Proc. Nail. Acad. Sci. USA
`Vol. 81, pp. 3180-3184, May 1984
`Immunology
`
`Generation of antibody diversity in the immune response of
`BALB/c mice to influenza virus hemagglutinin
`(antibody sequences/antibody genes/somatic mutation)
`DAVID MCKEAN*, KONRAD Huppit, MICHAEL BELL*, LOUIS STAUDTt, WALTER GERHARDt,
`AND MARTIN WEIGERTt§
`*Department of Immunology, Mayo Medical School, Rochester, MN 55901; tThe Institute for Cancer Research, 7701 Burholme Avenue, Philadelphia, PA
`19111; and tWistar Institute of Anatomy and Biology, 36th and Spruce Streets, Philadelphia, PA 19104
`
`Communicated by Matthew D. Scharff, January 30, 1984
`
`We have examined the amino-terminal se-
`ABSTRACT
`quence of the Kc light chains of a set of monoclonal antibodies
`specific for one of the major antigenic determinants (Sb) on
`the influenza virus PR8[A/PR/8/34(HlNl)J hemagglutinin
`molecule. This set was believed to be structurally related from
`earlier serological analysis that typed these K chains as mem-
`bers of the variable (V) region VK21 group [Staudt, L. M. &
`Gerhard, W. (1983) J. Exp. Med. 157, 678-704]. Our se-
`quence analysis confirms and extends this conclusion; all ex-
`amples of this set belong to a subgroup of the VK21 group,
`V,,21C. A special feature of this set of K light chains is that all
`examples were derived from the same mouse (designated H36).
`This sequence analysis along with the characterization of gene
`rearrangements at the K light chain loci of these hybridomas is
`consistent with the idea that certain members of this set are the
`progeny of one or two lymphocytes. Because of this potential
`clonal relationship, we can reach several conclusions about the
`diversity observed among these K light chains: (') the diversity
`is due to somatic mutation, (it) somatic mutations occur se-
`quentially and accumulate in the first complementarity-deter-
`mining region, and (iii) the extent of somatic variation in this
`sample is high, suggesting a somatic mutation rate of about
`10-3 per base pair per generation.
`
`Antibody diversity arises from several sources. Individuals
`inherit multiple variable (V) region gene segments for both
`heavy (VH) and light (VK, VA) chains, joining (J) gene seg-
`ments (JH, J,, Jh), and diversity (D) gene segments (DH).
`The initial antibody repertoire of an individual is a product of
`the combinatorial joining of these gene segments, i.e., V4s
`with JKS or different VH, DH, and JH combinations, that form
`complete V1 or VH genes. Errors committed during the pro-
`cess of joining contribute additional diversity to this reper-
`toire (reviewed in ref. 1). Finally, that somatic mutation fur-
`ther amplifies this germ-line repertoire seems to be estab-
`lished (2). The original evidence for somatic mutation
`favored a model by which point mutations accumulate se-
`quentially during cell division (3). Other models link somatic
`mutation with specific events during lymphocyte differentia-
`tion (4, 5) and propose cataclysmic mechanisms of mutagen-
`esis that introduce multiple amino acid substitutions in one
`step (6, 7). These models are based on the comparison of V
`region sequences of independently induced plasmacytoma
`and hybridoma antibodies to their putative germ-line coun-
`terparts. Hence, little can be concluded about the time
`course of somatic mutation.
`A better understanding of the nature of somatic mutation
`can be reached by comparisons of the V genes of a cell lin-
`eage. Scharff and colleague have analyzed certain mutants
`and revertants of the cell line S107 and conclude that the in
`
`vitro rate of mutation at the VH gene expressed in this plas-
`macytoma is significantly higher than that of nonimmuno-
`globulin genes (8). A possible in vivo analogy is described
`here: we have initiated a structural comparison of hybrid-
`omas derived from a single mouse and have identified a set
`or sets, the members of which may be clonally related. The
`pattern of variability observed so far suggests that somatic
`mutations accumulate sequentially and that in vivo somatic
`mutation occurs at a high rate.
`
`MATERIALS AND METHODS
`Anti-Hemagglutinin (HA) Hybridomas. Twenty-four days
`prior to fusion, a BALB/c mouse, H36, was primed by intra-
`peritoneal injection of 1000 hemagglutinating units of PR8.
`Three days prior to fusion, an intravenous injection of the
`same dose of virus was administered. The procedures for
`fusion, in vitro growth, and serological characterization of
`the H36 panel of hybridomas has been described (9). The
`fusion partner was Sp2/0-Agl4 (10).
`Protein Purification and Sequence Analysis. Hybridomas
`were grown in (BALB/c x NZB)F1 mice that had been Pris-
`tane-primed (Aldrich). Immunoglobulins were isolated from
`ascitic fluid by using protein A-Sepharose (Pharmacia) (11).
`IgG1 hybridoma antibodies were isolated by twice precipi-
`tating with 50% saturated ammonium sulfate and chromato-
`graphing on a Bio-Gel A-1.5m column (Bio-Rad) in 0.01 M
`sodium phosphate/0.9% sodium chloride/0.01% NaN3, pH
`7.2. Heavy and light chains were separated by Sephadex G-
`100 gel filtration in 6.0 M urea/1.0 M acetic acid after a 90-
`min reduction with 15 mM dithiothreitol and a 60-min alkyl-
`ation with 35 mM iodoacetamide in 0.2 M Tris-HCl (pH 8.0)
`at 5.0 mg/ml of protein.
`The strategy for the amino acid sequence determination of
`the VK21C light chains has been reported (12). Briefly, the
`amino-terminal 42-50 residues of each light chain were de-
`termined from sequence analysis of the intact light chain.
`Light chains (H36-15 and H36-18) that contained methionine
`at residue 33 were cleaved with cyanogen bromide, the frag-
`ments from residue 34 to residue 171 were purified on G-50
`Sephadex, and the amino-terminal 40 residues were identi-
`fied by sequence analysis. The light chains H36-5 and H36-7
`were cleaved at tryptophan residues with cyanogen bromide
`(13), the fragment from residue 36 to residue 144 was puri-
`fied on G-50 Sephadex, and the amino-terminal 40 residues
`were identified by sequence analysis. The tryptic fragments
`from residue 69 to residue 100 were isolated from light chains
`H36-5, H36-7, H36-15, and H36-18 on G-50 Sephadex and
`were sequenced completely.
`DNA Hybridization Analysis. The DNA probes were isolat-
`ed from a cloned 12.7-kb fragment containing the constant
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement"
`in accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`Abbreviations: V, variable; C, constant; D, diversity; J, joining;
`HA, hemagglutinin; kb, kilobase(s).
`§To whom reprint requests should be addressed.
`
`3180
`
`Lassen - Exhibit 1034, p. 1
`
`

`

`Immunology: McKean et aL
`
`Proc. Natl. Acad Sci. USA 81 (1984)
`
`3181
`
`(C) region K light chain locus (CK) (14). From this cloned
`BamHI fragment, a 0.9-kilobase (kb) EcoRI fragment (desig-
`nated pR1) was subcloned (14), and a 1.1-kb Xba I/HindIII
`fragment (designated IVS) was isolated from a low-tempera-
`ture-melting agarose (Sigma) gel. The physical maps of these
`probes are shown in Figs. 2 and 3 Lower. DNA was isolated
`from hybridomas, plasmacytomas, or 12- to 14-day mouse
`embryos (15). The DNA samples (usually 10 pug) were digest-
`ed to completion with the appropriate restriction enzyme(s)
`(Bethesda Research Laboratories), electrophoresed through
`0.7% agarose, transferred to nitrocellulose paper (Schleicher
`and Schuell), and hybridized to 32P-labeled DNA probes.
`Hybridizations were performed as originally described by
`Southern (16) with the modifications of Wahl et al. (17). Af-
`ter a final wash in 45 mM NaCl/4.5 mM sodium citrate/0. 1%
`sodium dodecyl sulfate for 1 hr at 650C, filters were exposed
`to x-ray film at -70'C.
`
`RESULTS
`The focus of this study is the structure of the K light chains
`from seven hybridomas (H36-1, -4, -5, -7, -15, -17, and -18)
`derived from an individual adult BALB/c mouse, H36. In
`particular, these antibodies bind to the same antigenic re-
`gion, Sb, on the HA molecule of influenza virus but can
`clearly be differentiated from each other by paratypic and
`idiotypic analysis. Tissue culture fluids from six of these hy-
`bridomas were serologically positive for the VK21C subgroup
`of K chains (9). Subsequent serological analysis with purified
`antibody showed that the H36-4 light chain is also cross-re-
`active with VK21C-specific antisera (18). The heavy chain
`isotype of each H36 antibody is shown in Fig. 1.
`A survey of 45 VK21 amino acid sequences from NZB and
`BALB/c plasmacytomas has so far divided the VK21 light
`chain group into eight subgroups, each subgroup being de-
`fined as a set of VK21 chains that share certain amino acid
`residues between positions 1 and 96. The closely related
`VK21B and -C subgroups are defined by the residues indicat-
`
`ed in Fig. 1, and VK21C in turn can be distinguished from
`VK21B by its own characteristic set of amino acid residues
`(12, 21, 22). All of the H36 V region sequences completed so
`far contain the VK21C-specific residues as well as the resi-
`dues shared by the VK21B and VK21C subgroups.
`These sequence data confirm the serological results in
`demonstrating that the H36 VK regions are members of the
`VK21C subgroup. The prototype sequences of the VK21 sub-
`groups were originally defined by recurrent sequences. For
`example, the VK21C prototype sequence (VK21C0) is the se-
`quence shared by 6 of the 12 VK21C light chains from plas-
`macytomas and the anti-HA hybridoma light chain H2-6C4
`(Fig. 1). Prototype sequences are thought to be encoded by
`the VK21 germ-line genes, a conclusion confirmed by the
`DNA sequence of several VK21 genes, including the authen-
`tic VK21CO (20). The H36 VK sequences each differ at multi-
`ple residues from the VK21C0 prototype sequence. Since
`each of the H36 VK sequences share the VK21C subgroup-
`specific residue(s), the substitutions found in the H36 se-
`quences must have resulted from somatic mutation of the
`V,21C0 gene. The only other possibility, namely that these
`sequences are mutants of a second "VK21C-like" germ-line
`gene, is unlikely, as none of the VK21 germ-line genes de-
`fined to date (20, 21) resembles such putative V,2JC-like
`gene products. Furthermore, independently isolated V,21C
`genes in either the unrearranged or aberrantly rearranged
`form code exactly for VK2JC0 (20, 23). Such surveys, being
`independent of light chain expression and hence free of con-
`straints due to selection, should have yielded such V,2JC-
`like genes. Hence, we believe the H-36 K chains are somatic
`mutants of the VK21C0 germ-line gene.
`K Light Chain Gene Rearrangements of the H36 Hybrid-
`omas. A relationship between certain H36 hybridomas has
`been established from the nature of rearrangements at the K
`light chain locus of these cell lines. These rearrangements
`are detectable by Southern blot analysis of hybridoma DNA
`digests using as probes segments of DNA near JK (IVS and
`
`1
`1
`2 2
`2
`2 3
`4
`3
`4
`1 ...5....0O....5 ....0.... 5.7....8.0 .... 5....0.... 5.
`
`1
`1
`1
`6
`9
`8
`8
`7
`5
`7
`5
`9
`6
`0
`0
`0
`.0.... 5....0.... 5....0.... 5....0....5....0.... 5....0 ....5..8
`
`CDRI
`
`CDR2
`
`CDR3
`
`IgA
`
`+
`+
`*
`*
`*
`+
`*
`+
`DIVLTQSPASLAVSLGQRATISCRASESVDSYGNSFMHWYQQKPGQPPKLLIYRASNLESG IPARFSGSGSRTDFTLTINPVEADDVATYYCQQSNEDFtYTFGGGTKLEIKR
`Iga--------------
`IKG2a.---..----------------------------------------------.-----__
`__F--S------
`___----------------_ F---S----
`-_
`-_- --S--
`
`_
`
`---------
`
`Till
`3741
`Al7
`2242
`10916
`M321
`T124
`8982
`8153
`
`H36-7
`H36-5
`H36-17
`H36-4
`H36-1
`H36-15
`H36-18
`
`IgA
`
`IgG3
`IgG3
`IgG3
`
`1-1
`
`,.Z..,
`
`ii--- t
`H-
`. ------ V- lFr DNI)
`- .----------
`IgGI
`r--------_-- -s--)---- -4
`. --------- ---------------
`IgG2a
`Ig G2a ----
`
`-
`
`-J fl-F I----------------------
`
`
`
`-----------______-
`
`____________________-____
`
`V J
`
`H2-6C4 IgM
`
`Amino-terminal sequences of the K light chains from H36 anti-HA (Sb) hybridomas and VK21C-expressing plasmacytomas. The Tll1
`FIG. 1.
`sequence (12), noted by the one letter code of Dayhoff (19), is identical to the V region sequence encoded by the VK21C germ-line gene (20).
`Dashed lines represent amino acid sequence identities with the VK21C prototype sequence. Overbars delineate the regions of the K light chain
`complementarity-determining (hypervariable) regions. Residues 30, 68, 74, and 80 (+) are the set of residues that is uniquely shared by the
`closely related VK21C and VK21B subgroups (21); residues 50, 58, 76, and 83 (*) are the set of residues that together defines the members of the
`VK21C subgroup. Encircled residues are substitutions unique to a given VK21C light chain; residues in boxes are substitutions shared between
`the light chains. Plasmacytoma sequences are taken from refs. 12, 21, and 22. M321 and T124 are Bence-Jones proteins. The PC8153 amino acid
`sequence is translated from the VK gene expressed in this plasmacytoma (unpublished data). The sequence of H2-6C4 is from ref. 12.
`
`Lassen - Exhibit 1034, p. 2
`
`

`

`3182
`
`Proc. Natl. Acad Sci. USA 81 (1984)
`Immunology: McKean etaLP
`
`Lo - ,;I m
`f,
`c)
`rl-
`U-
`7
`7
`7
`r) r) rPO rorOr
`n <
`T I I I II I Co
`
`*
`
`12.7kb -
`
`I-uCO
`U
`
`Em
`
`.I---
`
`F
`o
`u
`Li
`
`E
`0
`mn
`
`1 kb
`i a:FE
`o
`o
`u u
`LU
`Lu
`_IL
`
`IZZZZZZ_ ----rfl~r
`
`p R
`
`Detection of upstream (U) K light chain gene segments in
`FIG. 3.
`the H36 hybridomas. Southern blot of BamHI-digested DNA from
`hybridomas hybridized with the pRI probe. All hybridomas share a
`12.7-kb band corresponding in size to the K1, BamHI fragment that
`includes upstream DNA. This could originate from the H36 hybrid-
`omas and/or contaminating host tissue in these in vivo grown cell
`lines. H36-1, -4, -15, and -18 share a unique band (labeled U) hybrid-
`izing with the pR1 probe. The star denotes a faint, reproducible frag-
`ment in the H37 hybridoma.
`
`bridomas is significant because, in a survey of 75 plasmacy-
`tomas and hybridomas of independent origin, no three
`examples chosen at random had the same sizeK-fragment
`(unpublished data).
`The DNA upstream of JK in the H36 hybridomas was ana-
`lyzed with the pR1 fragment as a probe (Fig. 3). This region
`of the K locus is often retained in hybridomas and plasmacy-
`tomas and is found in BamHI fragments of a unique size rela-
`tive to the germ-line fragment (24). Four of the H36 series
`(H36-18, -15, -1, and -4) have indistinguishable upstream ele-
`ments. H36-7 has a unique upstream element. This is seen as
`a faint band (marked by a star in Fig. 3). Because the hybrid-
`ization to this fragment is weak, we believe this may be due
`to a contaminating subpopulation of cells in the H36-7 line.
`As in the case of K- fragments, the equivalent size of up-
`stream elements is a significant feature of this set of hybrid-
`omas as compared to plasmacytomas and hybridomas of in-
`dependent origin.
`For comparison, other PR8 HA-specific hybridomas have
`been analyzed (unpublished data). These include examples
`typed as VK21C but from different fusions and non-VK21C
`examples from the H36 fusion. The size of the K+ rearrange-
`ments of these VK21C examples is consistent with a VK21C
`rearrangement to JK gene segments other than JK2, whereas
`non-VK21C examples do not coincide with any VK2JC-JK re-
`arrangement. K- and upstream fragments are usually of dif-
`ferent sizes in independently derived examples. Certain hy-
`bridomas derived from the same fusion do share K- and up-
`stream fragments (H37-63 and H37-77 or H36-101, -12 and
`-6; ref. 9) as in the case of the examples from the H36 VK21C
`series.
`
`DISCUSSION
`In adult BALB/c mice, the secondary antibody repertoire to
`the influenza virus HA is large. Based on the repeat frequen-
`cy of isolation of paratypically indistinguishable hybridoma
`antibodies from different mice, one can estimate that
`BALB/c mice are able to produce at least 300 different anti-
`
`pR1, Figs. 2 and 3). The rearrangement bearing the produc-
`tive K light chain allele (K+) has been examined in BamHI/
`HindIII double digests of H36 hybridoma DNA. Since the
`VK2JC gene contains two BamHI sites that map to amino
`acid residues 60 and 95 of this subgroup (20), the K+ rear-
`rangement is seen as a 3- to 4-kb fragment by using the IVS
`probe (Fig. 2). Most of the H36 hybridomas have K+ rear-
`rangements that are indistinguishable from that of the VK21C
`(JK2) plasmacytoma PC3741 and clearly distinguishable from
`the VK21C gene rearranged to the other J!K genes (Fig. 2).
`This analysis further supports the premise that the VK2JC
`gene codes for the H36 light chains. That all examples are
`rearranged to JK2 also supports the relatedness of this set of
`antibodies since JK2 per se is not a prerequisite for HA(Sb)
`binding (i.e., the VK21C hybridoma, H2-6C4; Fig. 1). How-
`ever, the BamHI/HindIII rearrangements of two H36 hy-
`bridomas are not exactly the same: H36-17 is slightly larger
`and H36-7 is slightly smaller than the characteristic VK21C-
`JK2 rearrangement. The reason for the larger size of H36-17
`is a point mutation that obliterates the BamHI site coding for
`residue 95 (unpublished data). The reason for the smaller
`size of H36-7 is not known.
`This analysis (Figs. 2 and 4) also identifies the silent K light
`chain allele. Typically this allele is either unrearranged (K0)
`or aberrantly rearranged (K-). Three of the H36 hybridomas
`have indistinguishable K- alleles (H36-15, -1, and -4). Two
`examples (H36-5 and H36-18) have unique K- alleles, and
`two examples (H36-7 and -17) may have KO alleles or may
`have lost the K- chromosome. K- alleles in hybridomas and
`plasmacytomas of independent origin are usually on differ-
`ent-size IVS-positive fragments (24, 25). That the K- rear-
`rane.ements are of the same size in the H36-15. -1 and -4 hv-
`
`".% _.,..%
`
`U
`
`NJ
`z
`
`V
`UV
`NJr NJ
`7
`z
`
`N
`:1
`
`I
`
`rs-
`(LU
`ff
`I
`
`'
`LU
`VI
`a
`I
`
`-
`
`(0
`ml
`I
`
`LU
`rO
`
`LU
`rO
`I I
`
`LU
`r)
`
`I
`
`n
`m
`
`CI
`
`0S')
`
`a.
`
`WI;
`
`t.
`
`.00
`-0,..*, A.., -, 4
`
`AM ..:
`"10
`
`I+,.K~
`
`?21CJK1gII_
`V21CJ,<2
`
`V21CJK3-
`
`V2ICJK4-
`
`*
`
`S
`
`E
`I
`O
`cn
`
`0 U
`i-
`..,
`-I-
`x I-
`
`O
`D
`x
`
`0-0
`
`_0
`
`XI I I
`
`.. L...
`
`-.
`
`LI
`
`I T
`
`I
`JK
`
`kb
`
`--i
`
`CK
`
`m
`
`O
`
`I
`K
`
`'vs~IJ
`
`hybridomas and
`K light chain rearrangements in the H36 I
`FIG. 2.
`I-digested DNA
`plasmacytomas. Southern blots of BamHI/HindIII
`Lrison of VK21C
`were hybridized with the IVS probe. (Left) Compa
`plasmacytomas in which each plasmacytoma V,,2J(
`, gene has been
`shown to be joined to a different JK gene by aminc
`acid seauence
`(21, 22). (Right) Comparison of the H36 hybridoma
`,s. The K+ allele
`from H36-1, -4, -5, -17, and -18 was identified by the
`DNA sequence
`rearrangement
`of this rearrangement (unpublished data). The K'
`usually corresponds in size to the VC21C-JK2 rearra
`Lngement of the
`plasmacytoma, NZPC 3741. The K( fragment is ide
`ntified by com-
`parison to the BALB/c embryo DNA digest. The KO
`'fragment is ob-
`ccontaminating
`served in all DNAs in part because of host tissue
`these in vivo grown hybridomas and plasmacytoma
`,s.
`
`Lassen - Exhibit 1034, p. 3
`
`

`

`Immunology: McKean et al.
`
`Proc. Natl. Acad. Sci. USA 81 (1984)
`
`3183
`
`between the members of this set. Such parallel replacements
`are infrequent among VK21 or VA variants. This large number
`of common replacements could result from independent anti-
`gen selection for particular VK21C light chain sequences that
`are required (in conjunction with certain VH regions) for the
`formation of HA (Sb)-specific antibody combining sites.
`However, similar substitutions have not been observed so
`far in VK21C chains of anti-HA (Sb) antibodies from hybrid-
`omas of independent origin (H2-6C4, Fig. 1; unpublished
`data). Nevertheless, we cannot formally exclude this possi-
`bility because the isolated H chain of H36-15 could be shown
`to physically associate with an isolated VK21C0-JK2 chain
`(PC 3741), yet failed to produce HA (Sb)-binding antibodies
`(unpublished data).
`An alternative explanation for the common replacements
`is that these H36 hybridomas originated from just one or two
`lymphocyte precursors and that the shared substitutions rep-
`resent the sequential accumulation of mutations in their
`progeny. One possible genealogical relationship of the H36 K
`light chains to the ancestral V,,2JC0 germ-line gene is shown
`in Fig. 4. In this model the H36 sequences initially diverged
`from VK21C0 by the glycine-to-aspartic acid substitution at
`position 29. The branches of this tree represent substitutions
`acquired at early generations, and each K chain has a differ-
`ent terminus because all H36 K chains have one or more
`unique substitutions. By this model, the number of indepen-
`dent, parallel substitutions is reduced to two examples.
`That at least some of these H36 fusants may stem from a
`common precursor is suggested by the nature of rearrange-
`ments at the K locus other than the productive VK2JC-hJ2
`rearrangement. Two sorts of rearrangements have been ana-
`lyzed. One type is of the DNA upstream of the JK locus. This
`DNA is often retained in lymphocyte lines but in a unique
`context compared to germ-line DNA. As such rearranged
`DNA is the result of VK-to-JK recombination, the size of up-
`stream fragments is diverse among plasmacytomas and hy-
`bridomas of independent origin (24). The second type of K
`light chain rearrangement inspected is at the silent K allele.
`Typically this allele is either unrearranged (KO) or aberrantly
`rearranged (K-) (25). Since the K- rearrangement often re-
`sults from abortive VK rearrangements (23), these too are of
`diverse sizes in cell lines of independent origin. Hence, these
`types of rearrangements, when of identical size in cell lines,
`may have been inherited from a common ancestral lympho-
`cyte.
`Four hybridomas (H36-1, -4, -15, and -18) have indistin-
`guishable rearranged upstream DNA; of these, three (H36-1,
`-4, and -15) have indistinguishable K- rearrangements. H36-
`18 has a slightly smaller K- rearrangement; this may be a
`different K- form or a secondary deletion of the K- shared by
`H36-1, -4, and -15. The other H36 hybridomas (H36-5, -7,
`and -17) do not share these upstream DNA or K- rearrange-
`ments. Thus, by these criteria, we believe that at least H36-
`1, -4, -15, and -18 may be related; H36-5, -7, and -17 may be
`of independent origin(s), but their relatedness to each other
`is strongly suggested by shared amino acid substitutions. It
`is also possible that all seven examples are related through a
`common precursor but that mutations at restriction sites or
`rearrangements affecting upstream DNA or the silent K allele
`occurred during subsequent cell divisions.
`From this evidence of relatedness between at least two
`subsets of the H36 hybridomas, we favor the hypothesis that
`amino acid substitutions shared by the VK regions of a subset
`represent the sequential accumulation of somatic mutations.
`By this interpretation certain features of somatic mutation
`emerge:
`(0 Somatic mutation is an ongoing process occurring over
`many generations and during different stages of lymphocyte
`differentiation. Since the lineage comprising antibodies H36-
`1, -4, -15, and -18 (Fig. 4) includes examples with the same
`
`Genealogical relationship of H36 hybridoma VK21C light
`FIG. 4.
`chains. This tree shows the descent of the VK21C chains from the
`germ-line Va21C gene. The distances are proportional to the number
`of amino acid replacements from this germ-line sequence. Only the
`sequence to residue 33 (Fig. 1) has been used; hence, H36-5 and
`H36-7 are not separated. Further amino acid sequence analysis does
`separate this pair. This tree requires the fewest independent parallel
`replacements. In this case, valine-27b (VK21C°) to isoleucine (H36-4
`and -17) and methionine-33 (VK21C0) to isoleucine (H36-1 and -4) are
`assumed to have occurred independently. It is interesting to note
`that the former substitution has occurred twice and the latter has
`been observed once in V,,21 light chains from plasmacytomas. None
`of the other H36 substitutions have been seen in other VK21 variant
`chains. The set encircled are hybridomas thought to be related be-
`cause all four have an indistinguishable upstream element and three
`of these (H36-1, 4, and -15) have an indistinguishable K- rearrange-
`ment. H36-5, -7, and -17 may be of independent origin as indicated
`by the dashed lines.
`
`bodies binding to the antigenic site, Sb, on the HA (9). To
`examine how this diverse repertoire is generated, we have
`initiated a structural analysis of HA (Sb)-specific hybridoma
`antibodies that had been generated from a single donor
`mouse, H36. The seven H36 hybridomas were chosen on the
`basis of a previous serological analysis, which indicated that
`these antibodies used a K light chain of the VK21C subgroup
`and formed a closely related set whose individual members,
`however, could be differentiated from each other by para-
`typic and/or idiotypic analysis (9). The amino acid se-
`quences of these K chains confirms that all belong to the
`VK21C subgroup. Further, the size of the DNA endonuclease
`fragment bearing the productive VK gene shows that all ex-
`amples have rearranged the VK21C gene to the JK2 gene seg-
`ment. However, the VK sequences of the H36 K light chains
`differ from each other and from the gene product of the
`VK21C germ-line gene. Hence, we believe that the diversity
`within this set arises by somatic mutation.
`The extent and pattern of amino acid replacements is re-
`markable. The H36 VK21C regions that are nearly complete-
`ly sequenced have either seven or eight replacements com-
`pared to the VK21C germ-line sequence. This is more than
`the average number of substitutions (ca. 1) observed in a
`survey of X light chains or VK21 chains produced by mouse
`plasmacytomas (3, 21). The pattern of variability is also un-
`usual in that many of the amino acid substitutions are shared
`
`Lassen - Exhibit 1034, p. 4
`
`

`

`3184
`
`Immunology: McKean et al.
`
`Proc. Natl. Acad Sd USA 81 (1984)
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`8.
`
`9.
`
`heavy chain isotype (y2a) (Fig. 1), the mutational differences
`between these must have occurred after the switch from IgM
`to Ig2a expression. This lineage also includes examples with
`different isotypes (4yl or y2a, Fig. 1). As the switch from IgM
`to distinct isotypes appears to occur in a single step (ju to yl
`and A to y2a) (26), the mutations shared by these examples
`must have occurred prior to the isotype switch.
`(ii) The rate of somatic mutation seems to be high. If we
`assume that a precursor to these H36 mutants was selected
`during the primary immunization with PR8 because of the
`common glycine-to-aspartic acid substitution at position 29,
`then 6 to 7 replacements accumulated at the time of fusion
`(24 days after primary immunization). If these lymphocytes
`were dividing continuously with a generation time of 18 hr,
`the mutation rate would be in the range of 10-3 per base pair
`per generation. This estimate of a high mutation rate is con-
`sistent with the observations of Scharff and co-worker on
`the VH of the plasmacytoma S107 (8).
`(iii) The somatic mutations leading to amino acid replace-
`ments are clustered and are found mainly in complementar-
`ity-determining regions. In this respect the pattern of vari-
`ability is similar to that found in VA and VK21 plasmacytoma
`light chains, but the number of replacements is more exten-
`sive. To explain the pattern of variability in VA, it was pro-
`posed that antigen may act in the selection (selective expan-
`sion) of B cells expressing mutated immunoglobulin recep-
`tors that fit the antigen (3). The marked clustering of
`replacement mutations in complementarity-determining re-
`gions of the light chains of the H36 hybridoma antibody set
`agrees with this proposal and further suggests that antigen
`selection acts on sequentially arising single-point mutations
`throughout the development of a B-cell lineage. The few mu-
`tations in framework regions also may play a role in modify-
`ing antibody specificity or else may have been coselected
`along with complementarity-determining region replace-
`ments. The latter could occur if the rate of mutation of these
`genes is high.
`This work was supported by National Institutes of Health Grants
`GM-20964, CA-31638, CA-26297, AI-13989, and CA-06927 and by
`an appropriation from the Commonwealth of Pennsylvania. D.M. is
`the recipient of a Research Career Development Award, CA00586,
`from the National Cancer Institute.
`Tonegawa, 5. (1983) Nature (London) 302, 575-581.
`1.
`Baltimore, D. (1981) Cell 26, 295-296.
`2.
`
`Weigert, M., Cesari, I. M., Yonkovich, S. J. & Cohn, M.
`(1970) Nature (London) 228, 1045-1047.
`Brenner, S. & Milstein, C. (1966) Nature (London) 211, 242-
`243.
`Gearhart, P. J., Johnson, N. D., Douglas, R. & Hood, L.
`(1981) Nature (London) 291, 29-34.
`Kim, S., Davis, M., Sinn, E., Patten, P. & Hood, L. (1981)
`Cell 27, 573-581.
`Baltimore, D. (1981) Cell 24, 592-594.
`Cook, W. D. & Scharff, M. D. (1977) Proc. Natl. Acad. Sci.
`USA 74, 5687-5691.
`Staudt, L. M. & Gerhard, W. (1983) J. Exp. Med. 157, 687-
`704.
`Shulman, M., Wilde, C. D. & Kohler, G. (1978) Nature (Lon-
`don) 276, 269-270.
`Ey, P. L., Prowse, S. J. & Jenkins, C. R. (1978) Immuno-
`chemistry 15, 429-436.
`12. McKean, D. J., & Potter, M. (1979) in T and B Lymphocytes:
`Recognition and Function, ed. Bach, F. H. (Academic, New
`York), pp. 63-71.
`Huang, H. V., Bond, M. W., Hunkapiller, M. W. & Hood, L.
`(1983) Methods Enzymol. 91, 318-323.
`14. Van Ness, B. G., Weigert, M., Coleclough, C., Mather, E. L.,
`Kelley, D. E. & Perry, R. P. (1981) Cell 27, 593-602.
`Perry, R. P., Kelley, D. E., Schibler, U., Heubner, K. & Cro-
`ce, C. M. (1979) J. Cell. Physiol. 98, 553-560.
`Southern, E. M. (1975) J. Mol. Biol. 98, 503-517.
`Wahl, G. M., Stern, M. & Stark, G. R. (1979) Proc. Natl.
`Acad. Sci. USA 76, 3683-3687.
`Huppi, K., Julius, M., Staudt, L., Gerhard, W. & Weigert, M.
`(1984) Ann. Immunol. (Paris) 135c, 181-185.
`Dayhoff, M. (1972) Atlas of Protein Sequences and Structure
`(National Biomedical Research Foundation, Washington, DC),
`Vol. 5.
`Heinrich, G., Traunecker, A. & Tonegawa, S. (1984) J. Exp.
`Med. 159, 417-435.
`Weigert, M., Gatmaitan, L., Loh, E., Schilling, J. & Hood, L.
`(1978) Nature (London) 276, 785-790.
`Julius, M. (1982) Dissertation (University of Pennsylvania,
`Philadelphia, PA).
`Seidman, J. G., Nau, M. M., Norman, B., Kwan, S.-P.,
`Scharff, M. & Leder, P. (1980) Proc. Natl. Acad. Sci. USA 77,
`6022-6026.
`24. Van Ness, B. G., Coleclough, C., Perry, R. P. & Weigert, M.
`(1982) Proc. Natl. Acad. Sci. USA 79, 262-266.
`Coleclough, C., Perry, R. P., Kardalainen, K. & Weigert, M.
`(1981) Nature (London) 290, 372-378.
`Marcu, K. D., Lang, R. B., Stanton, L. W. & Harris, L. R.
`(1982) Nature (London) 298, 87-89.
`
`10.
`
`11.
`
`13.
`
`15.
`
`16.
`17.
`
`18.
`
`19.
`
`20.
`
`21.
`
`22.
`
`23.
`
`25.
`
`26.
`
`Lassen - Exhibit 1034, p. 5
`
`