi
`and L chains contributed to the active site in
`a specific manner. Results obtained in the pres-
`ent study support this conclusion. Noncovalent
`assembly of Po n chains and Da •/. chains re-
`sulted in hyhrid 6S suhunits that failed
`to
`complex with IgG. Identical results were ob-
`tained by noncovalent assembly of Da n chains
`and Po x chains. Da IgM had no detectable
`anti-IgG activity. However, assembly of hybrid
`subunits that formed soluble complexes with
`IgG from the chains of two IgM with anti-
`IgG activity indicated that L chains of the
`appropriate specificity can substitute for the
`homologous chain in the formation of the ac-
`tive site. Failure of the other hybrid from this
`pair of proteins to complex with IgG may
`reflect the limited nature of this phenomenon.
`These observations bear certain similarities to
`results obtained with antibodies of the IgG
`class. Significant
`levels of antibody activity
`were recovered on recombination of y and L
`chains from the same antibody preparation or
`pool (5, 10, 19-22). Much lower levels of ac-
`tivity were detected when antibody of a dif-
`ferent specificity (5, 10, 22), antibody of the
`same specificity from a different immunoglob-
`ulin pool (20), or inactive immunoglobulin
`(19, 21) served as the source of y or L chains.
`However, certain experiments indicate that L
`chains from IgG antibodies of the same spe-
`cificity can substitute for the homologous chain
`under appropriate conditions for recombination
`(14). Heterologous recombination between the
`
`Reassembly of IgM Heavy and Light Chains 645
`
`heavy chains of a mouse myeloma protein,
`which binds 2.4-dinitrophcnyl ligands, and I.
`chains from rabhit anti-DNP antibodies resulted
`in hyhrid molecules with structural and func-
`tional properties that closely resembled those
`of the parent molecules (12).
`Egorov et al. (6) demonstrated that subunits
`produced from monoclonal TgM by reduction
`and alkylation dissociated into half subunits at
`pH lower than 6.0. This dissociation was also
`observed in pH 8.6 buffer at protein concen-
`trations lower than 2 mg/ml. Tht latter phe-
`nomenon was similarly characterized by Sol-
`heim & Harboe (28). These observations fol-
`lowed the earlier demonstration by Frank &
`Humphrey (11) that half subunits were released
`from rabbit IgM anti-Forssman antibody by
`reduction and alkylation at low protein con-
`centrations. The possible significanceof thehalf-
`subunit structure in the in vivo assembly of
`the molecule has been indicated by its prom-
`inence as an intermediate in the assembly
`of IgMs by mouse plasma cell tumors (2, 18).
`Although the relationship of in vitro reassembly
`of immunoglohulin chains to their assembly
`subsequent to biosynthesis remains to be estab-
`lished, the spontaneous formation of half sub-
`units from isolated chains and reassembly of
`the latter into 7S subunits suggests that for-
`mation of the IgM monomer requires a minimum
`of energy or control. In contrast, the hetero-
`geneity of the polymeric products obtained on
`reassociation of native subunits (26) or recon-
`stituted subunits suggests a regulatory mecha-
`nism at the cellular level. Evidence has been ob-
`tained that such a mechanism may involve a
`disulfide-exchanging enzyme and J chain (4).
`
`ACKNOWLEDGEMENTS
`The authors are indebted to Drs. Stephan E.
`Ritzmann, William J. Hammack, and J. Claude
`Bennett for the pathologic blood plasma used
`in this study. This work was supported by
`grants from the National Institutes of Health,
`United States Public Health Service, and the
`John A. Hartford Foundation, Inc.
`
`Merck Ex. 1040, pg 1166
`
`

`
`646 R. E. Schrohenloher Gf R.B. Hester
`
`REFERENCES
`t. BjOrk, I. & Tanford, C. Recovery of native
`conformation of rabbit immunoglobulin G upon
`recombination of separately renatured heavy and
`light chains at near-neutral pH. Biochemistry 10,
`1289, 1971.
`2. Buxbaum, J. & Scharff, M.D. The synthesis,
`assembly, and secretion of gamma globulin by
`mouse myeloma cells. VI. Assembly of IgM pro-
`teins. J. exp. Med. 138, 278, 1973.
`3. Cecil, R. & McPhee, J.R. Sulfur chemistry of
`proteins. Advanc. Protein Chem. 14, 255, 1959.
`4. Delia Corte, E. & Parkhouse, R.M.E. Biosyn-
`thesis of immunoglobulin A (IgA) and immuno-
`globulin M (IgM). Requirement for J chain and
`a