| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Veterans Affairs Medical Center and Departments of Medicine and Microbiology, University of Mississippi Medical Center, Jackson, MS 39216; and
Tougaloo College, Tougaloo, MS 39174
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
200,000) that can bind C3b,
C4b, and C1q (1, 2, 3, 4). The latter molecules function in
innate and adaptive immunity by attaching to the targets of immune
reactions, focusing further complement activation at these sites and
opsonizing the targets for binding by erythrocyte CR1
(4, 5, 6). Complexes that have bound to CR1 are transferred
to phagocytes as erythrocytes traverse the liver and spleen
(7, 8, 9, 10). Quantitative expression of CR1 on erythrocytes is regulated by a genetic element that is linked to the site of a HindIII RFLP of the CR1 gene (11). Two alleles that are associated with either high (H allele) or low (L allele) expression of CR1 differ in having genomic HindIII fragments of 7.4 and 6.9 kb, respectively. Individuals homozygous for the H allele can have more than 1000 CR1 molecules per erythrocyte, whereas those homozygous for the L allele generally have fewer than 200 CR1 per cell; heterozygous individuals have intermediate CR1 expression. Correlation between the HindIII RFLP and the quantity of CR1 on erythrocytes has been confirmed in multiple populations on several continents (12, 13, 14, 15). A significant correlation was not observed, however, among African Americans, despite a trend in the expected direction (15).
As erythrocytes age in the circulation CR1 is lost through proteolysis (16, 17, 18), perhaps during transfer of bound immune complexes to reticuloendothelial phagocytes (19). In initial studies of age-fractionated erythrocytes, groups of donors homozygous for either the H or L allele did not appear to differ in their rates of loss of erythrocyte CR1 (20). It was subsequently found that donor-dependent differences in CR1 expression were already evident in reticulocytes and that erythrocyte CR1 loss was actually faster in donors with higher CR1 expression (21). However, erythroid cells develop in intimate contact with macrophages that may modify their surface proteins (22, 23), and circulating erythrocytes with low CR1 numbers are less likely to bind immune complexes (24, 25, 26) and transfer them to phagocytes. Thus, rapid proteolysis of CR1 during erythroid development could result in its more gradual loss from circulating erythrocytes. When age-fractionated erythrocytes from donors heterozygous both for the H and L alleles and for different CR1 structural allotypes were analyzed in Western blots, it was found that the product of the L allele may in fact be lost more rapidly during erythrocyte aging (15). The H and L allelic products of heterozygous donors would presumably have similar exposure to immune complex binding and transfer reactions. Further, analyzing these allelic products simultaneously in Western blots has the advantage of controlling internally for variations in erythrocyte fractionation and protein analysis. The fact that the H and L allelic products were lost from erythrocytes of one donor at apparently equal rates (15) may indicate that some individuals lack a relevant protease or that linkage disequilibrium between the HindIII RFLP and the element that regulates CR1 expression is incomplete.
The H and L alleles are identical in the sequences of their
transmembrane and cytoplasmic domains (15). Thus, any
amino acid substitutions causing allelic differences in susceptibility
to proteolysis would have to be extracellular. The extracellular domain
of CR1 is comprised of a series of long homologous repeats (LHRs), each
containing seven short consensus repeats (SCRs) (Fig. 1
) (27, 28, 29). The most common
structural allele of CR1 has four LHRs, but alleles known or presumed
to have three, five, or six LHRs also exist (1, 2). Very
high homology among the LHRs complicates efforts to amplify selected
coding regions specifically, and highly homologous repeats also span
extended regions of CR1 genomic DNA (30, 31, 32). In this
study, an approach was developed that allows amplification and direct
sequencing of overlapping fragments entirely spanning the CR1 coding
sequence. The complete coding sequence of representative H and L
alleles was determined, polymorphisms distinguishing these alleles were
identified, and haplotypes defined by these polymorphisms were analyzed
in groups of Caucasian and African American donors.
|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Southern blots of HindIII-digested genomic DNA were probed with the pUC18 subclone pHL1.4 (kindly provided by Dr. Winnie Wong), having an insert that hybridizes specifically to the genomic fragments that define the HindIII RFLP of CR1.
Reverse transcription-polymerase chain reaction
RNA was isolated with a Rapid Total RNA Isolation Kit (5[prime
3', Boulder, CO) from PBMC of two Caucasian donors, one homozygous
for the H allele of CR1 and having a mean of 1050 CR1 per erythrocyte
and the other homozygous for the L allele and having 140 CR1 per
erythrocyte. Cellular CR1 was quantitated by the binding of
125I-YZ-1 anti-CR1 (33, 34).
Overlapping fragments spanning the coding sequence of CR1 were
amplified by RT-PCR (Fig. 1
, Table I).
Four primary products (Fig. 1, fragments 1–4) were produced directly
from cDNA under the conditions shown (Table I). Amplification of
fragments 1 and 3 was performed with a GeneAmp XL RNA PCR Kit
(Perkin-Elmer, Foster City, CA); cDNA synthesis was primed by the
reverse amplimer shown for each fragment in Table I. cDNA for the
production of fragment 2 was synthesized with the SuperScript
Preamplification System (Life Technologies, Gaithersburg, MD) primed by
the reverse amplimer for this fragment. Amplification of fragment 2 was
performed with a GeneAmp XL PCR Kit and a "hot start" technique, as
described by the manufacturer for the control template DNA. The
conditions described for fragment 2 (Table I) did not yield a single,
homogenous product but produced a major fragment of 3910 bp that was
isolated from minor contaminating fragments by excision and
electroelution from agarose gels. Fragment 4 was prepared as described
(15).
|
Highly homologous repeats occurring throughout fragment 2
precluded the identification of internal sequencing primers or
amplimers that hybridized at unique sites within this fragment. Thus,
templates for secondary amplification reactions (Fig. 1, fragments
5–9) were produced by digesting the gel-purified fragment 2 of each
donor with either PstI (fragments 5–7) or
BsrDI (fragments 8 and 9) (New England BioLabs, Beverly,
MA). Desired restriction fragments were gel purified, and selected
regions were amplified under the conditions shown (Table I).
Sequence analysis
Products amplified from cDNA were analyzed by automated sequencing, with internal primers derived from the published sequence of CR1 (GenBank accession number Y00816) (29). Primers were spaced so that each cDNA region was covered by multiple sequencing reactions. Genomic sequences were analyzed with a Sequenase 7-Deaza-dGTP DNA Sequencing Kit (United States Biochemical, Cleveland, OH) or a SequiTherm EXCEL II DNA Sequencing Kit (Epicentre Technologies, Madison, WI).
Analysis of polymorphic sites in genomic DNA
Exons containing polymorphic sites detected in CR1 cDNA (Table II) were amplified from genomic DNA of
Caucasian and African American donors under the conditions shown in
Table I. Amplimers were derived from published sequences of introns
flanking these exons (GenBank accession numbers L17390–L17430)
(32). Polymorphisms at CR1 nucleotide positions 1360,
2078, and 4870 were analyzed by direct sequencing. The polymorphisms at
positions 3093 and 3650 alter restriction sites for BstNI
(New England BioLabs) and RsaI (Boehringer Mannheim,
Indianapolis, IN), respectively; genotypes at these sites were
determined by restriction analysis of PCR fragments spanning the sites.
BstNI fragments of the exon 19 PCR product shown in Table I
have the following sizes (base pairs): G3093: 566, 230, 91;
T3093: 566, 321. Multiple sets of primers and conditions
for amplifying exon 19 also produced a comigrating fragment that
resembled the G3093 allotypic fragment with respect to its
cleavage by BstNI. This finding undoubtedly reflects the
occurrence of highly homologous genomic repeats within CR1 or a
CR1-related sequence (28, 31, 35). Heterozygosity at
nucleotide 3093 was thus distinguished from homozygosity for
T3093 by differences in band density in ethidium
bromide-stained gels. Independent interpretations of the stained gels
by two individuals were completely concordant. RsaI
fragments of the exon 22 PCR product (Table I) have the following sizes
(base pairs): A3650: 520, 162; G3650: 458, 162,
62. The nucleotide substitution at position 5507 alters a restriction
site for MnlI (New England BioLabs). This site was analyzed
either by direct sequencing of the exon 33 PCR product described in
Table I or by restriction analysis of a 305-bp fragment produced with
the same forward primer but with the reverse primer
5'-GAACAGAAAGTTCACAGCGAGG-3'. MnlI fragments of
the latter product have the following sizes (base pairs):
C5507: 111, 80, 37, 33, 33, 11; G5507: 111, 80,
70, 33, 11; these fragments were resolved in nondenaturing 16%
polyacrylamide gels. The HindIII RFLP in CR1 intron 27 was
analyzed by Southern blotting or by a modification of described methods
(36). HindIII fragments of the intron 27 PCR
product shown in Table I have the following sizes (base pairs): H
allele, 1600, 84; L allele, 1,150, 465, 84.
|
Donors were classified by the investigators as African American or Caucasian based on appearance and were not questioned regarding their ancestry. Prior analysis suggests up to 25% European genetic admixture among African Americans (37). Genomic DNA was obtained from Caucasian donors in the Jackson, MS, and Boston, MA, areas. All African American donors were from Jackson, MS.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Several of these
polymorphisms had been detected previously (15, 32, 35),
and the C5507G substitution has been shown to be specific to the L
allele (15). Both donors in the current study were
homozygous for "T" at position 5654, which was previously reported
as "C" (27); thus, this polymorphism is not specific
with respect to the H and L alleles. The nucleotide substitutions at
positions 207, 981, 1356, and 2367 would not result in amino acid
substitutions, and their possible allele specificity was not
assessed.
The nucleotide substitutions at positions 3093, 3650, and 5507 in CR1
cDNA alter restriction sites for BstNI, RsaI, and
MnlI, respectively. These sites and that of the
HindIII RFLP were analyzed in genomic DNA of 85 Caucasian
and 75 African American donors (Table III). Findings regarding the
HindIII RFLP and C5507G polymorphisms of some of these
donors have been reported previously (11, 15). In the
samples analyzed, the frequencies of the L allele as identified by the
HindIII RFLP were 0.2 and 0.23 for Caucasians and African
Americans, respectively, similar to previous findings (11, 15). However, because these samples included a mixture of
Caucasian donors from the Jackson and Boston areas and because data for
the HindIII RFLP were already known for some donors, these
should not be considered random samples. The G3093T and A3650G
substitutions, like the C5507G substitution (15), were
largely specific to the L allele identified by the HindIII
RFLP. Thus, two predominant haplotypes were observed in both Caucasians
and African Americans: an H haplotype containing G3093,
A3650, and C5507 (encoding Gln1022,
His1208, and Pro1827); and an L haplotype
containing T3093, G3650, and G5507
(encoding His1022, Arg1208, and
Arg1827).
|
has one
allele bearing either the prototypic H or L haplotype, the likely
features of six less common haplotypes can be deduced (Table III). At
least one minor haplotype was observed mainly among Caucasians, whereas
another was seen only in African Americans. Each of the minor
haplotypes contains regions of continuity that resemble portions of the
major H and L haplotypes, suggesting that the minor haplotypes were
produced by recombination between prototypic H and L alleles.
Sequences occurring at the other three polymorphic sites shown in Table II, A1360G, T2078C, and A4870G, were analyzed in products amplified
from genomic DNA of donors homozygous for the H or L allele, as
indicated by the HindIII RFLP. In this more limited
analysis, each of the three polymorphisms appeared to be in linkage
disequilibrium with the HindIII allelic markers. "A" at
position 1360 was observed only in association with the H allele, being
present on 6 of 18 H alleles of Caucasians and 2 of 10 H alleles of
African Americans; the remaining H alleles and all of 18 L alleles
analyzed (10 Caucasian, 8 African American) had a "G" at this site.
Among Caucasians, there was complete concordance of T2078
and A4870 with the H allele, and of C2078 and
G4870 with the L allele (Table IV). In African Americans, similar allele
specificity was observed at position 2078 in 17 of 18 cases. In
contrast, "G" was present at position 4870 not only on all of 8 L
alleles of African Americans but also on 8 of 10 H alleles.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Six of the nucleotide differences between the H and L alleles correspond to predicted amino acid substitutions (Tables II–IV). At least two of these may be biologically insignificant. The substitution of Thr445 to alanine would be associated with slightly increased hydrophobicity within a predicted ß sheet region of SCR 7 (38). This SCR is not required for ligand binding (39, 40) and may serve principally as a "spacer" in the extended CR1 molecule. Thus, although it is possible that this substitution could cause exposure of a novel proteolytic cleavage site, it is perhaps more likely to be functionally neutral. I1615V is a conservative substitution at a conserved hydrophobic residue (38); the corresponding amino acid in multiple other SCRs of CR1 is either isoleucine or valine.
Two substitutions, I684T and Q1022H, lie in SCRs that participate in ligand binding (29, 39, 40, 41, 42). Replacement of the tandem isoleucines 683 and 684 by two glutamic acids was reported to reduce ligand binding by a soluble LHR-B construct (40). Whether a single change of Ile684 to threonine, an uncharged polar amino acid, might also affect receptor function is unknown. The position corresponding to Q1022H is occupied by glutamine in all four LHRs of both human and chimpanzee CR1, except in the human L allele (27, 28, 29, 43). An aspartic acid 2 residues upstream of this glutamine (or a corresponding aspartic acid) is important in C3b binding by both human CR1 and the chimpanzee erythrocyte complement receptor (44). Thus, local effects of changing Gln1022 to histidine, which is weakly basic and less hydrophilic than glutamine, could perhaps alter ligand binding by CR1. However, any functional consequences of either the I684T or Q1022H substitutions may be mitigated by the presence of two C3b-binding sites within each CR1 molecule (29) and by the clustered distribution of CR1 on erythrocytes (24, 25, 26, 45).
Each of the other two allele-specific amino acid substitutions, H1208R and P1827R, could have important effects on the structure and stability of CR1. Each would introduce a potential cleavage site for tryptic proteases. H1208R occurs in a region that, by analogy to SCRs 5 and 16 of factor H (46, 47), may be highly solvent exposed, and substitution to arginine would cause increased local hydrophilicity. Similarly, analysis of the SCRs of multiple proteins suggests that P1827R may lie within a surface-exposed turn (38, 48). A change to arginine, having a strongly basic side chain, would not only make this region more hydrophilic but might also be accompanied by a loss of spatial constraints otherwise imposed by proline. Arg1208 and Arg1827 are both present in the prototypic L allele of African Americans, in whom the HindIII RFLP does not correlate tightly with the quantity of CR1 on erythrocytes (15); thus, if either site is the target of a tryptic protease(s) that affects CR1 expression, then deficiency of this protease must be common in African Americans.
These surface-exposed arginine substitutions could perhaps produce novel antigenic epitopes, and thus could potentially contribute to CR1-specific blood group Ags such as the Knops, McCoy, Swain-Langley, and York Ags (49, 50). Antisera defining these blood groups recognize epitopes in both the A (four LHR) and B (five LHR) structural allotypes of CR1, and persons homozygous for each of these allotypes have been identified among producers of the antisera (49, 50, 51). The L allele of CR1 appears to be associated exclusively with the A structural allotype (20, 30). Thus, a person homozygous for the B allotype would also be homozygous for the H allele and would be unlikely to make antiserum against the H allelic product. The fact that these blood group antisera do recognize epitopes in the B (and hence the H) allotype makes it unlikely that they are specific for H/L-related amino acid substitutions.
The prototypic H and L haplotypes of CR1 (Table III) had
apparently become established before the divergence of the African and
European populations, which probably occurred 100,000–200,000 years
ago (52, 53, 54, 55). The minor haplotypes shown in Tables III and
IV show further evolution of these alleles in both populations,
involving recombination events and perhaps point mutations. All of the
minor haplotypes in Table III and the single L allele bearing T2078 in
Table IV could be the result of crossover events. On the other hand,
A1360 (text, above) and A4870 (Table IV), which
were observed largely among Caucasians, probably resulted
from point mutations of the H allele. Assuming up to 25% European
genetic admixture among African Americans (37), the latter
mutations may represent relatively recent changes that occurred in
European founders. Comparison of polymorphisms in the H and L alleles
to the corresponding sites in the sequence of 220-kDa chimpanzee CR1
(underlined nucleotides in Table II) shows identity between the H
allele and chimpanzee CR1 at the two amino acid substitutions in ligand
binding regions (I684T and Q1022H) and the two substitutions to
arginine (H1208R and P1827R). Thus, evolution of the L allele from an
ancestral CR1 sequence probably involved point mutations at each of
these sites.
Low expression of erythrocyte CR1 is associated with impaired clearance of immune complexes and with deposition of the complexes outside the reticuloendothelial system (9, 10, 56, 57). Although patients with diseases like systemic lupus erythematosus do not appear to have altered frequencies of the H and L alleles (12, 13, 14, 58, 59, 60), low CR1 expression determined by the L allele could result in increased tissue damage in a large number of inflammatory and infectious conditions. On the other hand, low CR1 expression may provide a selective advantage in other settings. In falciparum malaria, a parasite-encoded protein of infected erythrocytes, PfEMP1, has recently been shown to bind to CR1 of adjacent erythrocytes, producing an agglutination reaction that may cause increased vascular endothelial damage and more severe disease (61). Erythrocytes expressing low numbers of CR1 or having the Sl(a-) CR1 polymorphism, which is common in Africans, were found to have reduced binding to PfEMP1 in vitro (61). These considerations make it unlikely that the survival effects of the H and L alleles are entirely neutral. Thus observed frequencies of the L allele in a fairly limited range, from 0.19 to 0.27, in populations as diverse as those of New Delhi and Oulu, Finland (11, 12, 13, 14, 15, 62), may suggest that selective pressures in addition to falciparum malaria have contributed to a balanced polymorphism of the H and L alleles.
Our study has established the coding sequence of the H and L alleles of CR1, identified allele-specific amino acid substitutions that may be functionally significant, and provided information regarding the evolution of these alleles. The methods described may be useful in analyzing additional CR1 polymorphisms, such as those related to Sl(a-) and other blood group Ags. If the H and L alleles represent a balanced polymorphism, then identifying the selective pressures responsible for this balance may yield insights into important human diseases.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. James Wilson, VA Medical Center (151), 1500 East Woodrow Wilson Boulevard, Jackson, MS 39216.
3 Abbreviations used in this paper: CR1, complement receptor type 1; H allele, high allele; L allele, low allele; LHRs, long homologous repeats; SCRs, short consensus repeats.
Received for publication May 12, 1999. Accepted for publication August 18, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. FITNESS, S. FLOYD, D. K. WARNDORFF, L. SICHALI, L. MWAUNGULU, A. C. CRAMPIN, P. E. M. FINE, and A. V. S. HILL LARGE-SCALE CANDIDATE GENE STUDY OF LEPROSY SUSCEPTIBILITY IN THE KARONGA DISTRICT OF NORTHERN MALAWI Am J Trop Med Hyg, September 1, 2004; 71(3): 330 - 340. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. FITNESS, S. FLOYD, D. K. WARNDORFF, L. SICHALI, S. MALEMA, A. C. CRAMPIN, P. E. M. FINE, and A. V. S. HILL LARGE-SCALE CANDIDATE GENE STUDY OF TUBERCULOSIS SUSCEPTIBILITY IN THE KARONGA DISTRICT OF NORTHERN MALAWI Am J Trop Med Hyg, September 1, 2004; 71(3): 341 - 349. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Tamano, H. Ohi, S. Sudo, and Y. Tomino Quantitative polymorphism of complement receptor type 1 (CR1) in patients undergoing haemodialysis Nephrol. Dial. Transplant., June 1, 2004; 19(6): 1467 - 1473. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. A. Cockburn, M. J. Mackinnon, A. O'Donnell, S. J. Allen, J. M. Moulds, M. Baisor, M. Bockarie, J. C. Reeder, and J. A. Rowe From The Cover: A human complement receptor 1 polymorphism that reduces Plasmodium falciparum rosetting confers protection against severe malaria PNAS, January 6, 2004; 101(1): 272 - 277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Zorzetto, I. Ferrarotti, R. Trisolini, L. L. Agli, R. Scabini, M. Novo, A. De Silvestri, M. Patelli, M. Martinetti, M. Cuccia, et al. Complement Receptor 1 Gene Polymorphisms Are Associated with Idiopathic Pulmonary Fibrosis Am. J. Respir. Crit. Care Med., August 1, 2003; 168(3): 330 - 334. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. E. Sullivan, A. F. Jawad, L. M. Piliero, N. Kim, X. Luan, D. Goldman, and M. Petri Analysis of polymorphisms affecting immune complex handling in systemic lupus erythematosus Rheumatology, March 1, 2003; 42(3): 446 - 452. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. C. Iannuzzi, M. Maliarik, and B. A. Rybicki Nomination of a Candidate Susceptibility Gene in Sarcoidosis . The Complement Receptor 1 Gene Am. J. Respir. Cell Mol. Biol., July 1, 2002; 27(1): 3 - 7. [Full Text] [PDF]
|
|
|
|
|
|
M. Zorzetto, C. Bombieri, I. Ferrarotti, S. Medaglia, C. Agostini, C. Tinelli, G. Malerba, N. Carrabino, A. Beretta, L. Casali, et al. Complement Receptor 1 Gene Polymorphisms in Sarcoidosis Am. J. Respir. Cell Mol. Biol., July 1, 2002; 27(1): 17 - 23. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Moulds, P. A. Zimmerman, O. K. Doumbo, L. Kassambara, I. Sagara, D. A. Diallo, J. P. Atkinson, M. Krych-Goldberg, R. E. Hauhart, D. E. Hourcade, et al. Molecular identification of Knops blood group polymorphisms found in long homologous region D of complement receptor 1 Blood, May 1, 2001; 97(9): 2879 - 2885. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
