Faculty Bibliography

Amyloid fibrils were found at postmortem examination in a 70 year old woman with generalized amyloidosis associated with renal carcinoma (hypernephroma). Clinically, her amyloid disease presented as nephrotic syndrome. It was demonstrated by electrophoretic and amino acid sequence analysis studies that the amyloid fibrils contained AA protein identical to that found in amyloidosis associated with chronic inflammatory and infectious diseases as well as in the genetic form of familial Mediterranean fever.

The 11/2-yr course of idiopathic systemic amyloidosis in a 63-yr-old woman was characterized by inanition, subcutaneous ecchymoses and giant hepatomegaly, the liver weighing 8.5 kg at autopsy. Skeletal survey and bone marrow aspirate were normal. The major components of isolated amyloid fibrils were 16,000- and 23,000-dalton proteins. The 16,000-dalton component was shown by amino acid sequencing to be a fragment of the kappa (k)iv light chain, the first such case. These clinicochemical correlations suggest that isolated massive hepatomegaly may prove to be a hallmark of idiopathic amyloid light chain-related protein k amyloidosis.

Haemophilus influenzae is one of five bacterial species known to produce IgA proteases, enzymes that specifically cleave the human IgA1 heavy chain. Strains of H. influenzae produce three distinct types of IgA proteases that cleave different peptide bonds within the IgA1 hinge region. Type 1 protease cleaves the prolyl-seryl bond at position 231-232; type 2 protease cleaves the prolyl-threonyl bond at position 235-236, the same bond attacked by Neisseria gonorrhoeae and Neisseria meningitidis type 2 proteases. Type 3 protease yields a unique double Fd cleavage pattern; the exact peptide bonds cleaved have not been determined. The type of protease produced correlates with the serotype, but not with the biotype, of the isolate; serotypes A, B, D, and F produce primarily type 1 protease, whereas serotypes C and E produce only type 2 enzyme. Each nontypable strain yields one of the three protease types. These data further extend our knowledge of the extreme specificity of the IgA proteases and suggest that IgA protease type may be useful in the taxonomy and epidemiology of H. influenzae.

An antiserum prepared against a lambda-Bence Jones protein from a patient (SUT) who had multiple myeloma and amyloidosis had specificity for lambda-light chains of the chemically defined variable (V) region lambda-chain subgroup lambda VI. Sequence analyses of protein SUT and of five other lambda-light chains recognized immunologically as of the V lambda VI subgroup revealed that all six proteins had the N-terminal sequence characteristic for prototype lambda VI proteins. The isotypic nature of the V lambda VI subgroup was demonstrated immunochemically: lambda VI molecules were detected among light chains isolated from the IgG proteins of each of 12 normal individuals and lambda VI antigenic determinants were also detectable on the intact IgG proteins. The frequency of lambda VI molecules among lambda-type light chains is estimated to be approximately 5% based on the finding that 5 of 91 lambda Bence Jones proteins were of the V lambda VI subgroup. Proteins of the V lambda VI subgroup, in contrast to those of the other five chemically-classified lambda chain subgroup, appear to be preferentially associated with the amyloid process as evidenced by the fact that all six lambda VI proteins were obtained from patients with amyloidosis AL and, in addition, 5 of 42 lambda-type monoclonal immunoglobulins from patients with primary or myeloma-associated amyloidosis were classified by immunodiffusion analyses as having lambda VI-type light chains.

Human IgG subclasses 1, 2, and 4, as well as proteins of the IgG3 subclass that are allotype G3m (s+t+), bind avidly to staphylococcal protein A by means of their Fc portion. Proteins of the IgG3 subclass that are allotype G3m (s-t-) do not bind. The importance of a histidine residue at position 435 has been implicated from comparison of amino acid sequences of immunoglobulins that bind with those that do not bind to staphylococcal protein A, as well as from crystallographic data. Modification of histidines at a low concentration of diethylpyrocarbonate successfully and reversibly alters the binding of immunoglobulins to staphylococcal protein A with only minimal change in the antigenic properties. This method provides strong evidence for the critical importance of histidine in the binding of immunoglobulins to staphylococcal protein A.

Human gamma heavy chain disease (HCD) is characterized by the presence in serum of a short monoclonal Ig gamma chain unattached to light chains. Although most HCD proteins have internal deletions, in some the defect is NH2-terminal. The OMM gamma 3 HCD serum protein is of the latter type, having undergone an extensive NH2-terminal deletion with a sequence starting within the hinge. A cell line synthesizing the OMM protein has enabled us to study the biogenesis of the abnormal molecule. In vitro translation of isolated mRNA yields a protein containing a hydrophobic NH2-terminal leader sequence. In the intact cell, the precursor molecule is processed normally to yield a protein with an NH2-terminal sequence homologous to the beginning of the variable (V) region. The nucleotide sequence of cDNA prepared from the OMM mRNA encodes a 19-amino acid leader followed by the first 15 residues of the V region. An extensive internal deletion encompasses the remainder of the V and the entire CH1 domain. Immediately following the short V region, there is information in the cDNA for the entire normal hinge. The primary synthetic product is thus an internally deleted molecule that undergoes postsynthetic degradation to yield the NH2-terminally deleted serum protein. The structure of the OMM mRNA suggests that the protein abnormality results from a partial gene deletion rather than defective splicing.

Staphylococcal protein A binds molecules belonging to the IgG1, IgG2, and IgG4 subclasses. IgG3 proteins generally do not bind, except for those coded by the two gamma 3 alleles, which are G3m(u-): G3m(b0,b3,b5,s,v). G3m(u) is located in the CH2 domain. The difference between G3m(u-) and G3m(u+) IgG3 proteins correlates with the sequence at position 339 in the CH2 domain--Ala and Thr respectively. There is another structural difference in the CH3 domain which correlates with protein A binding and non-binding: all IgG proteins that bind protein A have His at position 435, whereas those that do not, have Arg at that position