Blackwell Science, LtdOxford, UKTRFTransfusion0041-11322004 American Association of Blood BanksNovember 20044411••••Review ArticleRh BLOOD GROUP SYSTEM REVIEWWESTHOFF

REVIEW

The Rh blood group system in review: A new face for

the next decade

Connie M. Westhoff

R
h is the most well-recognized blood group to D. This suggested the rationale for the development of
system after ABO, probably because of the Rh immune globulin (RhIG).4 Although immunoglobulin
dramatic presentation of a fetus suffering M (IgM) antibodies did not provide protection, immuno-
hemolytic disease of the newborn (HDN) fol- globulin G (IgG) anti-D was effective. By the early 1960s, a
lowing maternal alloimmunization to the D antigen. Even mere 20 years after the discovery of Rh incompatibility, an
individuals not associated with medicine have heard of effective treatment was available.
the “Rh factor” and are aware that it has some importance Despite their clinical importance, the extremely
in pregnancy. The earliest recorded description of the syn- hydrophobic nature of the Rh proteins made biochemical
drome dates to the 1600s from a French midwife who studies difficult, and the proteins were not successfully
attended the delivery of a set of twins, one of which was isolated until the late 1980s.5 This led to the cloning of the
hydropic and the other was jaundiced and died of kernict- genes in the 1990s6 and to major advances in our under-
erus.1 The agent responsible for the wide range of fetal standing of the Rh system. The molecular bases of most
symptoms, from mild jaundice to fetal demise, remained Rh antigens have been determined, and the RH gene
obscure until 1941. Levine and colleagues2 observed that structure explains why this system is so polymorphic. Spe-
the delivery of a stillborn fetus and the adverse reaction in cifically, the conventional Rh antigens are encoded by two
the mother to a blood transfusion from the father were genes, RHD and RHCE, but numerous gene conversion
related and were the result of an immune reaction to a events between them create hybrid genes. The resulting
paternal antigen. novel hybrid proteins containing regions of RhD joined to
Serologists’ relationship with the offending blood RhCE, or the converse, generate the myriad of different Rh
group system began when it was confused with a Rhesus antigens.
monkey red blood cell (RBC) protein, now termed LW, and The goal of this review is to highlight the insights
much argument and debate ensued over who should gained since the cloning of the genes, describe applica-
receive credit for its discovery.3 Becoming aware of the tions for RH molecular testing to clinical practice, intro-
antigen, however, was only the beginning of the story. This duce other members of the Rh family of proteins that are
blood group system would become notorious for its com- present in other tissues, and focus on the next piece of the
plexity, with numerous antigens and multiple nomencla- Rh puzzle, that is, efforts to determine the structure and
tures defining it. function of the Rh family of proteins.
Several seminal events characterized the history of
the Rh system. One of the most important was the obser-
TERMINOLOGY
vation that ABO mismatch between a mother and the
fetus had a partial protective effect against immunization Now that the genes and the proteins have been elucidated,
current Rh terminology attempts to distinguish these
from the antigens, which are referred to by the letter des-
ABBREVIATIONS: RhIG = Rh immune globulin; SNP = single-
ignations, D, C, c, E, e, etc. Capital letters and italics are
nucleotide polymorphism.
used when referring to the RH genes, which include RHD,
From the American Red Cross and the Department of Pathology RHCE, and RHAG, as well as the nonerythroid RHBG and
and Laboratory Medicine, University of Pennsylvania, RHCG. The different alleles of the RHCE gene are desig-
Philadelphia, Pennsylvania. nated RHce, RHCe, and RHcE, according to which antigens
Address reprint requests to: Connie M. Westhoff, SBB, PhD, they encode. The proteins are indicated as RhD, RhCE (or
Scientific Director, American Red Cross, 700 Spring Garden, according to the specific antigens they carry Rhce, RhCe,
Philadelphia, PA 19130; e-mail: westhoffc@usa.redcross.org or or RhcE), and RhAG, RhBG, and RhCG. Rh haplotypes are
westhoff@mail.med.upenn.edu. designated Dce, DCe, DcE, etc., or ce, Ce, cE when refer-
TRANSFUSION 2004;44:1663-1673. ring to a specific CE haplotype.

Volume 44, November 2004 TRANSFUSION 1663

WESTHOFF
RBC RhCE, RhD, and RhAG
Two genes (RHD, RHCE) in close proximity on chromo-
some 1 encode the Rh proteins, RhD and RhCE; one car-
ries the D antigen, and the other carries CE antigens in
various combinations (ce, Ce, cE, or CE).6-9 The genes are
97 percent identical, each has 10 exons and encode pro-
teins that differ by 32 to 35 amino acids. This is in contrast
to most blood group antigens that are encoded by single
genes with alleles that differ by only one or a few amino
acids. The large number of amino acid differences
explains why exposure to RhD can result in a potent
immune response in an RhD– individual.
RBCs express yet another Rh protein termed RhAG,
for Rh-associated glycoprotein, which carries one N-gly-
can chain.10 RhAG shares 38 percent identity with RhCE
and RhD, has the same predicted topology in the mem-
brane, and is encoded by a single gene on chromosome 6.
RhAG is not polymorphic, so it does not carry blood group
antigens, but it is important for targeting RhCE and RhD
to the membrane. Two molecules of RhAG associate with
two molecules of RhCE and/or RhD.11 RHAG mutations
are responsible for loss of expression of Rh antigens
(Rhnull), because without RhAG, neither RhCE nor RhD
reach the RBC membrane.12,13 RhAG is expressed early
during erythroid differentiation, being detectable on
CD34+ progenitors, whereas RhCE appears later, followed
by RhD.14 The timing of expression may reflect the assem-
bly of these proteins in the RBC membrane.
Variations of RhCE
RHCE encodes both C/c and E/e antigens on a single mol-
ecule.9 RHce probably represents the oldest human locus
and encodes the c and e antigens. RHCe arose from RHce
via gene conversion with exon 2 from RHD (Fig. 1A). This
resulted in three new amino acids in the protein, but only
the amino acid change at position 103 is predicted to be
extracellular (Fig. 2). The fact that the amino acids
encoded by exon 2 of RHCe are identical to those encoded
by exon 2 of RHD explains the expression of G antigen on
both RhD and RhC proteins.
The antigens CW and CX, which were thought to be
antithetical to C, are the result of single amino acid
changes on the first extracellular loop of RhCE (Fig. 2).15
These changes cause the weak expression of C antigen on
RBCs that express CW or CX.
The E and e antigens differ by one amino acid,
Pro226Ala, located on the fourth extracellular loop of the
protein (Fig. 2). A single point mutation acquired in exon
5 of RHce resulted in RHcE (Fig. 1A). The requirement for
e antigen expression, however, is more complex than the
226Ala polymorphism, and weak or altered expression is
frequently encountered in Black persons and people with
mixed ancestry.16 Variant e expression occurs on the RBCs
of the more than 30 percent of Black persons who are V+
1664 TRANSFUSION Volume 44, November 2004
and VS+. Both antigens are the result of a single Leu245Val
substitution located in the predicted eighth transmem-
brane segment of Rhce (Fig. 2),17 which causes a local
conformation change affecting expression of e. This hap-
lotype encoding 245Val and V and VS antigens is referred
to as ceS (Fig. 1B). The subsequent loss of V expression (the
V–VS+ phenotype) results from a Gly336Cys change on
this background.18 Loss of VS expression (the V+VS–
phenotype) is associated with additional amino acid
changes and is characteristic of the ceAR haplotype
(Fig. 1B).19 Altered e expression is also associated with a
Trp16Cys amino acid polymorphism in exon 1 of the RHce
gene.20
Antibodies. Individuals who are homozygous for
RHce genes that encode variant e antigens type as e+, but
they often make alloantibodies with e-like specificities.
The antibodies, designated anti-hrS, -hrB, -RH18, and
-RH34, are difficult to identify serologically and, impor-
tantly, they are clinically significant and have caused
transfusion fatalities.21 The prevalence of e variants in
Black persons, together with the incidence of sickle cell
disease requiring transfusion support often provided by
Caucasian donors with conventional RHce, make the
occurrence of alloanti-e in the e variants not uncommon.
Some of the RH genetic backgrounds of individuals who
make these antibodies have now been defined and
include the RHCE haplotypes designated ceS, ceAR, ceMO,
ceEK, and ceBI.21 All encode the Trp16Cys difference in
exon 1 and have additional changes, primarily localized to
exon 5. Homozygous ceS RBCs lack the e-related antigen
hrB, and RBCs from ceAR, ceMO, ceEK, and ceBI lack the
hrS antigen (Fig. 1B and summarized in Reid and Lomas-
Francis22). Importantly, because of the multiple molecular
backgrounds responsible for the hrS– and hrB– pheno-
types, some of which are not yet elucidated, the antibodies
they produce are not all serologically compatible. This
explains why it is difficult to find compatible blood for
patients with these antibodies, and often only rare deleted
D-- RBCs appear compatible. As an additional complica-
tion, these variant RHce can often be inherited with an
altered RHD, for example, DAR or DIIIA (Fig. 1B), so they
can also make anti-D. With the implementation of molec-
ular testing, patients with altered Rhce/RhD can now be
identified. The next challenge is to institute molecular
screening of minority donors to develop a registry of units
that are genotyped for these variants. This is critical to
meet the transfusion needs of these alloimmunized and
often critically ill patients, especially as rare D-- units are
in very limited supply and are not always compatible.
Other modifications of RHCE include the hybrids
DHAR, RN, and rG and several E/e variants23-27 (Fig. 1B).
Some of these can cause discrepancies in serologic typing
between monoclonal and polyclonal reagents. For exam-
ple, DHAR is found in people of German ancestry. The RBCs
are negative with polyclonal reagents and with some mon-
 Rh BLOOD GROUP SYSTEM REVIEW

(hrB-)

(hrS-)
Go(a+)

(hrS-)

Fig. 1. (A) Origin of the common RHCE genes. The 10 RHD exons are shown as black boxes, and the RHCE are in gray. RHCe and RHcE
have arisen from RHce by gene conversion or point mutation, respectively. See text for details. (B) Examples of some RHD and RHCE
gene rearrangements. Shown are those that are more frequent in Black and Caucasian persons. (C) Examples of some rearrangements
responsible for elevated expression of D antigen.

oclonal anti-D including the weak D test, but type as D+
with monoclonal anti-D containing IgM clone MS201
(Immucor, Norcross, GA) (Table 1). These individuals do
not carry RHD, but, instead, exon 5 of RHce has been
replaced by exon 5 from RHD (Fig. 1B). Not surprisingly,
they can be stimulated by conventional D antigen and
should be treated as D– for transfusion and RhIG prophy-
laxis. RN RBCs carry a hybrid RHCe in which exon 4 is
replaced with RHD, and they may type as e– or weak with
polyclonal reagents, but are indistinguishable from “nor-
mal” e+ RBCs with monoclonal anti-e.
Variations of RhD
D. D– is much more prevalent in Caucasian persons
of European descent (15%-17%), less likely in individuals
of African background (3%-5%), and rare in Asian popula-
tions (<0.1%).1 The distribution of D– individuals world-
wide attracted the early attention of population biologists,
who attempted to make predictions of human origins and
migrations based on Rh phenotype.28 How the D– pheno-
type reached the high proportions seen in European per-
sons remains a mystery. Genetic selection models dictate
that there must have been some advantage to individuals
who were heterozygous (balancing selection). What that
advantage might have been is obscure, as the only obvious
selection pressure is against the heterozygote, reflected in
morbidity and mortality of D/d infants born to D– alloim-
munized women.
It is now clear that the D– phenotype has arisen
numerous times on different genetic backgrounds. In

Volume 44, November 2004 TRANSFUSION 1665

WESTHOFF

TABLE 1. Reactivity of some Partial D and variant CE with current FDA licensed anti-D72 (and our observations)
Partial D Reactivity Method
DII, DIIIG-, DIIIa** DIIIb, DIVa, DIVb, DVa, DVc Positive IS Tube*
Positive GEL-D‡
DVI#, DVb, FPTT Negative IS Tube
Negative GEL-D
Positive IAT†
DBT Positive IS Tube (Gamma, Immucor)
Positive GEL-D
Negative IS Tube (Ortho)
Positive IAT (Ortho)
DHAR Negative IS Tube and IAT (Ortho)
Positive IS Tube Immucor (Series 4, MS201)
Positive GEL-D
Variant CE
Crawford (Rh43)§ Negative IS Tube and IAT (Ortho, Immucor)
Negative GEL-D
Positive IS Tube (Gamma, Dominion, RUM1)
Negative IAT (Gamma, Dominion)
* IS = immediate spin.
‡ GEL-D, Micro Typing Systems, Pompano Beach, FL; Ortho reagent—IgM monoclonal blended with polyclonal IgG; Gamma and Immucor
reagents—IgM monoclonal blend with monoclonal IgG.
† IAT = indirect antiglobulin test.
§ predominantly found in Blacks.
** most frequent partial D in Blacks.
# most frequent partial D in Caucasians.

Caucasian persons of European descent, it is primarily the

Fig. 2. Predicted structure of the RhCE, RhD, and RhAG proteins
in the RBC membrane. The amino acid differences are shown
as circles, and those that are extracellular on RhD are shown in
black dots. C/c differ by several amino acids but only Ser103Pro
on loop 2 is predicted to be extracellular. E/e differs by
Pro226Ala on loop 4, and CW and CX antigens are located on the
first extracellular loop of RhCE. The antigens V and VS result
from a Leu245Val change located in the predicted eighth trans-
membrane region of Rhce. RhAG does not carry Rh antigens.
result of deletion of the entire RHD gene.29 This deletion
occurred on a Dce (R0) haplotype; that is, the RHCE allele
carried with the deletion is ce. In Caucasians rare excep-
tions exist and include RHD that is not expressed because
of a premature stop codon, nucleotide insertions, point
mutations, or RHD/CE hybrids.30-33 Most of the exceptions
arose on the DCe (R1) or DcE (R2) background and result
in the less common Ce (r¢) or cE (r≤) haplotypes. D– phe-
notypes in African and Asian persons are often caused by
inactive or silent RHD, rather than a gene deletion. Asian
D– phenotypes result from mutations in RHD most often
associated with Ce, indicating that they originated on a
DCe (R1) haplotype. Many Asian persons, however, who
type as D– are Del and actually have low-level expression
of D, which is detectable only by adsorption and elution.
Del results from deletion of exon 9 or the missense muta-
tion 1227G>A.34,35 In D– African Black persons, 66 percent
have RHD genes that contain a 37-bp insertion, which
results in a premature stop codon, and 15 percent carry a
hybrid RHD-CE-D linked to ceS, termed (C) ceS, character-
ized by expression of weak C and no D antigen.36
Weak D. An estimated 0.2 to 1 percent of Caucasian
persons and a greater number of African-American per-
sons have reduced expression of the D antigen16 but the
number of samples classified as weak D depends on the
characteristics of the typing reagent. The molecular basis
of weak D expression is primarily the result of point muta-
tions that cause amino acid changes predicted to be intra-
cellular or in the transmembrane regions of RhD and not

1666 TRANSFUSION Volume 44, November 2004


Fig. 3. (A) Weak D results from single amino acid changes
located in the intracellular regions shown as open circles. Weak
D Type 1 (V270G) predominates in European persons. (B) Some
partial D also result from single amino acid changes, but these
are located on the extracellular loops and are shown as filled
circles. Also shown is #the location of weak D type 15, shown to
make anti-D, so more accurately classified as partial D.
on the outer surface of the RBC (Fig. 3A).37 These
mutations affect the efficiency of insertion, and therefore
the quantity of RhD protein in the membrane. This is
reflected in the reduced number of antigen sites on these
RBCs37 and explains why the indirect antiglobulin test
(IAT) may be required for detection. Many different
mutations, the most common being a Val270Gly, cause
weak D expression, and the mutations have been classi-
fied as Types 1 to 24. These changes probably do not alter
D epitopes, because the majority of individuals with a
weak D phenotype can safely receive D+ blood and do not
make anti-D. Two weak D, type 4.2 (also called DAR) and
type 15, however, have been reported to make anti-D.21,38
Partial D. RBCs with partial D antigens have histori-
cally been classified as such based on the fact that the
RBCs type as D+, but the individuals make anti-D when
alloimmunized. Many react in direct tests with anti-D
reagents, but some require the IAT, and the reactivity of
the same cells may differ, depending on the reagent man-
ufacturer (Table 1). The molecular bases of some partial D,
like weak D, are the result of point mutations in RHD that
cause single-amino-acid changes, but, in contrast to weak
D, these changes are located on the extracellular loop
regions of RhD (Fig. 3B). Some examples include DNB
(Gly355Ser), a frequent partial D in Europeans.40 Others
are shown in Fig. 3B (see Reid and Lomas-Francis22 for a
comprehensive list).
Rh BLOOD GROUP SYSTEM REVIEW
Most partial D RBCs result from the inheritance of a
hybrid gene with portions of RHD replaced by the corre-
sponding portions of RHCE (Fig. 1B) (reviewed in Huang13
and Avent and Reid41). Some of the replacements involve
single or multiple exons, whereas others involve short
regions encompassing several codons. Different rear-
rangements can also cause the same partial D category.
For example, DVI, which is the most common form of par-
tial D in Caucasian persons,42 can result from the replace-
ment of two, three, or four RHD exons with the
corresponding regions of RHCE (Fig. 1B).43 These rear-
rangements are designated as DVI Types 1, 2, and 3 to dis-
tinguish them. Additionally, the novel sequences of the
hybrid protein resulting from regions of RhD joined to
RhCE can generate new antigens, and DVI RBCs carry the
BARC antigen. Because individuals with partial D can
make anti-D, they should receive D– donor blood. In prac-
tice, many will type as D+ by direct tests and will only be
recognized after they have made anti-D.
Elevated D. Several phenotypes, including D–, Dc–,
and DCw–, have an enhanced expression of D antigen and
no, weak, or variant CE antigens, respectively.16 They are
caused by replacement of portions of RHCE by RHD, the
converse of the partial D rearrangements described above
(Fig. 1C and other examples in Reid and Lomas-Francis22).
The additional RHD sequences in RHCE, along with a nor-
mal RHD, explain the enhanced D and account for the
reduced or missing CE antigens. Immunized people with
these CE-depleted phenotypes make anti-Rh17.
Testing for weak D in transfusion practice. Weak D
red cells have stimulated the production of anti-D in D-
individuals.16,39 Testing for weak D in the donor setting is
straightforward; that is, donor center typing procedures
must detect and label all weak D RBCs and components
as D+. In contrast, transfusion service policies vary, and
the test for weak D is optional.44 Historically, the weak D
test was performed on transfusion recipients and on pre-
natal samples to conserve supplies of D– components and
to prevent unnecessary RhIG administration, respectively.
Several developments in the past decade, however, have
appropriately caused many to reevaluate their policy of
weak D testing in the transfusion and prenatal setting. For
example, the current generation of monoclonal anti-D
reagents contains an IgM component that causes direct
agglutination of many RBCs with reduced D antigen. Sam-
ples that would previously have been classified as weak D
type as D+ and patients receive Rh+ donor units, eliminat-
ing concerns regarding the unnecessary use of Rh– sup-
plies. In addition, some partial D and, importantly,
category DVI, the most common partial D in Caucasian
persons,42 react only in the IAT and not by direct testing
(Table 1). Because individuals with partial D RBCs make
anti-D, elimination of the IAT classifies them as Rh– for
transfusion and for RhIG administration and avoids the
potential for the production of anti-D. Furthermore, the
Volume 44, November 2004 TRANSFUSION 1667
WESTHOFF
distinctions between weak D and partial D have blurred
with the finding that some people with weak D RBCs have
produced anti-D21,38, suggesting that they would also be
better served by elimination of the IAT. These consider-
ations, along with the cost savings realized by elimination
of this extended testing step and the prospect of avoiding
the potentially serious error of misinterpreting the D type
when RBCs are coated with immunoglobulin (positive
direct antiglobulin test [DAT]), often result in a decision
to eliminate the IAT phase of D testing for transfusion
recipients and prenatal screening.
Ideally, tests to determine an individual’s D status
would distinguish those with RBCs that lack, or have
altered, D epitopes and are at risk of immunization to
conventional D from those who carry mutations that
reduce expression levels of D and are not at risk of pro-
ducing anti-D. Unfortunately, serologic reagents cannot
discriminate between these RBCs, and reactivity patterns
often reflect the characteristics of the reagent rather than
the characteristics of RhD expressed on RBCs. These lim-
itations suggest that molecular testing to determine D sta-
tus may be commonplace in the future.
Applications of RH molecular genotyping for
transfusion medicine practice
Molecular assays for blood group genes currently rely on
polymerase chain reaction (PCR) amplification to gener-
ate sufficient material for analysis. One of the most com-
mon approaches uses oligonucleotide primers that are
sequence-specific, with amplification indicating the pres-
ence of an allele and lack of amplification denoting the
absence of the allele. Another common method is PCR-
restriction fragment length polymorphism (RFLP) with
primers designed to amplify the region containing an
allele single-nucleotide polymorphism (SNP) followed by
digestion with a restriction enzyme to detect the SNP. The
size of the products, as determined by electrophoresis,
indicates the presence or absence of the SNP. Automated
platforms with fluorescein-based detection and micro-
chip technologies are being actively explored.
RH molecular testing can aid resolution of D typing
discrepancies. These often are the result of differences in
manufacturers’ reagents, but in the donor setting they can
be FDA reportable. Molecular genotyping can be used to
determine the phenotype of RBCs in patients who have
been recently transfused or whose RBCs are coated with
IgG. RH testing by molecular methods has applications in
the prenatal setting to determine paternal RHD zygosity,
to predict fetal D status to prevent invasive procedures,
and to determine D– status in the presence of a strongly
positive DAT.
Molecular genotyping in the prenatal setting
RHD zygosity determination. To determine paternal
RHD zygosity, assays are designed to probe the region,
1668 TRANSFUSION Volume 44, November 2004
termed Rhesus boxes, generated by deletion of RHD.45 A
PCR-RFLP assay45 can be discordant when testing individ-
uals from mixed ethnic groups46 and a PCR assay that
directly detects the hybrid region is often included.47
These tests, combined with a direct screen for the 37-bp
insert RHD pseudogene and for the RHD-CE-D hybrid
that causes D– in individuals of mixed ancestry are reliable
for determination of RHD zygosity.
Prediction of fetal risk for HDN. DNA analysis to
determine the D status of the fetus is important to deter-
mine when an infant is D– and is not at risk for HDN,
thereby preventing invasive and costly monitoring of the
mother during pregnancy. These assays rely on detecting
the presence or absence of portions of RHD. Fetal DNA
can be obtained by amniocentesis or villus sampling, but
fetal-derived DNA extracted from maternal plasma is
attractive as a noninvasive sample source.48 For the pre-
diction of HDN, samples from the mother, and the father
if known, are analyzed serologically and molecularly to
identify RHD– genotypes in the parents. This is important
for accurate interpretation in situations where families
carry inactive, rather than deleted, RHD.
Determination of D type when DAT-positive. Infants
born to D– mothers sometimes have a positive DAT, often
due to ABO mismatch, which can result in an invalid test
for weak D. If the infant cannot be shown to be weak
D–, the mother is a candidate for RhIG.44 Because RhIG is
not completely without risk (HCV and unknown variant
Creutzfeldt-Jakob disease), is costly, and is derived from
human source material that is in short supply, it is prudent
policy to avoid unnecessary use. Molecular testing to
assess the D– status of the infant can prevent needless
administration of RhIG.
In conclusion, RH molecular testing is an important
tool that, when combined with serologic methods, can
resolve discrepancies and ambiguous typing results, aid
the management of HDN, and improve transfusion safety
and outcomes for patients.
Function of the Rh/RhAG proteins
Clues to a possible function for the Rh/RhAG proteins ini-
tially came from structural predictions which suggested
they were transport proteins. Database searches for pro-
teins with similarities to RBC erythrocyte Rh/RhAG, how-
ever, were not successful until the genomic sequencing
of Caenorhabditis elegans revealed two Rh-like proteins.
These proteins were similar to ammonium transporters
from bacteria and yeast, thus linking the human proteins
to a family of transporters with ammonia or ammonium
as the substrate. Rh/RhAG protein analysis is a good
example of the power of comparative genomics, in which
the sequencing of the genomes of other organisms, along
with computer analysis, allows a testable hypothesis for
protein function.

Difficulty in expressing the Rh/RhAG proteins in
nonerythroid cells has hampered functional studies.
Xenopus oocytes are used for expression because of their
well-documented ability to reveal the function of ion
channels and transporters. Single-cell oocytes are micro-
injected with cRNA, and expression can be monitored
by Western blot. It was found that RhAG was readily
expressed in oocyte membranes and mediated transport
of the ammonium analog, [14C]methylammonium, a
radioactive substrate often used to study ammonium
transport.49 Transport of the analog was readily com-
peted by ammonium, demonstrating specificity. Func-
tional characterization revealed that uptake was
independent of the membrane potential and the Na+
gradient, but was dramatically stimulated by raising
extracellular pH or lowering intracellular pH. This sug-
gested that uptake was coupled to an outwardly directed
H+ gradient and led us to hypothesize that RhAG might
function by NH+/H+ countertransport.49
Additional evidence for RhAG-mediated transport of
ammonium comes from complementation experiments
in Saccharomyces cerevisiae yeast cells that are deficient in
ammonium transport. Studies reported by Marini and
coworkers50 and data from our laboratory51 showed that
expression of RhAG could rescue yeast cells lacking the
ammonium transport proteins (MEPS) when the cells
were grown on low ammonium. This is important because
the ability of a mammalian protein to functionally substi-
tute for the yeast protein is one of the most powerful indi-
cations that the proteins have a related function. Several
hundred mammalian proteins have now been shown to
function in yeast and to genetically complement mutants
in S. cerevisiae.
Complementation in yeast only occurs when the pH
value of the medium was raised to 6.1 to 6.5,51 which may
explain the failure of others (C.H. Huang, personal com-
munication) to see complementation, which has led to
Fig. 4. S. cerevisiae cells expressing RhAG (Clones 5 and 9) or with the empty vector
grown on medium with or without methylamine. A concentration of 50 mmol per L
methylamine is toxic to the cells, but expression of RhAG confers resistance.
Rh BLOOD GROUP SYSTEM REVIEW
criticism of the earlier report.52 But the most striking evi-
dence for RhAG-mediated transport of ammonium or
methylammonium was seen when RhAG was expressed in
wild-type yeast cells. RhAG dramatically rescued wild-
type cells from the toxic effects of high concentrations of
methylamine (Fig. 4). Rescue was associated with efflux
of methylammonium and was also pH-dependent, but in
contrast to complementation, was enhanced at acid pH
values. These data revealed that in yeast RhAG mediates
both uptake and efflux of substrate,50 and that the direc-
tion of transport depends on the pH value of the
medium.51
It has also been speculated that Rh/RhAG proteins
may mediate movement of CO2, based on the observation
that expression of RH1, one of two Rh-like proteins in the
blue green algae, Chlamydomonas reinhardtii, increased
when these cells were grown in 3 percent CO2 and
decreased when the cells were shifted to air.53 More than
2000 genes, however, are induced in C. reinhardtii when
shifted to growth in CO2. RH2 did not show this effect, and
no direct evidence for RH1-mediated movement of CO2 has
been demonstrated. This hypothesis, if substantiated with
direct transport data, could be reconciled with the ammo-
nium data if Rh/RhAG are found to mediate nonspecific
movement of uncharged molecules, like CO2 and NH3.
RhAG
Ascribing a physiologic role for an ammonia or ammo-
nium transporter in RBCs is rather speculative at present,
but it may be more complex than simply allowing the
RBC to survive transit through regions of high ammo-
nium in the kidney. Because total blood ammonia consists
of NH4+ and NH3 in equilibrium (NH4+ Æ NH3 + H+, pKa
9.25), at blood pH 7.4 most of the total blood ammonia
exists as charged NH4+ and about 1 percent as uncharged
NH3. The uncharged molecule NH3 is very toxic to brain
cells and mitochondria54 and any
increase in blood ammonium (NH4+)
levels will also increase the amount of
the toxic, uncharged NH3 molecule. Our
hypothesis is that RhAG serves as an
ammonia or ammonium scavenger,
keeping the total blood ammonia
(NH4+ + NH3) level low by trapping
ammonium (NH4+) inside the RBC. In
support, total ammonia levels in blood
plasma are low (0.1-0.2 mmol/L), but
total RBC ammonia levels are, on aver-
age, three times greater.55 The large total
RBC mass in the blood would enable
these cells to carry away a significant
amount of ammonia or ammonium,
and the bidirectional nature of RhAG
transport would allow ammonium to be
Volume 44, November 2004 TRANSFUSION 1669
WESTHOFF
exchanged in the liver and kidney, where other Rh/RhAG
homologs are found.
RhBG and RhCG
Nonerythroid Rh homologs, designated RhCG and RhBG,
have now been found in the kidney, liver, brain, and
skin56,57 and these proteins have the same predicted 12-
transmembrane structure as erythroid RhAG/Rh.
When expressed in oocytes, kidney RhBG and RhCG
also mediate transport of ammonium.58 Significantly, in
the kidney, ammonium ions act as expendable cations
that facilitate excretion of acids, and renal ammonium
metabolism and transport are critical for acid-base bal-
ance.59,60 In the collecting segment and collecting duct,
where large amounts of ammonium are excreted, RhBG
and RhCG are found on the basolateral and apical mem-
branes, respectively, of the intercalated cells.61 These
localization studies suggest that RhBG and RhCG are
ideally situated to mediate transepithelial movement of
ammonium from the interstitium to the lumen of the
collecting duct. In support, mouse collecting duct
(mIMCD-3) cells, which show polarized expression of
these proteins, demonstrate transporter-mediated move-
ment of ammonium, and the transport characteristics
are consistent with electroneutral NH4+/H+ exchange
activity.62
The liver is a critical organ for ammonium metabo-
lism. Localization studies in the liver reveal that RhBG is
found on the basolateral membrane in perivenous hepa-
tocytes, where it may function in ammonium uptake.
RhCG is present in bile duct epithelial cells, where it is
ideally positioned to contribute to ammonium secretion
into the bile fluid.63
In conclusion, RhCG and RhBG have been localized
to regions where ammonium production and elimination
are critical in mammalian tissues, strongly suggesting a
role for these proteins in ammonium homeostasis. Future
functional studies of the kidney, liver, and brain Rh
homologs, along with the RBC RhAG/Rh proteins, prom-
ises to lead to development of a unifying hypothesis of
ammonium transport in mammals by the Rh family of
proteins.
Function of RhCE and RhD
If RBC RhAG functions as an ammonium transporter,
what is the function of the more recently evolved Rh pro-
teins RhCE/D? Because RhCE and RhD have arisen from
RhAG, it might be assumed then that they also function
as transporters, but assays for a transport function are
hindered by their dependence on RhAG for membrane
expression. To determine whether RhCE could enhance,
modulate, or influence RhAG-mediated transport, trans-
port was measured in oocytes injected with equimolar
1670 TRANSFUSION Volume 44, November 2004
RhAG and RhCE cRNAs. Coexpression of RhCE/RhAG did
not influence the rate or total substrate accumulated in
oocytes compared to that seen with expression of RhAG
alone (unpublished observations). Although additional
studies are necessary to determine whether RhCE/D are
involved in membrane transport, our hypothesis is that
RhCE/D may have lost transport function and are evolving
to have a structural role in the RBC membrane. Evidence
in support of an alternative role for RhCE/D in the RBC
comes from DIVERGE analysis (personal observations)
and evidence from others64,65 that indicate the RhCE/D
proteins are rapidly evolving, suggesting that their func-
tion may be changing.
A structural role for the RBC Rh proteins is suggested
from the RBC shape defects, fragility, and the resulting
anemia seen in patients with Rhnull disease.66,67 Rh/RhAG
are part of a membrane protein complex that includes
CD47, also known as integrin-associated protein, glyco-
phorin B, and LW (ICAM-4) (reviewed in Avent and Reid41
and Huang et al.68). Band 3 may also be associated with
the complex.69 The Rhnull defect suggests that there is an
important membrane-cytoskeletal attachment site in RBC
that may be mediated by Rh, RhAG, or a member of the
Rh complex. Recent studies reveal that the Rh complex is
linked to the membrane skeleton through CD47–protein
4.2 interactions70 and through a novel Rh/RhAG-ankryin
connection.71
SIGNIFICANCE AND SUMMARY
The genetic basis of the Rh blood group proteins has been
intensely investigated in the past decade, and the origin
of most of the antigens has now been determined. At
present, routine RH testing by molecular methods is ham-
pered because of the large number of Rh polymorphisms.
More than 77 RHD and 22 RHCE gene variants have been
reported (summarized in Reid and Lomas-Francis22), and
additional variants are still being discovered. Molecular
testing, however, as an adjunct to serologic methods,
improves transfusion safety and outcomes, and the chal-
lenge for the next decade is to develop automated plat-
forms that sample several regions of the genes for
unequivocal interpretation.
The discovery that the Rh family of proteins is
involved in ammonia or ammonium transport and is ide-
ally positioned in key tissues essential for ammonium
elimination is a significant finding because it was long
assumed that the high membrane permeability of ammo-
nia (NH3) would obviate the need for specific transport
pathways in mammalian cells. This is reminiscent of the
discovery of the function of the Colton blood group pro-
tein as the first member of a family of water transporters
(aquaporins) and of the Kidd blood group protein as the
first member of a family of urea transporters. The move-
ment of water and urea were also originally thought to

occur only by passive movement through the lipid bilayer.
Like the aquaporins and the urea transporters, homologs
of the Rh proteins have been found in other tissues and in
many organisms including the sponge, the slime mold,
the fruit fly, fish, and the frog (summarized in Huang et
al.68), indicating that they have an essential and conserved
function throughout evolution.
RBC membrane protein–cytoskeleton and protein–
protein interactions are an active area of investigation.
Although the major attachment sites between the RBC
cytoskeleton and the lipid bilayer are understood to be
through glycophorin C and band 3, new evidence for an
additional attachment site, mediated by the Rh complex,
is indeed exciting. Future studies are needed to deter-
mine the protein–protein associations and the dynamics
of the assembly of the Rh–membrane complex to under-
stand the Rhnull defect. Importantly, the findings above
highlight the major contributions that the study of the
function of blood group proteins continues to make to
biology and physiology and emphasize the opportunities
for research within the profession of transfusion medi-
cine. The recent revelations about the role of the Rh/
RhAG proteins in RBC membrane structural integrity and
their transport function have taken the field of Rh well
beyond consideration as simply a family of blood group
antigens. Blood groups are not just for blood bankers
anymore.
REFERENCES
1. Race RR, Sanger R. Blood groups in man. 6th ed. Oxford:
Blackwell; 1975.
2. Levine P, Burnham L, Katzin WM, Vogel P. The role of isoim-
munization in the pathogenesis of erythroblastosis fetalis.
Am J Obstet Gynecol 1941;42:925-37.
3. Rosenfeld R. Who discovered Rh? A personal glimpse of the
Levine-Wiener argument. Transfusion 1989;29:355-7.
4. Mollison PL, Hughes-Jones NC, Lindsay M, Wessely J.
Suppression of primary RH immunization by passively-
administered antibody: experiments in volunteers.
Vox Sang 1969;16:421-39.
5. Agre P, Saboori AM, Asimos A, Smith BL. Purification and
partial characterization of the Mr 30,000 integral membrane
protein associated with the erythrocyte Rh(D) antigen. J Biol
Chem 1987;262:17497-503.
6. Cherif-Zahar B, Bloy C, Le Van Kim C, et al. Molecular
cloning and protein structure of a human blood group Rh
polypeptide. Proc Natl Acad Sci U S A 1990;87:6243-7.
7. Cherif-Zahar B, Mattei MG, Le Van Kim C, Bailly P, Cartron
JP, Colin Y. Localization of the human Rh blood group gene
structure to chromosome 1p34.3-1p36.1 region by in situ
hybridization. Hum Genet 1991;86:398-400.
8. Arce MA, Thompson ES, Wagner S, Coyne KE, Ferdman BA,
Lublin DM. Molecular cloning of RhD cDNA derived from a
Rh BLOOD GROUP SYSTEM REVIEW
gene present in RhD-positive, but not RhD-negative individ-
uals. Blood 1993;82:651-5.
9. Simsek S, de Jong CAM, Cuijpers HTM, et al. Sequence anal-
ysis of cDNA derived from reticulocyte mRNAs coding for Rh
polypeptides and demonstration of E/e and C/c polymor-
phism. Vox Sang 1994;67:203-9.
10. Ridgwell K, Spurr NK, Laguda B, et al. Isolation of cDNA
clones for a 50 kDa glycoprotein of the human erythrocyte
membrane associated with Rh (Rhesus) blood-group anti-
gen expression. Biochem J 1992;287:223-8.
11. Eyers SAC, Ridgwell K, Mawby WJ, Tanner MJA. Topology
and organization of human Rh (Rhesus) blood group-related
polypeptides. J Biol Chem 1994;269:6417-23.
12. Cherif-Zahar B, Raynal V, Gane P, et al. Candidate gene
acting as a suppressor of the RH locus in most cases of Rh-
deficiency. Nat Genet 1996;12:168-73.
13. Huang CH. Molecular insights into the Rh protein family
and associated antigens. Curr Opin Hematol 1997;4:94-
103.
14. Southcott MJ, Tanner MJ, Anstee DJ. The expression of
human blood group antigens during erythropoiesis in a cell
culture system. Blood 1999;93:4425-35.
15. Mouro I, Colin Y, Sistonen P, et al. Molecular basis of the
RhCW (Rh8) and RhCX (Rh9) blood group specificities.
Blood 1995;86:1196-201.
16. Daniels G. Human blood groups. 2nd ed. Cambridge (MA):
Blackwell Science; 2002.
17. Faas BHW, Beckers EAM, Wildoer P, et al. Molecular back-
gound of VS and weak C expression in blacks. Transfusion
1997;37:38-44.
18. Daniels GL, Faas BH, Green CA, et al. The VS and V blood
group polymorphisms in Africans: a serologic and molecular
analysis. Transfusion 1998;38:951-8.
19. Hemker MB, Ligthart PC, Berger L, et al. DAR, a new RhD
variant involving exons 4, 5, and 7, often in linkage with
ceAR, a new Rhce variant frequently found in African Blacks.
Blood 1999;94:4337-42.
20. Westhoff CM, Silberstein LE, Wylie DE, Reid ME. 16Cys
encoded by the RHce gene is associated with altered expres-
sion of the e antigen and is frequent in the Ro haplotype.
Br J Haematol 2001;113:666-71.
21. Noizat-Pirenne F, Lee K, Pennec PY, et al. Rare RHCE phe-
notypes in black individuals of Afro-Caribbean origin: iden-
tification and transfusion safety. Blood 2002;100:4223-31.
22. Reid ME, Lomas-Francis C. The blood group antigen facts
book. 2nd ed. San Diego: Academic Press; 2004.
23. Mouro I, Colin Y, Gane P, et al. Molecular analysis of blood
group Rh transcripts from a rGr variant. Br J Haematol 1996;
93:472-4.
24. Noizat-Pirenne F, Mouro I, Gane P, et al. Heterogeneity of
blood group RhE variants revealed by serological analysis
and molecular alteration of the RHCE gene and transcript.
Br J Haematol 1998;103:429-36.
25. Beckers EA, Porcelijn L, Ligthart P, et al. The RoHar antigenic
complex is associated with a limited number of D epitopes
Volume 44, November 2004 TRANSFUSION 1671
WESTHOFF
and alloanti-D production: a study of three unrelated per-
sons and their families. Transfusion 1996;36:104-8.
26. Wallace M, Lomas-Francis C, Tippett P. The D antigen char-
acteristic of RoHar is a partial D antigen. Vox Sang 1996;70:
169-72.
27. Rouillac C, Gane P, Cartron J, et al. Molecular basis of the
altered antigenic expression of RhD in weak D(Du) and RhC/
e in RN phenotypes. Blood 1996;87:4853-61.
28. Cavalli-Sforza LL. Genes, peoples and languages. Sci Am
1991;265:104-10.
29. Colin Y, Cherif-Zahar B, Le Van Kim C, et al. Genetic basis of
the RhD-positive and RhD-negative blood group polymor-
phism as determined by Southern analysis. Blood 1991;78:
2747-52.
30. Avent ND, Martin PG, Armstrong-Fisher SS, et al. Evidence
of genetic diversity underlying Rh D-, weak D (Du), and par-
tial D phenotypes as determined by multiplex polymerase
chain reaction analysis of the RHD gene. Blood 1997;89:
2568-77.
31. Andrews KT, Wolter LC, Saul A, Hyland CA. The RhD- trait
in a white patient with the RhCCee phenotype attributed to
a four-nucleotide deletion in the RHD gene. Blood 1998;92:
1839-40.
32. Wagner FF, Frohmajer A, Flegel WA. RHD positive haplo-
types in D negative Europeans. BMC Genet 2001;2:10.
33. Huang CH. Alteration of RH gene structure and expression
in human dCCee and DCW-red blood cells: phenotypic
homozygosity versus genotypic heterozygosity. Blood 1996;
88:2326-33.
34. Chang JG, Wang JC, Yang TY, et al. Human RhDel is caused
by a deletion of 1,013 bp between introns 8 and 9 including
exon 9 of RHD gene [letter]. Blood 1998;92:2602-4.
35. Shao CP, Maas JH, Su YQ, Kohler M, Legler TJ. Molecular
background of Rh D-positive, D-negative, D (el) and weak D
phenotypes in Chinese. Vox Sang 2002;83:156-61.
36. Singleton BK, Green CA, Avent ND, et al. The presence of an
RHD pseudogene containing a 37 base pair duplication and
a nonsense mutation in Africans with the Rh D-negative
blood group phenotype. Blood 2000;95:12-8.
37. Wagner FF, Gassner C, Muller TH, et al. Molecular basis of
weak D phenotypes. Blood 1999;93:385-93.
38. Wagner FF, Frohmajer A, Ladewig B, et al. Weak D alleles
express distinct phenotypes. Blood 2000;95:2699-708.
39. Flegel WA, Khull SR, Wagner FF. Primary anti-D
immunization by weak D type 2 RBCs. Transfusion
2000;40:428-34.
40. Wagner FF, Eicher NI, Jorgensen JR, Lonicer CB, Flegel WA.
DNB: a partial D with anti-D frequent in Central Europe.
Blood 2002;100:2253-6.
41. Avent ND, Reid ME. The Rh blood group system: a review.
Blood 2000;95:375-87.
42. Beck ML, Harding J. Incidence of D category VI among Du
donors in the USA. Transfusion 1991; 31(Suppl):25.
43. Wagner FF, Gassner C, Muller TH, et al. Three molecular
structures cause rhesus D category VI phenotypes with
1672 TRANSFUSION Volume 44, November 2004
distinct immunohematologic features. Blood 1998;91:2157-
68.
44. Fridey JL. Standards for blood banks and transfusion ser-
vices. 23rd ed. Bethesda: American Association of Blood
Banks; 2003.
45. Wagner FF, Flegel WA. RHD gene deletion occurred in the
Rhesus box. Blood 2000;95:3662-8.
46. Matheson KA, Denomme GA. Novel 3¢Rhesus box sequences
confound RHD zygosity assignment. Transfusion 2002;42:
645-50.
47. Chiu RW, Murphy MF, Fidler C, et al. Determination of RhD
zygosity: comparison of a double amplification refractory
mutation system approach and a multiplex real-time
quantitative PCR approach. Clin Chem 2001;47:667-72.
48. Lo YM, Hjelm NM, Fidler C, et al. Prenatal diagnosis of fetal
RhD status by molecular analysis of maternal plasma.
N Engl J Med 1998;339:1734-8.
49. Westhoff CM, Ferreri-Jacobia M, Mak DOD, Foskett JK. Iden-
tification of the erythrocyte Rh blood group glycoprotein as
a mammalian ammonium transporter. J Biol Chem 2002;
277:12499-502.
50. Marini AM, Matassi G, Raynal V, et al. The human
Rhesus-associated RhAG protein and a kidney homologue
promote ammonium transport in yeast. Nat Genet
2000;26:341-4.
51. Westhoff CM, Siegel DL, Burd CG, Foskett JK. Mechanism of
genetic complementation of ammonium transport in yeast
by human erythrocyte Rh-associated glycoprotein. J Biol
Chem 2004;279:17443-8.
52. Soupene E, Ramirez RM, Kustu S. Evidence that fungal MEP
proteins mediate diffusion of the uncharged species NH3
across the cytoplasmic membrane. Mol Cell Biol 2001;21:
5733-41.
53. Soupene E, King N, Feild E, et al. Rhesus expression in a
green alga is regulated by CO(2). Proc Natl Acad Sci U S A
2002;99:7769-73.
54. Kosenko E, Kaminsky YG, Felipo V, Minana MD, Grisolia S.
Chronic hyperammonemia prevents changes in brain
energy and ammonia metabolites induced by acute ammo-
nium intoxication. Biochim Biophys Acta 1993;1180:321-6.
55. Huizenga JR, Tangerman A, Gips CH. Determination of
ammonia in biological fluids. Ann Clin Biochem 1994;31:
529-43.
56. Liu Z, Chen Y, Mo R, et al. Characterization of human RhCG
and mouse Rhcg as novel nonerythroid Rh glycoprotein
homologues predominantly expressed in kidney and testis.
J Biol Chem 2000;275:25641-51.
57. Liu Z, Peng J, Mo R, Hui CC, Huang CH. Rh type B glycopro-
tein is a new member of the Rh superfamily and a putative
ammonia transporter in mammals. J Biol Chem 2001;276:
1424-33.
58. Westhoff CM, Weiner ID, Foskett JK. Transport characteris-
tics of the putative ammonium transporters, RhB Glycopro-
tein (RhBG) and RhC Glycoprotein (RhCG) when expressed
in the Xenopus oocyte. J Am Soc Nephrol 2003;14:SA-P0011.

59. Knepper MA, Packer R, Good DW. Ammonium transport in
the kidney. Physiol Rev 1989;69:179-249.
60. Frank AE, Wingo CS, Andrews PM, et al. Mechanisms
through which ammonia regulates cortical collecting duct
net proton secreation. Am J Phys 2002;282:F1120-8.
61. Verlander JW, Miller RT, Frank AE, et al. Localization of the
ammonium transporter proteins RhBG and RhCG in mouse
kidney. Am J Physiol 2003;284:F323-37.
62. Handlogten ME, Hong SP, Westhoff CM, Weiner ID. Basolat-
eral ammonium transport by the mouse inner medullary
collecting duct, mIMCD-3, cell. Am J Physiol 2004;287:F628-
38.
63. Weiner ID, Verlander JW. Renal and hepatic expression of
the ammonium transporter proteins, Rh B glycoprotein and
Rh C glycoprotein. Acta Physiol Scand 2003;179:331-8.
64. Matassi G, Cherif-Zahar B, Pesole G, Raynal V, Cartron JP.
The members of the RH gene family (RH50 and RH30)
followed different evolutionary pathways. J Mol Evol
1999;48:151-9.
65. Kitano T, Sumiyama K, Shiroishi T, Saitou N. Conserved
evolution of the Rh50 gene compared to its homologous Rh
blood group gene. Biochem Biophy Res Commun 1998;
249:78-85.
Rh BLOOD GROUP SYSTEM REVIEW
66. Sturgeon P. Hematological observations on the
anemia associated with blood type Rhnull. Blood 1970;
36:310-20.
67. Ballas S, Clark MR, Mohandas N, et al. Red cell membranes
and cation deficiency in Rh null syndrome. Blood 1984;63:
1046-55.
68. Huang CH, Liu PZ, Chen JG. Molecular biology and genetics
of the Rh blood group system. Semin Hematol 2000;37:150-
65.
69. Bruce LJ, Beckmann R, Ribeiro ML, et al. A band 3-based
macrocomplex of integral and peripheral proteins in the
RBC membrane. Blood 2003;101:4180-8.
70. Dahl KN, Westhoff CM, Discher DE. Fractional attachment
of CD47 (IAP) to the erythrocyte cytoskeleton and visual
colocalization with Rh protein complexes. Blood 2003;101:
1194-9.
71. Nicolas V, Le Van Kim C, Gane P, et al. Rh-RhAG/ankyrin-R,
a new interaction site between the membrane bilayer and
the red cell skeleton, is impaired by Rh (null)-associated
mutation. J Biol Chem 2003;278:25526-33.
72. Judd WJ, Moulds M, Schlanser G. Reactivity of FDA-
approved monoclonal anti-D reagents with partial D RBCs.
Transfusion 2002;42(Suppl):20.
Volume 44, November 2004 TRANSFUSION 1673