Molecular cloning and characterization of CFT1, a developmentally regulated avian alpha(1,3)-fucosyltransferase gene.

Although coordinate expression of carbohydrate epitopes during development is well described, mechanisms which regulate this expression remain largely unknown. In this study we demonstrate that developing chicken B cells express the LewisX terminal oligosaccharide structure in a stage-specific manner. To examine regulation of this expression, we have cloned and expressed the chicken α(1,3)-fucosyltransferase gene involved in LewisX biosynthesis, naming it chicken fucosyltransferase 1 (CFT1). CFT1 is characterized by a single long open reading frame of 356 amino acids encoding a type II transmembrane glycoprotein. The domain structure and predicted amino acid sequence are highly conserved between CFT1 and mammalian FucTIV genes (52.8% and 46.3% identity to mouse and human respectively). In vitro CFT1 fucosyltransferase activity utilizes LacNAc > 3′sialyl-LacNAc acceptors with almost no utilization of other neutral type II (lactose, 2-fucosyllactose), or type I (lacto-N-biose I) acceptors. CFT1-transfected cells make cell surface LewisX (COS-7) and LewisX + VIM-2 structures (Chinese hamster ovary). CFT1 gene expression is tissue-specific and includes embryonic thymus and bursa. Furthermore, expression of the CFT1 gene and cell surface LewisX structures are closely linked during B cell development. These findings reveal the evolutionary conservation between nonmammalian and mammalian α(1,3)-fucosyltransferase genes and demonstrate a role for fucosyltransferase gene regulation in the developmental expression of oligosaccharide structures.

a ligand for the selectin receptors and directly mediates cell to cell adhesion (32). It has been proposed that Lewis X has similar developmentally important adhesive function (33,34).
We and others have demonstrated that Lewis X (Le X ) 1 and sialyl Lewis X (sLe X ) are coordinately expressed during chicken B lymphocyte development (35)(36)(37)(38). Unlike mammals, avian B cell development occurs synchronously within a single primary lymphoid organ over a single discrete period of time, simplifying the analysis of stage-specific expression (39). Avian B cell progenitors exist only during embryonic life and home from the dorsal mesenchyme to the primary B lymphoid organ, the bursa of Fabricius, in a single wave starting from embryonic day 7.5 (e7.5) through e15. The single wave of homing results in the synchronous progression of bursal lymphocytes through development. This is followed by a stage of rapid cellular proliferation and immunoglobulin gene diversification (starting at e15), and then by emigration of mature B cells to the periphery (starting at hatch (e21)). After hatch, the bursa progressively becomes a secondary lymphoid organ until it involutes at puberty (Ͼ12 weeks).
We have reported that chicken B cell progenitors express sLe X but not Le X (35). sLe X mediates adhesion of B cell progenitors to bursal stromal structures in vitro and may be involved in progenitor cell homing to or within (to nascent follicles) the bursa. As the lymphoid progenitors begin to proliferate and diversify the immunoglobulin gene locus, their surface glycosylation switches from sLe X -positive/Le X -negative to sLe X -negative/Le X -positive "bright." As bursal lymphocytes mature and return to a resting state, Le X expression is down-regulated (Le X -positive "intermediate"). B cells that have emigrated from the bursa to peripheral secondary lymphoid organs are also Le X -positive intermediate.
The coordinate expression and potential biological functions of sLe X and Le X in developing B cells suggest their construction is directly regulated. Transcriptional regulation of glycosyltransferase genes (40,41) or enzymatic competition for substrate have been proposed as general mechanisms (42). A key regulatory enzyme in B cell sLe X and Le X biosynthesis is likely to be the ␣(1,3)-fucosyltransferase which adds the final fucose residue (43), as is the case for the ␣(1,3)-fucosyltransferase involved in construction of the selectin binding epitope (44 -48). Molecular cloning has identified five human ␣(1,3)-fucosyltransferases genes (FucTIII through FucTVII) with homolo-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  gous sequence but differing activity, acceptor substrate requirements and tissue distribution (40, 45-47, 49 -52). We have used low stringency hybridization to the human FucTIV gene to clone out an avian ␣(1,3)-fucosyltransferase gene, naming it chicken fucosyltransferase 1 (CFT1). CFT1 shares a 46.3% overall amino acid identity to human FucTIV. Biochemical and transfection studies demonstrated that the CFT gene product can construct Le X and Vim2 determinants. CFT1 mRNA is expressed in specific tissues, including the embryonic bursa and thymus. Expression of the CFT1 gene during bursal lymphocyte development closely correlates with surface Le X expression, providing evidence for developmental regulation of a terminal oligosaccharide by ␣(1,3)-fucosyltransferase gene expression.

EXPERIMENTAL PROCEDURES
Animals and Cell Lines-Animal care, preparation, and cell isolation have been previously described (35). Briefly, chickens used in these experiments were Hyline SC birds, an F 1 cross between two inbred B 2 chicken strains (Hyline International, Dallas Center, IA). Fertile eggs were incubated at 39°C and embryos staged by time of incubation. Bursal and splenic mononuclear cells were isolated as a single-cell suspension and purified for mononuclear cells by centrifugation over Ficoll-Hypaque (Pharmacia, Uppsala, Sweden).
Flow Cytofluorometric Analysis-Cells were stained as described previously (35). Briefly, 5 ϫ 10 5 cells/sample were resuspended in 100 l of Dulbecco's modified Eagle's medium ϩ 2% fetal calf serum ϩ 10 mM HEPES ϩ 0.1% azide (staining medium) with the appropriate dilution of antibody and incubated on ice for 30 min. Cells were washed with 2 ml of staining medium twice and stained with secondary antibody if necessary. Cells were again washed twice with staining medium and analyzed immediately using a Becton-Dickinson FACScan.
CFT1 Cloning-Isolation of clones from a genomic phage library has been described previously (53). Briefly, the 1.3-kb NotI-NheI fragment of the human ␣(1,3)-fucosyltransferase gene FucTIV (51) was labeled by random hexamer priming and used to probe nitrocellulose lifts from a primary plating of a chicken genomic library (Fix, the generous gift of K. Conklin, University of Minnesota). Lifts were prehybridized and hybridized as described previously (53). After hybridization, lifts were washed for 10 min twice at room temperature in 2 ϫ SSC, 0.1% SDS and then once for 20 min in 0.2 ϫ SSC, 0.1% SDS at 42°C. The lifts were then exposed to x-ray film for 12 h at Ϫ70°C with intensifying screens. Positively hybridizing plaques were isolated by sequential dilution purification. An approximately 15-kb XbaI fragment was subcloned into Bluescript SK (Stratagene, La Jolla, CA). Subsequently, a 2.6-kb PstI fragment and a 4.4-kb SacI fragment, which both cross-hybridized with the human probe were subcloned. Southern blot (normal stringency washes: 10 min twice at room temperature in 2 ϫ SSC, 0.1% SDS; 20 min twice in 0.2 ϫ SSC, 0.1% SDS at 56°C) and sequence analysis of the CFT1 subclones were undertaken as described FIG. 1. A, flow cytometric analysis of Le X expression on developing chicken B cells. Lymphocytes were isolated from the bursa of Fabricius of 13-day-old embryos (e13), 18-day-old embryos (e18), hatch (e21), and 2, 7, and 10 weeks post-hatch birds. Bursal lymphocytes were then stained with anti-Le X mAb SSEA-1 3 goat anti-mouse IgM FITC and analyzed by flow cytometry. The data are plotted as cell size (x axis, as a measure of proliferation) versus Le X staining (y axis, log 10 fluorescence). B, comparison of Le X expression on bursal and splenic lymphocytes 2 weeks posthatch. Lymphocytes from the bursa and spleen of a single 2-week-old bird were isolated and stained with anti-Le X mAb. Both sets of cells analyzed with the same flow cytometer PMT voltage settings to allow for cell size and fluorescent intensity comparison. Splenic lymphocytes were also analyzed by two-color staining for Le X (stained with CD15 FITC mAb) versus surface immunoglobulin (Ig, stained with rabbit anti-chicken Ig phycoerythrin) expression. Stained cells were analyzed on a Becton Dickinson FACScan.
previously (53). DNA and protein sequence analysis was performed using the LaserGene analysis package (DNASTAR Inc., Madison, WI).
Enzymatic Assays-Fucosyltransferase assays from transfected cell lysates were performed as described previously (51). Briefly, cell extracts containing 1% Triton X-100 were prepared from transfected COS-7 cells. Preliminary assays with 3 mM GDP-[ 14 C]fucose and Nacetyllactosamine were used to determine the amount of cell extract necessary to yield a linear reaction, typically 10 g. Fucosyltransferase assays were performed in a total volume of 20 l and contained 50 mM sodium cacodylate (pH 6.2), 5 mM ATP, 10 mM fucose, 20 mM MnCl 2 , 3 mM GDP-[ 14 C]fucose, and 10 g of cell extract. Acceptor substrates N-acetyllactosamine, lactose, 2Ј fucosyllactose, lacto-N-biose I, and 3Јsialyl-N-acetyllactosamine were added to a final concentration of 20 mM. The reaction products were separated from GDP-[ 14 C]-fucose by column chromatography through Dowex 1-X2-400 (formate form) for neutral acceptor substrates or Dowex 1-X2-200 (phosphate form) for sialylated acceptor substrates. The reaction products were collected in the column flow-through fraction and quantified by liquid scintillation counting.
CFT1 Transfection-The 4.4-kb SacI fragment of CFT1 was subcloned into the mammalian expression vector pcDNA1 (Invitrogen, San Diego) in both the sense and antisense orientations. These constructs were then transfected into COS-7 cells using the DEAE-dextran procedure as described previously (49). 72 h after transfection the cells were examined for Lewis X and sialyl-Lewis X surface expression by antibody staining and flow cytometric analysis.
CHO cells were stably transfected with control vector pCDM8, fucosyltransferase vectors pcDNA1/CFT1, pCDM8/hFTIII (human FucTIII; Ref. 49), or pCDM8/hFTIV (human FucTIV; Ref. 51) utilizing the calcium phosphate procedure (51). A sample of each fucosyltransferasetransfected population was stained with ␣-VIM-2 antibody and the positive staining population was sorted by flow cytometry (FACStar, Becton-Dickinson). After several passages in culture the sorted transfectants were analyzed by flow cytometry for the expression of fucosylated epitopes.
Northern Analysis-RNA was extracted from bursal lymphocytes of specific developmental time points and whole organs from day 18 chick embryos and subjected to Northern blot analysis as described previously (53). Individual RNA samples were equalized to one another through visualization on ethidium bromide-agarose gels and serial dilution. These blots were probed initially with the 2.6-kb PstI fragment of CFT1 labeled by nick translation. The blots were then probed with the ␤-actin gene to confirm equal loading and integrity of the mRNA.

Lewis X Expression during Chicken B Lymphocyte Development-
The stage-specific expression of Le X during bursal lymphocyte development is shown in Fig. 1A. By embryonic day 13 (e13), B cell progenitors are homing to and colonizing the bursal rudiment. As can be seen, these cells are Le X -negative (sLe X -positive; see Ref. 35) and are not proliferating based on cell size. Starting at e15 and seen clearly by e18, these cells begin proliferating and simultaneously up-regulate expression of Le X while losing sLe X expression (see Ref. 35). Based on the frequency of gene conversion events, cells undergoing active Ig gene diversification are contained in the proliferating Le X bright population (36). These cells are also all surface Ig (sIg)positive (data not shown) and represent the developing immature bursal lymphocyte population. The Le X bright B cell phenotype was never found outside of the bursa, providing further evidence these are developing, immature lymphocytes. The actively cycling Le X bright population continues to expand until hatch (e21) where they are the majority of cells in the bursa. At this point a second population of noncycling, sIg ϩ/Le X intermediate cells begins to develop. These cells have the same phenotype as mature B cells found in the periphery (see below). Over the next 12 weeks posthatch, the Le X intermediate population becomes progressively larger with a reciprocal diminution of Le X bright population, consistent with the conversion of the bursa from a primary to secondary B cell organ.
Mature B cells that have emigrated from the bursa into the periphery can be found in the spleen. As shown in Fig. 1B, at 2 weeks posthatch a distinct noncycling Le X intermediate popu-lation can be detected in both the bursa and the spleen. Twocolor flow cytometric analysis demonstrates that all the surface Ig-positive cells in the spleen are of the Le X intermediate phenotype. Splenic B cells also express low levels of sLe X (data not show). These data clearly demonstrate stage-specific Le X expression on developing B cells: Le X "negative" progenitor 3 Le X bright immature bursal lymphocyte 3 Le X intermediate mature lymphocyte.
Molecular Cloning of the Chicken ␣(1,3)-Fucosyltransferase Gene, CFT1-The expression pattern of Le X (and sLe X ) during B cell ontogeny is consistent with direct regulation of their biosynthesis, potentially at the genetic level. To examine this possibility, we used low stringency hybridization to the human FucTIV cDNA (51) to clone potential genomic chicken fucosyltransferase genes. Since the nucleotide sequences of all the ␣(1,3)-fucosyltransferases cloned out to date (all mammalian) are well conserved (47), the same is likely to be true across a wider evolutionary divergence.
Four identical genomic clones were recovered ( Fig. 2A). The solid rectangle indicates the region that cross-hybridizes to human FucTIV. The 2.6-kb PstI fragment containing this region was used to probe a chicken genomic Southern blot under normal stringency hybridization conditions (Fig. 2B), demonstrating the putative fucosyltransferase gene is single-copy. Lower stringency hybridization revealed the possibility of related genes (data not shown).
Sequence analysis of the 2.6-kb PstI fragment (Fig. 3A)  been reported for other ␣(1,3)-fucosyltransferases (45,54). There are three in-frame AUG start codons but only one conforms to the Kozak consensus sequence. Utilizing this start site, the reading frame predicts a 356-amino acid protein of 41.4 kDa. A 24-amino acid N-terminal hydrophilic sequence precedes the putative hydrophobic transmembrane domain (29 amino acids flanked by basic residues) (Fig. 3B). The C-terminal domain contains two N-linked glycosylation sites (aa 81 and 150) colinear to the predicted N-linked glycosylation sites in human and murine FucTIV. A consensus polyadenylation signal is found 1080 base pairs 3Ј of the termination codon (base pair 2320). Hydrophobicity analysis of the CFT1 polypeptide predicts a type II membrane glycoprotein (Fig. 3C), consistent with all previously cloned ␣(1,3)-fucosyltransferases (40, 45-47, 49 -52, 55).
Despite the evolutionary distance, the predicted amino acid sequence is well conserved with a 46.3% overall identity to human FucTIV and 52.8% to murine FucTIV (Fig. 4). Homology to the other mammalian ␣(1,3)-fucosyltransferase genes ranges from 38% to 41%. Comparison of the more highly conserved C-terminal catalytic domains (51,55) increases the identity to 63.4% (human FucTIV) and 61.8% (murine FucTIV), respectively. As expected, the areas of least similarity are in the N termini, including the putative cytoplasmic domain (CFT1 aa 1-21, where mouse and human also have low homology) and transmembrane domain (aa 22-51). There is also a 32-amino acid gap (versus murine FucTIV) at CFT1 residue 116 that spans a poorly conserved region between human and murine FucTIV (35% identity versus 75% overall) and another 6-aa gap at CFT1 residue 168. Surprisingly, nearly identical gaps are reported in human FucTIII and FucTVII when aligned against human FucTIV (46), suggesting insertional events during the evolution of mammalian FucTIV away from the ancestral ␣(1,3)-fucosyltransferase gene.
Biochemical Characterization of the CFT1 Gene Product-The high degree of sequence identity of CFT1 to human and murine FucTIV strongly suggests CFT1 encodes a fucosyltransferase. To test this, the 4.4-kb SacI fragment was sub-  cloned (sense and antisense) into the eukaryotic expression vector pcDNA1 and transfected into COS-7 cells. Lysates from cells transfected with the vector alone did not display any fucosyltransferase activity (data not shown). In contrast, fucosyltransferase activity was readily detected in lysates from cells transfected with the CFT1 construct (Table I). The neutral type II disaccharide substrate N-acetyllactosamine (LacNAc) was most efficiently utilized whereas other neutral type II substrates (lactose, 2-fucosyllactose) were not. The type I acceptor lacto-N-biose I was not used at all, suggesting that fucose transfer occurs to an open 3-position (N-acetyllactosamine) but not to an open 4-position (lacto-N-biose I) on the accepting glucose. CFT1 activity was also capable of utilizing the sialylated acceptor 3Ј-sialyl-LacNAc, although less efficiently than LacNAc. CFT1's spectrum of acceptor specificity thus most closely resembles that of murine FucTIV (55).
The ability of the CFT1 gene product to direct synthesis of cell surface oligosaccharide groups was tested in transfected COS-7 and CHO cells. Both cell lines have the appropriate glycosylation status for these experiments as they have no detectable ␣(1,3)or ␣(1,4)-fucosyltransferase activity but do express the neutral and sialylated type II precursors necessary to construct Lewis X , Lewis A , sialyl-Lewis X , sialyl-Lewis A , and VIM-2 determinants (51). Additionally, it has been noted that synthesis of sialylated structures may be more readily detected in CHO versus COS cells (51,52). Transfection of the CFT sense construct into COS-7 cells (Fig. 5, solid line) directed the synthesis of the Lewis X epitope, whereas transfection with the antisense construct or vector alone (dotted line) had no effect. No expression of Lewis A , sialyl-Lewis X , sialyl-Lewis A , and VIM-2 was detected, similar to what has been reported for both murine and human FucTIV (45,50,51,55). However, CFT1 transfection into CHO cells demonstrated significant expression of the sialylated fucosylated epitope VIM-2 (NeuNAc␣2 3 3Gal␤1 3 4GlcNAc␤1 3 3Gal␤1 3 4[Fuc␣(1 3 3)]GlcNAc-R) with modest expression of Le X (Table II). This is more evident following flow cytometric sorting for VIM-2-positive cells. Like COS-7, CFT1-transfected CHO cells do not express Lewis a , sialyl-Lewis X , or sialyl-Lewis A .
Tissue-and Developmental Stage-specific Expression of the CFT1 Gene-The preceding findings supports the conclusion that CFT1 encodes an ␣(1,3)-fucosyltransferase homologous to human and murine FucTIV. Northern blot analysis was used to determine any correlation between CFT1 gene expression and B cell expression of Le X . At day 18 of embryogenesis, CFT1 is expressed as a single ϳ4-kb mRNA transcript in the brain, eye, gizzard, thymus, bursa, and spleen (Fig. 6A). CFT1 expression in the brain and eye may be involved neuronal development (9, 56 -58) and expression in the gizzard may be similar to murine FucTIV expression in gastric/colonic epithelium (55). At e18, the chicken spleen is primarily a myeloid organ (data not shown) and CFT1 expression is consistent with "myeloid" expression of murine and human FucTIV (45,51,55). Finally, CFT1 expression in developing T and B lymphocytes is demonstrated by detection of CFT1 transcripts in the bursa and thymus.
Bursal lymphocyte mRNA from different developmental time points were analyzed to determine any linkage between Lewis X and CFT1 gene expression. At embryonic day 18, the plurality of bursal lymphocytes are Le X bright (Fig. 1A) and CFT1 gene expression is readily detected (Fig. 6B). At hatch (e21), almost all the bursal lymphocytes are Le X bright, correlating with an increase in CFT1 gene expression. By 12 weeks posthatch, significant down-regulation of Le X on mature lymphocytes is reflected by a similar down-regulation in CFT1 expression (message is detected following prolonged exposure). Similarly, fractionation of bursal lymphocytes by cell size (via counterflow centrifugation) demonstrated highest CFT1 expression in the larger, actively proliferating cells (data not shown). These finding supports direct regulation of Le X expression during bursal lymphocyte development by CFT1 gene expression.

DISCUSSION
In this study we have demonstrated that the Lewis X terminal oligosaccharide group is expressed in a stage-specific manner during chicken B cell development. To address how this expression is regulated, we have cloned the chicken ␣(1,3)fucosyltransferase gene CFT1 through low stringency hybridization to the human FucTIV gene. A single long open reading frame predicts a type II transmembrane protein of 356 amino acids. The predicted CFT1 protein is very similar to murine and human FucTIV in both amino acid sequence and domain structure. In vitro biochemical characterization of CFT1 fucosyltransferase activity reveals an acceptor preference for Lac-NAc Ͼ 3Ј-sialyl-LacNAc with almost no utilization of other neutral type II (lactose, 2-fucosyllactose), or type I (lacto-Nbiose I) acceptors. Transfection of a CFT1 expression construct into COS-7 and CHO cells results in the synthesis of cell  surface Lewis X and VIM-2 structures. Northern blot analysis demonstrates tissue-specific CFT1 gene expression, including in the embryonic thymus and bursa. Further analysis establishes that the expression of the CFT1 gene and cell surface Le X structures are linked during B cell development. Together, these findings provide evidence that coordinate expression of oligosaccharide structures during development may be regulated by expression of key glycosyltransferase genes. Coordinate expression of the Lewis X carbohydrate structure during development has been well described during hematopoiesis (20), neural development (9, 56 -58), and mouse embryogenesis (17,59). We have also demonstrated stage-specific Le X expression during B cell development (this report and Refs. 35 and 36). It has been proposed that Le X has a direct biological role by mediating cell-cell adhesion through homotypic Le X -Le X interactions (33,34). How the expression of these structures is regulated is unknown, although transcriptional regulation of glycosyltransferase genes (40) and competition between enzymes for substrate (42) have been proposed as general mechanisms. Transcriptional control of glycosyltransferase activity has been suggested during thymocyte development (60) and E-selectin ligand synthesis in T cells (48).
We hypothesized that genetic regulation of the last enzyme in the biosynthetic pathway, the ␣(1,3)-fucosyltransferase (41), plays a role in controlling bursal lymphocyte Le X expression. The predominance of Le X without Lewis A or sialylated structures suggested that this fucosyltransferase is the chicken homologue of mammalian FucTIV (45, 50, 51, 55). Low stringency screening with human FucTIV recovered four identical chicken genomic clones, although Southern analysis suggests the presence of other genes related to this single copy locus.
CFT1 is the name we have given the single long open reading frame contained in the cross-hybridizing 4.4-kb SacI fragment. CFT1 has all the canonical sequence/structural motifs described for the cloned mammalian ␣(1,3)-fucosyltransferases (45,54), including high GC content (particularly at the 5Ј end), a monocistronic coding region predicting a type II transmembrane glycoprotein and two N-linked glycosylation sites, which map colinearly to sites in human and mouse FucTIV. CFT1 amino acid sequence is surprisingly homologous to human and mouse FucTIV (particularly the C-terminal catalytic domain) given the evolutionary distance between species. Despite this similarity CFT1 does not contain the unique 32-aa insertion found in FucTIV when compared to the other mammalian FucT genes (46). Phylogenetic analysis indicates that FucTVII diverged first from the ancestral ␣(1,3)-fucosyltransferase gene, followed by a divergence of CFT1 from the chromosome 19 (FucTIII, FucTV, FucTVI) subfamily, and finally divergence of CFT1 from mammalian FucTIV (40, 45-47, 49 -52).
There is strong evidence that developmentally regulated CFT1 gene expression controls the construction of bursal lymphocyte Le X determinants. Under hybridization conditions that only detect the CFT1 gene on Southern analysis, CFT1 mRNA is readily found in the embryonic bursa when there is high B cell Le X expression. Purified bursal lymphocytes from different developmental time points demonstrate a strong correlation between CFT1 gene expression and Le X surface expression. Direct regulation of the Le X epitope in turn suggests a direct biological function.
Whether CFT1 is also involved in the stage-specific expression of the sLe X determinant during chicken B cell development is less clear. Human FucTIV has been variably reported to make or not make sLe X (50,51,61). CFT1 does not make sLe X when transfected into COS-7 or CHO cells, only VIM-2. Interestingly, VIM-2 expression was never detected in chicken bursa, spleen, or bone marrow cells during development, suggesting additional complexities not reflected in transfected cell lines. However, developing myeloid cells in the embryonic spleen are sLe X bright/Le X -negative and have high CFT1 gene expression by Northern analysis. 2 It is unlikely this actually represents cross-hybridization to splenic FucTVII homologue given the hybridization conditions used. It seems more likely that sLe X is created through the addition of a 2,3-linked sialic acid to Le X by a sialyltransferase, which itself may be developmentally regulated. This is supported by the observation that stable transfection of CFT1 into the bursal B stem cell line DT40, which only expresses sLe X , does not result in new Le X or Vim-2 expression. 3 Other factors (specific substrates, activity of other enzymes, etc.) may also play a role.
Cloning of CFT1 reveals remarkable conservation of structure between nonmammalian and mammalian ␣(1,3)-fucosyltransferase genes despite a Ͼ300 million-year evolutionary divergence. The coupled expression of CFT1 mRNA and bursal lymphocyte surface Le X further supports the model of oligosac-  6. A, Northern blot analysis of CFT1 tissue expression. Total RNA was extracted from various tissues of day 18 embryos, separated on a formaldehyde-agarose gel, and transferred to nitrocellulose membrane. The membrane was then probed with 32 P-labeled 2.6-kb PstI CFT1 gene fragment, washed under normal stringency conditions, and exposed. B, expression of CFT1 mRNA during B cell development. Bursal lymphocytes were isolated from day 18 embryos (day 18), hatch (embryonic day 21), and 12 week posthatch birds and total RNA extracted. Ethidium bromide visualization and serial dilution were used to equalize mRNA concentration (Total RNA). Equal amounts of mRNA were separated on a formaldehyde-agarose gel, transferred to a nitrocellulose membrane, and serially probed with 32 P-labeled CFT1 (2.6-kb PstI fragment) and ␤-actin cDNA (to demonstrate mRNA integrity) gene fragments. charide expression regulation through genetic control of key glycosyltransferase genes. Given the complexity of their biosynthesis, this undoubtedly represents only one of many regulatory mechanisms involved in coordinate expression of oligosaccharides during development.