COMPLEMENT SYSTEM 
John P. Atkinson, M.D., Investigator 
Complement is a recognition and effector im- 
mune system that evolved to protect the host from 
infectious organisms. The biology of the comple- 
ment proteins that become bound to antigens and 
the regulatory and receptor glycoproteins with 
which these attached fragments interact is the cen- 
tral focus of Dr. Atkinson's research. The metabo- 
lism of such immune complexes has an important 
relationship to many inflammatory diseases of 
humans. 
I. Sulfation of the Fourth Component of Comple- 
ment (C4). 
Sulfation of tyrosine residues is a recently recog- 
nized biosynthetic modification of many plasma 
proteins. Effects of this site-specific modification on 
protein function are not known, but the activity of 
several peptides, such as cholecystokinin, is greatly 
augmented by sulfation. C4, a component that at- 
taches to antigens during complement activation, is 
one of the few proteins in which sites and stoichi- 
ometry of tyrosine sulfation have been character- 
ized. Decreased hemolytic activity of nonsulfated 
C4 was demonstrated and resulted from impaired 
interaction with complement subcomponent Cls, 
the protease that physiologically activates C4. Puri- 
fied Cls was able to cleave nonsulfated C4, but 
~10-fold higher concentrations of Cls were re- 
quired for that cleavage than for equivalent cleav- 
age of sulfated C4. These results, with C4 as a para- 
digm, suggest that sulfation of tyrosine residues can 
have major effects on the activity of proteins partici- 
pating in protein-protein interactions. 
II. Regulators of Complement Activation (RCA) 
Gene Cluster. 
A recent advance in which Dr. Atkinson's labora- 
tory has played a major role is the identification of a 
multigene family of complement regulatory and 
receptor glycoproteins. In the past year Dr. Atkin- 
son and his colleagues identified two plasma regu- 
latory proteins [factor H (C3b binding) and C4- 
binding protein (C4b binding)], two receptors 
(CRl and CR2), and two widely distributed regula- 
tory membrane proteins [decay-accelerating factor 
(DAF) and membrane cofactor protein (MCP)]. 
These proteins are related functionally because 
they bind the opsonic fragments of complement. 
C3b and C4b; structurally they are related because 
they contain, beginning at their amino terminus, 
multiple copies of an ~60-amino acid cysteine-rich 
repeat; genetically they are related because their 
structural genes are closely linked on the long arm 
of chromosome 1. The laboratory discovered the 
MCP molecule, cloned and sequenced three of 
these proteins (CRl, DAF, and MCP), and most re- 
cently found that the order of five of these proteins 
on an 800 kb fragment is MCP-CRl-CR2-DAF-C4bp. 
A. C3b/C4b receptor or complement receptor type 
one (CRl). Translation of human CRl mRNA in a 
cell-free system and by Xenopus oocytes was accom- 
plished. These experiments suggested that a cotrans- 
lational modification of CRl structure occurred, 
probably involving a proteolytic cleavage event. 
The CRl protein exhibits an interesting struc- 
tural organization. The entire extracellular portion 
of the mature receptor is composed of a tandemly 
repeated amino acid motif or complement control 
protein (CCP) repeat of 59-72 residues in length, 
extending for 30 CCPs in the case of the most com- 
mon CRl polymorphic form. Four polymorphic 
variants of human CRl have been identified at the 
protein level. Their reduced forms exhibit molecu- 
lar weights of 220,000, 250,000, 190,000, and 
280,000. Classical genetic and molecular studies in- 
dicate that each variant is the product of a different 
allele. Tryptic peptide mapping, mRNA studies, and 
genomic analysis indicate that these forms could 
differ by multiples of an internal repeat. The com- 
mon polymorphic form (M^ 220,000) exhibits an in- 
ternal repetition that is seven CCPs in length and is 
reiterated four times from the 1st to the 28th CCP 
It is hypothesized that the polymorphic variants are 
the result of unequal crossing-over within such re- 
peated regions. 
In the course of studying CRl, several overlap- 
ping cosmid clones were isolated that together 
carry a CRl-like sequence that encompasses 40 kb 
and contains at least 10 potential exons. The CRl- 
like sequence, which may encode an unreported 
protein, exhibits 95% homology to CRl at the nu- 
cleotide level and 91% homology at the amino acid 
level and is likely to have arisen by duplication of at 
least a portion of an ancestral CRl gene. A compari- 
son between the CRl-like sequence and CRl indi- 
cates specific examples of identical nucleotide sub- 
stitutions at corresponding sites in the most highly 
Continued 
359 
