Eur. J. Immunol. 1980. 20: 467-476
`
`Immunoglobulin M mutants
`
`467
`
`11 Kohler, G., Hengartner, H. and Milstein, C., Protides Biol. Fluids
`Proc. Colloq. 1978. 25: 545.
`12 Galfrt, G., Milstein, C. and Wright, B. W., Nature 1979.277: 131.
`13 Galfrt, G., Howe, S. C., Milstein, C., Butcher, G. W. and How-
`ard, J. C., Nature 1977. 266: 550.
`14 Littlefield, J. W., Science 1964. 145: 709.
`15 Cotton, R. G. H., Secher, D. S. and Milstein, C., Eur. J.
`Immunol. 1973. 3: 136.
`16 Mechler, B. and Vassalli, P., J. Cell Biol. 1975. 67: 1.
`17 Cory, S. and Adams, J. M., J. Mol. Biol. 1975. 99: 579.
`18 Pelham, H. R. B. and Jackson, R. J., Eur. J. Biochem. 1976. 67:
`247.
`19 Lisowska-Bernstein, B., Lamm, M. E. and Vassalli, P., Proc. Nar.
`Acad. Sci. USA 1970. 66: 425.
`20 Baglioni, C. and Liberti, P., Mol. Biol. Rep. 1974. I: 329.
`21 Okuyama, A,, McInnes, J., Green, M. and Pestka, S., Biochem.
`Biophys. Res. Commun. 1977. 77: 347.
`
`22 Sakano, H., Rogers, J. H., Hiippi, K., Brack, C., Traunecker, A,,
`Maki, R., Wall, R. and Tonegawa, S., Nature 1979. 277: 627.
`23 Early, P. W., Davis, M. M., Kaback, D. B., Davidson, N. and
`Hood, L., Proc. Nat. Acad. Sci. USA 1979. 76: 857.
`24 Morrison, S. L., Eur. J. Immunol. 1978. 8: 194.
`25 Fleischman, J. M., Porter, R. R. and Press, E. M., Biochem. J .
`1963. 88: 200.
`26 Margulies, D. H., Kuehl, W. M. and Scharff, M. D., Cell 1976. 8:
`405.
`27 Kohler, G. and Milstein, C., Nature 1975. 256: 495.
`28 Mosmann, T. R., Baumal, R. and Williamson, A. R., Eur. J.
`Itnmunol. 1979. 9: 511.
`29 Kohler, G., Hengartner, H. and Shulrnan, M. J., Eur. J. Immunol.
`1978. 8: 82.
`30 Howard, J. C., Butcher, G. W., GalfrC, G., Milstein, C. and Mil-
`stein, C. P., Immunol. Rev. 1979. 47: 139.
`
`Georges Kohler and Marc J. Shulman'
`
`Immunoglobulin M mutants
`
`Basel Institute for Immunology, Basel
`
`Mutants of a hybridoma line secreting IgM with anti-trinitrophenyl specificity were
`enriched by coupling the hapten to the cell surface and incubating the cells in the
`presence of complement. In this way, cells secreting wild-type Ig commit suicide,
`whereas cells secreting IgM with reduced lytic activity preferentially survive.
`Mutants have been isolated which showed no or reduced synthesis of either the heavy
`or the light chain of the IgM molecule, mutants with internal and carboxy-terminal
`deletions of the p chain, and a mutant with an electrophoretically silent change of the
`p chain that reduces the affinity of the antibody for the hapten tenfold. The frequency
`to lo-'. All
`of mutants in the original population was estimated to be about
`mutants were stable with reversion frequencies of about
`or less.
`
`1 Introduction
`
`Immunoglobulin (Ig) production involves many steps: activa-
`tion of the Ig genes, transcription of these genes into RNA,
`RNA processing and translation into the Ig heavy (H) and
`light (L) chains, H and L chain assembly, glycosylation, and
`secretion. The resulting mature Ig themselves have several
`functions such as antigen binding, and, depending on the Ig
`class, complement fixation or Fc receptor binding.
`
`Many of the phenomena related to Ig production and function
`can be conveniently studied in tissue culture, using cell lines
`derived from myelomas and hybridomas for which Ig com-
`prises a major part of the total protein. Furthermore, using
`such cell lines, the biochemical basis of these phenomena can
`
`[I 27351
`Present address: Laboratory of Cell Biology, Rheumatic Disease
`Unit, Wellesley Hospital, Toronto, Ontario M4Y 1 J 3, Canada
`
`Correspondence: Georges Kohler, Basel Institute for Immunology,
`487 Grenzacher StraBe, Postfach, CH-4005 Basel 5, Switzerland
`
`Abbreviations: DF, IF, BSS-G, T medium, PBS and HAT: see Sect. 2.1
`SDS-PAGE: Sodium dodecyl sulfate polyacrylamide gel electro-
`phoresis TNP: 2,4,6-Trinitrophenyl TNBS: 2,4,6-Trinitrobenzene sul-
`fonic acid IEF: Isoelectric focusing SRC: Sheep red cells PFC: Plaque-
`forming cells C: Complement FACS: Fluorescence-activated cell
`sorter
`
`0 Verlag Chemie, GmbH, D-6940 Weinheim, 1980
`
`be elucidated by comparing the constituents of normal cells
`and mutants that are defective in some aspect of Ig production
`and function.
`
`Studying mutants affecting Ig synthesis in hybridoma cells has
`several advantages over other mammalian systems. The struc-
`ture of Ig and its basic gene organization [ 1 4 ] are known. The
`large amount of secreted Ig makes it easy to collect enough
`material for biochemical studies, and antigen-specific plaque
`assays allow the detection of single antibody-secreting cells.
`Another advantage is the fact that the synthesis of most other
`proteins results from the expression of both of the two
`chromosomes of diploid cells, and for such proteins most
`mutations will, therefore, be recessive. Only one of the two
`chromosomes directs Ig synthesis [7]. The other chromosome
`is silent, a phenomenon called allelic exclusion. Therefore,
`one could expect mutants in the Ig genes to be more easily
`obtainable than mutants defective in other proteins. This also
`implies that mutations in regulatory genes affecting Ig synthe-
`sis must be of the rare dominant type or they will remain
`unnoticed. Therefore, all mutants including the so-called reg-
`ulatory ones affecting, for example, the amounts of Ig, will
`most likely be alterations in the genes coding for the Ig L and
`H chains.
`
`The isolation of cells mutated in genes for nonessential pro-
`teins, such as Ig, has so far required single cell-screening
`techniques. For example, Scharff and his colleagues [8, 91 dis-
`cerned mutant clones making altered Ig by testing for anti-
`
`0014-298018010606-0467$02.5010
`
`Mylan v. Genentech
`IPR2016-00710
`Merck Ex. 1118, Pg. 1
`
`

`

`468
`
`G. Kohler and M. J. Shulrnan
`
`Eur. J. Immunol. 1980. 10: 467-476
`
`body-dependent precipitation around colonies in agar. Cotton
`et al. [lo] tested the isoelectric focusing (IEF) pattern of Ig
`secreted by individual myeloma clones. The maximal screen-
`ing capacity is about lo5 individual clones in the former exam-
`ple and lo4 individual clones in the latter. More recently, the
`fluorescence-activated cell sorter (FACS) has been used to
`enrich for rare variants representing one cell among lo4 in the
`initial culture [ll]. We have devised a selection procedure
`which depends on the secretion of cytolytic antibody and takes
`advantage of its antigen-binding specificity. In short, the anti-
`gen is coupled to the surface of the antibody-secreting cell, and
`the cells are then incubated in the presence of complement.
`Cells secreting wild-type antibodies commit suicide, whereas
`mutants which, e.g., secrete antibodies with reduced antigen-
`binding capacity survive and are thus enriched. Using this
`selection procedure in conjunction with single cell-screening
`techniques, we were able to screen relatively rapidly and easily
`about lo6 cells and to derive five distinct mutant cell lines.
`
`2 Materials and methods
`
`2.1 Media
`
`DF is Dulbecco's modified Eagle's medium (Gibco, Grand
`Island, NY), supplemented to contain 100 unitsiinl penicillin
`and streptomycin, 15% heat-inactivated fetal bovine serum
`2-mercaptoethanol. IF is Iscove's modified
`and 3.5 x
`Dulbecco's medium (Gibco), supplemented as above for DF
`medium. BSS-G contains 0.14 g CaC12, 8 g NaCl, 0.4 g KCl,
`0.2 g MgS04.7 H20, 0.2 g MgC12.6 H20, 0.06 g KHZP04,
`0.24 g Na2HPO4-2H20, and 0.01 g phenol red, per liter.
`T medium containing 10 g tryptone (Difco, Detroit, MI), 5 g
`NaCl, and 0.05 g yeast extract, per liter, was used for growth
`of E. Coli C3000 and plaquing of phage f 2. PBS contains 8.5 g
`NaCl, 1.28 g Na2HP04. 2 H20, and 0.386 g NaH2P04. H20,
`per liter. HAT indicates that the medium contains 13.6 mg
`hypoxanthine, 0.174 mg aminopterin, and 3.87 mg thymidine,
`per liter.
`
`2.2 Culture conditions
`
`Cells were grown in DF or IF medium. When it became neces-
`sary to grow cells either at very low density or in the presence
`of a large excess of dead cells, peritoneal exudate cells were
`added at lo5 cells/ml. These cells were prepared by flushing the
`peritoneum of a BALB/c mouse with 5 ml BSS-G. The result-
`ing cell suspension was centrifuged and washed twice in BSS-G
`and finally resuspended in medium at lo5 cells/ml.
`
`the mixtures contained rabbit anti-mouse p serum to develop
`the plaques [14]. Plaques were visible after several hours' incu-
`bation at 37°C. In general, plaques were larger in Cunning-
`ham chambers than in agarose. However, the mutant igx-215
`makes very small plaques in agarose, but no plaques (<
`plaque/cell) in Cunningham chambers.
`
`2.4 Preparation of radiolabeled Ig and analysis of
`Ig production
`
`Extracellular preparations were obtained by growing cells
`overnight in leucine-free D F medium supplemented to contain
`5 pCi (= 185 kBq) 14C-leucine/ml and 5% dialyzed fetal
`bovine serum. Cultures were centrifuged, and the supernatant
`was used as a source of secreted (extracellular) Ig. Intracellu-
`lar material was prepared by incubating the cells for 2 h in the
`above medium. Cells were harvested by centrifugation and
`lysed by the addition of 0.5% Nonidet-P 40. Where indicated,
`the extra- or intracellular material was reacted with a rabbit
`anti-mouse IgM serum to purify and concentrate the IgM.
`
`Radiolabeled Ig was analyzed by sodium dodecyl sulfate poly-
`acrylamide gel electrophoresis (SDS-PAGE) [15], IEF [16]
`and by two-dimensional IEF/SDS-PAGE [ 171. IgM production
`was also assayed by measuring lysis of protein A (Pharmacia,
`Uppsala, Sweden)-coupled red cells [14] using specific rabbit
`sera donated by Dr. F. Melchers. TNP-specific IgM was
`assayed by lysis or agglutination of TNP-coupled red cells [13].
`The sensitivity of the hemagglutination assay was increased by
`including p-specific rabbit serum (Litton Bionetics Inc., Ken-
`sington, MD). Plates for spot tests of hemolysis titer were
`prepared as described for plaquing.
`
`2.5 TNP-phage inactivation
`
`For the detection of anti-TNP activity, TNP-coupled phage f 2
`(kindly donated by Dr. H. Pohlit) was incubated with culture
`supernatants at various dilutions for 2 h at 37 "C. For measur-
`ing the affinity of the IgM for the free TNP, we used the
`method of Haimovich and Sela [18], whereby the fraction of
`IgM bound to the TNP hapten is measured by the inhibition of
`TNP-phage inactivation. The hapten TNP-caproic acid (pro-
`vided by Dr. H.-R. Kiefer) at various dilutions was preincu-
`bated for 10 min at room temperature with an appropriate
`dilution of culture supernatant. TNP-phages were then added
`and the incubation continued for 7 min at room temperature.
`After incubation, an aliquot was titered for plaques on E. Coli
`C 3000.
`
`2.3 Plaquing
`
`2.6 Classification of Ig from Sp6
`
`Plaque-forming cells (PFC) were titered either in agarose or
`Cunningham chambers [12]. For titering 2,4,6-trinitrophenyl
`(TNP)-specific cells by agarose plaquing, 2 ml of 0.6% agarose
`in PBS was added to a mixture containing 0.1 ml 25% TNP-
`coupled red cells [13], 0.1 ml guinea pig serum (@RAY,
`Behringwerke, Marburg/Lahn, FRG) and the cells to be
`titered. For plaquing in Cunningham chambers, a suspension
`containing 60 pl of the cells to be titered, 10 pl guinea pig
`serum, 20 pl of 25% TNP-coupled red cells was introduced
`into a chamber. For titering IgM-secreting cells, protein A-
`coupled red cells were used in place of the TNP-red cells, and
`
`The Ig of the hybridoma Sp 6 has been classified as IgM (n) for
`the following reasons. On SDS-PAGE, unreduced Ig has the
`mobility corresponding to a molecular mass of 920 kD, while
`the reduced Ig yields an H chain for which the mobility corre-
`sponds to 70 kD. Furthermore, the Ig from this cell line gives
`rise to direct hemolysis which is not enhanced by y-specific
`antisera and also effects direct agglutination of TNP-coupled
`red cells, and this agglutination is enhanced by p-specific anti-
`sera. The L chain has been classified as x by Ouchterlony tests
`using anti-x and -1 antisera kindly donated by Dr. A. Kelus.
`To avoid interference in this analysis by the myeloma n chain,
`
`Merck Ex. 1118, Pg. 2
`
`

`

`Eur. J. Immunol. 1980.10: 467-476
`
`Immunoglobulin M mutants
`
`469
`
`secreted material from the mutant line igm-10, which contains
`only the L chain (see Sect. 3), was reacted with these sera.
`Precipitation lines with anti-x but not with anti-h were seen.
`
`2.7 Fusion for complementation
`
`Cells of the indicated type (4.5 x lo6) grown in DF medium
`were mixed together, centrifuged (200 x g for 10 min), resus-
`pended in serum-free medium and recentrifuged. The cell pel-
`let was gently resuspended in 0.15 ml of a solution composed
`of 43% polyethylene glycol 1500 (British Drug Houses, Poole,
`GB), 43% serum-free Dulbecco's modified Eagle's medium
`(Gibco) and 14% dimethyl sulfoxide, and incubated for 2 rnin
`at 37°C. Then, 0.15 ml serum-free medium, at 37"C, was
`added over a 1-min interval, after which 20 ml serum-free
`medium at 37°C was added. The cells were incubated for
`10 min, then centrifuged and resuspended
`in 30 ml DF
`medium and incubated for 18 h. At this point, plaque-forming
`cells were titered in Cunningham chambers (3 X lo5 cells/
`chamber).
`
`2.8 Selection of mutants
`
`The mutants described here were derived from a hybridoma
`cell line that secretes IgM specific for the hapten TNP. The
`selection system involves first, attaching TNP to the mem-
`brane of the cells. Wild-type cells then secrete IgM that binds
`to the TNP on their own membrane, thus rendering the cells
`sensitive to complement (C) killing. By contrast, mutants
`unable to make TNP-specific IgM that can fix C survive the
`TNPiC treatment. The selection method described in detail
`below was devised after studying the effects of varying the
`temperature and times of incubation and TNP concentration
`and has since been used to generate mutants of the Sp603
`clone. The mutants of Sp602 were obtained using this selec-
`tion method under somewhat different conditions [2,4,6-trini-
`trobenzene sulfonic acid (TNBS) at 1.5 mglml instead of 1 mg/
`ml, incubated first cycle at approximately 23"C, second cycle
`at 4"C, both cycles for 20 min].
`To select the mutant cells, 1 x lo7 Sp 603 cells were washed in
`PBS and resuspended at 5 x lo6 cells/ml in 2 ml PBS contain-
`ing 1 mg/ml TNBS (British Drug Houses) and incubated at
`22-23 "C for 10 min, to allow TNP to attach to the cell surface.
`The cells were then diluted to 25 ml with cold PBS, cen-
`trifuged, washed once in PBS and then resuspended at 5 X 10'
`cellslml in DF medium containing 1% methyl cellulose
`(Methacel4000, Hoffmann-La Roche, Basel) and 10% guinea
`pig serum as a C source. This mixture was then incubated for
`30 min at 37°C to allow the wild-type cells to secrete IgM
`which would then bind to the TNP on the cell surface, thus
`rendering these cells sensitive to C-mediated killing. After this
`incubation, the cells were diluted to 150 ml with PBS, cen-
`trifuged, washed once in PBS and then resuspended in 6 ml
`medium. Because very few (< 1%) of the cells survived, 10'
`peritoneal exudate cellslml were added as fillericleaner cells.
`The cells were incubated for one day so that the dead cells
`would not make a plaque, after which time the cells were
`tested for their capacity to produce TNP or protein A plaques.
`The surviving cells were then grown to about lo7 cells and the
`selection repeated as above. In these subsequent TNP/C selec-
`tions, killing was not as extensive as for the original untreated
`population because the mutants comprised a greater fraction
`
`of the cells. For this reason, cells were resuspended in 15 ml
`medium without peritoneal exudate cells.
`
`To devise this selection system, we tested the effects of varying
`the TNBS concentration and incubation temperature. We
`monitored the effectiveness of the selection by treating a mix-
`ture of cells composed of a majority of Sp603 cells and a
`minority of other hybridoma cells, either the Spl/HL, Sp 2/HL,
`or Sp3/HL (which secrete sheep red cell (SRC)-specific IgM,
`IgG2,,, and IgGl, respectively) or the mutant igm-20 that has
`lost its capacity to synthesize the TNP-specific x chain (see
`Sect. 3). The extent of the enrichment for the non-TNP-
`specific cells, following the TNP/C treatment, was monitored
`by measuring the plaques on the appropriate indicator red
`cells (TNP-horse red cells, -SRC, or protein A-red cells). The
`extent of killing after the selective treatment was variable, and
`the enrichment observed was in the range of 20 to 200-fold.
`
`As described in Sect. 3, approximately 1% of the surviving
`cells secreted IgM which failed to lyse TNP-coupled red cells.
`To isolate these presumptive IgM-secreting mutants, the cells
`were cloned by one of three methods: (a) in methyl cellulose,
`over protein A-coupled red cells, in which case the IgM-pro-
`ducing clones made plaques, (b) in soft agar, in which case the
`IgM-producing clones were recognized by overlaying the col-
`onies with anti-y antiserum which caused the formation of a
`precipitate around the IgM-secreting colonies, or (c) by limit-
`ing dilution in microtiter wells, in which case the IgM-secreting
`clones were scored by testing whether culture fluid would lyse
`protein A-coupled red cells in the presence of anti-y develop-
`ing serum and C.
`
`2.9 Isolation and cloning of mutants
`
`The mutants igk-1, igm-12, and igm-21 were obtained from the
`Sp602 clone as follows. The Sp602 cells that survived two
`cycles of TNP coupling and C treatment were cloned and
`plaqued in methyl cellulose in 35-mm dishes, as follows. A
`1-ml layer containing 0.5% indubiose in DF medium (igk-1,
`igm-12) or IF medium (igm-21) was used to provide a flat
`surface. Over this was layered 0.2 ml of 0.3% agarose (L'In-
`dustrie Biologique FranGaise S. A., Genevillier, France) in the
`corresponding medium containing 10 yl C (lyophilized guinea
`pig C: BRAY, Behringwerke) and 10 ~1 SRC coupled with
`protein A (for the isolation of igk-i and igm-12) or TNP (for
`the isolation of igm-21). Finally, 1 ml of 1% methyl cellulose
`in medium containing the surviving Sp602 cells was added.
`Plates were incubated for 2-3 days, after which time the PFC
`were transferred with a micropipet to 15 pl medium containing
`lo5 peritoneal exudate cells/ml. These isolates were then incu-
`bated and diluted until they had grown to lo6 cells. Isolates
`secreting Ig which differed from wild type were then recloned
`in soft agar [lo].
`
`The mutant igx-215 was obtained from the Sp603 clone as
`follows. After subjecting the Sp 603 cells to three cycles of
`TNP coupling and C treatment, we attempted to enrich for
`these IgM producers by using the FACS: cells were exposed to
`a rabbit antiserum and then stained with a fluorescein-labeled
`sheep anti-rabbit antiserum. Approximately 1% of the cells
`were stained, passed through the sorter, and the brightest 2%
`of the cells were cultured further. Whereas approximately
`50% of the cells in this preparation were stained by the anti-y
`treatment, only 1% of the cells secreted IgM, as measured by
`
`Merck Ex. 1118, Pg. 3
`
`

`

`470
`
`G. Kohler and M. J. Shulman
`
`Eur. J. Immunol. 1980. 10: 467-476
`
`the frequency of protein A plaques, implying that the FACS
`sorting led to no enrichment for IgM-secreting cells. Neverthe-
`less, this preparation was cloned by plating the cells in 35 mm
`dishes in 0.2 ml 0.22% agarose in D F medium layered over
`1.5 ml 0.5% agarose in D F medium. After 1 week, the col-
`onies were scored for IgM secretion by overlayering 0.2 ml
`containing 0.17% agarose, 1.4-fold concentrated Dulbecco's
`modified Eagle's medium and 8 pl rabbit anti-y antiserum.
`After 1 day, colonies surrounded by a precipitate were trans-
`ferred to 0.1 ml DF medium and grown to mass culture.
`
`Culture fluid from 15 such isolates lysed protein A- but not
`TNP-coupled red cells; in SDS-PAGE, the H and L chains
`appeared to be normal. The IEF pattern of 6 of these mutants
`was the same as for wild type. We supposed that these isolates
`all derived from the same initial variant cell and limited our
`analysis to reclones igx-209 and igx-215 of one such isolate.
`
`The mutant igm-692 was isolated as follows. The Sp603 cells
`which survived 4 cycles of TNP coupling and C treatment were
`diluted to a concentration of 20 cellsiml, in D F medium sup-
`plemented to contain 2000 HAT-sensitive BW 5147 cells/ml
`added as filler cells. This mixture was distributed in 0.1 ml
`aliquots in microtiter wells. After 3-4 days, the HAT-sensitive
`BW 5147 cells were killed by adding 0.1 ml HAT-sup-
`plemented D F medium. Supernatants were tested for IgM by
`spotting them onto a protein A-SRC indicator layer. After
`about 1 week, the hybridoma cells had grown to high density
`and were transferred to fresh HAT-containing medium. For
`the following 2 weeks, they were diluted in HT-containing DF
`medium and finally, in normal DF medium. Three mutant
`isolates were found to have the property of secreting an H
`chain of the same higher than normal mobility in SDS-PAGE.
`One such isolate, igm-692, was selected for further analysis.
`
`3 Results
`3.1 Derivation of the wild-type cell line
`
`the low secretion of K. In contrast to the secreted material
`which contains a low amount of K, comparable amounts of K
`and L are observed in intracellular preparations. Similarly,
`cell-free in vitro synthesis directed by mRNA of Sp602 cells
`yielded equal amounts of K and L (Hirama and Kohler,
`unpublished results). These results suggest that the low level
`of K secretion does not reflect a defect in K synthesis. The K
`chain can only be secreted when attached to an H chain [19],
`whereas L can also be secreted alone. The differences in
`amount of K secreted by the reclones of Fig. 1 might, there-
`fore, be explained either by enhanced degradation of K or by
`reduced affinity of K for the H chains.
`
`Figure I. Ig from Sp6 clones. Radio-
`labeled culture supernatants of Sp 6 clones
`were reduced and analyzed by SDS-
`PAGE as described in Sect. 2.4.
`
`3.2 Enrichment for mutants altered in Ig production
`
`The selection scheme is designed to enrich for mutants that
`secrete either less IgM or else IgM that is reduced in TNP
`binding or C fixation (see Sect. 2.8). To effect this selection,
`the population of cells is treated so that TNP is covalently
`bound to the cell membrane. Under conditions of limited dif-
`fusion, the wild-type cells secrete IgM that binds to the TNP
`on their own membrane, so that in the presence of C, the wild-
`type cells are killed. By contrast, the mutants survive this
`treatment and are thus enriched.
`
`I
`
`Fig. 1 shows the SDS-PAGE pattern of Ig from 7 TNP-specific
`. -
`clones derived from the original Sp6 pot. Two kinds of sub-
`clones were observed: those, such as C2, which secrete com-
`parable amounts of the TNP-specific (L) and nonspecific (K)
`light chains, and those, such as Al, which secrete less K. For
`several reasons, we consider the C 2 type to be the parental
`phenotype and A1 a subsequent variant: most hybridomas
`obtained with X63-Ag8 are of the C2, rather than A1 type.
`Furthermore, of 8 reclones of Sp 6-C 2, 6 showed the parental
`C 2 phenotype, whereas 2 secreted a low level of K, similar to
`the Sp6-A1 phenotype. By contrast, all of the 48 reclones of
`Sp6-A1 had the low level K phenotype. One clone (Sp6-AI-
`14) did not express the y chain and was recloned twice. From
`this second recloning, we obtained the clones Sp602 and
`Sp 603 from which the mutants described below were derived
`(see also Scheme 1).
`In the course of these experiments, we have made several
`observations which might help to understand the reason for
`
`cells were titered by measuring plaques on TNP-coupled red
`').
`
`Randomly picked clones of twice-treated Sp 602 revealed that
`the non-plaque-forming population consisted of cells which do
`not express the p chain (see description of igm-10). Prior to the
`selection, these y- variants would be expected to represent
`about 1% of the cells [20]; in fact, screening of 185 reclones of
`untreated Sp 603 revealed one such y- variant. After several
`cycles of this selective treatment, the frequency of IgM-pro-
`ducing (protein A PFC) among the y- cells was 1% (Table 1,
`bottom line). This fraction includes mutant cells making
`altered IgM as well as wild-type cells. The frequency of wild-
`type cells is estimated by TNP plaque formation to be less than
`0.003% for the Sp603 populations. Assuming that the ratio of
`mutant cells to y- cells is the same before and after selection,
`we estimate the frequency of the mutants among wild-type
`cells in the original population to be:
`(Mutantdp-) X (y-/wild type) = 0.01 x 0.01 = 1 x
`
`Merck Ex. 1118, Pg. 4
`
`

`

`Eur. J. Immunol. 1980. 10: 467-476
`
`Table 1. Effect of suicide selection")
`
`Line
`
`sp 602
`
`s p 603
`
`Selection
`round
`
`0
`1
`2
`0
`1
`2
`3
`4
`
`Survival
`
`TNP plaques
`
`( %)
`-
`0.2
`30
`
`n.d.
`n.d.
`n.d.
`n.d.
`
`(%)
`
`100
`4
`1
`100
`56
`3.5
`0.w
`5 0.003
`
`Prot. A
`plaques
`(Q)
`
`100
`6
`4
`1 (X)
`70
`7
`4
`I
`
`a) The population of Sp 602 or Sp 603 cells was subjected to the TNP/
`C selection (see Sect. 2.8). The plaque-forming capacity is ex-
`pressed as plaquesicell surviving the selection, divided by plaques/
`cell of the untreated population. Plaques were measured in Cun-
`ningham chambers. The protein A plaques reveal cells secreting
`IgM irrespective of its antigen-binding capacity; TNP plaques
`depend on cells secreting TNP-specific IgM. n.d. = not deter-
`mined.
`
`3.3 Isolation of variants
`
`As described above, the selected population was comprised
`mainly of p- variants, and in order to reveal IgM-secreting
`clones, several approaches have been used. In the case of the
`twice-treated Sp 602 population (Table l), mutants still mak-
`ing IgM were sought by plating the cells in methyl cellulose for
`protein A plaques with p-specific developing serum. Assuming
`that there might be mutants making IgM that retained some
`TNP-binding activity, cells were also plated for TNP plaques
`(see Sect. 2.3). Of 15 clones derived from TNP plaques, only
`
`Table 2. Properties of the mutant cell lines")
`
`Immunoglobulin M mutants
`
`471
`
`one (igm-21) behaved differently from the wild-type line. Of
`the 20 clones picked from the center of protein A plaques, 12
`could be grown to mass culture. Three kinds of variants could
`be isolated (igk-1, isolated 10 times, igk-20, isolated once, and
`igm-12, isolated once, see also Scheme 1).
`After 4 cycles of suicide selection of Sp603, 26 clones were
`selected reacting with specific rabbit anti-mouse-p serum (see
`Sect. 2.3). Of these, 5 were found to be of the igk-1 type, and
`3 of the igk-20 type. Two new kinds of variants were found;
`igx-215 (isolated 15 X ) and igm-692 (isolated 3 X). All isolates
`were recloned before analysis. The mutants were then sub-
`jected to various tests to measure IgM production and antigen
`binding (Table 2 and Fig. 2).
`
`3.4 Mutants affecting L chain production
`
`igk-20. The IgM secreted by this cell line has no TNP-binding
`activity when assayed by hemagglutination or hemolysis of
`TNP-coated SRC, or by the more sensitive assay of inactiva-
`tion of TNP-conjugated phage f 2 (Table 2). SDS-PAGE
`analysis of intra- and extracellular reduced material indicates
`that this mutant makes no TNP-specific L chain (Fig. 2 a). By
`contrast, the myeloma K chain is synthesized and secreted in
`amounts characteristic of the parent Sp6MLGK (C2 type,
`Fig. 1) cell line. In Sect. 4, we shall return to the question of
`why the mutations expressing no L chain should occur less
`frequently than mutants making no p chain.
`
`igk-1. This mutant makes a small amount of the TNP-specific
`L chain, intra- as well as extracellularly, and secretes larger
`amounts of K than the parental lines (Figs. 2 and 3). Among
`64 clones derived from igk-1, two (igk-1-1, igk-1-2) had lost p
`expression (Fig. 3). The result shows that the K chain has
`retained its parental trait not to be secreted as free K.
`
`Line
`
`Prot. A lysis
`
`sp 603
`
`igm-12
`igm-692
`igx-215
`igx-209
`igk-l
`igk-20
`igm-I0
`
`1:25
`
`1 :2'
`1 :23
`1:2"
`1:P
`1 :25
`1 :25
`none
`
`TNP aggl.
`Direct
`Indirect
`(%,
`(%)
`loo
`(1:26)
`2s
`< 6
`< 2
`< 2
`< 2
`< 2
`< 2
`
`100
`(1:21")
`50
`6
`3
`1
`2
`<0.2
`<0.2
`
`TNP lysis
`
`TNP plaques
`
`(c/.)
`
`100%
`(1:2")
`< 2%"
`< 2%
`< 0.4%
`< 0.4%
`(0.4%
`< 0.3%
`< 0.4%
`
`1 h'
`
`2 x 10 I
`< 1 x lo->
`1 x 1 0 I
`4 x 10-j
`1 x 10-5
`3 x 10
`n.d.
`
`TNP-phage
`inactivation
`(5%)
`
`Frequency of
`occurrence
`in Sp602 in Sp603
`
`100
`
`80
`8
`60
`40
`5
`< 0.2
`< 0.2
`
`1/12
`o/ 12
`n.d.
`0/12
`10/12
`1/12
`n.d.
`
`0126
`3/26
`n.d.
`15/36
`5/26
`3/26
`n.d.
`
`a) The radiolabeled culture supernatants analyzed in Fig. 2 were tested as follows. For the columns protein A lysis, TNP
`aggl., and TNP lysis, the dilution endpoint of lysis or agglutination was measured. For TNP-phage inactivation, the
`supernatant concentration resulting after 2 h incubation at 37°C in SO% inactivation of TNP-coupled phage f 2 was taken.
`To compare the activity per unit IgM for mutant and wild-type culture supernatants, the protein A lysis titer was taken as a
`measure of total IgM concentration. The values presented for TNP aggl., TNP lysis, TNP-phage inactivation were
`calculated as:
`Mutant titerimutant protein A lysis titer
`x 100.
`Wild-type titedwild-type protein A lysis titer
`The numbers in parentheses are the wild-type titers. TNP plaques give the plaquesicell obtained by the mutant cells
`divided by the TNP plaqueskell of the wild-type culture. Plaquing was done in agarose. The last 2 columns indicate how
`often a particular mutant phenotype was revealed in the primary isolates after the suicide selection.
`b) 160 plaques/103 cells.
`c) 1-376 TNP-lysis activity was observed in other preparations.
`
`Merck Ex. 1118, Pg. 5
`
`

`

`472
`
`G. Kohler and M. J. Shulman
`
`Eur. J. Immunol. 1980. 10: 467-476
`
`In contrast to igk-20 IgM, the TNP-binding activity is easily
`detected by
`the inactivation of TNP-coupled phage f 2
`(Table 2); from this, we argue that igk-1 must be making some
`TNP-specific L. We have also analyzed the L chains secreted
`by igk-1 in two-dimensional electrophoresis using first IEF and
`then SDS-PAGE. Material moving to the region characteristic
`of L and K on IEF also moves to the characteristic region on
`SDS-PAGE implying that both chains are indistinguishable
`from the corresponding parental L and K (Fig. 4). The igk-1
`cell line gave TNP-specific plaques at an efficiency lo5 lower
`than the parental cell line suggesting that this mutant is quite
`stable (Table 2).
`
`3.5 Mutants affecting p chains
`
`igm-10. Both intra- and extracellular preparations were found
`by SDS-PAGE to lack p (Fig. 2). The myeloma K chain is not
`secreted except when bound to an H chain [19] and, therefore,
`is observed only in intracellular preparations but not in the
`secreted material. Membrane staining with anti-mouse
`p-specific, fluoresceinated rabbit serum did not reveal any
`positive cells among 5000 cells examined.
`
`igm-21. We have analyzed the Ig made by the mutant igm-21
`in several ways. Extra- and intracellular preparations were
`reacted with rabbit anti-mouse IgM serum, which was largely
`specific for the mouse p chain. SDS-PAGE of the reduced
`precipitates indicated that igm-21 makes a smaller amount of 1
`
`Figure 2. Extra- and intracellular Ig of mutant clones. Radiolabeled
`culture supernatants (extra) and intracellular preparations were
`reacted with rabbit anti-mouse IgM. The precipitate was solubilized
`and analyzed by SDS-PAGE under (a) reducing and (b) nonreducing
`conditions. (The mutant igx-833 is an isolate that behaved like igx-215;
`Sp 2515-1-Ag5 was included as IgG reference protein.)
`
`Figure 3. Ig production by igk-1. Radiolabeled culture supernatants
`(extra) and intracellular preparations were reacted with rabbit anti-
`mouse IgM, and the resulting preclpitates were reduced and analyzed
`by SDS-PAGE.
`
`Figure 4. Two-dimensional gel electrophoresis of igk-1. First dimen-
`sion IEF from pH 5 (left) to 9 (right) of the radiolabeled and reduced
`igk-1 Ig. The igm-10 L chain is shown as reference. Both samples were
`run side by side on an IEF glass plate, fixed, dried and autoradio-
`graphed. The strip of gel containing the igk-1 Ig was located with the
`help of the autoradiograph of the IEF pattern, peeled off the plate,
`and subjected to SDS-PAGE under reducing conditions [ 171.
`
`chain (Fig. 5). Most L chains were not precipitated by the
`antiserum and remained in the supernatant. Small aliquots of
`the supernatants wer