The JI
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     
 

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Messmer, B. T.
Right arrow Articles by Thaler, D. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Messmer, B. T.
Right arrow Articles by Thaler, D. S.
The Journal of Immunology, 1999, 162: 2184-2192.
Copyright © 1999 by The American Association of Immunologists

Two Human Neonatal IgM Antibodies Encoded by Different Variable-Region Genes Bind the Same Linear Peptide: Evidence for a Stereotyped Repertoire of Epitope Recognition1

Bradley T. Messmer*, James J. Sullivan*, Nicholas Chiorazzi{dagger}, Toby C. Rodman* and David S. Thaler2,*

* Sackler Laboratory for Molecular Genetics and Informatics, The Rockefeller University, New York, NY, 10021; and {dagger} Department of Rheumatology and Clinical Immunology, North Shore University Hospital, New York University School of Medicine, Manhasset, NY 11030


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Two monoclonal IgM Abs have been produced from lymphocytes isolated from two human umbilical cord bloods. These mAbs recognize a conformational epitope present in a CNBr digestion fraction of lactoferrin. Linear epitopes recognized by each mAb were selected from several phage display peptide libraries. In each case, phages displaying a peptide with a motif defined by [WF],G,[EQS],N were recovered. Phages displaying that motif bound equally well to either mAb but did not bind to control IgM. A peptide bearing this motif competed with the phage-displayed peptides for binding to either mAb. The same peptide also competes with a component of the CNBr digestion fraction of lactoferrin for Ab binding in ELISA. The Abs use different families of VH, JH, and VK gene cassettes but use the same JK cassette. All segments are virtually identical to their germline gene counterparts. This work provides further evidence that certain innate specificities are stereotyped among individuals.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Human serum contains Abs that have not arisen as a result of direct immunization with foreign Ag or pathogenic exposure 1 . These Abs have been called "natural" 2 or "innate" 3 . Several diverse roles for these naturally occurring Abs have been proposed, including a homeostatic role as part of a connectivity scheme 4, 5 ; a primary defensive role 6 , perhaps in conjunction with the complement system 7 ; and a role (through their loss or incorrect activation) in autoimmune disease 1 . It has also been suggested that natural Abs may have acquired several of these functions during the course of the immune system’s evolution 8 . Natural Abs are often, though not exclusively 9 , IgM class, and some cross-react with a variety of self and nonself Ags 10 . IgM class Abs do not cross the placenta in significant amounts during normal pregnancy 11 , presumably due to their large pentameric structure and the lack of a specific transporter mechanism as for IgG 12 . Therefore, the IgM Abs circulating in the umbilical cord and neonate are predominantly the product of neonatal lymphocytes (although 1 in 104–105 nucleated cells are of maternal origin 13, 14, 15). Therefore, in a normal, noninfected neonate the IgM Abs are predominately naturally occurring in that they were not stimulated by foreign Ags (excepting the possibility of Abs developed against maternally transferred Ags).

Several lines of evidence support the possibility that the naturally occurring Ab repertoire in human neonates contains certain stereotyped specificities. Homogenous tissue extracts separated on a one-dimensional gel and subsequently immunostained with umbilical cord blood sera have a pattern of reactivities conserved among cord blood samples but variable in adult sera 16, 17, 18, 19 . Furthermore, cord blood sera from many individuals contain significant titers of IgM reactive with a specific peptide epitope found on protamine and HIV Tat protein 20, 21 . Ig variable-region gene-usage bias has been noted and proposed as another factor potentially limiting the neonatal repertoire 22, 23, 24, 25 .

Here, we report another type of evidence that individuals have stereotypical natural Ab specificities. Two cord blood B cell hybridomas from different individuals were constructed. Each hybridoma secreted IgM Abs reactive with a conformational epitope of lactoferrin inferred to be present on sperm heads 26 . We used the method of screening phage-displayed peptides 27, 28, 29, 30 to identify linear peptides that bind to the Abs. In both cases, the Abs bound to phages displaying a specific peptide motif, indicating that the two hybridomas secrete Abs with near identical specificity. Furthermore, the peptide motif competed with the conformational epitope of lactoferrin for Ab binding. The sequences of the expressed rearranged heavy (H)3 and light (L) chain variable-region genes used by each hybridoma have revealed that there are virtually no somatic mutations and that the two hybridoma lines use different germline VH, JH, and VK genes to achieve this same specificity.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Libraries and Phage

Three phage-displayed peptide libraries were purchased from New England Biolabs (Beverly, MA). The random peptide is displayed at the N terminus of the gene III protein (pIII). There is a small leader sequence separating the peptide from the protein as shown: Ph.D.7, NH2-XXXXXXX-GGGS-pIII; Ph.D.12, NH2-XXXXXXXXXXXX-GGGS-pIII; Ph.D.C7C, NH2-ACXXXXXXXC-GGGS-pIII. These libraries were built in M13 mp18 and as such contain the lacZ{alpha} fragment. They will not form blue plaques on S26lacZ- 31 when plated on media containing Xgal. All phage stocks were grown on strain K91, an S26 derivative that has been cured of {lambda}.

Strain cclac+ f1 is an f1 phage containing a lacZ insertion. This phage forms blue plaques when plated on a lacZ- strain on media containing Xgal. This phage was a gift from Dr. C. Cupples and is described elsewhere 32 . All phages discussed here that result from a panning will be identified by a label as follows: {phi}(target Ab)-(library)x, where x is a letter unique to that phage clone. For example, a candidate from a Ph.D.7 panning against mAbRWL1 would appear as {phi}RWL1-7e. Where appropriate, the amino acid sequence of the displayed peptide will also be given.

Production of hybridomas and characterization of mAbRWL1 and mAbRWL2

mAbRWL1 was produced from a cord blood B cell hybridoma and originally characterized as binding to a fraction of human lactoferrin as previously described 26 . mAbRWL2 was produced by the same methodology, described briefly here. Mononuclear cells were isolated from cord blood (from different individuals for mAbRWL1 and mAbRWL2) by density gradient centrifugation and, after transformation with EBV, fused with parental cell line HMMA by the method of Chiorazzi et al. 33, 34 . Monoclonal hybridoma cell lines RWL1 and RWL2 secreted IgM reactive with fraction 7B of lactoferrin determined by ELISA as described 26 . IgM was purified from cell culture supernatants of both cell lines as follows. Whole-cell culture supernatant was concentrated to 1/100 original volume and dialyzed overnight in ddH2O buffered with HBSS, pH 7.2. The remaining culture supernatant was then passed through a centricon C100 column to remove all proteins <100 kDa. A sephacryl S100 fast flow column was prepared as per manufacture instructions (Pharmacia, Bridgewater, NJ) and washed with 10x volume of pH 7.2 PBS without CaCl2 or MgCl2. Then, 5 ml of concentrated culture supernatant was applied to the column bed and eluted at 1 ml/h for ~62 h. Fractions were read spectrophotometrically at 280 nm, revealing a sharp peak within eluants from 36–48 h. This peak was then pooled and reconcentrated to a volume of 2 ml and reapplied to the column at a flow rate of 2 ml/h. Fractions were read as above, revealing a strong peak in the eluants from 20–24 h. SDS-PAGE with 2-ME confirmed the presence of a 28-kDa L chain and 72-kDa H chain band.

mAbRWL1 and mAbRWL2 were tested by described methods 9 against a panel of Ags consisting of porcine insulin, porcine thyroglobulin, porcine actin, single-stranded calf thymus DNA, phosphocholine, human IgG (Sigma, St. Louis, MO), and fraction 7B of lactoferrin 26 . The two mAbs only gave a positive signal against fraction 7B of lactoferrin (data not shown) and are thus considered to be monospecific.

Rearranged Ig variable-region sequencing

Total RNA was isolated using Ultraspec RNA (Biotecx Laboratories, Houston, TX) according to the manufacturer’s instructions. First, 1 µg of RNA was reverse transcribed to cDNA using 200 U of Moloney murine leukemia virus reverse transcriptase (Life Technologies, Grand Island, NY), 1 U of RNase inhibitor (5 Prime 3 Prime, Boulder, CO) and 20 pmol of oligo dT primer in a total volume of 20 µl. The reaction mixture was incubated at 42°C for 1 h, heated at 65°C for 10 min to stop the reactions, and then diluted to a final volume of 100 µl.

To determine the VH and VL gene families used by the cell lines, 2 µl of cDNA were amplified using a sense VH or VL family-specific framework region (FR) primer in conjunction with the appropriate antisense CH or CL primer 35 . The reactions were conducted in 50 µl using 20 pmol of each primer and cycled with a 9600 GeneAmp System (Perkin-Elmer Cetus, Emeryville, CA) as follows: denaturation at 94°C for 45 s; annealing at 65°C for 45 s; and extension at 72°C for 45 s. After 35 cycles, extension was continued at 72°C for an additional 10 min.

The DNA sequence of the VH and VL genes were then determined by reamplifying 5 µl of the original cDNA using the appropriate VH or VL leader and CH or CL primers. These reactions were conducted as follows: denaturation at 94°C for 45 s; annealing at 62°C for 30 s; and extension at 72°C for 45 s. After 30–32 cycles, extension was continued at 72°C for an additional 10 min. PCR products were sequenced directly after purification with Wizard PCR Preps (Promega, Madison, WI) using an automated sequenator (Applied Biosystems, Foster City, CA). cDNAs were prepared and amplified from at least two different samples of each cell line. Each PCR product was subject to three separate sequencing runs. Sequences were compared with those in the V BASE sequence directory 36 using MacVector software, version 6.0 (Eastman Kodak, New Haven, CT).

Panning

For each target/library combination, a separate well of a 96-well high-capacity ELISA plate (catalogue number 25805-96, Corning, Corning, NY) was coated with 1–5 µg of target Ab in 100 µl PBS, pH 7.4. Plates were left at 4°C overnight. Coating solution was removed, and 200 µl blocking buffer (0.5% BSA) was added. Plates were left at room temperature for at least 1 h. Blocking buffer was removed, and the wells were washed six times with 200 µl PBS plus 0.1% Tween 20. A total of 1011 library phages in 100 µl PBS were added, and the plates were left on the bench for at least 1 h. Wells were washed 10 times with PBS plus 0.1% Tween 20. 100 µl of elution buffer (0.2 M glycine-HCl, pH 2.2) was added and removed after 20 min then neutralized with 15 µl Tris-HCl, pH 9.6. The eluant was titerd and 50 µl was added to 100 µl of a fresh overnight K91 culture. After at least 15 min on the bench, the infected culture was added to 5 ml Luria-Bertani medium and grown at 37°C with shaking for ~10 h. The following morning the cultures were spun at 5000 x g for 5 min, and the supernatant filter was sterilized by passage through a 0.45-µm syringe tip filter. Then, 1.0 ml of supernatant was mixed with 250 µl polyethylene glycol (PEG) solution (30% PEG m.w. 8000, 1.6 M NaCl) in a small centrifuge tube, left at room temperature for at least 20 min, and spun at 14,000 x g for 5 min. The supernatant was removed, and the tube was respun for 30 s. The last trace of PEG solution was removed by pipet, and the cloudy phage pellet was resuspended in 100 µl PBS. Then, 10 µl of this concentrated phage was used for the next round of panning. In the third round, 10 µl of cclac+ f1 phage (1012/ml) were added with the library phage, thus giving a roughly 1:1 ratio between library phage and cclac+ f1. The third round eluant was titrated on K38lacZ- with Xgal and isopropyl ß-D-thiogalactoside. Library phages form white plaques while the cclac+ phages form blue plaques. A white:blue plaque ratio greater than 100:1 was usually indicative of a successful panning.

Candidate phage sequencing and analysis by eluant phage titer

A dilution series of the final-round eluant was plated on a lawn of K91. Individual plaques were picked with a pasteur pipet from a plate with well dispersed plaques, vortexed in 5 ml of Luria Bertani medium, and incubated with shaking at 37°C for 10–18 h. The phage containing supernatant was filtered as above. Between 1.5 and 3.0 ml was used to prepare ssDNA with a Qiagen (Valencia, CA) qiaspin kit as per manufacturer instructions, and this was submitted for sequencing at the Rockefeller University Sequencing Facility.

Individual phage candidates were tested for binding specificity as follows. First, 1 ml of the filtered phage was concentrated as above. Concentrated phages were mixed with roughly equal numbers of cclac+ phages (which form blue plaques when plated with Xgal) in PBS and panned against mAbRWL1, mAbRWL2, human IgM (myeloma; purchased from Jackson Immunoresearch Laboratories, West Grove PA), and a well with no target by the procedure above. Successful candidates showed a large white:blue enrichment in the eluant from the specific mAb-coated well but not from either control well when plated on media containing Xgal (20 µg/ml).

Candidate phage analysis by ELISA

Individual wells of a 96-well high-capacity ELISA plate were coated with 100 µl of ~1010/ml phage candidates and left at 4°C overnight. The following day, the wells were blocked with 200 µl blocking buffer (5% BSA) and left at room temperature for 2 h. The plates were washed, and 100 µl of a 100 ng/ml Ab solution was placed in each well. After 1 h at room temperature, the plates were washed, and 100 µl of a 1:1000 dilution of anti-human IgM-Peroxidase (Sigma) was added to each well. After 1 h, the plates were washed, and 200 µl of o-phenylenediamine dihydrochloride (OPD) substrate (Sigmafast tablets as per instructions) was added. Development proceeded for 30 min before the ODs were measured on a plate reader (Emax Multi Plate Reader, Molecular Devices, Palo Alto, CA) at 450 nm.

Peptides

All peptides were purchased from Biosynthesis (Lewisville, TX) at the "discovery" scale. Peptides were analyzed by mass spectrometry by the manufacturer and the spectra provided. Table IGo shows the amino acid sequence of the various peptides used in this study.


View this table:
[in this window]
[in a new window]
  		Table I. Peptide sequences

 

   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Ig variable-region sequence analysis of hybridoma cell line RWL1 and RWL2

cDNA was prepared from cloned hybridoma cell lines RWL1 and RWL2 by the RT-PCR technique described in Materials and Methods, and the rearranged variable regions of the H and L chains were sequenced. The two cell lines express different rearranged VH genes derived from separate VH families (RWL1, VH4–39 (previously DP79); and RWL2, VH3–09 (previously DP31)) and consequently are significantly different from each other in both nucleotide (Fig. 1Go) and amino acid sequence (Fig. 2Go). However, when compared with the appropriate ancestral germline gene counterpart, only rare differences are noted (Fig. 1Go). These could represent heretofore unidentified allelic polymorphisms or somatic mutations. Similarly, the two lines express different VK genes (RWL1, PPK21; and RWL2, DPK24) derived from different VK families (VKIII vs VKIV, respectively), that are also very similar to the germline counterparts (Fig. 1Go). Although the JH segments in the two lines differed, they both used identical JK4 genes with identical amino acid sequences. In both instances, there was insufficient similarity to the described D genes to make a D segment assignment by the method of Corbett 37 for either rearranged VHDJH.



View larger version (27K):
[in this window]
[in a new window]
  		FIGURE 1. Nucleotide sequences of rearranged variable regions of mAbRWL1 (A) and mAbRWL2 (B) and nearest germline gene from VBASE database. Dots indicate identical nucleotides. Where there are differences between the Ab sequence and the corresponding germline, the germline nucleotide is indicated below.

 


View larger version (56K):
[in this window]
[in a new window]
  		FIGURE 2. Amino acid sequences of the H and L chains of mAbRWL1 and mAbRWL2. Darkly shaded residues are identical between the two sequences, while light shading indicates that the residues are members of the same functional group. Alignment was done by the Clustal method, and the CDRs were labeled according to the Kabat scheme (38). The asterisk at position 28 of mAbRWL2 indicates the only amino acid substitution resulting from a deviation of the nucleotide sequence from germline. The germline encoded residue is a serine (S).

 
Thus, the expressed rearranged Ig V genes of these two cell lines are very different based primarily on the use of distinct V gene segments and not on the basis of somatic alterations. The amino acid sequences are also very different, particularly in the complementarity determining region (CDR) (Fig. 2Go). Therefore, differences in the specificities of the Abs coded by these two sets of genes would not be surprising.

Epitope definition by phage display

Three phage-displayed peptide libraries obtained from New England Biolabs (Ph.D.-7, Ph.D.-12, and Ph.D.-C7C) were independently "panned" (see Materials and Methods) against each mAb. After three rounds of panning, individual plaques were used to grow phage cultures for sequencing and testing. In total, six independent pannings were conducted (3 libraries x 2 mAbs). The displayed peptide sequences derived from the nucleotide sequences of the phage candidates are shown in Fig. 3Go.



View larger version (61K):
[in this window]
[in a new window]
  		FIGURE 3. Amino acid sequences of displayed peptides present on phages recovered from the third round of panning against mAbRWL1 and mAbRWL2. Numbers after a sequence indicate the number of separate clones recovered with that sequence. Sequences are aligned by the consensus motif, which appears in the shaded regions.

 
The motif [WF],G,[EQS],N clearly dominates the population of recovered phages. The conserved motif most often appears at the N terminus of the displayed peptide. The N terminus of the peptide is the N terminus of the pIII in the Ph.D. 7 and Ph.D. 12 libraries. In the cysteine-constrained library, the N terminus of the peptide is followed by a cysteine and an N-terminal alanine. Thus in either case the displayed motif shows a bias for the most N-terminal position possible. A likely explanation is that N-terminal sequences are further from the phage body and thus more accessible for interaction with the Ab.

Tryptophan (W) and phenylalanine (F) both contain aromatic rings although tryptophan is also polar. Their apparent interchangability is most likely due to that aromatic ring. The different residues that occupy the third position of the motif show some similarity as well. Glutamine (Q) and glutamic acid (E) have side chains with acidic groups, but serine (S) has a polar hydroxy group. There may be influences from those residues outside the core motif, but testing such a hypothesis was beyond the scope of this work.

Phages bearing the consensus motif are specific for mAbRWL1 and mAbRWL2

Six phage candidates, one from each panning, were tested for specific binding to mAbRWL1 and mAbRWL2 (Fig. 4GoA). Each phage candidate displays a unique peptide that contains the consensus motif, as shown in Fig. 4Go. Candidate phages were panned (see Materials and Methods) against mAbRWL1, mAbRWL2, IgM myeloma, and blank (no Ab) wells. A roughly equal number of cclac+ f1 were included as an indicator of the background level of nonspecific binding. The titer of the eluant from wells coated with either mAb was at least 100-fold greater than that from the control wells for all six candidate phages. This indicates that phage-displayed peptide sequences selected as binding to one of the mAbs also bind the other but do not bind to unrelated IgM. The level of nonspecific phage binding, indicated by the eluant titer of cclac+ f1, was the same regardless of the presence of Ab on the well, suggesting that the greater eluant titer from the mAb-coated wells is due to a specific interaction with the displayed peptides.



View larger version (49K):
[in this window]
[in a new window]
  		FIGURE 4. Specificity of phage clones identified as binding mAbRWL1 and mAbRWL2. A, 100 µl of a roughly equal mix of candidate phage and cclac+ f1, both at ~1010/ml, were added to wells coated overnight with 1 µg/ml of Ab and subsequently blocked with 0.5% BSA. In all cases, an equal number of cclac+ f1 phages were included and eluted at a titer of ~105/ml. The sequence to the right is the amino acid sequence of the peptide displayed by that phage. Error bars are ± SD of three experiments. B, Wells were coated with 100 µl of phage at 1010/ml overnight and subsequently blocked with 5% BSA. Primary Ab (mAbRWL1, mAbRWL2, or human IgM myeloma) was added at a concentration of 100 ng/ml. Anti-human IgM (µ-chain specific) horseradish peroxidase conjugate was then added at a dilution of 1:1000. Development was done with OPD tablet substrate for 30 min. {phi}CLL57-7a is a phage from the same 7-mer library that had been selected against an unrelated target. Phage nomenclature is defined in Materials and Methods section.

 
A second assay confirms and extends the results above. While demonstrating the specificity of the phage candidates, the results above do not unequivocally show that the phages are interacting with the mAbs and not some other component of the coating solution (i.e., a contaminant in the Ab preparation). This possibility was addressed by an ELISA in which the mAbs and IgM myeloma were each used as a primary reagent with the phages as targets (Fig. 4GoB). Ab binding was detected with {alpha}IgM-horseradish peroxidase as described in Materials and Methods. Two control phages were included: cclac+ f1, which has no displayed peptide, and {phi}CLL57-7a, which displays a peptide selected in an unrelated experiment. Each of the six candidate phages gave a positive signal when mAbRWL1 or mAbRWL2 was used as primary reagent, but gave no signal above background when the IgM myeloma was used. Furthermore, none of the Abs gave a signal above background against cclac+ or {phi}CLL57-7a, reaffirming the specificity of mAbRWL1 and mAbRWL2. The congruent results of these two experimental approaches strongly support the interpretation that these phages are directly and specifically interacting with these mAbs.

Phage-displayed peptide motif functions as a free peptide

It is possible that the phage-displayed peptide motif requires some structural contribution from the phage body. If such were the case, a small peptide containing the motif would lack the activity of the phage-displayed motif. Several peptides (Table IGo) were tested for their ability to inhibit consensus peptide-bearing phage {phi}RWL1-C7b (with displayed peptide sequence FGENTGY; consensus motif in bold) binding to the mAbs over a wide concentration range. Peptide D017-1 (sequence WGENNYTSQIRP) contains the consensus motif, while peptide D058-1(sequence GPRPPLNHALHS) has an unrelated sequence and is a control. As shown in Fig. 5GoA, peptide D017-1 inhibited phage-displayed peptide binding in a concentration-dependent manner, while the peptide D058-1 had no effect at any concentration. Neither peptide had any effect on the level of nonspecific binding as indicated by the constant level of cclac+ f1 phage binding.



View larger version (28K):
[in this window]
[in a new window]
  		FIGURE 5. Free peptides compete with phage-displayed peptides for Ab binding. A, Pannings were done essentially as described in Materials and Methods. After blocking the coated wells with 0.5% BSA, cclac+ phage (1010/ml) and peptide were added in 100 µl to the wells. After 1 h, {phi}RWL1-C7b was added to a final concentration of 1010/ml. B, Wells were coated with 1 µg/ml mAbRWL1 overnight and blocked with 0.5% BSA. Peptide was added as before at a concentration of 10 µg/ml. cclac+ phages were included at ~1010/ml and eluted at 2–5 x 105/ml in all cases. Error bars are ± SD of three experiments. C, Wells were coated with 100 µl of 100 µg/ml peptide. After blocking, 1 µg/ml mAbRWL1 or mAbRWL2 was added. Detection was by anti-human IgM-horseradish peroxidase conjugate with OPD as the substrate.

 
Peptide D017-1 also inhibited binding to mAbRWL1 by five other phages displaying different peptides, all containing the consensus motif, as shown in Fig. 5GoB. The recovered phage titer from the candidate testing protocol was no higher than background for each of these phages in the presence of peptide D017-1 at a concentration of 10 µg/ml.

Peptides were also tested in a direct ELISA in which mAbRWL1 and mAbRWL2 were used as primary Abs against wells coated with various peptides (Fig. 5GoC). Only peptide D017-1 gave a signal above background. Furthermore, that signal was found to be dependent on the peptide concentration (data not shown). Peptide D531-1 (PRIQSTYNNEGW) has the reverse amino acid sequence of peptide D017-1 and does not give a positive signal. Thus the Ab binding is dependent on the order of amino acids in peptide D017-1 and not simply the amino acid composition.

A peptide containing the consensus motif competes with fraction 7B for binding to mAbRWL1

Competitive ELISAs were performed to determine whether peptide D017-1 could compete with the previously identified ligand of mAbRWL1 and mAbRWL2, CNBr digestion fraction 7B of lactoferrin 26 . Peptide D017-1 blocks mAbRWL1 from binding to fraction 7B in a concentration-dependent manner (Fig. 6Go), while peptides D017-2, D118, and C977 have no effect. Similar results have been obtained with mAbRWL2 (data not shown). The sequence of C977 is taken from the SWISS-PROT database sequence of human lactoferrin and is the only region of lactoferrin to contain a sequence similar to the consensus motif identified here.



View larger version (19K):
[in this window]
[in a new window]
  		FIGURE 6. Phage-derived peptides compete with the CNBr fraction 7B of lactoferrin. A, Peptide D017-1 blocks binding of mAbRWL1 to fraction 7B of lactoferrin over a wide range of Ab concentration. Wells were coated overnight with 100 µl of 0.5 µg/ml fraction 7B. Plates were blocked with 3% BSA. Dilutions of Ab were premixed with peptide at 10 µg/ml for 1 h before being put on blocked wells. Detection was by anti-human IgM-horseradish peroxidase conjugate with OPD substrate for development. B, Peptide D017-1 blocks binding of mAbRWL1 to fraction 7B of lactoferrin but control peptides do not. Error bars are ± SD of three experiments.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Here we describe two monoclonal IgM Abs, derived from the cord blood of different individuals, that specifically recognize the same linear phage-displayed epitope. The two Abs each selected phage bearing the same consensus motif from several phage-displayed peptide libraries. The exact amino sequences of the selected peptides were not the same. This difference could reflect differences in the fine specificity of the two Abs. On the other hand, it could be a consequence of sampling in the peptide display library. The peptide library has a diversity of ~109. The consensus motif is four amino acids in length and is thus estimated to be present in ~6 x 104 unique members of the library. A peptide sequence initially selected on phages binds specifically to both mAbs as free peptide D017-1 while several other peptides, including one with the reverse amino acid sequence, do not. Peptide D017-1 also competes for mAb binding with the conformational epitope in fraction 7B of CNBr digested lactoferrin 26 . A sequence similar to the consensus motif is present in the sequence of lactoferrin. However, when a peptide is made with that sequence (C977) it does not compete with fraction 7B of lactoferrin for Ab binding. We propose that our linear peptide motif is a "mimotope" of a conformational epitope present in fraction 7B.

The exact site of the peptide-Ab interaction has not been completely defined. None of the tested phage containing the consensus sequence bind human IgM myeloma, indicating that the constant regions of the µ H chain are not responsible for this binding. Furthermore, because the VH and VL genes of these two mAbs are derived from distinct VH and VL gene families it also seems unlikely that variable FR similarities can be responsible for the peptide interactions in a superantigenic fashion, although this cannot be completely excluded 38, 39, 40, 41 . Finally, the sharing of JK4 segments alone seems unlikely to explain these interactions, especially because the junctional residues between the VL and JL are different in each case.

Therefore, it is more likely that the different CDRs are able to form binding regions with chemical and steric characteristics sufficiently similar to give rise to nearly identical ligand selection. Examples of such mimicry have been documented at high resolution. Two Abs composed of different gene cassettes and with CDRs of different amino acid composition and length were found to bind Guinea Fowl lysozyme by contacting the same twelve amino acid epitope via different configurations of the Ags side chains 42 . In another case, two mAbs that bound to neuraminidase were found to recognize epitopes that overlapped by 80%. Using different amino acid sequences in the CDRs responsible for Ag contact, these two Abs were found to create similar interactions with the shared epitope using different amino acids and different side-chain geometries 43 . It is noteworthy that in the former case, side-chain flexibility of the Ag contributed to the induced fit required for specific binding. The phage-displayed peptide sequences presented here may behave similarly.

There is no evidence of significant somatic mutation in the sequence of the Ab genes. This is consistent with the conventional view that IgM sequences contain few somatic mutations with notable exceptions 44 . The deviations from the germline sequences that occur at the VDJ and VJ junctions are most likely the product of recombinational and exonuclease activities 45 . There are also few amino acids that resulted from N-addition. This is not surprising as it has previously been shown that there is limited N-terminal addition in neonatal Ig genes in the human 46 and almost none in the mouse 47 . This suggests that the specificity exhibited by these Abs is inherent in these arrangements of germline genes and is not the product of a mutational history. Moreover, the two mAbs use the same JK family gene but use different genes for the VH, JH, and VK and have very different CDRs at the amino acid level. That they then recognize the same peptide epitope implies that there is a degree of redundancy in the inherent specificities of germline configuration Abs. Alternatively, the JK or shared FR of the L chains may play a key role in determining the shared specificities of mAbRWL1 and mAbRWL2.

We are intrigued by the possibility that individuals are born already synthesizing an Ab repertoire that recognizes a particular set of epitopes. Several lines of evidence support such a hypothesis of stereotyped specificities but each line of evidence has significant limitations.

The results presented in this paper are limited to the analysis of two mAbs that recognize a single epitope. As such it is somewhat tenuous to generalize from these findings. However, previous studies have shown that adult and cord blood serum from many individuals contain Abs that react with the CNBr digestion fraction 7B of lactoferrin 26, 48 . Therefore, we propose that the cord blood of most, if not all, individuals will contain Abs reactive with this epitope. We further propose that this epitope specificity is one of many that are present in serum from most, if not all, individuals at birth. One of the strongest lines of evidence for stereotyped specificities among the cord blood IgM repertoire comes from Kazatchkine and colleagues 18 . They have generated profiles of autoreactive Abs present in sera by separating homogenous tissue extracts on a polyacrylamide gel to resolve protein bands. Whole serum is then used as a primary immunostain and detected with labeled anti-IgM Abs. The resulting densiometric profiles from different individuals can be compared. They find that IgM profiles generated from cord blood serum are essentially invariant from individual to individual. The primary limitation of this approach is that it is not known what Ags or epitopes react with the serum Abs to generate the signals. The method does not distinguish whether many monoreactive or a few polyreactive Abs are responsible for the various peaks in their profiles.

Numerous studies have addressed the possibility that the fetal and neonatal repertoires could be restricted as part of a developmental program of gene usage bias. Preferential use of JH proximal VH genes 22 in mouse pre-B cells and the prevalence of diversity segment DHQ52, as well as significantly restricted VH gene usage among second trimester fetal humans 23 , have been noted. However, the restrictions placed on the early human fetal repertoire are not as apparent in the cord blood 24, 25 , although the cord blood {kappa} repertoire has been found to overuse a small subset of VK genes 49 .

Neonatal B cells are mostly CD5+, whereas the adult contains a greater proportion of CD5- B cells 50 . When VH gene usage by adult CD5+ and CD5- peripheral B cells was examined, there was not a significant difference between the two populations. In both cases, the VH3 family genes were found at a slightly greater frequency than expected by random selection 51 . However, the CD5- population contains the "sister" cells (B1b) of the CD5+ (B1a) class as wells as the conventional B2 cells 9 . When the CD5- population was separated into B2 and B1b populations (by the level of CD45RA expression and CD5 mRNA levels) and compared with the CD5+ B1a population, it was found that the B2 population was more restricted to VH3 family genes whereas the B1a and B1b populations use both VH3 and VH4 family genes 52 . EBV-transformed B cell clones that secreted polyspecific Abs were not found to have significant VH or VL gene usage pattern differences when compared with clones of unknown specificity 53 . Analysis of B cells at different stages of ontogeny has suggested that there may be sequential mechanisms responsible for some of these observed biases 54 . It will most likely be some time before there is a complete understanding of the mechanistic origins and patterns of variable-region gene use bias.

There is a significant limitation to analysis of Ig variable-region gene sequences when one is concerned with the repertoire of Ab specificities. It is not currently possible to predict the Ag specificity of an Ab by looking at its variable-region sequence. Even if the protein-folding problem were solved, it would still not be likely that predictions of epitope specificity could always be made from variable-region gene sequences. Variable-region sequences and epitope specificities do not exist as discreet linked pairs. There is not a strict one-to-one correspondence between an Ab and a single epitope. On the one hand, a single variable-region sequence may be capable of binding several diverse Ags as is the case for polyspecific Abs 1 . On the other hand, the data presented here and in the lysozyme example discussed above 42 imply that Abs that differ in variable-region gene usage and the amino acid sequence of the CDRs can share a near identical epitope specificity. Thus the number of different variable-region sequences present may not accurately reflect the number of different epitope specificities embodied therein, and the error could be in either direction. These complications lie at a key juncture in our understanding of immune responses to pathogens as well as the tendency to fall prey to deleterious autoreactive Abs.

In conclusion, we have presented further evidence that human neonates have naturally occurring Abs of defined specificity. Furthermore, our evidence from two mAbs derived from different individuals’ cord blood indicates that such defined specificities can be generated by multiple combinations of essentially germline configuration genes. Elucidating the physiologically relevance of this redundantly encoded specificity remains a challenging problem.


   Acknowledgments
 
We thank Peter Model and Marjorie Russel for bacterial strains and considerable phage knowledge. We also thank Joshua Lederberg, Davorka Bilicic, and Michel Nussenzweig for encouragement and insight during the course of this work and David Haig, Lucio Luzatto, Kenji Adzuma, and Sylvie Beaulieu for comments on the manuscript.


   Footnotes
 
1 This work was supported by National Institutes of Health Virology Training Grant CA09673-22 to Rockefeller University (B.T.M.), the Harry Winston Foundation (T.C.R. and J.J.S.), National Institutes of Health Grant AI 10811 (N.C.), and the Sloan Foundation (D.S.T.). Back

2 Address correspondence and reprint requests to Dr. David S. Thaler, Sackler Laboratory of Molecular Genetics and Informatics, The Rockefeller University, 1230 York Avenue, New York, NY 10021-6399. E-mail address: Back

3 Abbreviations used in this paper: H, heavy; L, light; pIII, gene III peptide; FR, framework region; OPD, o-phenylenediamine dihydrochloride; CDR, complementarity determining region. Back

Received for publication September 8, 1998. Accepted for publication November 10, 1998.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Casali, P., E. W. Schettino. 1996. Structure and function of natural antibodies. Curr. Top. Microbiol. Immunol. 210:167.[Medline]
  2. Boyden, S. V.. 1966. Natural antibodies and the immune response. Adv. Immunol. 5:1.[Medline]
  3. Rodman, T. C., S. E. To, J. J. Sullivan, R. Winston. 1997. Innate natural antibodies: primary roles indicated by specific epitopes. Hum. Immunol. 55:87.[Medline]
  4. Holmberg, D., A. Andersson, L. Carlsson, S. Forsgren. 1989. Establishment and functional implications of B-cell connectivity. Immunol. Rev. 110:89.[Medline]
  5. Perelson, A. S., G. Weisbuch. 1997. Immunology for physicists. Rev. Mod. Phys. 69:1219.
  6. Casali, P., A. L. Notkins. 1989. Probing the human B-cell repertoire with EBV: polyreactive antibodies and CD5+ B lymphocytes. Annu. Rev. Immunol. 7:513.[Medline]
  7. Thornton, B. P., V. Vetvicka, G. D. Ross. 1994. Natural antibody and complement-mediated antigen processing and presentation by B lymphocytes. J. Immunol. 152:1727.[Abstract]
  8. Bouvet, J. P., G. Dighiero. 1998. From natural polyreactive autoantibodies to a la carte monoreactive antibodies to infectious agents: is it a small world after all?. Infect. Immun. 66:1.[Free Full Text]
  9. Kasaian, M. T., H. Ikematsu, P. Casali. 1992. Identification and analysis of a novel human surface CD5- B lymphocyte subset producing natural antibodies. J. Immunol. 148:2690.[Abstract]
  10. Casali, P., A. L. Notkins. 1989. CD5+ B lymphocytes, polyreactive antibodies and the human B-cell repertoire. Immunol. Today 10:364.[Medline]
  11. Goudeau, A., B. Yvonnet, G. Lesage, F. Barin, F. Denis, P. Coursaget, J. P. Chiron, I. Diop Mar. 1983. Lack of anti-HBc IgM in neonates with HBsAg carrier mothers argues against transplacental transmission of hepatitis B virus infection. Lancet 2:1103.[Medline]
  12. Malek, A., R. Sager, P. Kuhn, K. H. Nicolaides, H. Schneider. 1996. Evolution of maternofetal transport of immunoglobulins during human pregnancy. Am. J. Reprod. Immunol. 36:248.
  13. Petit, T., M. Dommergues, G. Socie, Y. Dumez, E. Gluckman, O. Brison. 1997. Detection of maternal cells in human fetal blood during the third trimester of pregnancy using allele-specific PCR amplification. Br. J. Haematol. 98:767.[Medline]
  14. Petit, T., E. Gluckman, E. Carosella, Y. Brossard, O. Brison, G. Socie. 1995. A highly sensitive polymerase chain reaction method reveals the ubiquitous presence of maternal cells in human umbilical cord blood. Exp. Hematol. 23:1601.[Medline]
  15. Scaradavou, A., C. Carrier, N. Mollen, C. Stevens, P. Rubinstein. 1996. Detection of maternal DNA in placental/umbilical cord blood by locus-specific amplification of the noninherited maternal HLA gene. Blood 88:1494.[Abstract/Free Full Text]
  16. Lacroix-Desmazes, S., L. Mouthon, A. Coutinho, M. D. Kazatchkine. 1995. Analysis of the natural human IgG antibody repertoire: life-long stability of reactivities towards self antigens contrasts with age-dependent diversification of reactivities against bacterial antigens. Eur. J. Immunol. 25:2598.[Medline]
  17. Mouthon, L., M. Haury, S. Lacroix-Desmazes, C. Barreau, A. Coutinho, M. D. Kazatchkine. 1995. Analysis of the normal human IgG antibody repertoire: evidence that IgG autoantibodies of healthy adults recognize a limited and conserved set of protein antigens in homologous tissues. J. Immunol. 154:5769.[Abstract]
  18. Mouthon, L., A. Nobrega, N. Nicolas, S. V. Kaveri, C. Barreau, A. Coutinho, M. D. Kazatchkine. 1995. Invariance and restriction toward a limited set of self-antigens characterize neonatal IgM antibody repertoires and prevail in autoreactive repertoires of healthy adults. Proc. Natl. Acad. Sci. USA 92:3839.[Abstract/Free Full Text]
  19. Mouthon, L., S. Lacroix-Desmazes, A. Nobrega, C. Barreau, A. Coutinho, M. D. Kazatchkine. 1996. The self-reactive antibody repertoire of normal human serum IgM is acquired in early childhood and remains conserved throughout life. Scand. J. Immunol. 44:243.[Medline]
  20. Rodman, T. C., F. H. Pruslin, Y. Chauhan, S. E. To, R. Winston. 1988. Protamine-reactive natural IgM antibodies in human sera: characterization of the epitope demonstrates specificity of antigenic recognition; occurrence indicates obscurity of origin and function. J. Exp. Med. 167:1228.[Abstract/Free Full Text]
  21. Rodman, T. C., F. H. Pruslin. 1990. Identification of a low-affinity subset of protamine-reactive IgM antibodies present in normal, deficient in AIDS, sera: implications for HIV latency. Clin. Immunol. Immunopathol. 57:430.[Medline]
  22. Yancopoulos, G. D., S. V. Desiderio, M. Paskind, J. F. Kearney, D. Baltimore, F. W. Alt. 1984. Preferential utilization of the most JH-proximal VH gene segments in pre-B-cell lines. Nature 311:727.[Medline]
  23. Jr Schroeder, H. W., J. L. Hillson, R. M. Perlmutter. 1987. Early restriction of the human antibody repertoire. Science 238:791.[Abstract/Free Full Text]
  24. Mortari, F., J. Y. Wang, Jr H. W. Schroeder. 1993. Human cord blood antibody repertoire: mixed population of VH gene segments and CDR3 distribution in the expressed C{alpha} and C{gamma} repertoires. J. Immunol. 150:1348.[Abstract]
  25. Mortari, F., J. A. Newton, J. Y. Wang, Jr H. W. Schroeder. 1992. The human cord blood antibody repertoire: frequent usage of the VH7 gene family. Eur. J. Immunol. 22:241.[Medline]
  26. Rodman, T. C., R. Winston, J. J. Sullivan, X. Yan, N. Chiorazzi. 1997. An innate natural antibody is reactive with a cryptic sequence of lactoferrin exposed on sperm head surface. Proc. Soc. Exp. Biol. Med. 216:404.[Medline]
  27. McGregor, D.. 1996. Selection of proteins and peptides from libraries displayed on filamentous bacteriophage. Mol. Biotechnol. 6:155.[Medline]
  28. Cortese, R., F. Felici, G. Galfre, A. Luzzago, P. Monaci, A. Nicosia. 1994. Epitope discovery using peptide libraries displayed on phage. Trends Biotechnol. 12:262.[Medline]
  29. Smith, G. P., J. K. Scott. 1993. Libraries of peptides and proteins displayed on filamentous phage. Methods Enzymol. 217:228.[Medline]
  30. Scott, J. K., G. P. Smith. 1990. Searching for peptide ligands with an epitope library. Science 249:386.[Abstract/Free Full Text]
  31. Bachman, B. J. 1987. Derivations of Genotypes of Some Mutant Derivatives of Escherichia coli K-12. In Escherichia coli and Salmonella typhimurium Cellular and Molecular Biology, Vol. 2. F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger, eds. American Society for Microbiology, Washington D.C., p. 1190.
  32. Cupples, C. G., J. H. Miller. 1988. Effects of amino acid substitutions at the active site in Escherichia coli ß-galactosidase. Genetics 120:637.[Abstract/Free Full Text]
  33. Chiorazzi, N., R. L. Wasserman, H. G. Kunkel. 1982. Use of Epstein-Barr virus-transformed B cell lines for the generation of immunoglobulin-producing human B cell hybridomas. J. Exp. Med. 156:930.[Abstract/Free Full Text]
  34. Chiorazzi, N.. 1992. Generation of stable human autoantibody-secreting B cell hybridomas. Mol. Biol. Rep. 16:65.[Medline]
  35. Hashimoto, S., M. Dono, M. Wakai, S. L. Allen, S. M. Lichtman, P. Schulman, V. P. Vinciguerra, M. Ferrarini, J. Silver, N. Chiorazzi. 1995. Somatic diversification and selection of immunoglobulin heavy and light chain variable region genes in IgG+ CD5+ chronic lymphocytic leukemia B cells. J. Exp. Med. 181:1507.[Abstract/Free Full Text]
  36. Tomlinson, I. M., S. C. Williams, S. J. Corbett, J. B. L. Cox, G. Winter. 1996. V BASE Sequence Directory Medical Research Council Centre for Protein Engineering, Cambridge, U.K.
  37. Corbett, S. J., I. M. Tomlinson, E. L. L. Sonnhammer, D. Buck, G. Winter. 1997. Sequence of the human immunoglobulin diversity (D) segment locus: a systematic analysis provides no evidence for the use of DIR segments, inverted D segments, "minor" D segments or D-D recombination. J. Mol. Biol. 270:587.[Medline]
  38. Kabat, E. A., T. T. Wu, H. M. Perry, K. S. Gottesman, C. Foeller. 1991. Sequences of Proteins of Immunological Interest National Institutes of Health, Bethesda.
  39. Silverman, G. J., R. Pires, J. P. Bouvet. 1996. An endogenous sialoprotein and a bacterial B cell superantigen compete in their VH family-specific binding interactions with human Igs. J. Immunol. 157:4496.[Abstract]
  40. Domiati-Saad, R., P. E. Lipsky. 1997. B cell superantigens: potential modifiers of the normal human B cell repertoire. Int. Rev. Immunol. 14:309.[Medline]
  41. Levinson, A. I., L. Kozlowski, Y. Zheng, L. Wheatley. 1995. B-cell superantigens: definition and potential impact on the immune response. J. Clin. Immunol. 15:26S.[Medline]
  42. Lescar, J., M. Pellegrini, H. Souchon, D. Tello, R. J. Poljak, N. Peterson, M. Greene, P. M. Alzari. 1995. Crystal structure of a cross-reaction complex between Fab F9.13.7 and guinea fowl lysozyme. J. Biol. Chem. 270:18067.[Abstract/Free Full Text]
  43. Malby, R. L., W. R. Tulip, V. R. Harley, J. L. McKimm-Breschkin, W. G. Laver, R. G. Webster, P. M. Colman. 1994. The structure of a complex between the NC10 antibody and influenza virus neuraminidase and comparison with the overlapping binding site of the NC41 antibody. Structure 2:733.[Medline]
  44. Wagner, S. D., V. Martinelli, L. Luzzatto. 1994. Similar patterns of V{kappa} gene usage but different degrees of somatic mutation in hairy cell leukemia, prolymphocytic leukemia, Waldenstrom’s macroglobulinemia, and myeloma. Blood 83:3647.[Abstract/Free Full Text]
  45. Carlsson, L., C. Overmo, D. Holmberg. 1992. Selection against N-region diversity in immunoglobulin heavy chain variable regions during the development of pre-immune B cell repertoires. Int. Immunol. 4:549.[Abstract/Free Full Text]
  46. Gu, H., I. Forster, K. Rajewsky. 1990. Sequence homologies, N sequence insertion and JH gene utilization in VHDJH joining: implications for the joining mechanism and the ontogenetic timing of Ly1 B cell and B-CLL progenitor generation. EMBO J. 9:2133.[Medline]
  47. Feeney, A. J.. 1990. Lack of N regions in fetal and neonatal mouse immunoglobulin V-D-J junctional sequences. J. Exp. Med. 172:1377.[Abstract/Free Full Text]
  48. Manchester, K. M., R. Winston, T. C. Rodman. 1997. Lactoferrin-reactive natural antibodies. Ann. NY Acad. Sci. 815:475.[Medline]
  49. Feeney, A. J., G. Lugo, G. Escuro. 1997. Human cord blood {kappa} repertoire. J. Immunol. 158:3761.[Abstract]
  50. Hardy, R. R., K. Hayakawa. 1994. CD5 B cells, a fetal B cell lineage. Adv. Immunol. 55:297.[Medline]
  51. Brezinschek, H. P., S. J. Foster, R. I. Brezinschek, T. Dorner, R. Domiati-Saad, P. E. Lipsky. 1997. Analysis of the human VH gene repertoire: differential effects of selection and somatic hypermutation on human peripheral CD5(+)/IgM+ and CD5(-)/IgM+ B cells. J. Clin. Invest. 99:2488.[Medline]
  52. Schettino, E. W., S. K. Chai, M. T. Kasaian, Jr H. W. Schroeder, P. Casali. 1997. VHDJH gene sequences and antigen reactivity of monoclonal antibodies produced by human B-1 cells: evidence for somatic selection. J. Immunol. 158:2477.[Abstract]
  53. Guigou, V., B. Guilbert, D. Moinier, C. Tonnelle, L. Boubli, S. Avrameas, M. Fougereau, F. Fumoux. 1991. Ig repertoire of human polyspecific antibodies and B cell ontogeny. J. Immunol. 146:1368.[Abstract]
  54. Kraj, P., S. P. Rao, A. M. Glas, R. R. Hardy, E. C. Milner, L. E. Silberstein. 1997. The human heavy chain Ig V region gene repertoire is biased at all stages of B cell ontogeny, including early pre-B cells. J. Immunol. 158:5824.[Abstract]



This article has been cited by other articles:


Home page
 BloodHome page
K. Stamatopoulos, C. Belessi, C. Moreno, M. Boudjograh, G. Guida, T. Smilevska, L. Belhoul, S. Stella, N. Stavroyianni, M. Crespo, et al.
Over 20% of patients with chronic lymphocytic leukemia carry stereotyped receptors: pathogenetic implications and clinical correlations
Blood, January 1, 2007; 109(1): 259 - 270.
[Abstract] [Full Text] [PDF]

Home page
 BloodHome page
B. T. Messmer, E. Albesiano, D. Messmer, and N. Chiorazzi
The pattern and distribution of immunoglobulin VH gene mutations in chronic lymphocytic leukemia B cells are consistent with the canonical somatic hypermutation process
Blood, May 1, 2004; 103(9): 3490 - 3495.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Messmer, B. T.
Right arrow Articles by Thaler, D. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Messmer, B. T.
Right arrow Articles by Thaler, D. S.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS