Molecular cloning of DNA complementary to bovine growth hormone mRNA.

We have cloned DNA complementary to mRNA coding for bovine growth hormone (bGH). Double-stranded DNA complementary to bovine pituitary mRNA was inserted into the Pst I site of plasmid pBR322 by the dC x dG tailing technique and amplified in E. coli x 1776. A recombinant plasmid containing bGH cDNA ws identified by hybridization to cloned rat growth hormone cDNA. It contains the entire coding and 3'-untranslated regions and 31 bases in the 5'-untranslated region. Nucleotide sequence analysis determined the sequence of the 26-amino acid signal peptide and confirmed the published amino acid sequence of the secreted hormone at all but 2 residues. Codon usage is nonrandom, with 81.7% of the codons ending in G or C. The nucleotide sequence of bGH mRNA is 83.9% homologous with rat GH mRNA and 76.5% homologous with human GH mRNA, while the respective amino acid sequence homologies are 83.5% and 66.8%.

Corp. to J. D. B. 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.
shortly after killing at a local abattoir arld were frozen immediately in liquid NZ. Total RNA was prepared by extraction with guanidine thiocyanate, centrifugation through 5.7 M CsCl (6, 71, phenol extraction, and ethanol precipitation. Polyadenylated RNA (poly(A)+ RNA) was prepared by oligo(dT)-cellulose chromatography using a m o acation (8) of the method of Aviv and Leder (9). Integrity of the poly@)+ RNA was seseased by cell-free translation using a rabbit reticulocyte system (8, 10). Bovine growth pre-hormone was immunoprecipitated using an heterologous anti-ovine GH antiserum and prepared by adsorption to formalin-fixed S. aureus Cowan strain I, as described previously (10, 11). Cell-free translation products and immunoprecipitates were dieplayed on NaDodSOd-acrylamide gels (12).
Preparation of Double-stranded cDNA for Cloning-Polyadenylated RNA was reverse-transcribed into single-stranded complementary DNA using reverse transcriptase by a modification (8) of the method of Monahan et al. (13). After alkali digestion of the RNA, phenol extraction, Sephadex G-50 chromatography, and ethanol precipitation, the single-stranded cDNA was used to self-prime the synthesis by reverse transcriptase of a complementary DNA strand. The single-stranded "hairpin loop" at the 5' end of the first cDNA strand, as well as any other single-stranded DNA regions, was digested briefly with S, nuclease (14) and the double-stranded cDNA was again purified by phenol extraction, Sephadex C-50 chromatography, and ethanol precipitation. Single-stranded tracts of homopolymeric deoxycytidine were then added to the 3' ends of the cDNA using terminal transferase (Enzo Biochemicals) and the potassium cacodylate buffer system of Roychoudhury et al. (15). Plasmid pBR322 (16) was cut with Pst I and similarly "tailed with deoxyguanosine. Twenty nanograms of tailed cDNA was hybridized to 50 ng of tailed pBR322 in 50-4 reactions as described previously (17). This mixture was then used directly to transform bacteria under P2, EK 2 conditions, in accordance with NIH Guidelines. E. coli .1776 were rendered permeable to DNA by incubation in 75 m~ CaCIz,5 mM MgC12, IO m~ Tris, pH 7.5, for 20 min at 4°C. Plasmid and bacteria were incubated for 60 min at 4"C, then 2 min at 41"C, mixed with 10 volumes of nutrient broth, incubated 2 to 3 h at 37"C, and plated onto agar containing 15 pg of tetracycline/ml. Identification and Preparation of pBP348-E. coli harboring recombinant plasmids were identified by antibiotic sensitivities and screened for the presence of cloned DNA by colony hybridization to freshly prepared bovine pituitary [32P]cDNA (18). Plasmid DNA was prepared from selected colonies, cut with Pst I, electrophoresed on IS agarose, stained with ethidium bromide, and photographed, then transferred to nitrocellulose Nters by the method of Southern (19). The presence of GH gene sequences in the transferred DNA was assessed by hybridization to cloned, full length rat GH cDNA (20) labeled by nick translation to -10' cpm/pg of DNA (21). One clone, pBP348 (plasmid bovine pituitary, 348th clone obtained), which hybridized with the nick-translated rGH cDNA, was chosen for further study.
DNA Sequence Analysis-Plasmid pBP348 was cut with Pst I, and the phosphate on the 5' ends of the DNA fragments was removed with alkaline phosphatase and replaced with [32P]phosphate using polynucleotide kinase. Subsequent cutting with a variety of other restriction endonucleases, polyacrylamide gel electrophoresis, and staining and autoradiography of the bands of DNA provided a restriction map of the cloned DNA. A large batch of pBP348 was then prepared and cut with Pst I, Puu 11, or Sau 3A, labeled with ["PI-dATP and polynucleotide kinase, and then cut with other enzymes to yield DNA fragments labeled at a single end. These fragments were prepared by elution from polyacrylamide gel and sequenced by the modified (17) method of Maxam and Gilbert (22). molecular weight bands testifies to the integrity of the poly(A)' RNA preparation. The two prominent bands having molecular weights of about 25,000 and 24,000 are pre-prolactin and growth pre-hormone, respectively, as shown by the specific immunoprecipitations in the adjacent lanes. This result is similar to that obtained with the human GH and Prl (23). As assayed by densitometric scanning of the autoradiograph of the total lysate, mRNAs for GH and Prl represent 12.6 and 18.0% of total cow pituitary polyadenylated mRNA, respectively.

Preparation and Danslation
Synthesis and Cloning of cDNA-Total pituitary poly(A)' RNA was used as template for the synthesis of cDNA with reverse transcriptase using oligo(dT) primers for first strand synthesis and self-priming for the second strand. Efficiencies of reverse transcription were generally 6 to 10% for the first strand and 75 to 100% for the second. When aliquots of doublestranded cDNA were digested with a variety of restriction endonucleases and displayed on acrylamide gels, discrete bands were seen suggesting that a small number of cDNA species predominated (20,23). This cDNA was then trimmed with SI, tailed with dCMP, and ligated into the Pst I site of pBR322 as described under "Materials and Methods." Transformation efficiencies with such recombinant plasmids were about 5 colonies/ng of pBR322 used, whereas supercoiled pBR322 gave efficiencies of 1000 to 4000 colonies/ng.
Nucleotide Sequence of bGH cDNA-A colony containing bGH gene sequences was identified by hybridization to cloned rat GH cDNA as described under "Materials and Methods." Restriction enzyme analysis of the DNA cloned in pBP348 showed that it contained two internal Pst I sites and two Puu I1 sites, which divided the cDNA into pieces of convenient sue for sequencing. The two Puu I1 sites subtend a DNA fragment of 489 base pairs; a Puu I1 fragment of this sue was seen in the analytical digests of the original cDNA. The cloned cDNA contains 831 base pairs including 31 base pairs in the 5'-untranslated region, the entire region coding for the prehormone (651 base pairs), the entire 3'-untranslated region (104 base pairs), a brief stretch of poly (A), and the dC -dG tails (Fig. 2). The amino acid sequence predicted by the nucleotide sequence differs from the published amino acid sequence of bGH (5,24) only at positions 47 and 66, where aspartic acid and glutamic acid are replaced by their respective amides. Bovine growth hormone has previously been noted to contain either valine or leucine a t position 127 (24,25); the individual mRNA molecule leading to pBP348 coded for leucine at this position. The amino acid sequence of the signal peptide was deduced from the nucleotide sequence preceding the codon for amino acid 1. We have assigned this position to alanine, although bGH isolated from bovine pituitaries may have either alanine or phenylalanine at its NH2 terminus (5). Two methionine codons, the first one presumably being the initiation site for translation, are found in the same reading frame at positions -25 and -26. The positions of the methionines and 9 other amino acids in the pre-sequence have been

FIG. 2.
Nucleotide sequence of bovine growth hormone mRNA and ita derived amino acid sequence as determined by the nucleotide sequence of the cDNA contained in pBP348.

TABLE I Amino acid and nucleotide homologies among bouine, human, and rat growth pre-hormones
The entire coding regions, including the "pre-sequences," as well as the termination codons, were compared. Gaps in the amino acid sequence were introduced as shown in Fig. 3; where gaps were not needed in a paired comparison, they were eliminated, accounting for the differences in the lengths of the compared sequences. Times of evolutionary divergence, in millions of years before the present, were calculated using a unit evolutionary period = 4 (30).  Homologies among bovine, rat, and human growth pre-hormones. Gaps were introduced at the amino acid residues shown to permit matching of regions of obvious homology. The arrangement was used to calculate the data in Table I. reported previously (26), and are confumed by our sequence. Hydrophobic amino acids predominate in the pre-sequence (including 6 leucine and 4 alanine residues) as reported for signal peptides of several other secretory hormones (17,20,23). Based on the deduced amino acid sequence, we calculate the molecular weight of bovine growth pre-hormone to be 24562.30.
The 5"untranslated region of the cloned bGH cDNA contains 31 nucleotides without apparent pallindromes or inverted repeats to suggest secondary structures. The 3'-untranslated region contains three long pallindromic regions (Fig. 2), the first of which (5' U-G-C-C-C-C-U-C-C-C-C-C-G-T 3') is remarkably similar to pallindromic sequences rich in cytosine in the 3'-untranslated regions of mRNAs for hGH (23), rGH (20), and hCS (27). The significance of such conserved nontranslated pallindromes remains unknown.

DISCUSSION
Although non-primate growth hormones are inactive in man (28, 29), they exhibit considerable amino acid homology. We have compared the nucleotide sequences of the coding regions from the mRNAs for bovine, rat (20), and human (23) growth pre-hormones (Table I). One or two single codon gaps were introduced in each sequence as shown in Fig. 3 to permit alignment of long sequences with obvious homology. The coding regions of the bovine and rat mRNAs are very similar, having 83.9% homology, whereas the coding regions of human and bovine or human and rat GH mRNAs are only 76.5 and 74.9% homologous, respectively. The homology in the amino acid sequences of bovine and rat growth pre-hormones is 83.5%, similar to the nucleotide homology; this contrasts with the other two paired comparisons, where the amino acid homology is about 10% less than the nucleotide homology. Of the 36 amino acids differing in bovine and rat GHs, 27 (75.0%) are replaced by "favorable" substitutions, i.e. those which occur more frequently than random chance would predict (5), whereas only 59 to 64% of the substitutions are favorable in the other two paired comparisons.
These findings might suggest that cattle and rats are evolutionarily more closely related than human beings are with either of the other species. Wilson et al. have described a parameter called the unit evolutionary period, defined as the length of time in millions of years of evolution required for 1% non-homology to arise in the same protein in two different species (30). Using a unit evolutionary period of 4 for growth hormone (30), we calculate that the evolutionary precursors to cattle and rats diverged about 66 milLon years ago, while the evolutionary precursors to men diverged from those to rats and cattle about 140 million years ago. As Wilson et al. calculated units of evolutionary period for proteins rather than their pre-proteins, it could be argued that the pre-sequence should be eliminated from these calculations. This would reduce the time of divergence of rats and cattle to 48 million years ago, but would not change the time of divergence of human precursors from the other two species.
The 3"untranslated region of bGH mRNA is similar to the 3"untranslated regions of mRNAs coding for hGH and rGH. We have not attempted to calculate homologies as it is unclear where to introduce the required gaps in the sequences without the additional guidance of an amino acid sequence. However, the relatedness of these sequences can be seen in Fig. 4. The homologies are clustered in three regions, of which the frst two contain long pallindromes. The third homologous region includes precisely the same 12-nucleotide sequence 5' U-C-C-

U-A-A-U-A-A-A-A-U 3' in all three species. This last homologous region contains the 5' A-A-U-A-A-A 3' found in the 3' untranslated region of most eukaryotic mRNAs (31).
Excluding the region of poly(A) and the "tails" of enzymatically added dC or dG, the cloned cDNA contains 787 nucleotides, of which 468 (59.5%) are either C or G. This contrasts sharply with 42% G + C content of vertebrate DNAs in general (32). This is due to non-random codon selection in bGH mRNA; 81.7% of which end in G or C. G and C are highly favored in the third position of the codons for rat (20) and human (23) growth pre-hormones and human chorionic somat,omammotropin (27), but are only slightly favored in rat (17) and bovine' Prl. The reasons for this preference for G and C in the third position of codons in mRNAs for growth hormone is unknown,