Secretion-Impaired Heavy Chains Assembled Antibody Complexes

Home , Secretion assay

Restoration of Ig Secretion: Mutation of Germline-Encoded Residues in T15L Chains Leads to Secretion of Free Light Chains and Assembled Antibody Complexes Bearing This information is current as Secretion-Impaired Heavy Chains of October 1, 2021. Elizabeth A. Whitcomb, Tammy M. Martin and Marvin B. Rittenberg J Immunol 2003; 170:1903-1909; ;

doi: 10.4049/jimmunol.170.4.1903 Downloaded from http://www.jimmunol.org/content/170/4/1903

References This article cites 35 articles, 16 of which you can access for free at: http://www.jimmunol.org/content/170/4/1903.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

by guest on October 1, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2003 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Restoration of Ig Secretion: Mutation of Germline-Encoded Residues in T15L Chains Leads to Secretion of Free Light Chains and Assembled Antibody Complexes Bearing Secretion-Impaired Heavy Chains1

Elizabeth A. Whitcomb,2 Tammy M. Martin,3 and Marvin B. Rittenberg

We previously described T15H chain mutants that were impaired in assembly with L chain and in ability to be secreted from the cell. The unmutated T15L chain is unusual in that it is secretion-impaired in the absence of assembly with H chain. The T15L chain preferentially pairs with T15H in vivo, suggesting that if we introduced mutations that would allow secretion of free T15L chain, they might also lead to the secretion of the complex with the defective H chain. We mutated four positions in the germline T15L

that had amino acids infrequently found in other ␬-chains. Mutation to the most frequently occurring amino acid at three of the Downloaded from four positions allowed secretion of free L chain, while the combination of two secretion-restoring mutations was synergistic. Coexpression of secretion-restored mutant L chains with the secretion-defective mutant H chains rescued secretion of the assem-

bled H2L2 complex, suggesting that during somatic hypermutation in vivo, deleterious mutations at the H chain may be com- pensated by mutations on the L chain. To our knowledge, this is the first example of mutations in IgL chains that are able to restore secretion-defective H chains to secretion competence in mammalian cells. The Journal of Immunology, 2003, 170: 1903Ð1909. http://www.jimmunol.org/ mmunoglobulins have long served as models for heteromeric may occur in the germinal center might be able to restore secretion protein assembly and secretion, although these processes are function to the secretion-impaired mutant H chains. Because the I still not completely understood. For example, the basis for T15L chain appears to be relatively unique in its need to pair with quality control of proper assembly that determines whether an Ab the T15H chain and is secretion-impaired on its own, we reasoned molecule is destined for secretion or degradation remains to be that if we could mutate T15L to secretion competence, the modi- elucidated. Complete understanding of the process of Ig secretion fied L chain might complement the secretion-impaired mutant would bear not only on B cell function, but also on assembly, T15H chains and restore secretion. Based on a strategy used pre- transport, and secretion of other proteins as well. Our approach to viously to stabilize Ig proteins expressed in Escherichia coli (3), these questions has been to study the effects on secretion of mu- we identified four positions in the V␬22 sequence that contained by guest on October 1, 2021 ␬ tations introduced into the VH complementarity-determining re- amino acids present in less than 2% of all other -chains in the gion (CDR)4 2 of murine anti-phosphocholine (PC) Abs (1, 2). We database (4). This suggested that the presence of these uncommon had begun studies of mutation to assess effects on Ag recognition amino acids in V␬22 could be responsible for its inability to be and found ϳ50% of multisite H chain CDR (HCDR)2 mutants had secreted in the absence of H chain. To test this possibility, we impaired Ag binding. Surprisingly, we also learned that a signif- introduced point mutations to convert these residues to the amino icant number of mutants (10%) were not secreted from the cell. acids most commonly found in other ␬-chains. Three of the four

This contribution of VH CDR2 to Ig secretion had not been rec- single mutants demonstrated a modest, but signiﬁcant level of se- ognized previously. cretion of free T15L chain, while the combination of two secre- Because somatic mutation occurs on both H and L chains, we tion-restoring mutations was synergistic. Importantly, when the hypothesized that compensatory mutations in the L chain such as mutant L chains were coexpressed with the T15 low secretor H chains, Ab secretion was restored. These results suggest that mutations in the L chain occurring in vivo could have the potential to Department of Molecular Microbiology and Immunology, Oregon Health & Science compensate for deleterious H chain mutations and maintain the University, Portland, OR 97239 functional viability of the cell. Received for publication August 19, 2002. Accepted for publication December 13, 2002. Materials and Methods The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance Mutagenesis, transfection, and cell culture with 18 U.S.C. Section 1734 solely to indicate this fact. T15L-C␬-pSV2-neo vector was a kind gift from S. Morrison (Department 1 This work was supported by National Institutes of Health Grants AI-14985 and of Microbiology and Molecular Genetics, University of California, Los AI-26827 to M.B.R. and a grant from the Tartar Trust Foundation to E.A.W. Angeles, CA) (1). Site-directed mutagenesis was performed as described 2 Address correspondence and reprint requests to Dr. Elizabeth A. Whitcomb, De- using the pTZ18U vector (1). Oligonucleotides used for mutagenesis: Ile56 partment of Molecular Microbiology and Immunology, Mail Code L220, Oregon to Ser (I56 Ͼ S), 5Ј-CGA TAC AGT GGG GTC C-3Ј; Thr84 to Ala (T84 Ͼ Health & Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239. A), 5Ј-AA GAC CTC GCA CAT TAT-3Ј; Lys17 to Glu (K17 Ͼ E), 5Ј-G E-mail address: [email protected] ACA GCA AGT GAG AAG GTC ACC-3Ј; Ser16 to Gly (S16 Ͼ G), 5Ј-CT 3 Current address: Department of Ophthalmology, Oregon Health & Science Univer- GTG ACA GCA GGT AAG AAG GTC ACC-3Ј; Ser16 to Gly and Lys17 sity, Casey Eye Institute, 3375 SW Terwilliger Boulevard, Portland, OR 97239. to Glu (S16 Ͼ G ϩ K17 Ͼ E), 5Ј-CT GTG ACA GCA GGT GAG AAG Ј 4 Abbreviations used in this paper: CDR, complementarity-determining region; AP, GTC ACC-3 . Mutated codons are highlighted in bold, and speciﬁc mu- 56 alkaline phosphatase; FW, framework region; HCDR, H chain CDR; PC, phospho- tations are underlined. For construction of double mutants expressing Ile choline; wt, wild type. to Ser and Thr84 to Ala (I56 Ͼ S ϩ T84 Ͼ A) or Lys17 to Glu and Thr84 to

Ala (K17 Ͼ E ϩ T84 Ͼ A), the second mutation was added to the pTZ18U graphs or IP Lab Gel software (version 1.5; Analytics, Vienna, VA) for clone containing the single T84 Ͼ A mutation. All mutants were sequenced phosphorimager scans. in pTZ18U before subcloning into pSV2-neo for transfection. SP2/0, a mouse hybridoma fusion partner line that does not express Statistics endogenous H or L chains (5), was transfected with the T15L mutant con- Statistical signiﬁcance was determined by unpaired Student’s t test using structs by lipofectin (Life Technologies, Grand Island, NY) or electropo- Multistat software (version 1.12; Biosoft, Ferguson, MO). ration (6), and stably transfected clones were selected by G418 resistance (Geneticin; Life Technologies). SP2/0 expressing the wild-type (wt) T15L Modeling of Ig proteins chain was previously described (7). Lines expressing the low secretor H chains were previously described (8) and were transfected using lipofectin. Fig. 4 was created using the model of T15 (9) with RasMol software (version All cells were propagated in IMDM (Life Technologies) containing 20% 2.7.1; RasMac Molecular Graphics, University of California, Berkeley, CA). FBS (HyClone, Logan, UT), as described (1). Results Secretion assay Single point mutations partially restore secretion of the T15L Quantitation of Ig secretion was performed essentially as described (7). chain: a synergistic effect of two mutations Brieﬂy, 1 ϫ 106 cells from transfectants were washed and cultured in 1 ml A number of mutations at particular residues have been shown to of fresh IMDM at 37¡C. After 4 h, supernatants were collected and cells affect secretion of free L chains (11Ð14). Although T15L preserves were lysed in 1 ml of PBS containing 25 mM iodoacetamide, 20 ␮g/ml soybean trypsin inhibitor, 50 ␮g/ml PMSF, and 0.25% Nonidet P-40 (all the secretion-permissive residues previously described with one reagents from Sigma-Aldrich, St. Louis, MO). Dilutions of lysates and exception, a conservative change from Gly to Ser at position 16, it supernatants were assayed for Ig content by sandwich ELISA. For L chain- cannot be secreted unless assembled with H chain. T15L expresses only determinations, goat anti-mouse ␬ antiserum was used to coat the the unmutated V␬22 gene rearranged to J␬5; thus, we compared

␬ Downloaded from plates, and alkaline phosphatase (AP)-coupled goat anti-mouse antiserum ␬ (Southern Biotechnology Associates, Birmingham, AL) was used for de- the germline V 22 sequence (15) with the database of over 5000 tection. For H chain determinations, rabbit anti-IgG2b and AP-coupled expressed and germline V␬ sequences and identified a number of anti-IgG2b (Zymed Laboratories, South San Francisco, CA) were used. For positions containing residues present in less than 2% of all V␬ determinations of assembled Ig, plates were coated with anti-␬ reagent and genes (4). Mutations were made at four positions: two in frame- detected with AP-coupled anti-IgG2b. All samples were compared with a work region (FW) 1, S16 Ͼ G and K17 Ͼ E; one in CDR2, I56 Ͼ standard T15-IgG2b-purified Ab (1) to calculate ng/ml of Ig protein. For L 84 Ͼ chain-only transfectants, percentage of secretion was determined by com- S; and one in FW3, T A. Position 16 was shown previously to paring the amount present in the supernatant after the 4-h culture with the have an effect on ␭ L chain secretion, in which a nonconservative http://www.jimmunol.org/ total amount of L chain present in the cell lysate plus the amount in the change from Gly to Arg abolished secretion of free L chains (11). supernatant. Residue frequency in other ␬-chains as well as the most commonly Binding of Ig to PC-containing Ags occurring residue and frequency are shown in Fig. 1A. Single point mutations as well as the double mutants I56 Ͼ S ϩ T84 Ͼ A, Tissue culture supernatants isolated from the transfectants containing both S16 Ͼ G ϩ K17 Ͼ E, and K17 Ͼ E ϩ T84 Ͼ A were made and the mutant H and L chains were assayed for binding to PC-containing Ags by ELISA using AP-coupled anti-IgG2b for detection. Binding to PC-hi- stable transfectants were generated. stone or Streptococcus pneumoniae (R36A) was performed as described At least four individual transfectants expressing the mutations (9). Binding to Trichinella spiralis or Ascaris suum Ags was performed as were isolated and L chain secretion was initially analyzed using a by guest on October 1, 2021 described (10). Percentage of wt binding was determined by calculating the quantitative ELISA. Percentage of secretion was determined by concentration of mutant Ab needed to give an OD of 0.5 in the ELISA, then comparing the amount of L chain in the supernatant with the total comparing that figure with the concentration of wt Ab needed to get the same OD. Binding to PC-histone was inhibited using free PC and the PC amount of L chain in the lysate plus supernatant. Although less analogs, nitrophenylphosphocholine and ␣-glycerophosphocholine, at con- than 1% of the total wtT15L chain was in the supernatant after a centrations ranging from 1 ϫ 10Ϫ2 to 1 ϫ 10Ϫ6 M. 4-h incubation in fresh medium, all transfectants expressing the single mutations S16 Ͼ G, K17 Ͼ E, or T84 Ͼ A secreted a sig- Metabolic labeling and immunoprecipitation nificant amount of L chain in the supernatant ( p Ͻ 0.001 vs wt, Cells were plated at a density of 2.5 ϫ 105Ð1 ϫ 106/ml and allowed to Fig. 1B). The percentage of secretion for each mutant was similar grow to 60Ð80% confluence in six-well culture plates (Falcon; BD Bio- in different transfectants expressing intracellular L chain at a wide sciences, Franklin Lakes, NJ). Cells were starved of methionine and cysteine by incubation in methionine/cysteine-free DMEM (Sigma-Aldrich) range of concentrations; thus, the increase in L chain in the super- for 30 min. Cells were pulse labeled for 15 min with 0.4 mCi 35S Express natant was due to the effect of the mutations and was independent protein labeling mix (PerkinElmer Life Sciences, Boston, MA) in methi- of intracellular L chain concentration. The I56 Ͼ S mutation had no onine/cysteine-free DMEM supplemented with 10% FBS. After the pulse, significant impact on secretion on its own and may have an an- cells were washed in PBS and chased into complete IMDM. At appropriate tagonistic effect when paired with the secretion-restoring mutation chase time points (see Results), the supernatants were harvested and cells 84 Ͼ 56 Ͼ ϩ were washed with PBS before lysis in Tris, sodium, azide buffer (10 mM T A. Although the secretion of the double mutant I S 84 Ͼ 56 Ͼ Tris-HCl, pH 8.0, 140 mM NaCl, 0.025% NaN3) supplemented with 50 T A is significantly above wt or the single mutant I S and ␮g/ml PMSF and 1% Nonidet P-40. Lysates were incubated on ice for 1 h, below that of the single mutant T84 Ͼ A, this putative antagonistic and cell debris was pelleted and discarded. effect was not consistently observed in all the clones, as nearly Lysate and supernatant samples were immunoprecipitated using rabbit ␬ one-third of the clones (6 of 19) were above the minimum secre- anti-mouse antiserum (Cappell/ICN, Costa Mesa, CA, or Cortex Bio- 84 chemicals, San Leandro, CA), followed by protein A-Sepharose (Amer- tion observed for T Ͼ A. The combination of two secretion- sham Pharmacia Biotech, Piscataway, NJ). For assembly of IgG2b, super- restoring mutations was synergistic with the double mutants natants were harvested and precipitated with protein A-Sepharose. (K17 Ͼ E ϩ T84 Ͼ A and S16 Ͼ G ϩ K17 Ͼ E) secreting Ͼ2.5 Immunoprecipitates were washed twice in lysis buffer, followed by one times more L chain than was predicted if the effects of the indi- wash in Tris, sodium, azide and a final wash in 50 mM Tris, pH 6.7. For 16 Ͼ ϩ analysis of samples under reducing conditions, immunoprecipitates were vidual mutations were simply additive (predicted S G 17 17 84 eluted by heating in 40 ␮l SDS sample buffer (2% SDS, 10% glycerol, 60 K Ͼ E ϭ 10.1%, observed ϭ 29%; predicted K Ͼ E ϩ T Ͼ mM Tris, pH 6.8, 0.005% bromphenol blue) containing 100 mM DTT. A ϭ 9.5%, observed ϭ 24.8%; Fig. 1B). Other secretion-compe- Samples analyzed under nonreducing conditions were eluted in 40 ␮lof tent ␬ (non-V␬22) and ␭ L chains exhibit a range of secretion in SDS sample buffer without DTT. Samples were heated to 85¡C for 10 min ϩ before analysis on 10 or 12% SDS polyacrylamide gels. this assay (12Ð65% of total L chain (lysate supernatant) in su- Pulse-chase experiments were quantitated using NIH Image software pernatant after 4 h, n ϭ 4 different L chains; data not shown). Thus, (version 1.61; National Institutes of Health, Bethesda, MD) for autoradio- the degree of rescue of T15 mutant L chains resembles that of The Journal of Immunology 1905 Downloaded from

FIGURE 1. Infrequently occurring amino acids can affect the secretion of T15L. A, The germline V␬22 sequence (15) was compared with the http://www.jimmunol.org/ database of expressed and germline V␬ sequences (4). The residues in T15L at positions 16, 17, 56, and 84 occur in less than 2% of other V␬ sequences. The percentage in other ␬-chains (shown in parentheses) as well as the most commonly occurring residue and its frequency (4) are shown. FIGURE 2. Secretion profile of L chains. Cells were pulse labeled with B, Secretion of SP2/0 transfectants expressing wt or mutant L chains. [35S]methionine/cysteine for 15 min and chased up to 6.5 h. Lysates and Transfectants were cultured in fresh medium for 4 h, and amount of L chain supernatants were immunoprecipitated with rabbit anti-mouse ␬ antiserum, present in the supernatant or in the cell lysate was determined by quanti- followed by protein A-Sepharose. Immunoprecipitates were analyzed un- tative ELISA. Percentage of secretion was determined by comparing the der reducing SDS-PAGE conditions. Three independent transfectants were amount of L chain in the supernatant with the total amount in the super- analyzed for each mutant with similar results. A, Representative autora- natant plus the amount in the cell lysate. Data represent the average (ϩSD) by guest on October 1, 2021 diographs or phosphorimager scans of pulse-chase analysis. B, Quantita- p Ͻ 0.001 vs wtT15L; #, p Ͻ 0.001 ,ء .of at least four separate transfectants tion of L chain label in the supernatant for experiments shown in A. Per- vs either single mutant. centage of label in the supernatant was calculated by comparing the amount detected in the supernatant with the total labeled L chain detected imme- diately after the pulse (t ϭ 0). The I56 Ͼ S mutant (not shown) was at other secretion-competent L chains expressed by myeloma and background for all time points. hybridoma cells. To confirm that the increases in secretion observed in the double mutants S16 Ͼ G ϩ K17 Ͼ E and K17 Ͼ E ϩ T84 Ͼ A were indeed secretion assays, which compared the amount in the supernatant synergistic, a pulse-chase analysis was performed to determine the with the steady state level of intracellular L chain (Fig. 1B). secretion of L chain labeled during the pulse. Transfectants expressing steady state levels of intracellular L chain similar to the Mutations that restore secretion of free L chains can wtT15L cell line were pulse labeled for 15 min and chased up to complement secretion-impairing H chain mutations 6.5 h. Cell lysates and supernatants were immunoprecipitated and We previously described mutations in the CDR2 of the T15H analyzed under reducing SDS-PAGE (Fig. 2A). The percentage of chain that impaired Ab secretion (7). Because somatic mutation in label in the supernatant was graphed by comparing the amount of vivo can occur at both the H and L chain loci, we asked whether labeled L chain in the supernatant with the amount of labeled L the secretion-restoring mutations in the T15L chain could com- chain in the lysate at time ϭ 0 (Fig. 2B). As in previous assays, the pensate for the H chain secretion defect. The restored mutant L wtT15L and I56 Ͼ S mutant showed no detectable L chain in the chain S16 Ͼ G ϩ K17 Ͼ E was expressed in cells with the low supernatant over the 6.5-h chase. The secretion-restored single mu- secretor H chains. Stable transfectants were obtained, and secre- tant transfectants S16 Ͼ G, K17 Ͼ E, and T84 Ͼ A exhibited a tion was determined by incubating cells in fresh medium for 4 h. modest amount of secretion with 5Ð10% of the total label in the The amount of Ig was quantified by sandwich ELISA. The amount supernatant at the 6.5-h chase time point, similar to the levels of assembled H ϩ L chain in the supernatants of transfectants observed in the initial assays (compare Fig. 1B with Fig. 2B). The expressing either the wtT15L chain or the S16 Ͼ G ϩ K17 Ͼ E double mutants (S16 Ͼ G ϩ K17 Ͼ E and K17 Ͼ E ϩ T84 Ͼ A) mutant L chain is shown in Fig. 3, A–C. The secretion-restored secreted 35Ð50% of the total label into the supernatant during this mutant L chain was able to rescue secretion of the three mutant H same time period (Fig. 2B), confirming the synergistic increase chains tested; multiple transfectants were screened, and restored seen in the initial assays (Fig. 1B). The pulse-chase experiments secretion was consistently observed. The M153 transfectant ex- compare the amount of secreted L chain with the amount synthe- pressing the wtT15L chain did consistently secrete a low amount sized only during the pulse period. Thus, it is not surprising that of Ig into the supernatant, but the level was 10 times less than the the secretion seems more efficient than that observed in the initial transfectant expressing the mutant L chain, although similar 1906 RESTORATION OF Ig SECRETION

Fully assembled IgG2b consists of an H2L2 heterotetramer. The low secretor mutants expressing the wtT15L chain are impaired in

assembly, arresting at the H2L stage (8); thus, we wished to determine whether the secretion-restored mutants were secreting fully assembled IgG2b. Cells were pulse labeled for 15 min and chased for 4 h. Ig from the supernatant was immunoprecipitated with protein A-Sepharose, which will precipitate H chains along

with any assembled L chains. Fully assembled H2L2 IgG2b was detected in the restored mutants running at the same electro- phoretic mobility as the wtT15 (Fig. 3D); thus, the mutations in the L chain allow complete assembly with the mutant H chains and restore transport of completely assembled IgG2b. Free H chain was observed in the supernatant of the M153 transfectant. This is most likely due to labeled H chains assembling with, but not di-

sulﬁde linked to, unlabeled L chains. The intracellular t1/2 of L

chains is 1Ð2 h (6), while the intracellular t1/2 of the secretion- impaired H chains can be Ͼ20 h (8); thus, it is likely that L chains synthesized after the pulse are assembling with H chains synthesized during the pulse. Indeed, unpaired H and L chains were detected in supernatants collected from this mutant by Western blot Downloaded from under nonreducing conditions after immunoprecipitation with protein A-Sepharose (data not shown). We also tested whether the secreted Ab expressing the double- mutant L chain could bind Ag. The secretion-restored M153 mutant expressing the double-mutant L chain S16 Ͼ G ϩ K17 Ͼ E was able to bind the PC-containing Ags: PC-histone, S. pneu- http://www.jimmunol.org/ moniae (R36A), A. suum, and T. spiralis, although at less than 10% of the wtT15 Ab. The secretion-restored M164 mutant did not bind any of the PC Ags, while the secretion-restored M241 did show detectable binding to PC-histone and S. pneumoniae (R36A), although at a level Ͻ1% of wtT15 binding (Table I). These results suggest that while the secretion-competent L chains may fully as-

semble and allow the mutant H2L2 complex to be secreted, the combining site is not identical with that of the wt Ab. Neverthe- by guest on October 1, 2021 less, binding of the secretion-restored M153 Ab to PC-histone was inhibitable by free PC and the PC analogs, nitrophenylphosphocholine and ␣-glycerophosphocholine, suggesting that its fine specificity still resembles that of wtT15 and was not altered ap- FIGURE 3. Secretion-restored L chain also restores secretion of mutant preciably by the mutations in either the H or L chains (data not H chains. A–C, Transfectants were cultured in fresh medium for 4 h. shown). Amount of assembled H ϩ L was determined by sandwich ELISA using goat anti-mouse ␬ to coat the plates and AP-coupled rabbit anti-IgG2b for Discussion detection. Purified wtT15 Ab was used for a standard curve to determine In this study, we report that mutation of uncommon residues con- ng/ml of Ig. This experiment performed in triplicate is representative of at served in the T15L germline sequence permits the secretion of free least four independent experiments. A, M164 H chain; B, M241 H chain; T15L chains and also allows the L chains to assemble with and C, M153 H chain. D, Secretion of fully assembled Ig in the restored mu- restore the secretion of impaired mutant H chains. The T15L chain tants expressing the double-mutant L chain S16 Ͼ G ϩ K17 Ͼ E. Cells were pulse labeled for 15 min and chased for 4 h. Ig from the supernatant was is unusual for several reasons: 1) it is secretion impaired in the immunoprecipitated with protein A-Sepharose and examined under non- absence of H chain assembly; 2) to our knowledge, it has been reducing conditions to preserve Ig assembly.

Table I. Ag-binding characteristics of secretion-restored Iga amounts of intracellular Ig (both H and L chain) were observed (Fig. 3C). In most experiments, Ig (either as individual H or L H Chain Mutant M153 M164 M241 chain or as assembled IgG) in the supernatants of the M164 and Ag M241 transfectants expressing the wtT15L chain was undetect- PC-histone 7.8% N.B.b Ͻ1%c able. We also expressed the single mutant L chains (S16 Ͼ G and S. pneumoniae 2.8% N.B. Ͻ1%c K17 Ͼ E) with the mutant H chains. All mutated L chains did T. spiralis 4.1% N.B. N.B. Ͻ c restore secretion, although the degree of restoration was variable A. suum 10% N.B. N.B. (data not shown). Variation in secretion could perhaps be due to a Binding shown is percentage of wt. The concentration of wtT15 Ig needed to ϭ differences in concentrations of intracellular H or L chain protein. obtain OD480 0.5 was calculated and compared with the amount of mutant Ig needed to obtain the same OD. Wt and mutant Ig were compared at the same time An excess of L chain might promote the secretion of free L chain point on the same plate. b No detectable binding at 100 ng/ml. Binding of the wtT15 was detectable at 0.76 at the expense of H2L2 tetramers, whereas low levels of intracel- ng/ml for PC-histone. S. pneumoniae, and T. spiralis, and 12.5 ng/ml for A. suum. lular L chain might lead to partial assembly; incompletely assem- c Detectable binding at 100 ng/ml; however, too low to accurately compare with bled complexes would not be secreted. wtT15. The Journal of Immunology 1907 observed in vivo only in the immune response to PC-containing Ags; and 3) it has been observed in vivo associated only with the T15H chain. In this study, we set out to determine the molecular basis for the secretion impairment of T15L. Mutation of three single residues restored secretion modestly, and the combination of two secretion-restoring residues was synergistic; thus, multiple residues contribute to the secretion impairment of T15L. Addition- ally, we asked whether the secretion-restoring L chain mutations could affect the assembly of T15L with mutant H chains that are secretion impaired because T15H and T15L associations appear to be preferred in vivo. We found that secretion-restored T15L chains could assemble with and allow the secretion of H chains that could not fully assemble with the unmutated secretion-impaired T15L chain. Mutation of germline residues permits secretion of free L chains Although most L chains can be secreted in the absence of assembly with H chain, a handful have been described that are secretion impaired (11Ð13, 16). Most of these secretion-impaired L chains are defective as a result of replacement mutations. Several residues Downloaded from in FW1 and FW3 have been identified as critical for L chain secretion: Gly16 in FW1 (11), and Phe62 and Tyr/Phe87 in FW3 (12, 13). Other residues in FW3 have also been suggested to be involved in L chain secretion (aa 57Ð65) (12). T15L, which has no mutations from the germline V␬22, is secretion impaired in the absence of H chain assembly, yet preserves the amino acids pre- http://www.jimmunol.org/ viously identified to affect secretion of free L chains in FW3 (positions 57Ð65 and 87) and has a conservative substitution at position 16 from Gly to Ser, suggesting involvement of other residues in L chain secretion and the T15L defect. We demonstrate that secretion of T15L can be achieved by the replacement of its infrequently occurring amino acids. Mutation of three positions to the most commonly occurring residue (S16 Ͼ G, K17 Ͼ E, T84 Ͼ A) allowed secretion of free L chain, while one mutation (I56 Ͼ S) by guest on October 1, 2021 had no effect by itself, but may be antagonistic when combined with a secretion-restoring mutation (T84 Ͼ A). Thus, in addition to stability of expression in E. coli (3), our data indicate that sequence statistics can often, but not always, predict an effect on secretion in mammalian cells. Additionally, the combination of two secretion- restoring mutations can synergistically increase L chain secretion, confirming that multiple amino acids contribute to the secretion impairment of wtT15L. Structural consequences of secretion-restoring mutations Two mutations in FW1 of the T15L chain allow modest secretion of free L chain on their own and act synergistically to increase secretion when expressed together. The conservative change at S16 Ͼ G is at the pivot point of the ␤ turn in FW1 (Fig. 4, A and B); the preferred Gly would most likely be more flexible. Mutation of this position in a ␭ L chain from Gly to Arg prevented secretion FIGURE 4. Modeling of secretion-restoring mutations. A, Space-filling 16 model of the wtT15 V region. V is in gray; V is in white. The hapten PC (11). Conservation of Gly is high (97%), indicating its impor- H L is black, as are the L chain secretion-restoring positions 16, 17, and 84. B, tance in L chain structure (4). Position 17 is an acidic amino acid Backbone model of FW1 ␤ turn. Thr14, Ser16, and Lys17 are shown in stick in the majority of L chains (79%), but a Lys in T15L; an acidic conformation in dark gray. C, Backbone model of hydrophobic core of the residue at this position may influence L chain stability by forming 84 62 VL domain. Residues Thr and Ser are shown in stick conformation in a hydrogen bond with Ser/Thr at position 14 (Fig. 4B) (17). Thus, dark gray. D, Space-filling model of wtT15 combining site. H chain CDR2 converting Lys to Glu in T15L restores the potential for this ap- is shown in light gray. The L chain secretion-restoring positions 16, 17, and parently critical bond. Mutations at position 17 have been observed 84 are shown in black. in pathogenic or unstable L chains (17). The observation that the double-mutant L chain with both FW1 mutations is secreted ro- bustly provides further evidence that the stability of FW1 is im- frequently observed residue (4); thus, small hydrophobic residues portant for overall L chain secretion function. appear preferred. Position 84 comprises part of the hydrophobic 84 Ͼ The FW3 mutation T A permits modest secretion on its own core of the VL domain (18); thus, the larger polar Thr at residue 84 and acts synergistically when expressed with the K17 Ͼ E muta- is likely to require a major alteration in the overall folding of the 84 tion. Position 84 is most often Ala, with Gly the second most VL domain. Ala was shown to be one of the earliest residues to 1908 RESTORATION OF Ig SECRETION

␬ become inaccessible during folding of another murine -chain Fv ditionally, cooperative folding of isolated VH and VL domains has (19). Additionally, Thr in place of the conserved Ala at position 84 also been observed (23, 30). in a human ␬-chain proved to be destabilizing (20). Phe62 also The endoplasmic reticulum resident chaperones BiP and GRP94 contributes to the stabilization of this hydrophobic core (18), and associate with both secreted and nonsecreted Ig proteins (16, 31Ð others have shown replacing a polar Ser for the conserved Phe 33). We have shown that the low secretor H chains have an in- impairs L chain secretion (see Fig. 4C) (12). creased association with these chaperones as compared with the secretion-competent wtT15H chain when coexpressed with the Secretion-restoring L chain mutations restore H chain secretion wtT15L chain (8). Preliminary evidence suggests the levels of BiP We previously described T15H chains with mutations in CDR2 and GRP94 associated with the H chains are similar in the secre- that were impaired in Ab assembly and secretion (7). Because tion-restored mutants and their secretion-impaired counterparts introducing secretion-competent L chains D16L or J558L did not (data not shown); thus, the mechanism for secretion impairment restore secretion in the T15 HCDR2 secretion-impaired mutants and restoration is not due directly to the degree of chaperone as- (21), it was surprising that mutations in T15L that allowed secre- sociation. However, as L chain displaces BiP on the H chain dur- tion of free L chains could restore the H chain mutants. The low ing Ig assembly (28, 29), it is possible that displacement is more secretor H chains are unable to fully assemble with the wtT15L efﬁcient by the secretion-restored L chains, thus promoting the

chain arresting at the H2L intermediate (8). However, assembly of association of H/L over the H chain/BiP interaction. the H2L2 complex is complete when the low secretor H chains are Others have shown that mutation of L chain residues can in- paired with the secretion-restored L chains, suggesting that the crease the secretion and folding of Ig fragments in E. coli (22, 34).

secretion-restoring potential of the L chains may rest in their abil- VL FW1 residues have frequently been implicated in the stability ity to allow complete assembly. of Fv proteins (23, 35, 36). Wall and Pluckthun (23) showed that Downloaded from The L chain mutations that restore the H chain mutants are distal mutation of four FW1 residues in T15L, including S16 Ͼ G, in- to the H/L interface; thus, the mechanism for restoration is most creased folding and secretion of the T15Fv. These studies have likely due to long range effects (see Fig. 4D). Although the ex- shed considerable light on the residues important for folding, se- change at position 17 does alter the charge (Lys to Glu), a single cretion, and stability of Ig fragments expressed in E. coli. How- mutation at position 16 (from Ser to Gly) can at least partially ever, these Igs are secreted efficiently when expressed as whole restore secretion of the mutant H chains (data not shown); thus, a Abs in mammalian cells. Thus, our studies add new information to http://www.jimmunol.org/ change in surface charge at position 17 may not explain the re- the growing body of work on Ig folding and secretion. stored secretion. We speculate that the changes in the ␤ turn of Our data indicate that deleterious H chain mutations can be T15L FW1 affect the structure of L chain CDR1 and its interaction overcome by compensatory L chain mutations. This may be a with HCDR3. Deletion of four residues in the apex of HCDR3 mechanism whereby B cells in the germinal center are not wasted restored secretion of these H chain mutants when expressed with due to secretion-impairing H chain mutations. We have evidence the wtT15L chain (21). The deleted residues in HCDR3 make mul- suggesting that secretion-impairing H chain mutations occur in tiple contacts with L chain CDR1 (see Fig. 4D). Because the se- vivo during an immune response (G. D. Wiens, M. Brown, and cretion defect may be in part due to inefficient assembly with L M. B. Rittenberg, manuscript in preparation). Whether restorative by guest on October 1, 2021 chain (8), the interaction of these regions may be critical. The mutations are also observed is currently under investigation. Ad- CDR1 of T15L is quite long, being 17 aa; shortening the 17-res- ditionally, it will be of interest to see whether the secretion-re- idue CDR1 in the McPC603 L chain by 6 aa resulted in a consid- stored L chains can also restore secretion of impaired FW2 T15H erable increase in stability (22), although the sequence was altered chain mutants previously described (2). Dul and Argon (12) reported as well as the length. Thus, it is possible that the FW1 changes that mutation of an L chain V region could impair secretion of as- stabilize CDR1 and promote the assembly with H chain. Further- sembled Ab. To our knowledge, our results represent the first example more, others have shown that mutations within L chain FW1 can of L chain mutations that can rescue secretion-defective Abs. increase the assembly and secretion of Fv proteins (22, 23) and whole Ab molecules (24). Acknowledgments Examination of L chains carrying mutations that prevented se- We thank Jennifer Owen for excellent technical assistance, and Dr. Greg cretion led to the hypothesis that a secretion signal was encoded Wiens, McKay Brown, and Susan Stevens for critical review of this within the VL domain by conserved amino acids in FW1 and FW3 manuscript. (12). These amino acids (residues Gly16, Pro59, Arg61, and Phe62) are distal from the H/L interface, and it was suggested they may be important for secretion of free L chains as well as complexed Ig References (12). Our data are partially consistent with this hypothesis, as the 1. Chen, C., V. A. Roberts, and M. B. Rittenberg. 1992. Generation and analysis of random point mutations in an antibody CDR2 sequence: many mutated antibodies secretion-competent double-mutant T15L chains that restore the lose their ability to bind antigen. J. Exp. Med. 176:855. secretion of the mutant H chains preserve all of these residues, 2. Wiens, G. D., K. A. Heldwein, M. P. Stenzel-Poore, and M. B. Rittenberg. 1997. while the secretion-impaired wtT15L has the conservative substi- Somatic mutation in VH complementarity-determining region 2 and framework 16 region 2: differential effects of antigen binding and Ig secretion. J. Immunol. tution at G Ͼ S. However, other L chains that preserve these 159:1293. residues (Gly16, Pro59, Arg61, and Phe62), such as D16L and 3. Steipe, B., B. Schiller, A. Pluckthun, and S. Steinbacher. 1994. Sequence statis- J558L, do not permit secretion of our mutant H chains (21), while tics reliably predict stabilizing mutations in a protein domain. J. Mol. Biol. 240: 188. 17 Ͼ 16 the secretion-restored K E single mutant that has Ser was 4. Kabat, E. A., G. Johnson, and T. T. Wu. 2002. The Kabat database of sequences able to restore secretion. Thus, the pairing of T15H with T15L of proteins of immunological interest. http://immuno.bme.nwu.edu. most likely allows both the V and V domains to fold more 5. Shulman, M., C. D. Wilde, and G. Kohler. 1978. A better cell line for making H L hybridomas secreting specific antibodies. Nature 276:269. effectively than other H/L pairings. This was also suggested by 6. O’Hare, T., G. D. Wiens, E. A. Whitcomb, C. A. Enns, and M. B. Rittenberg. Wall and Pluckthun (23), using VH and VL fragments expressed in 1999. Proteasome involvement in the degradation of unassembled immunoglob- E. coli. Although initial folding of H and L chains occurs cotrans- ulin light chains. J. Immunol. 163:11. 7. Chen, C., T. M. Martin, S. Stevens, and M. B. Rittenberg. 1994. Defective se- lationally, there is considerable evidence that H/L assembly allows cretion of an immunoglobulin caused by mutations in the heavy chain comple- for additional folding to complete the H2L2 complex (25Ð29). Ad- mentarity determining region 2. J. Exp. Med. 180:577. The Journal of Immunology 1909

8. Martin, T. M., G. D. Wiens, and M. B. Rittenberg. 1998. Inefficient assembly and 24. Saldanha, J. W., A. C. R. Martin, and O. J. P. Leger. 1999. A single backmutation intracellular accumulation of antibodies with mutations in VH CDR2. J. Immunol. in the human kIV framework of a previously unsuccessfully humanized antibody 160:5963. restores the binding activity and increases the secretion in COS cells. Mol. Im- 9. Brown, M., M. B. Rittenberg, and V. A. Roberts. 1996. Tolerance to single, but munol. 36:709. not multiple, amino acid replacements in antibody VH regions: a means of min- 25. Bigelow, C. C., B. R. Smith, and K. J. Dorrington. 1974. Equilibrium and kinetic imizing B cell wastage from somatic hypermutation? J. Immunol. 156:3285. aspects of subunit association in immunoglobulin G. Biochemistry 13:4602. 10. Brown, M., G. D. Wiens, T. O’Hare, M. P. Stenzel-Poore, and M. B. Rittenberg. 26. Alexandru, I., D. I. C. Kells, K. J. Dorrington, and M. Klein. 1980. Non-covalent 1999. Replacements in the exposed loop of the T15 antibody VH CDR2 affect association of heavy and light chains of human immunoglobulin G: studies using carrier recognition of PC-containing pathogens. Mol. Immunol. 36:205. light chain labelled with a fluorescent probe. Mol. Immunol. 17:1351. 11. Wu, G. E., N. Hozumi, and H. Murialdo. 1983. Secretion of a ␭2 immunoglobulin 27. Kaloff, C. R., and I. G. Haas. 1995. Coordination of immunoglobulin chain fold- chain is prevented by a single amino acid substitution in its variable region. Cell ing and immunoglobulin chain assembly is essential for the formation of func- 33:77. tional IgG. Immunity 2:629. 12. Dul, J. L., and Y. Argon. 1990. A single amino acid substitution in the variable region of the light chain specifically blocks immunoglobulin secretion. Proc. 28. Lee, Y. K., J. W. Brewer, R. Hellman, and L. M. Hendershot. 1999. BiP and Natl. Acad. Sci. USA 87:8135. immunoglobulin light chain cooperate to control the folding of heavy chain and 13. Dul, J. L., O. R. Burrone, and Y. Argon. 1992. A conditional secretory mutant in ensure the fidelity of immunoglobulin assembly. Mol. Biol. Cell 10:2209. an Ig L chain is caused by replacement of a tyrosine/phenylalanine 87 with 29. Vanhove, M., Y.-K. Usherwood, and L. M. Herdershot. 2001. Unassembled Ig histidine. J. Immunol. 149:1927. heavy chains do not cycle from BiP in vivo but require light chains to trigger their 14. Horwitz, A. H., R. Nadell, F. Preugschat, and M. Better. 1994. Chimeric immu- release. Immunity 15:105. noglobulin light chains are secreted at different levels: influence of framework-1 30. Worn, A., and A. Pluckthun. 1998. Mutual stabilization of VL and VH in single- amino acids. Mol. Immunol. 31:683. chain antibody fragments, investigated with mutants engineered for stability. Bio- 15. Whitcomb, E. A., and P. H. Brodeur. 1998. Rearrangement and selection in the chemistry 37:13120. developing V␬ repertoire of the mouse: an analysis of the usage of two V␬ exons. 31. Knittler, M. R., S. Dirks, and I. G. Haas. 1995. Molecular chaperones involved J. Immunol. 160:4904. in protein degradation in the endoplasmic reticulum: quantitative interaction of 16. Knittler, M. R., and I. G. Haas. 1992. Interaction of BiP with newly synthesized the heat shock cognate protein BiP with partially folded immunoglobulin light immunoglobulin light chain molecules: cycles of sequential binding and release. chains that are degraded in the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA Downloaded from EMBO J. 11:1573. 92:1764. 17. Stevens, F. J., and Y. Argon. 1999. Pathogenic light chains and the B-cell rep- 32. Melnick, J., J. L. Dul, and Y. Argon. 1994. Sequential interaction of the chap- ertoire. Immunol. Today 20:451. erones BiP and GRP94 with immunoglobulin chains in the endoplasmic reticu- 18. Chothia, C., I. Gelfand, and A. Kister. 1998. Structural determinants in the se- lum. Nature 370:373. quences of immunoglobulin variable domain. J. Mol. Biol. 278:457. 33. Hendershot, L., J. Wei, J. Gaut, J. Melnick, S. Aviel, and Y. Argon. 1996. 19. Freund, C., A. Honegger, P. Hunziker, T. A. Holak, and A. Pluckthun. 1996. Inhibition of immunoglobulin folding and secretion by dominant negative BiP Folding nuclei of the scFV fragment of an antibody. Biochemistry 35:8457. ATPase mutants. Proc. Natl. Acad. Sci. USA 93:5269. 20. Hurle, M. R., L. R. Helms, L. Li, and R. Wetzel. 1994. A role for destabilizing 34. Frisch, C., H. Komar, A. Schmidt, G. Kleeman, A. Reinhardt, E. Pohl, I. Uson, amino acid replacements in light-chain amyloidosis. Proc. Natl. Acad. Sci. USA http://www.jimmunol.org/ 91:5446. T. R. Schneider, and H.-J. Fritz. 1996. Contribution of the intramolecular disul- 21. Martin, T. M., C. Kowalczyk, S. Stevens, G. D. Wiens, M. P. Stenzel-Poore, and fide bridge to the folding and stability of REIv, the variable domain of a human ␬ M. B. Rittenberg. 1996. Deletion in HCDR3 rescues T15 antibody mutants from immunoglobulin light chain. Folding and Design 1:431. a secretion defect caused by mutations in HCDR2. J. Immunol. 157:4341 35. Benhar, I., and I. Pastan. 1995. Identification of residues that stabilize the single- 22. Ohage, E., and B. Steipe. 1999. Intrabody construction and expression. I. The chain Fv of monoclonal antibodies B3. J. Biol. Chem. 270:23373. critical role of VL domain stability. J. Mol. Biol. 291:1119. 36. Jermutus, L., A. Honegger, F. Schwesinger, J. Hanes, and A. Pluckthun. 2001. 23. Wall, J. G., and A. Pluckthun. 1999. The hierarchy of mutations influencing the Tailoring in vitro evolution for protein affinity or stability. Proc. Natl. Acad. Sci. folding of antibody domains in Escherichia coli. Protein Eng. 12:605. USA 98:75. by guest on October 1, 2021