Cloning and characterization of a bovine genomic fragment homologous to epidermal growth factor genes

Abstract

Epidermal growth factor (EGF) is a potent mitogen for a variety of cell types. The 53-amino acid mature EGF protein is encoded by sequences in exons 20 and 21 of a gene spanning over 110 kb. In this study, we report the cloning and characterization of 7.5 kb of bovine genomic sequence homologous to exon 19 through 21 from EGF genes from other mammalian species. The cloned gene fragment had an unusual sequence composition in the form of an in-frame TGA codon in the coding sequence. The sequence was expressed at low levels in kidney tissue and the corresponding cDNA contained the TGA codon. The level of similarity between the bovine exonic sequence and the human, porcine, murine, feline, and canine corresponding sequences varied from 64% to 73%; however, when only sequences encoding the mature EGF protein were compared, the level of similarity between the bovine sequence and the sequence from these species was 59% to 66%. The sequence similarity of the deduced mature protein was lower (34% to 39%) than the sequence similarity of the deduced propeptide. Although the cloned sequences could originate from a bovine EGF pseudogene, the possibility exists that they originate from the functional EGF gene. An as yet unidentified mechanism to by-pass the stop codon would allow the synthesis of a functional EGF protein. Alternatively, the cloned sequence could originate from an EGF-like gene.

Introduction

Epidermal growth factor (EGF) is a short peptide of 53 amino acids that is derived from a much larger precursor protein of approximately 1200 amino acids. In humans, the precursor protein is encoded by a gene spanning 110 kb of DNA where 24 short exons are separated by introns of various sizes. The mature human EGF protein is produced by excision of amino acid residues 977 to 1029 of the precursor protein, which are encoded by 95 nucleotides at the 3’ end of exon 20 and 64 nucleotides at the 5’ end of exon 21 (1).

Epidermal growth factor is a potent regulator of cell growth and0 differentiation in several tissues (2) and has been implicated in the regulation of several biological processes, such as neoplasia (3), wound repair (4), gastrointestinal maturation (5), and nutrient absorption (6,7). Moreover, the potential of EGF as a therapeutic agent to treat gastrointestinal injury has been demonstrated (8), and oral EGF therapy may also inhibit colonization of the intestinal tract by pathogens (9).

Despite the diverse biological effects of EGF and its potential application as a treatment for enhancing intestinal recovery following infection in young farm animals (10), bovine EGF complementary or genomic DNA sequences have not yet been cloned. Cloning and sequencing of sequences encoding bovine EGF would provide homologous probes, primers, and antibodies for the study of EGF gene expression in the bovine. Moreover, it would be desirable to use bovine EGF to evaluate the effects of the growth factor on intestinal health and function, for example in calves. The sequences from the human (GenBank accession number AC 004050) and mouse EGF genes (GenBank accession number AC098732) are known. As well, cDNA sequences encoding the precursor protein have been cloned from mouse, rat, pig, human (1,11–13), dog, cat (GenBank accession numbers AB049597 and AB050947, respectively), and horse (sequence encoding the mature protein only, 14). The level of similarity in the DNA sequences encoding the mature EGF protein and the deduced protein sequences varies between these species from 70% to 92% and 58% to 94%, respectively. In this paper, we report the cloning and characterization of bovine genomic sequences homologous to exons 19 through 21 from EGF genes from other species. The DNA and deduced protein sequences are compared to EGF sequences from other mammalian species.

Materials and methods

Genomic DNA isolation

Genomic DNA was isolated from bovine (Bos taurus) blood using a salt extraction method (15). Briefly, nucleated cells obtained from coagulated blood were resuspended in 3 mL of nuclei lysis buffer (10 mM Tris-HCl, 400 mM NaCl, and 2 mM Na₂ EDTA, pH 8.2). The lysates were incubated overnight with 200 μL of 10% sodium dodecyl sulfate (SDS) and 500 μL of protease K solution (1 mg proteinase K [Invitrogen, Carlsbad, California, USA] in 1% SDS and 2 mM Na₂ EDTA). After digestion, 1 mL of 6 M NaCl was added, the tubes were shaken vigorously for 15 s and then centrifuged at 1000 × g for 15 min. The supernatants were transferred to clean tubes, 2 volumes of absolute ethanol was added to precipitate the DNA. The DNA strands were removed with a pipette and transferred to Tris buffer.

Amplification, cloning, and sequencing of a 1.5 kb bovine genomic fragment homologous to EGF genes

Approximately 300 ng of bovine genomic DNA was used as template in a polymerase chain reaction (PCR) containing 200 nM each of primers #1 and #2 (see Table I for all primer sequences), 100 μM deoxynucleotide triphosphates (dNTPs) (Roche Molecular Biochemicals, Montreal, Quebec), and 1.5 U Taq DNA polymerase (Life Technologies, Rockville, Maryland, USA) in 1 × reaction buffer containing 1.5 mM MgCl₂. After an initial denaturation step of 3 min at 96°C, the PCR was carried out for 35 cycles of a denaturing step of 30 s at 94°C, followed by an annealing step of 15 s at 57°C, an extension step of 2 min at 72°C, and a final extension step of 5 min at 72°C. The PCR product was run in a gel and a band of approximately 1.5 kb was cut out, purified and used as template for a second PCR in the same conditions as above. The 1.5 kb PCR product was purified (QIAquick gel extraction kit; Qiagen, Valencia, California, USA) and ligated to a cloning vector (pGEMT-EASY TA; Promega Corporation, Madison, Wisconsin, USA). The construct was transformed into Escherichia coli competent cells (MAX Efficiency DH5α; Life Technologies).

Table I.

Primers used to amplify bovine epidermal growth factor (EGF) sequences

		Location
Primers	Sequence	Position	Accession number
#1	5′ GAC ACA/C/G/T TGC/T ACA AAT ACA/C/G/T GAG GG 3′	Nucleotides 3215–3237 of human cDNA	(X04571)
#2	5′ CC AT/CC A/GTG GAG GCA GTA CCC A/GTC 3′	Nucleotides 3377–3399 of human cDNA	(X0457)
#3	5′ GACCATTCTCTTACTTGCCT 3′	Nucleotides 1160–1179 of bovine 1.5 kb fragment	(AY303683)
#4	5′ AGAGGGTCTCACTGAACAAC 3′	Nucleotides 1047–1066 of bovine 1.5 kb fragment	(AY303683)
#5	5′ TGTCTCACCTTGGGAAGAATGG 3′	Nucleotides 1389–1410 of bovine 1.5 kb fragment	(AY303683)
#6	5′ CTGCCAAGTTCTCCCAGCGG 3′	Nucleotides 351–370 of bovine 1.5 kb fragment	(AY303683)
#7	5′ GAGCTCTGCACAGCCCGA 3′	Nucleotides 7368–7385 of bovine 7.5 kb fragment	(AY192564)
#8	5′ GACGGGGCGTGCACAGA 3′	Nucleotides 502–518 of bovine 7.5 kb fragment	(AY192564)
#9	5′ TGGCAGGAGAACAGGTTGG 3′	Nucleotides 2057–2075 of bovine 7.5 kb fragment	(AY192564)
#10	5′ GCCAACCTGTTCTCCTGCCACTGTCCCATTGGCTA 3′	Nucleotides 2056–2075 and 7228–7242 of bovine 7.5 kb fragment	(AY192564)

Template DNA was extracted from overnight cultures of transformed E. coli (QIAprep Spin Miniprep Kit; Qiagen). Samples were prepared (BIGDYE Terminator Cycle Sequencing Kit; Applied Biosystems, Foster City, California, USA) for analysis on a DNA sequencing system (model #373A; Applied Biosystems). Overlapping sequences were generated by primer walking and both strands were sequenced. The sequence data was then analyzed using computer software (SEQUENCHER software; Gene Codes Corporation, Ann Arbour, Michigan, USA).

Inverse PCR for the amplification of a 7.5 kb bovine genomic fragment homologous to EGF genes

Template preparation

Bovine genomic DNA was digested with EcoRI, BamHI, HindIII, and SacI (New England BioLabs, Beverly, Massachusetts, USA) in separate reactions. After inactivation of the enzymes, the DNA was diluted to a concentration of 1 μg/mL in water and incubated with 30 U of T4 DNA ligase (Promega Corporation) and 1 × ligase buffer in a volume of 1 mL for 16 h at 16°C. At this low concentration, intramolecular ligation is favored; therefore, DNA circles are generated. After ligation, the DNA was cleaned (QIAquick PCR purification kit; Qiagen) and eluted in a final volume of 30 μL of 10 mM Tris. The volume was reduced to 10 μL by evaporation.

Primer design

Primers based on the sequence of the 1.5 kb bovine genomic fragment obtained previously were designed so that the extension proceeds outwards and the products contain sequences upstream and downstream of the known sequence (16,17). Primers #3 (nucleotides 1160 to 1179 of the 1.5 kb fragment, see Table I for sequence and Figure 1 for location of the primers) and #4 (nucleotides 1047 to 1066 of the 1.5 kb fragment) were used for the first amplification. Primers #5 (nucleotides 1389 to 1410 of the 1.5 kb fragment) and #6 (nucleotides 351 to 370 of the 1.5 kb fragment) were used to amplify a fragment internal to the fragment generated with primers #3 and #4.

Figure 1.

Organization of the 1.5 and 7.5 kb fragments of the EGF gene cloned in the present study. Exons are shown as open boxes and the intron sizes (horizontal lines) are not indicated to scale. (A) 1.5 kb fragment, approximate positions of primers #1 and #2, used to obtain that fragment and approximate positions of primers #3 to #6, used to obtained the 7.5 kb fragment. (B) 7.5 kb fragment of the bovine epidermal growth factor (EGF) gene and positions of primers #7 to #10, used to obtain the cDNA sequence. Primer #10 spans exon 20 and 21. The position of the in-frame TGA codon is shown.

Polymerase chain reaction

Approximately 300 ng of ligated DNA from each digestion was used as template in the PCRs. The reaction mixtures contained 1 × PCR buffer (LA PCR BUFFER II; PanVera Corporation, Madison, Wisconsin, USA), 2.5 mM MgCl₂, 1.6 mM dNTPs (PanVera Corporation), 0.2 μM of each primers #3 and #4, and 1.25 U Taq polymerase (TAKARA LA TAQ POLYMERASE; PanVera Corporation). After an initial denaturing step of 3 min at 96°C, the PCR cycle, which was repeated 35 times, consisted of a denaturing step of 30 s at 94°C, an annealing step of 15 s at 61°C, and an extension step of 7 min at 68°C, followed by a final extension step of 5 min at 72°C. One microlitre of the amplification products were used directly for 30 cycles of double nested amplifications with primers #5 and #6 in the same conditions as above except that the reaction volume was 50 μL instead of 25 μL.

The products from both the first and double nested amplifications were run in a 0.7% agarose gel. The products generated from the amplification of DNA digested with SacI were selected for further analysis because the first amplification reaction resulted in a single faint band of approximately 7.5 kb and the nested amplification resulted in a greater amount of DNA of slightly smaller size. The product from the nested amplification was cloned and both strands were sequenced, overlapping sequences were generated by primer walking. Computer software (Bioedit; T.A. Hall Software, Raleigh, North Carolina, USA) was used to perform sequence alignments (18). A phylogenetic tree showing the relationship between DNA sequences encoding the mature EGF proteins was constructed using the neighbour-joining method (BioEdit) and bootstrap values were obtained using 100 replicates of the original data set.

Southern blot analysis

Approximately 35 μg of bovine genomic DNA was digested with BamHI, EcoRI, HindIII, SacI, and XbaI in separate reactions. The digested DNA was separated in a 0.7% agarose gel at 30 V for 16 h. The gel was stained, photographed, and the DNA was depurinated in 0.2 N HCl for 10 min then denatured in 1.5 M NaCl/0.5 N NaOH. After neutralization, the DNA was transferred to a positively charged nylon membrane (Roche Molecular Biochemicals) by upward capillarity with 10 × SSC as the transfer buffer. Following transfer, the membrane was baked at 80°C for 3 h.

Twenty-five nanograms of the 1.5 kb fragment was labelled with 50 μCi of α³²PdCTP (Amersham Biopharmacia Biotech, Baie d’Urfé, Quebec) using a random primer labelling kit (Roche Molecular Biochemicals). The probe was purified using the NUC-TRAP column (Stratagen, La Jolla, California, USA). Prehybridization was carried out at 42°C for 1 h in a volume of 10 mL of 50% deionized formamide, 1% SDS, 1 M NaCl, 10% dextran sulfate, and 1 mg of denatured salmon sperm DNA. The probe was then added and hybridization was carried out for 12 h at 42°C. The membrane was washed twice in 2 × SSC/1% SDS at room temperature for 30 min; twice in 0.2 × SSC/0.1% SDS at 60°C for 1 h, and twice in 0.1 × SSC/0.1% SDS at 60°C for 30 min. The membrane was exposed to film (KODAK BIOMAX MS film; Kodak Canada, Toronto, Ontario) for 3 d.

Complementary DNA isolation

One microgram of bovine kidney mRNA (Clontech, Palo Alto, California, USA) was reverse transcribed with 10 pmol of primer #7 in a reaction containing 1 × reverse transcription buffer, 10 mM DTT, 1 mM dNTP mix, 40 U RNAse inhibitor, and 15 U of reverse transcriptase (Thermoscript; Life Technologies) for 1 h at 55°C. The reverse transcriptase was inactivated at 85°C for 5 min, 100 U RNAse H was added, the reaction was incubated at 37°C for 20 min, and then heated at 70°C for 10 min.The cDNA was then cleaned (QIAquick PCR purification kit;Qiagen).

Five microlitres of the reverse transcription reaction was used as template for PCR with 1 × PCR buffer (containing 1.5 mM MgCl₂), 0.2 mM of each dNTP, 0.4 μM of each primer # 7 and #8 and 0.63 U Taq polymerase (TAKARA Taq DNA POLYMERASE; PanVera Corporation). After an initial denaturation step at 94°C for 4 min, the PCR cycle, which was repeated 50 times, consisted of a denaturation step of 30 s at 94°C, an annealing step of 30 s at 59°C, an extension step of 30 s at 72°C, followed by a final extension step of 10 min at 72°C.

Several products were observed on the agarose gel after staining. A band of slightly more than 400 base pairs (bp) was cut out of the gel; purified; and the DNA was used as template for 2 separate PCR reactions, with primers # 8 and # 9 and the other with primers #7 and #10, in the same conditions as described above except that only 35 cycles were performed. Primer #10 spans sequences in exon 20 and exon 21 (Figure 1). The fragments obtained were cloned and sequenced on both strands.

Results

Cloning and characterization of a 1.5 kb bovine genomic fragment homologous to EGF genes

In attempt to obtain the DNA sequence encoding the mature bovine EGF protein, we tested several primer pairs where the forward primer was designed from the human, mouse, or both, EGF sequence upstream of the sequence encoding the mature EGF protein and the reverse primer was designed based on the sequence downstream of the mature region of the human EGF gene. None of the primer pairs tested could amplify bovine EGF sequences. We then identified a highly conserved region located upstream of the mature EGF protein in the human, mouse, rat, pig, and horse precursor protein encoding the amino acid sequence “CTNTEG.” This sequence was used to design the degenerate primer #1 (see Table I for primer sequences). As well, the reverse primer (primer #2) was designed from a very conserved region within the mature protein “DGYCLHDG.” With these 2 primers, a 1456 bp fragment was obtained, cloned, and sequenced (GenBank accession number AY303683). This fragment had significant homology to EGF DNA sequences from other species and was comprised of 56 bp from exon 19; a 1320 bp intron; and 80 bp of exon 20, including 30 bp encoding the first 10 amino acids of the NH₂-terminus of the putative bovine mature EGF protein (Figure 1).

The initial strategy to clone the DNA sequence encoding the remainder of the putative bovine mature EGF protein was to use a bovine specific forward primer designed from the sequence obtained above and a reverse primer designed from the sequence of other species downstream of the mature region. However, this strategy was not successful; it was thus suspected that the bovine EGF sequence downstream of the mature region was too different from the other EGFs to allow for successful amplification. Inverse PCR, which amplifies unknown flanking sequences, was then used (16,17).

Cloning and characterization of a 7.5 kb bovine genomic fragment homologous to EGF genes

Amplification of DNA digested with SacI and circularized, as described in the Materials and methods section, produced a single band of approximately 7.5 kb on agarose gel and the nested amplification of that material produced a very intense band of smaller size, as expected, which then was cloned and sequenced (GenBank accession number AY192564). This fragment had significant homology with EGF sequences from other species and encompassed 483 nucleotides of intron, exon 19, the 1320 bp intron identified previously, exon 20, a 4.8 kb intron, and 158 nucleotides of exon 21. Therefore, this fragment contained sequences homologous to those identified previously plus additional sequences, including those encoding the rest of the putative bovine mature EGF protein. Exon 19 is 124 bp in bovines and humans (Genbank accession number AC004050), while exon 20 is 145 and 149 bp in bovines and humans, respectively. There is more variation in the sizes of introns: the intron between exons 19 and 20 is 1362 and 1320 bp in humans and bovines, respectively; and the intron between exons 20 and 21 is 4799 bp and 5171 bp in humans and bovines, respectively. There appears to be an insertion of a SINE sequence in the bovine intron; however, the rest of the intronic sequence shows significant similarity to the human intron.

Figure 2 shows the alignment of the bovine exonic sequence with the corresponding human sequence. Analysis of the bovine sequence revealed an in-frame TGA codon in exon 19 (positions 537 to 539 of the 7.5 kb fragment; positions 51 to 53 on Figure 2). In the EGF from other species, the corresponding codon is TGC, which encodes for cysteine. The overall level of similarity between the bovine, human, porcine, murine, feline, and canine exonic nucleic acid sequences obtained when sequences were compared in pairs (BLAST, version 2) (19) is shown in Table II. The level of identity varied among the different regions of the sequences. For example, the levels of similarity between sequences upstream of sequences encoding the mature proteins were higher than the levels of similarity between sequences encoding mature proteins (Table II). The phylogenetic tree derived from DNA sequences encoding the mature EGF proteins revealed that the bovine sequence is divergent from EGF sequences from other species, but it clusters with EGF sequences rather than with sequences from other members of the EGF family (Figure 3).

Figure 2.

Alignment of the bovine epidermal growth factor (EGF) exonic sequences from the 7.5 kb fragment with the corresponding human sequence (GenBank X04571). Differences in sequence identity are indicated. The in-frame TGA (frame 3) codon is in position 51 to 53 of the bovine sequence. The mature EGF proteins are encoded by sequences from position 177 to 335 (underlined) in human and bovine.

Table II.

Levels of similarity of different regions of the cloned bovine DNA sequence with the corresponding regions in epidermal growth factor (EGF) genes from other mammalian species

	Region of the sequence
Species	Overall	Upstream of mature region	Mature region
Human	70%	82%	64%
Porcine	73%	85%	66%
Murine	64%	72%	59%
Feline	70%	78%	64%
Canine	69%	85%	64%
Equine	—	—	64%a

^aThe bovine DNA sequence encoding the mature protein had also 64% identity to the horse sequence (GenBank accession number S73527) on 144 nucleic acids when 9 gaps were introduced

Figure 3.

Unrooted neighbour-joining tree showing phylogenetic relationship among DNA sequences encoding mature epidermal growth factor (EGF) proteins, bovine betacellulin (NM_173896), human epiregulin (NM_001432), and human transforming growth factor alpha (TGFalpha, X70340).

The alignment of the deduced bovine protein with the human, porcine, murine, feline, and canine corresponding sequences is shown in Figure 4. The overall level of similarity between the bovine deduced protein sequence and the corresponding sequences from these species is shown in Table III. As was the case with the nucleic acid sequence, the levels of similarity varied among the different regions of the sequence (Table III).

Figure 4.

Alignment of the bovine epidermal growth factor (EGF) deduced protein sequence with the corresponding porcine (AAK18830), human (CAA28240), feline (AB050947), canine (AB049597), and murine (0907234A) sequences. The putative selenocysteine residue (*) in the bovine sequence is in position 17. The mature protein sequence is from position 59 to 111 (underlined).

Table III.

Level of similarity of different regions of the deduced bovine protein sequence with the corresponding regions in epidermal growth factor (EGF) protein from other mammalian species

	Region of the sequence
Species	Overall	Upstream of mature region	Mature region
Human	49%	66%	39%
Porcine	51%	75%	36%
Murine	46%	61%	36%
Feline	49%	71%	34%
Canine	48%	69%	34%
Equine	—	—	31%a

^aLevel of similarity on 48 amino acids

Southern blot analysis

Figure 5 shows a Southern blot of bovine genomic DNA digested with the restriction enzymes BamHI, EcoRI, HindIII, SacI, and XbaI and hybridized with the 1.5 kb fragment labelled with α³²PdCTP. The probe hybridized strongly to a single band in each lane ranging in size from approximately 4 to 8 kb.

Figure 5.

Southern blot of bovine genomic DNA hybridized with the 1.5 kb epidermal growth factor (EGF) genomic fragment labeled with ³²PdCTP. Lanes 1 to 5 are bovine genomic DNA digested with XbaI, SacI, HindIII, EcoRI, and BamHI.

Complementary DNA isolation

A bovine cDNA sequence was cloned from kidney mRNA (GenBank accession number AY195611). The sequence was 98% identical to the corresponding genomic sequence and also contained the in-frame TGA codon upstream of the sequence encoding the putative mature EGF protein. When compared to the deduced protein from the genomic sequence, 2 amino acid changes in the mature protein sequence deduced from the cDNA sequence were evident: glutamine instead of arginine in position 18, and glutamine instead of histidine in position 32.

Discussion

In this paper, we report for the first time, the cloning and characterization of a bovine genomic fragment where the sequence was homologous to exon 19 through 21 from EGF genes from other mammalian species. If the fragment obtained in the present study originates from the functional bovine EGF gene, it would contain the sequence encoding the bovine mature EGF protein.

Although the bovine sequence shares homology with mammalian EGF genes, it presents an unusual primary sequence in the form of an in-frame TGA termination codon in exon 19. In EGF genes from other species, the corresponding codon is TGC, which encodes for cystein. The presence of a termination codon could be interpreted as an indication that the cloned DNA is a portion of a pseudogene; one that is expressed, at least in the bovine kidney, since we obtained a sequence of cDNA containing the TGA codon from kidney mRNA. We have also cloned sheep genomic DNA fragments homologous to EGF genes (John and Bilodeau-Goeseels, unpublished; GenBank accession numbers AY195612 and AY195614). There is a TGA codon in exon 19, as in the bovine sequence and the sequence encoding the ovine mature protein contains 3 additional stop codons. To our knowledge, there are no reports of EGF pseudogenes in other mammalian species; however, a comparison (BLAST) using the human EGF mRNA sequence as the query resulted in the mouse EGF gene on chromosome 3, as expected, but also a secondary hit on mouse chromosome 16 (Accession number NT_039625). Interestingly, in that sequence, the region corresponding to the sequence encoding the mature protein is not separated by an intron as in functional EGF genes; therefore, it is possible that the mouse genome contains a processed EGF pseudogene.

Even though the fragment cloned contained an in-frame termination codon, we cannot rule out that it originated from a functional gene: the bovine EGF gene or a gene encoding an EGF-like protein. The lower level of similarity between the bovine sequence and the corresponding sequences from other mammalian species compared to the level of similarity between these species and the fact that the level of similarity between the bovine and other mammalian DNA and protein sequences is higher for exonic sequences upstream of the mature region, suggests that the sequence obtained would originate from an EGF-like gene. The authors are not aware of any examples in genes or proteins where there is more divergence between species in the mature protein than in the propeptide. To determine whether the homology between bovine and other EGF sequences in various locations on the gene is relatively high, as in sequences upstream of the mature region, or lower, as in the mature region, a fragment containing 107 nucleotides of exon 3, an intron, and 142 nucleotides of exon 4 was cloned (Prenevost and Bilodeau-Goeseels, unpublished). The similarity to corresponding nucleotide sequences from other mammalian species was over 80%.

Even if it is unlikely, there is a possibility that the fragment is from the functional EGF gene. Several attempts to clone bovine EGF genomic or cDNA sequences using primers designed based on the human, mouse, or porcine sequences (we hypothesized that if there was another functional EGF gene, its sequence would be more homologous to the EGF gene from these species) were unsuccessful, suggesting that the fragment obtained is from the only bovine EGF gene. Moreover, the structure of the fragment, which contains introns of the same sizes, at the same locations as in other EGF genes, and with significant levels of identities compared to the introns of the human EGF gene, also suggests that the fragment that was cloned could have originated from a bovine EGF gene rather than from a gene encoding an EGF-like protein. As well, the phylogenetic tree shows that even though the bovine nucleic acid encoding the mature protein is different from other EGF sequences, it is more related to them than to other members of the EGF superfamily. When the phylogenetic analysis was repeated using the neighbour-joining with the Kimura-2 correction, the maximum likelihood, and the maximum parsimony methods, similar trees were obtained even when the analysis was rooted (data not shown).

Southern blot analysis of bovine genomic DNA showed hybridization of the EGF probe to a single band in each digest. However, the results are inconclusive since for 2 out of the 3 restriction digests where the sizes of the hybridizing fragments could be predicted, the sizes of the bands were higher than expected (8 kb instead of 5.3 and 2.8 kb). This could be due to incomplete digestion or slower mobility due to the very large amount of DNA loaded (35 μg).

If the fragment originated from the functional gene, there would be a mechanism to bypass the stop codon in order to generate a functional protein. Alternative splicing or RNA editing are unlikely mechanisms since the cloned cDNA contained the corresponding exon 19 sequence, including the TGA codon. In addition to termination, the codon TGA has also been shown to code for the incorporation of selenocysteine, the 21st amino acid (20,21). In eukaryotes, recoding of the opal nonsense codon TGA to selenocysteine is directed by a selenocysteine insertion sequence in the 3′-untranslated region (22) and data from constructs containing sequences from selenoprotein genes indicate that termination and selenocysteine incorporation occur in parallel in selenoprotein mRNA (23). We have not obtained sequences from the 3′-untranslated region of the bovine gene cloned in the present study; therefore, it is not known if the required selenocysteine insertion sequence is present. Although no EGF proteins are known as being selenoproteins, there are examples of proteins that are selenoproteins in some species but not in others (24,25). We determined that the sheep genomic sequence we also identified (see above) is from a putative EGF pseudogene since it contains a TAG codon, which only encodes termination.

One important question is whether the protein encoded by the fragment we obtained would have EGF activity. While this manuscript was being reviewed, we became aware of a US patent (26) that discloses a sequence that the authors believe is the bovine EGF mature protein sequence. This sequence is identical to our deduced mature protein sequence except for an arginine residue (position 76, Figure 3), which is replaced by a histidine residue. Moreover, the Collier sequence includes 4 extra amino acids at the COOH terminus, which are identical to the last 4 amino acids of the precursor protein before the start of the mature protein, positions 55 to 58 (Figure 3). This protein stimulated the growth of dairy and beef heifers or cows mammary parenchyma, as was the case with human and mouse EGF demonstrating that a protein with a similar sequence (significantly different from other EGFs) as the one deduced from the DNA sequence obtained in the present study has biological activity. In the patent, there is no description of how this protein was isolated and no disclosure of DNA sequence. A further literature search revealed only 1 paper by the same authors describing the purification of an EGF-like protein, which stimulated the proliferation of bovine mammary epithelial cells (27). Only the amino acid composition of the protein is provided and it does not correspond to the amino acid composition from the deduced bovine sequence from the present study nor the protein sequence disclosed in the patent.

The cDNA sequence obtained in the present study was 98% identical to the corresponding genomic sequence. The differences in the DNA sequence result in 2 amino acid changes in the mature protein. This could be due to different animals, or animals of different breeds being used as sources of genomic DNA and mRNA. The EGF mRNA levels have been shown to be 1/400 and 1/200 in mouse kidney and male submaxillary gland, respectively (28). In the present study, 1 μg of bovine kidney mRNA was used in the PCR; previous attempts using total RNA were unsuccessful, suggesting that the level of expression in bovine kidney may be lower than in mouse kidney. Similarly, we were not successful at amplifying EGF sequences from bovine submaxillary gland or mammary gland RNA. These results and the fact that no EST identical to the cloned sequences are available (there are over 300 000 bovine EST currently available in GenBank), suggests that the expression levels of the putative bovine EGF gene may be very low. In that regard, few studies have examined the presence of the EGF protein or mRNA in bovine tissues. For example, bovine EGF sequences were amplified from bovine early embryos using primers based on the porcine sequence (29). We were not successful at amplifying bovine EGF genomic or cDNA sequences using the same primers. In another study, EGF cDNA sequence was amplified from bovine endometrium, the sequence was reported to be 97% homologous to porcine EGF (30). To our knowledge, this sequence was not submitted to Genbank.

In conclusion, we have obtained the sequence of 7.5 kb of bovine genomic DNA homologous to exon 19 through 21 from EGF genes from other mammalian species. The nucleotide and deduced protein sequences were significantly different from EGFs from other mammalian species. There was a TGA codon in the coding sequence suggesting that the sequence cloned originated from a pseudogene. However, the cloned sequence could potentially originate from the functional bovine EGF gene. Further investigations are required to determine whether the bovine gene translates into a functional protein and, if so, what mechanism is used to bypass the termination codon. In vitro expression studies are underway and the availability of recombinant bovine putative EGF protein will allow the characterization of its structure, receptor binding capacity, and biological activity.