Báo cáo khoa học: Cloning of a rat gene encoding the histo-blood group A enzyme Tissue expression of the gene and of the A and B antigens potx

Cloning of a rat gene encoding the histo-blood group A enzymeTissue expression of the gene and of the A and B antigens Anne Cailleau-Thomas1, Be´atrice Le Moullac-Vaidye1, Je´zabel Roche

Cloning of a rat gene encoding the histo-blood group A enzyme

Tissue expression of the gene and of the A and B antigens

Anne Cailleau-Thomas1, Be´atrice Le Moullac-Vaidye1, Je´zabel Rocher,1Danie`le Bouhours2,

Claude Szpirer3and Jacques Le Pendu1

1

INSERM U419, Institut de Biologie, Nantes, France;2INSERM U539, Faculte´ de Me´decine, Nantes Cedex, France;

3

IBMM, Universite´ Libre de Bruxelles, Gosselies, Belgium

The complete coding sequence of a BDIX rat gene

homo-logous to the human ABO gene was determined

Identifi-cation of the exon–intron boundaries, obtained by

comparison of the coding sequence with rat genomic

sequences from data banks, revealed that the rat gene

structure is identical to that of the human ABO gene It

localizes to rat chromosome 3 (q11-q12), a region

homolo-gous to human 9q34 Phylogenetic analysis of a set of

sequences available for the various members of the same

gene family confirmed that the rat sequence belongs to the

ABO gene cluster The cDNA was transfected in CHO cells

already stably transfected with an a1,2fucosyltransferase in

order to express H oligosaccharide acceptors Analysis of the

transfectants by flow cytometry indicated that A but not B

epitopes were synthesized Direct assay of the enzyme

activity using 2¢ fucosyllactose as acceptor confirmed the strong UDP-GalNAc:Fuca1,2GalaGalNAc transferase (A transferase) activity of the enzyme product and allowed detection of a small UDP-Gal:Fuca1,2GalaGal transferase (B transferase) activity The presence of the mRNA and of the A and B antigens was searched in various BDIX rat tissues There was a general good concordance between the presence of the mRNA and that of the A antigen Tissue distributions of the A and B antigens in the homozygous BDIX rat strain were largely different, indicating that these antigens cannot be synthesized by alleles of the same gene in this rat inbred strain

Keywords: ABO; N-acetylgalactosaminyltransferase; histo-blood group; antigen; rat

Histo-blood group antigens A and B are oligosaccharides

carried by glycolipids and glycoproteins or present as free

oligosaccharides in some biological fluids such as milk or

urine The immunodominant A and B epitopes correspond

to the trisaccharides GalNAca1,3(Fuca1,2)Galb- and

Gala1,3(Fuca1,2)Galb-, respectively In humans, the ABO

gene is polymorphic with A alleles encoding A transferases,

B alleles encoding B transferases and O alleles encoding

inactive products The A transferases catalyse the transfer of

an N-acetylgalactosamine to acceptor H substrates

(Fuca1,2Galb-) whereas the B transferases catalyse the

transfer of a galactose to the same substrates [1] ABH

antigens are found in many species and have a wide tissue

distribution Their main sites of expression appear to be

epithelia in contact with the external environment such as

the gut, the higher respiratory tract and the genito-urinary

tract In some primates, they are present on the vascular

endothelium of all tissues and in chimpanzee, gorilla and man they are additionally present on erythrocytes, hence their name blood group antigens [2]

The molecular genetic basis of the human ABO alleles has been elucidated Although many mutations have been described to date, only some of these are functionally relevant [3] Functional analyses have been performed to determine which amino-acids are responsible for the A or

B enzyme activities These studies revealed that the amino-acids at positions 266 and 268 and to a lesser extent at position 235 were critical in determining whether the enzyme transfers an N-acetylgalactosamine, a galactose or both [4] For example, the presence of a glycine at position

268 allows the transfer of an N-acetylgalactosamine (GalNAc) But this activity is modulated by amino-acids

at the other two positions as, depending on these, transfer

of Gal may become possible in addition to the transfer of GalNAc

The biological meaning of the ABO phenotypes is still largely obscure Yet, the above mentioned tissue distribu-tion and some associadistribu-tions between the ABO polymor-phism and infectious diseases suggest a role in the interaction with pathogens For example, a strong associ-ation is found between blood group O and susceptibility to cholera [5] Moreover, various strains of bacteria can adhere

to either A, B or H antigens, suggesting that microbes use them as receptors and that their host range may be influenced by the individual blood group phenotype [6] Histo-blood group antigens or structurally related carbo-hydrates may be present on pathogens Protective antibod-ies directed against these structures may therefore be

Correspondence to J Le Pendu, Inserm U419,

Institut de Biologie, 9 Quai Moncousu, F-44093, Nantes, France.

Fax: + 33 240 08 40 82, Tel.: + 33 240 08 40 99,

E-mail: jlependu@nantes.inserm.fr

Abbreviations: A transferase, UDP-GalNAc:Fuca1,2GalaGalNAc

transferase; B transferase, UDP-Gal:Fuca1,2GalaGal transferase;

CHO, Chinese hamster ovary; GalNAc, N-acetylgalactosamine;

Gal, galactose; Fuc, fucose; FITC, fluorescein isothiocyanate;

AEC, 3-amino-9-ethylcarbazol.

Enzymes: UDP-GalNAc:Fuca1,2GalaGalNAc transferase

(EC 2.4.1.40).

(Received 22 March 2002, revised 12 June 2002, accepted 5 July 2002)

differentially generated by the host, depending on the blood

group phenotype [7] The differential host range for

adhesion of pathogens and the presence of protective

antibodies mean that all individuals would not be equally

sensitive to a given pathogen This would provide a

mechanism of protection against the pathogenic strain at

the level of the population and a selective force to maintain

polymorphism at the ABO locus [6]

In the present work, we report the isolation of a BDIX rat

cDNA homologous to the human ABO sequences encoding

for an A histo-blood group enzyme Examination of the

tissue distribution of the corresponding mRNA and of the

A and B antigens in the BDIX strain of rat was performed,

allowing some aspects of the ABO genetics in this species to

be considered

M A T E R I A L S A N D M E T H O D S

Cloning of a rat ABO-like cDNA

An EMBL-3 rat genomic DNA phage library (Clontech)

was screened with a cDNA probe (P718), derived from the

human blood group A transferase and corresponding to

nucleotides 148–865 of the human cDNA sequence [8]

Positive recombinants were purified, digested with EcoRI

and analyzed by Southern Blot using the pb718 probe Two

hybridization-positive fragments (4.0 and 3.5 kb) were

obtained and digested by HinfI The products were ligated

into the pUC18 vector (Pharmacia) and sequenced One

clone, from the 4 kb fragment digest, contained an insert of

406 bp, a stretch of which was 84.4% similar to exon 6 of

the human blood group A gene coding sequence The

complete coding sequence was obtained by RACE/PCR

using primers deduced from this fragment To this end, total

RNA from BDIX rat stomach was extracted using the SV

Total RNA isolation kit from Promega This RNA

preparation was used to obtain mRNA using the Oligotex

kit from Qiagen Double stranded cDNA synthesis, adaptor

ligation and RACE/PCR were performed using the

Clon-tech Marathon cDNA Amplification kit Elongation in the

5¢ direction was performed using an inverted primer

deduced from the 406 bp fragment homologous to the

human A gene (GTCGATGTTGAAGGTCCCCTCCCA

GATG) and the adaptor AP1 primer provided by the

supplier The second nested PCR was performed using a

nested inverted primer deduced from the same fragment

sequence (TCCCAGATGATGGGAGCCACGCCAA

GG) and the nested adaptor AP2 primer from the supplier

Synthesis in the 3¢ direction was performed by the same

method using the following first primer

(CTTGTCTTCACTCCTTGGCTGGCTCCCAT) and

the adaptor AP1 primer, followed by a second nested

PCR with the second nested primer (CATCTGG

GAGGGGACCTTCAACATCGAC) The PCR was run

with the Advantage Polymerase (Clontech) and the

man-ufacturer’s touchdown-RACE program The PCR products

were ligated into the pUC18 vector and sequenced To

obtain the full coding fragment, stomach cDNA was PCR

amplified using the following primers, deduced from the

sequence of the 5¢ and 3¢ RACE products: AC

CATCCCGGGCCTTGCATGGA (forward) and GCTA

CAGGTACCGCCTCTCCAA (reverse) The product was

ligated into the pUC18 vector and sequenced

Sequence analysis Multiple alignments were performed with the CLUSTALW program [9] Genetic distances of the complete deduced peptide sequences were calculated by the
Neighbor joining method from thePHILIPprogram Approximate location of transmembrane regions were determined using theTMHMM, TMAPandTOPPRED2 programs Determination of the exon/ intron boundaries were obtained by analysis of the rat genomic sequences available in the NCBI database The programs used are all available from http://www.infobio-gen.fr

Chromosome localization The Abo gene was first assigned to a rat chromosome using

a panel of standard rat X mouse cell hybrids that segregate rat chromosomes [10] The hybrids were typed by PCR with the following primers: 5¢-GGAGCAGCTGGAGT CATG-3¢ and 5¢-GGTCATCCTGTATCCTTCA-3¢ (the 5¢ end of these primers corresponds to the positions 163 and

270, respectively, of [35104745] from the Rattus norvegicus WGS trace database) For regional localization, the panel of rat· hamster radiation cell hybrids [11] was typed in the same manner The mapping results were obtained from the rat radiation hybrid map server at the Otsuka GEN Research Institute (http://ratmap.ims.u-tokyo.ac.jp/menu/ RH.html) [11]

RT-PCR analysis Total RNAs (1 lg) from various rat tissues listed below were prepared using the SV Total RNA isolation System kit from Promega and reverse transcribed at 42C with the M-MLV reverse transcriptase from Promega Contaminat-ing DNA had been removed by digestion with RNAse-free DNAseI (10 unitsÆlg)1RNA) for 15 min at room temper-ature Amplification of the cDNA corresponding to the A enzyme cDNA was performed with the following primers: CAGACGGATGTCCAGAAAGTTG and: GCTACAG GTACCGCCTCTCCAA Amplification was performed using the Advantage Polymerase (Clontech) with initial denaturation at 94C 3 min, followed by 30 cycles of 94 C

30 s, 64C 45 s, 68 C 2 min The amplification yields a product of 929 bp The control of cDNA quality was performed by amplification of glyceraldehyde phosphate dehydrogenase (GAPDH)

Transfection of CHO cells The complete coding sequence of the rat A gene was inserted into the pDR2 eukaryotic expression vector (Clontech) deleted of the sequences lying between the EcoRV and ClaI sites Chinese hamster ovary carcinoma cells, CHO cells, are devoid of a1,2fucosyltransferase activity and therefore of ABH antigens They were first transfected using LipofectAMINTMwith the rat a1,2fuco-syltransferases cDNA FTB (GenBank accession number AF131238) in the pBK-CMV expression vector (Gibco, Paisley, UK) according to the manufacturer’s instructions Twenty-four hours later, fresh medium was added and 48 h later, selective medium containing 0.6 mgÆmL)1 G418 (Gibco) was added Transfected cells expressing a1,2-linked

fucose residues were selected by flow cytometry, using

fluorescein isothiocyanate (FITC)-labeled UEA-I, after

cloning by limiting dilutions as previously described [12]

For each cell line, a strongly expressing clone was selected

and transfected a second time using the same procedure

with the rat A enzyme cDNA in the pDR2 vector

Forty-eight hours later, cells were cultured in selective medium

containing 0.6 mgÆmL)1 hygromycin After cloning by

limiting dilutions, strongly A antigen expressing

transfec-tants were selected Control transfected cells were prepared

by transfection with the empty vectors These stable

transfectants were cultured in RPMI 1640, 10% fetal

bovine serum, 2 mM L-glutamine, free nucleotides

(10 lgÆmL)1), 100 UÆmL)1 penicillin and 100 lgÆmL)1

streptomycin (Gibco) supplemented with 0.25 mgÆmL)1

G418 and 0.2 mgÆmL)1hygromycin They were cultured at

confluence after dispersal with 0.025% trypsin in 0.02%

EDTA Cells were routinely checked for mycoplasma

contamination by Hoescht 33258 (Sigma, St Louis, MO,

USA) labeling

Cytofluorimetric analysis

Viable cells (2· 105per well) were incubated with

antibod-ies at the appropriate dilutions in NaCl/Picontaining 0.1%

gelatin for 1 h at 4C Optimal concentrations of antibodies

were chosen after serial dilutions to obtain the strongest

positive signal without cell death The anti-A mAb 3–3 A

was obtained from J Bara (INSERM U482, Villejuif,

France) It recognizes all types of A antigens [13] and does

not show any detectable cross reactivity with B epitopes as

judged from enzyme immunoassay with synthetic

oligosac-charides and immunostaining of human tissues of known

ABO phenotypes The anti-B mAb ED3 is a gift from

A Martin (CRTS, Rennes, France) It recognizes all types

of B and shows no detectable cross-reactivity with

A epitopes or with the Gala1,3Gal epitope [14] Following

the first incubation with monoclonal antibodies, after three

washes with the same buffer, a second 30 min incubation

was performed with the FITC-labeled anti-(mouse Ig) Ig

under the same conditions After washings in the same

buffer, fluorescence analysis was performed on a FACScan

(Beckton–Dickinson) using theCELLQUESTprogram

Detection of enzyme activity

Confluent transfected CHO cells were rinsed with ice-cold

NaCl/Pi, pH 7.2, then recovered by scraping After

washing with ice-cold NaCl/Pi, cells were solubilized in

50 mM cacodylate pH 7.0, containing 2% (v/v) Triton

X-100 on ice for 30 min Following a centrifugation at

13 000 g for 10 min, the supernatant was collected and used

as a crude enzyme preparation Protein concentration was

determined using bicinchoninic acid The reaction mixture

contained: 50 lg protein extracts, 30 mM MnCl2, 5 mM

ATP, 10 mM NaN3, 5 mM 2¢ fucosyllactose and 20 lM

UDP-D-[14C]N-acetylgalactosamine (55 mCiÆmmol)1, ICN,

Costa Mesa, CA, USA) or 20 lM UDP-D-[14C]galactose

(278 mCiÆmmol)1, NEN, Chemical Center,

Dreieichen-dain, Germany) in a final volume of 50 lL and was

incubated at 37C for 16 h After incubation, the reaction

mixture was quenched with 750 lL distilled water and

applied to an AG1-X8 column, chloride form, 100–200

mesh (Bio-Rad, Hercules, CA, USA) The radiolabeled product was then eluted with 1 mL water and counted in

5 mL scintillation liquid (Ready SafeTM, Beckman, Palo Alto, CA, USA) Background levels of radioactivity were obtained from controls without exogenous acceptor Values obtained for the controls were then subtracted from those obtained for the assays

Immunohistological analysis Tissues from 2- to 3-month old rats were collected and immediately frozen or paraffin embedded Sections (5 lm) were prepared and washed in NaCl/Pi Endogenous peroxidase was inhibited using methanol/H2O2 0.3% for

20 min Sections were then washed in NaCl/Pifor 5 min and covered with NaCl/Pi/BSA 1% for 20 min at room temperature in a moist chamber After washing in NaCl/Pi, sections were covered with either the primary antibodies diluted in NaCl/Pi/BSA 1% and left at 4C overnight Sections were then rinsed thrice with NaCl/Piand incubated with biotinylated anti-(mouse IgG) Ig (Vector Laboratories, Burlingame, CA, USA) diluted at 1/100 for 60 min at room temperature After washing in NaCl/Pi, the sections were covered with peroxidase-conjugated avidin (Vector Labo-ratories) diluted at 1 : 1000 for 45 min, washed with NaCl/

Piand reactions were revealed with 3-amino-9-ethylcarbazol (AEC) Counterstaining was performed with Mayer’s hemalun

R E S U L T S

Sequence analysis

A coding sequence homologous to the human ABO coding sequences was isolated (GenBank accession AF264018) It possesses an open reading frame of 1047 bp, 77% identical with the human A gene coding region Comparisons of this sequence with genomic sequences available in the rat genome data bank allowed determination of the exon/ intron boundaries as they conform to the GT-AG consensus rule (Fig 1) The gene organization appears similar to that

of the human ABO gene with seven exons [15] The same analysis was performed on the mouse A/B gene orthologous

to the human ABO gene [16] This gene lacks exon 4 At the amino-acid level, the rat sequence shows 70, 71 and 77% identity with the human, pig and mouse sequences, respec-tively (Fig 2A, Table 1) Like all glycosyltransferases the

Fig 1 Comparison of the rat Abo gene exonic structure with those of the human and mouse orthologs Organization of the human and mouse genes has been previously reported Exons are represented by boxes numbered in bold Nucleotide numbers limiting the exons are noted in superscript For the mouse gene, a search in the mouse genomic data bank confirmed the absence of one exon, corresponding to exon 4 of the other two species, as the genomic sequence separating mouse exons

3 and 4 revealed no potential exonic sequence with homology to the human and rat exon 4.

rat enzyme presents a short N-terminal intracytoplasmic

domain, a transmembrane domain followed by a stem

region and a catalytic domain, the latter domain presenting

the highest identity among the four sequences Of note, the

presence of a conserved potential N-glycosylation site at

the beginning of the catalytic domain Previous studies of

the human ABO transferases underscored the major

importance of the amino-acid at position 268 with a glycine

determining A activity whereas an alanine determines B

activity A glycine is present at the homologous position in

the rat sequence (position 263), suggesting a potential A

enzyme activity

At present, the ABO gene family is known to comprise four members, the ABO gene itself, the a3galactosyltrans-ferase (pseudo B) gene [17], the aN-acetylgalactosaminyl-transferase or Forssman synthetase gene [18] and the a-galactosyltransferase iGb3 synthetase gene [19] Phylo-genetic analysis of this gene family revealed a clear distinction between the four genes, with the new rat sequence falling within the ABO cluster (Fig 2B)

Chromosome localization of the rat Abo gene The gene was first assigned to rat chromosome 3, using a panel of 16 standard rat X mouse cell hybrid clones segregating rat chromosomes No discordant clone was obtained for chromosome 3, while at least two discordant clones were counted for each other chromosome (data not shown) Chromosome placement by radiation hybrid map-ping confirmed this result, with a precise localization between D3Rat54 (at 58cR, lod score¼ 5.67) and D3Rat50 (at 62cR, lod score¼ 5.12) This position corre-sponds to the centromeric region of the chromosome (bands 3q11–q12) This rat chromosome region is known to be homologous to the human region 9q34 [20,21], where the human ABO gene resides [22]

Determination of the enzyme activity

In order to study the functional characteristics of the rat A-like gene, the cDNA was transfected in CHO cells already stably transfected with an a2fucosyltransferase cDNA, namely the rat FTB [23] The presence of this fucosyltrans-ferase in CHO cells allows expression of H histo-blood group structures which are compulsory precursors of the A and B antigens Stable doubly transfected cells were isolated and tested by flow cytometry for their expression of A or B antigens (Fig 3) A positive control cell line (MT-450) known to constitutively express both antigens indicated that the antibodies readily detected their respective epitopes when present [24] The doubly transfected CHO cells were strongly labeled by the anti-A reagent, but not significantly

Fig 2 Comparison of the amino acid sequences of the A or cis A/B

transferases in four species (A) and phylogenetic analysis of the ABO

gene family (B) (A) Identical residues are marked by arrows The

transmembrane regions are highlighted in grey, the conserved

N-gly-cosylation site is boxed and the amino-acids corresponding to positions

266 and 268 of the human sequence are labeled in white on black

boxes The numbering corresponds to that of a consensus sequence.

(B) Genetic distances were calculated from CLUSTALW multiple

align-ments using the
neighbour joining method from the sequences listed

in Table 1 The scale bar represents the number of substitutions per site

for a unit branch length.

Table 1 Identification of the ABO gene family sequences used for the phylogenetic analysis.

Enzyme Species GenBank/EBI

A transferase Human J05175

A transferase Rat AF264018

A transferase Pig AF050177

A (cis A/B) transferase Mouse AB041039 A-likea Human M65082 Gal transferase Platyrrhini S71333 Gal transferase Marmoset A56480 Gal transferase Cow J04989 Gal transferase Pig L36152 Gal transferase Mouse M85153 Gal transferase Rat AF520589 IGb3 synthetase Human AL513327 IGb3 synthetase Rat AF246543 Forssman synthetasea Human AF163572 Forssman synthetase Dog CFU66140

a Pseudogenes.

by the anti-B reagent, suggesting that the new rat gene

encodes an enzyme with the catalytic activity of the A

histo-blood group transferase To confirm this result, the enzyme

activity was directly assayed on cell extracts of CHO

transfectants (Fig 4) No transfer of galactose or

N-acetylgalactosamine could be detected on extracts from

the control FTB transfected cells using 2¢ fucosyllactose as

acceptor However, cell extracts from the double

transfec-tants showed a high N-acetylgalactosaminyltransferase

activity and a weak galactosyltransferase activity These

results indicate that the new rat enzyme is indeed an A

histo-blood group transferase with a small B transferase activity

Tissue expression of the mRNA and of the A

and B antigens in the rat

An RT-PCR analysis was performed to determine the

tissue expression of the rat Abo gene Primers were chosen

to encompass different exons so as to confirm

amplifica-tion of cDNA and lack of contaminaamplifica-tion by genomic DNA A band at the expected size was detected in various tissues as indicated in Table 2 A strong signal was obtained in the oesophagus, the stomach, the colon, the

Fig 3 Cytofluorimetric analysis of cell surface

A and B antigens of stably transfected CHO cells CHO cells previously transfected with the rat FTB cDNA encoding an a1,2fucosyl-transferase were cotransfected with the rat A enzyme cDNA Stable transfectants were iso-lated and one of these clones was used for analysis The MT-450 cell line, a mammary carcinoma cell line from the w/Fu rat strain, was used as positive control The A and B antigens were detected using the 3–3 A and ED3 Mabs, respectively The Logs of fluo-rescence intensities are plotted against cell numbers Negative controls were performed in absence of primary antibody.

Fig 4 Enzymatic assays of CHO transfectant cell extracts The same

stable transfectants of CHO cells used for the cytofluorimetric analysis

of the A or B antigens expression (Fig 3) were used to assay the

enzyme activity Cell extracts were prepared as described in the

Materials and methods section Activities were determined using

2¢ fucosyllactose as acceptor substrate and either UDP-[ 14 C]galactose

or UDP-[ 14 C]N-acetylgalactosamine as donor substrates The product

of the reaction was separated on AG1-X8 anion exchange columns.

The background was determined in absence of acceptor substrate and

its value was deduced from the values obtained in the presence of the

acceptor Values of the specific activities are given in

pmolÆh)1Æmg protein)1of either [ 14 C]galactose or [ 14

C]N-acetylgalac-tosamine transferred.

Table 2 Tissue distribution of the Abo mRNA and of the A and B antigens in the BDIX rat Transcripts were detected by RT-PCR from total RNA extracts of various rat tissues using specific primers An indication of the intensity of the detected band is given, with +++ corresponding to the strongest signal and – to no detectable signal The

A and B antigens were detected by immunohistochemistry on frozen tissue sections using well characterized specific monoclonal antibodies.

In tissues positive for both A and B antigens, the cellular distribution

of the two antigens is not always identical (see text for details) The labeling by the anti-A antibody was the same on paraffin embedded sections, but no B antigen was detected on such sections ND ¼ not done.

Tissue mRNA A antigen B antigen

Oesophagus ++ ++ –

Small intestine – – –

Large intestine +++ +++ ++

Parotid gland ND +++ – Submaxillary gland +– ++ ++

Urinary bladder ++ – –

Seminal vesicle ND +++ – Thyroid gland ND +++ – Parathyroid gland ND +++ –

kidney, the urinary bladder, the uterus and the thymus A

weaker signal was obtained from the pancreas and very

weak, barely detectable, signals were visible from a

salivary gland, muscle and spleen The presence of this

mRNA was compared with that of the A and B

histo-blood group antigens in the various rat tissues The A

antigen appeared to have a wider distribution than the B

antigen as shown in Table 2 and Fig 5 There was a

general agreement between the presence of the A antigen

and that of the Abo gene mRNA Indeed, the A antigen

was expressed in the oesophagus, the stomach, the large

intestine (Fig 5a), the pancreas, the uterus (Fig 5g), the

seminal vesicle (Fig 5h) and the thymus (Fig 5e), while it

was essentially absent from the small intestine like the

mRNA, although a few glands were positive (Fig 5c)

However, a strong antigen expression was noted in the

submaxillary gland whereas the mRNA was barely

detected Conversely the mRNA was readily detected in

the kidney and the urinary bladder whereas the antigen

was not The B antigen was detected in fewer tissues than

the A antigen It could not be found in the tongue, the

oesophagus, the parotid gland, the uterus, the seminal

vesicle, the thyroid and parathyroid glands and the

thymus where the A antigen was present Inversely, the B antigen but not the A antigen could be detected in the kidney In this organ its distribution was limited to some tubules of the medulla (Fig 5f) In those tissues where both the A and B antigens were present, their cellular distribution differed For example, in the stomach, the A antigen was mainly present on cells of the neck area and on the parietal cells of the gastric glands whereas the

B antigen was present on the pits epithelial cells and the chief cells of the glands In the large intestine and the caecum, the A antigen was strongly expressed throughout the mucosa whereas presence of the B antigen was restricted to the surface epithelium (Fig 5a,b) In the submaxillary gland, the A antigen was expressed in the secretory cells and the B antigen in the duct cells The A and B antigens were only coexpressed in the pancreas acinar cells (Fig 5d) and in some sebaceous glands of the skin

D I S C U S S I O N

A rat cDNA homologous to the human ABO gene has been cloned It encodes a protein with a strong A transferase and

a small B transferase activity in vitro The human A enzyme also is known to possess a small B transferase activity [25] Yet, this activity is not comparable with that of cis A/B human alleles or of the mouse enzyme which are char-acterized by their ability to transfer galactose and N-acetylgalactosamine about equally well [16] That the rat enzyme described here is truly an A enzyme is confirmed

by the fact that upon transfection into CHO cells, A antigen was readily detected whereas B antigen was not In addition, there were BDIX rat tissues strongly expressing A antigen and no detectable B antigen

The ABO gene structure is conserved between rat, pig and man while one exon (exon 4) has been lost in the mouse

As the mouse cDNA encodes an active enzyme, it is clear that this exon is not required for enzyme activity This is not surprising as it corresponds to a part of the stem region of the enzyme Nevertheless, it could affect the type of structures used by the mouse enzyme as acceptor substrates, i.e glycolipids vs glycoproteins or the type of precursor Previous detailed studies aimed at defining the amino-acid residues that determine the A or B specificity of the human enzymes have been performed by site-directed mutagenesis From this work, it could be concluded that residues at positions 266 and 268 were critical and that the residue at position 235 was influencial [3] At the position correspond-ing to amino-acid 235 of the human sequence, the rat and pig A enzymes have a glycine residue like the human A and mouse cis A/B enzymes Of the three positions, 268 is considered the most important and, as noted above, a glycine at this position characterizes the A activity The rat sequence reported herein has a glycine at the position equivalent to human 268 in accordance with its A enzymatic activity Similar to the pig enzyme [26], but at variance with the human A enzymes, it has an alanine at position 266 In man, as well as in anthropoid apes, a leucine is present at this position, confirming that it is indeed less important than the amino-acid at position 268 in determining the A or B specificity of the enzyme

Phylogenetic analysis including sequences of the four known members of the ABO gene family confirmed that

Fig 5 Immunohistochemical analysis of the A and B antigens

expres-sion in BDIX rat tissues Frozen BDIX rat tissues sections were

incu-bated with an anti-A (a, c, e, g, h) or an anti-B (b, d, f) Mab and their

binding was detected as described in the
materials and methods

section In the large intestine, the A antigen is detected throughout the

mucosa (a) whereas the B antigen is restricted to the surface epithelium

(b) In the small intestine, the A and B antigens are absent except for a

few glands displaying A antigen (c) In the pancreas, both antigens are

present on the acinar cells as illustrated for the B antigen (d) In the

thymus some medullary epithelial cells express the A epitopes (e) In

the kidney medulla, some tubules are B positive (f) The A antigen is

present in the uterine epithelium (g) and the seminal vesicle (h).

these four genes can be clearly separated across various

species In addition, despite the small number of sequences

available for two of the genes, genetic distances among

species for each gene were quite similar, suggesting that they

evolved at about equal rates From this, one can speculate

that the four members of the ABO gene family are

submitted to the same kind of selective pressure

We generally observed a good concordance between the

presence of the mRNA and of the A antigen in tissues

Nevertheless, there were some discrepancies in tissues such

as the kidney and the urinary bladder where the mRNA was

easily detected but which did not express A or B epitopes

The presence of a glycosyltransferase mRNA and the

corresponding glycan structure may not always correlate as

the mRNA may not necessarily be translated or the

appropriate precursor glycans may not be available In the

present case, H antigen, the precursor of the A or B antigens

is not synthesized in the rat urinary bladder epithelium (data

not shown) Some H antigen should be present in the kidney

as B antigen was detected This is in accordance with the fact

that the rat FTA a1,2fucosyltransferase mRNA can be

detected in the BDIX rat kidney [23] One would thus expect

to find some A antigen in the rat kidney However, the A

enzyme mRNA may not be expressed in the same cells as

the FTA mRNA Further studies by in situ hybridization

are required to clarify this point

In humans, A or B antigens are present on glycolipids

as well as on glycoproteins and based on various types of

precursors [27] In rats, the A and B antigens have been

characterized on glycolipids [28–30] A polymorphism of

the expression of the A-active glycolipids has been found

among various strains of inbred rats [31] The A antigen is

also present on glycoproteins as it can be detected by

Western blotting on various bands from the transfected

cells (data not shown) or other A positive cell types from

BDIX rats [32,33] and as it has been characterized on

glycopeptides from Sprague–Dawley rats At variance, the

B antigen could not be found on such glycopeptides [34]

In addition, the A antigen is still strongly detected on

paraffin embedded rat tissue sections [35] Paraffin

embedding-deparaffination is known to remove

glyco-lipids, therefore the remaining antigenic activity should

correspond to glycoproteins, while frozen sections contain

both glycoproteins and glycolipids In the present study,

the B reactivity was only detected on frozen sections and

not on paraffin embedded sections, suggesting that it is

restrited to glycolipids The tissues that express most ABH

antigens in humans are quite similar to those that express

A antigen in the rat although there are some notable

differences Unlike humans, rats do not present these

antigens on erythrocytes, the vascular endothelium or the

epidermis There are also some regional differences Most

strains of rats, like the BDIX strain used in this study, do

not express the A antigen in the small intestine although

they express H antigen Inversely, A and H antigens are

present in the large intestine down to the rectum whereas

they are absent from the human rectal and distal colonic

epithelia [36] Another major difference between humans

and rats in terms of tissue expression of the A and B

antigens is that in man A and B epitopes are expressed in

the same cell types whereas in the rat they are not It had

been observed earlier that B, but not A, antigen is

developmentally expressed on rat cochlear hairy cells and

on olfactory cells [37,38] In the present study, we noticed that with the exception of the pancreas and the skin, in rat organs, the A and B antigens never codistributed at the cellular level It is unlikely that mAb ED3, the anti-B that

we used, detected the related pseudo-B or Gala1,3Gal epitope as the antibody did not react with the synthetic disaccharide (data not shown) and as this potentially cross-reactive epitope is expressed on the rat vascular endothelium to which mAb ED3 did not bind Neverthe-less, the possibility that mAb ED3 detects another B-like structure cannot be completely eliminated BDIX is a rat strain that has been generated in the 1940s and is inbred Therefore, in these animals, the B antigen cannot be synthesized by the enzyme product of an allele at the Abo locus It is to be expected that another gene encodes a galactosyltransferase with B histo-blood group activity The enzyme activity of the rat pseudo B a3galactosyl-transferase has not been reported as yet One possibility is that, unlike in other species, this enzyme could have the ability to transfer a galactose residue in a1,3 linkage to a1,2fucosylated precursor glycolipids It could also be that one of the two other known members of the family, namely the iGb3 synthetase and the Forssman synthetase, could have a dual specificity Alternatively, there could exist another gene in the rat genome encoding a B-like blood group transferase that would act exclusively on glycolipids These possibilities remain to be tested Availability of the rat A gene sequence will allow the search of genetic polymorphisms at the Abo locus in this species Comparisons of the tissue expression across species,

as well as of the sequences and genetic polymorphisms of various mammals should help understanding the biological significance of ABO antigens during evolution The results presented here should be useful in such future analyses

A C K N O W L E D G E M E N T S

The authors are grateful to Drs J Bara and A Martin for their generous gift of antibodies, to Dr M Cle´ment for help with the pictures,

to Mrs P Fichet and S Minaut for great animal care and to Pascale Van Vooren for excellent technical assistance They thank Dr J.-F Bouhours for helpul discussions The work was supported by grants from the Association for International Cancer Research (AICR), the Association pour la Recherche sur le Cancer (ARC) and the Fund for Scientific Medical Research (FRSM, Belgium) C S is a Research Director of the National Fund for Scientific Research (FNRS, Belgium).

R E F E R E N C E S

1 Watkins, W.M (1999) A half century of blood-group antigen research Some personal recollections Trends Glycosci Glycotech.

11, 391–411.

2 Oriol, R., Mollicone, R., Couillin, P., Dalix, A.M & Candelier, J.J (1992) Genetic regulation of the expression of ABH and Lewis antigens in tissues APMIS 100 (Suppl 27), 28–38.

3 Yamamoto, F (2000) Molecular genetics of ABO VoxSang 78, 91–103.

4 Yamamoto, F & McNeill, P.D (1996) Amino acid residue at codon 268 determines both activity and nucleotide-sugar donor substrate specificity of human histo-blood group A and B trans-ferases J Biol Chem 271, 10515–10520.

5 Glass, R.I., Holmgren, J., Haley, C.E., Khan, M.R., Svennerholm, A.M., Stoll, B.J., Belayet Hossain, K.M., Black, R.E., Yunus, M.

& Barua, D (1985) Predisposition for cholera of individuals with

O blood group Possible evolutionary significance Am J

Epi-demiol 121, 791–796.

6 Marionneau, S., Cailleau-Thomas, A., Rocher, J., Le

Moullac-Vaidye, B., Ruvoe¨n-clouet, N., Cle´ment, M & Le Pendu, J (2001)

ABH and Lewis histo-blood group antigens, a model for the

meaning of oligosaccharide diversity in the face of a changing

world Biochimie 83, 565–573.

7 Preece, A.F., Strahan, K.M., Devitt, J., Yamamoto, F.F &

Gustavson, K (2002) Expression of ABO or related antigenic

carbohydrates on viral envelopes leads to neutralization in the

presence of serum containing specific natural antibodies and

complement Blood 99, 2477–2482.

8 Yamamoto, F., Marken, J., Tsuji, T., White, T., Clausen, H &

Hakomori, S.I (1990) Cloning and characterization of DNA

complementary to human UDP-GalNAc:

Fuca1–2Gala1–3Gal-NAc transferase (histo-blood group A transferase) mRNA J Biol.

Chem 265, 1146–1151.

9 Thompson, J.D., Higgins, D.G & Gibson, T.J (1994) ClustalW:

improving the sensitivity of progressive multiple sequence

align-ment through sequence weighting, position specific gap penalties

and weight matrix choice Nucleic Acids Res 22, 4673–4680.

10 Szpirer, J Levan, G Tho¨rn, M & Szpirer, C (1984) Gene

map-ping in the rat by mouse-rat cell hybridization: synteny of the

albumin and alpha-foetoprotein genes and assignment to

chro-mosome 14 Cytogenet Cell Genet 38, 142–149.

11 Watanabe, T.K Bihoreau, M.T McCarthy, L.C Kiguwa, S.L.

Hishigaki, H Tsuji, A Browne, J Yamasaki, Y

Mizoguchi-Miyakita, A Oga, K Ono, T Okuno, S Kanemoto, N

Takah-ashi, E Tomita, K HayTakah-ashi, H Adachi, M Webber, C Davis, M.

Kiel, S Knights, C Smith, A Critcher, R Miller, J Thangarajah,

T Day, P.J.R Hudson, J.R Irie, Y Takagi, T Nakamura, Y.

Goodfellow, P.N Lathrop, G.M Tanigami, A & James, M.R.

(1999) A radiation hybrid map of the rat genome containing 5,255

markers Nat Genet 22, 27–36.

12 Goupille, C Marionneau, S Bureau, V Hallouin, F Meichenin,

M Rocher, J & Le Pendu, J (2000) a1,2fucosyltransferase

increases resistance to apoptosis of rat colon carcinoma cells.

Glycobiol 10, 375–382.

13 Bara, J Gautier, R Le Pendu, J & Oriol, R (1988)

Immuno-chemical characterization of mucins Polypeptide (M1) and

poly-saccharide (A and Leb) antigens Biochem J 254, 185–193.

14 Le Pendu, J Le Cabellec, M & Bara, J (1997)

Immunohisto-logical analysis of antibodies against ABH and other

glycocon-jugates in normal human pyloric and duodenal mucosae Transfus

Clin Biol 1, 41–46.

15 Yamamoto, F McNeill, P.D & Hakomori, S.I (1995) Genomic

organization of the human histo-blood group ABO genes.

Glycobiol 5, 51–58.

16 Yamamoto, M Lin, X.-H Kominato, Y Hata, Y Noda, R.

Saitou, N & Yamamoto, F (2001) Murine equivalent of the

human histo-blood group ABO gene is a cis-AB gene that encodes

a glycosyltransferase with both A and B transferase activity.

J Biol Chem 276, 13701–13708.

17 Joziasse, D.H Shaper, J.H Van den Eijnden, D.H Van Tunen,

A.J & Shaper, N.L (1989) Bovine alpha1–3galactosyltransferase:

isolation and characterization of a cDNA clone Identification of

homologous sequences in human genomic DNA J Biol Chem.

264, 14290–14297.

18 Xu, H & Storch, T., Yu.M Elliott, S.P & Haslam, D.B (1999)

Characterization of the human Forssman synthetase gene J Biol.

Chem 274, 29390–29398.

19 Keusch, J.J Manzella, S.M Nyame, K.A Cummings, R.D &

Baenziger, J.U (2000) Expression cloning of a new member of the

ABO blood group glycosyltransferases, iGb 3 synthase, that directs

the synthesis of isoglobo-glycosphingolipids J Biol Chem 275,

25308–25314.

20 Nilsson, S Helou, K Walentinsson, A Szpirer, C Nerman, O & Stahl, F (2001) Rat-mouse and rat-human comparative maps based on gene homology and high-resolution zoo-FISH Genomics

74, 287–298.

21 RATMAP Rat Genome Database, http://ratmap.Generalgu.se/

22 Henske, E.P., Ozelius, L., Anderson, M.A & Kwiatkowski, D.J (1992) A radiation-reduced hybrid cell line containing 5 Mb/17 cM

of human DNA from 9q34 Genomics 13, 841–844.

23 Bureau, V., Marionneau, S., Cailleau-Thomas, A., Le Moullac-Vaydie, B., Liehr, T & Le Pendu, J (2001) Comparison of the three rat GDP- L -fucose: b- D -galactoside 2-a- L -fucosyltransferases FTA, FTB and FTC Eur J Biochem 268, 1006–1019.

24 Sleeman, J.P., Kim, U., Le Pendu, J., Howells, N., Coquerelle, T., Ponta, H & Herrlich, P (1999) Inhibition of MT-450 rat mam-mary tumour growth by antibodies recognising subtypes of blood group antigen B Oncogene 18, 4485–4494.

25 Yates, A.D & Watkins, W.M (1982) The biosynthesis of blood group B determinants by the blood group A gene specified a-3-N-acetylgalactosaminyltransferase Biochem Biophys Res Com-mun 109, 958–965.

26 Meijerink, E., Neuenschwander, S., Dinter, A., Yerle, M., Stranzinger, G & Vogeli, P (2001) Isolation of a porcine UDP-GalNAc transferase cDNA mapping to the region of the blood group EAA locus on pig chromosome 1 Anim Genet 32, 132–138.

27 Clausen, H & Hakomori, S.I (1989) ABH and related histo-blood group antigens; immunochemical differences in carrier iso-types and their distribution VoxSang 56, 1–20.

28 Breimer, M.E., Hansson, G.C., Karlsson, K.A & Leffler, H (1982) Glycosphingolipids of rat tissues Different composition of epithelial and nonepithelial cells of small intestine J Biol Chem.

257, 557–568.

29 Hansson, G.C., Bouhours, J.F & Angstro¨m, J (1987) Char-acterization of neutral blood group B-active glycosphingolipids of rat gastric mucosa J Biol Chem 262, 13135–13141.

30 Hansson, G.C (1983) The structure of two blood group A-active glycosphingolipids with 12 sugars and a branched chain present

in the epithelial cells of rat small intestine J Biol Chem 258, 9612–9615.

31 Bouhours, D., Hansson, G.C & Bouhours, J.F (1995) Structure and genetic polymorphism of blood group A-active glycoshingo-lipids of the rat large intestine Biochem Biophys Acta 1255, 131–140.

32 Laferte, S., Prokopishyn, N.L., Moyana, T & Bird, R.P (1995) Monoclonal antibody recognizing a determinant on type 2 chain blood group A and B oligosaccharides detects oncodevelopmental changes in azoxymethane-induced rat colon tumors and human colon cancer cell lines J Cell Biochem 57, 101–119.

33 Me´noret, A., Otry, C., Labarrie`re, N., Breimer, M.E., Piller, F., Meflah, K & Le Pendu, J (1995) The expression of carbohydrate blood group antigens correlates with heat resistance J Cell Sci.

108, 1691–1701.

34 Norrsell, H., Bengtsson, J., Jovall, P.A & Hansson, G.C (1992) N-linked glycopeptides with blood group determinants lacking neuraminic acid from the epithelial cells of rat small and large intestine Eur J Biochem 203, 285–293.

35 Hallouin, F., Goupille, C., Le Cabellec, M., Bara, J & Le Pendu,

J (1997) Expression of A and H blood-group and of CD44 anti-gens during chemical rat colonic carcinogenesis Glycoconj J 14, 801–808.

36 Ravn, V & Dabelsteen, E (2000) Tissue distribution of histo-blood group antigens APMIS 108, 1–28.

37 Gil-Loyzaga, P., Pujol, R., Mollicone, R., Dalix, A.M & Oriol, R (1989) Appearance of B and H blood group antigens in the developing cochlear hair cells Cell Tissue Res 257, 17–21.

38 Astic, L., Le Pendu, J., Mollicone, R., Saucher, D & Oriol, R (1989) Cellular expression of H and B antigens in the rat olfactory system during development J Comp Neurol 289, 386–394.