Functional Analysis of a Mutation Occurring between the Two In-frame AUG Codons of Human Angiotensinogen*

Angiotensinogen (ANG) is the specific substrate of the renin-angiotensin system, a major participant in blood pressure control. We have identified a natural mutation at the −30 amino acid position of the angiotensinogen signal peptide, in which an arginine is replaced by a proline (R−30P). Heterozygous individuals with R−30P showed a tendency to lowered plasma angiotensinogen level (1563 ng of ANG I/ml (range 1129–1941)) compared with normal individuals in the family (1892 ng of ANG I/ml (range 1603–2072)). Human angiotensinogen mRNA has two in-phase translation initiation codons (AUG) starting upstream 39 and 66 nucleotides from the cap site. R−30P occurs in a cluster of basic residues adjacent to the first AUG codon that may affect intracellular sorting of the nascent protein. Pulse-chase experiments in transiently transfected cultured cells revealed that the R−30P mutation was associated with reduced amounts of both intra- and extracellular protein. In a cell-free system, we found that two forms of native angiotensinogen were generated by alternative initiation of translation at either AUG codon. Alteration of either the first or second AUG codons abolished the synthesis of the longer and the shorter form of native angiotensinogen, respectively. Furthermore, the rate of secretion of the shorter form was lower than that of the longer form. By transplanting angiotensinogen signal peptide onto green fluorescence protein, however, we found that both forms of the signal peptide could target green fluorescence protein, normally localized in the cytoplasm, to the secretory pathway. Although the R−30P mutation may not affect intracellular sorting of angiotensinogen in a qualitative manner, it leads to a quantitative reduction in the net secretion of mature angiotensinogen through decreased translocation or increased residence time in the endoplasmic reticulum.

The renin-angiotensin system affects salt and water homeostasis, vascular tone, and blood pressure, and each component of the system represents a potential candidate in the etiology of human essential hypertension. In previous reports (1-3), we have provided genetic evidence based on linkage and association studies that molecular variants of the angiotensinogen gene (AGT) constitute inherited predisposition to essential hypertension in humans. In this report we document a novel mutation in the signal peptide of AGT and investigate the potential impact of the mutation on protein secretion.
Angiotensinogen is the only known natural substrate for the aspartyl proteinase renin. The specific interaction between renin and angiotensinogen is the initial and rate-limiting step of the enzymatic cascade that generates angiotensin I, which is further processed to an active octapeptide, angiotensin II, by angiotensin-converting enzyme. Angiotensinogen is a heterogeneous glycoprotein with a molecular mass of around 60 kDa, constitutively secreted mainly from liver into plasma and extracellular fluid (4). In general, secretory proteins are synthesized as precursors with amino-terminal signal sequences that are cleaved out by specific proteases during maturation. The signal peptide plays an important role in translation by ribosomes; nascent protein is bound via the signal peptide to the signal recognition particle, which guides the complex to the endoplasmic reticulum (ER). 1 Translocation of the protein into the ER is followed by post-translational modifications such as glycosylation, signal peptide cleavage, disulfide bond formation, and folding (5,6). The sequences of signal peptides are heterogeneous, but typically three conserved regions have been recognized and shown to be essential for protein export as follows: (i) an amino-terminal positively charged region, (ii) a central hydrophobic region, and (iii) a more polar carboxylterminal region (5). The amino-terminal amino acid sequence deduced from the cDNA of human angiotensinogen revealed the presence of a signal peptide of 33 amino acids containing all three consensus regions (7,8). The 5Ј segment of human angiotensinogen mRNA contains two AUG codons in the same reading frame. We have identified a mutation leading to the substitution of proline for arginine at the Ϫ30 amino acid position of the angiotensinogen signal peptide (RϪ30P). This substitution affects a cluster of basic residues, and it has been suggested that such motifs may play a direct role in the association between signal sequence and the membrane of the endoplasmic reticulum through electrostatic interaction with negatively charged phospholipids (9). Furthermore, Lotteau et al. (10) demonstrated that the two amino-terminal amino acid sequences of human invariant chain (Ii), a type II membrane glycoprotein, resulted from alternative initiation codons determining the intracellular localization of the protein. The two distinct forms of Ii, Ii31 and Ii33, are generated by the use of alternative initiation codons and, consequently, differ only in their amino-terminal sequence (10,11). Ii31 was transported rapidly out of the ER after synthesis, whereas Ii33 was maintained in the ER (10,12). These authors suggested that the amino-terminal cluster of basic residues adjacent to the first AUG codon might serve as a signal for retention of Ii33 in the endoplasmic reticulum. In this report, we investigate the functional significance of the two AUG codons and the functional impact of the RϪ30P mutation on angiotensinogen processing in the secretory pathway.

MATERIALS AND METHODS
Mutation Screening-DNA samples from patients with essential hypertension were subjected to molecular screening of the angiotensinogen gene by single strand conformational polymorphism analysis. Single strand conformational polymorphism screening was performed on PCR-amplified segments spanning all five exons as well as 1 kilobase pair of the promoter region of angiotensinogen as described (1). In a patient presenting a unique electrophoretic variant, individual bands were excised from the gel and subjected to direct sequencing on an ABI model 373A sequencer. Plasma Angiotensinogen Measurement-Plasma level of angiotensinogen was measured by a two-step procedure in which conversion of angiotensinogen to angiotensin I by human renin was followed by measurement of angiotensin I by radioimmunoassay (13).
Expression of Human Angiotensinogen in COS-1 Cells-The entire coding sequence of human angiotensinogen (provided by Dr. Nakanishi at Kyoto University, Japan) was placed under control of the major late promoter of adenovirus in the expression vector pMT2 (Genetics Institute, Cambridge, MA) (wild type-ANG/pMT2). The first or the second ATG codon was replaced by an ATC codon for ATC2-ANG/pMT2 or for ATC1-ANG/pMT2, respectively, by PCR-based site-directed mutagenesis (14). A newly identified natural mutant, RϪ30P, was introduced into the same vector. These constructs were transfected into COS-1 cells (ATCC CRL1650, Rockville, MD) by electroporation with a Gene-pulsar (Bio-Rad), and pulse-chase experiments were performed as described previously (2,15). Transfected cells were pulse-labeled with 100 Ci/ml Tran 35 S-label (ICN, Costa Mesa, CA), for 10 min at 37°C and chased for 6 h with Dulbecco's modified Eagle's medium supplemented with excess methionine (20 mg/ml) and cysteine (20 mg/ml). The cells and conditioned media were collected at the time indicated. The cells were lysed by two freeze-thaw cycles in a lysis buffer (25 mM Tris-HCl, 50 mM NaCl, 2% Nonidet P-40, 0.2% SDS, 0.5% sodium deoxycholate, and 1 mM phenylmethylsulfonyl fluoride, pH 7.4). The lysed cells were cleared by centrifugation at 18,000 rpm for 20 min. Immunoprecipitation was performed according to a standard protocol with anti-human angiotensinogen polyclonal antibody (a gift from Dr. Tewksbury at Marshfield Foundation, Marshfield, WI) (16). The samples were separated on SDS-PAGE gels, and the signal was measured on a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) or exposed to Kodak x-ray film.
In Vitro Transcription, Translation, and Translocation-The cDNA of human angiotensinogen was placed under the control of the T7 promoter in the pBluescript SK(ϩ) vector (Stratagene, La Jolla, CA) (wild type-ANG/pBS). Three constructs were generated by PCR-based site-directed mutagenesis (14) as follows: (i) ATC2-ANG/pBS (the second ATG codon was changed to an ATC), (ii) ATC1-ANG/pBS (the first ATG codon was changed to an ATC), and (iii) RϪ30P-ANG/pBS (an arginine at the Ϫ30 position was changed to a proline). In vitro transcription and translation reactions were performed with TNT T7 quickcoupled transcription/translation system (Promega). Translation products labeled in the presence of canine microsomal membrane (Promega) and [ 35 S]methionine (Amersham Pharmacia Biotech) were subjected to SDS-PAGE gel analysis followed by autoradiography.
Enzymatic Deglycosylation of Recombinant Angiotensinogen-The radiolabeled angiotensinogen obtained through pulse-chase experiments or in vitro translation experiments was subjected to immunoprecipitation as described. Immunoprecipitated materials were dissolved in a reaction buffer (50 mM sodium phosphate buffer, 1 mM EDTA, 1% Nonidet P-40, 1% 2-mercaptoethanol, pH 7.4), and the samples were treated with endoglycosidase H, N-glycosidase F, or O-glycosidase (Roche Molecular Biochemicals) for 24 h at 37°C. After enzyme treatment, the samples were separated on SDS-PAGE gel followed by autoradiography.
Subcellular Localization of Recombinant Angiotensinogen Signal Peptide/EGFP Chimera and Immunoblot Analysis-The recombinant chimeric molecule between angiotensinogen signal peptide and enhanced green fluorescent protein (EGFP) was constructed by inserting the signal sequence of the human angiotensinogen gene into the 5Ј end of EGFP cDNA of the pEGFP-N13 vector (CLONTECH, Palo Alto, CA). Site-directed mutagenesis in the first or second ATG codons of the angiotensinogen signal peptide was carried out by the primer-overlap method (14). Three constructs were generated: (i) wild type-ANG/ pEGFP, (ii) ATC2-ANG/pEGFP (the second ATG codon was changed to an ATC), and (iii) ATC1-ANG/pEGFP (the first ATG codon was changed to an ATC) (Fig. 4A).
COS-1 cells were transfected with angiotensinogen signal peptide/ EGFP by a calcium phosphate method using modified bovine serum (Stratagene). 48 h after transfection, the cells were fixed with 4% paraformaldehyde in PBS for 10 min on ice. After rinsing three times with PBS, the cells were permeabilized with 1% Triton X-100 in PBS for FIG. 1. The R؊30P mutation in a pedigree. A, signal peptide sequence of wild type and RϪ30P mutant angiotensinogen. Numbers above the amino acids indicate their position relative to the beginning of the mature form of angiotensinogen. B, pedigree with the RϪ30P mutation. Plasma angiotensinogen concentration, expressed as nanograms of ANG I/ml, was measured by indirect assay after renin reaction. HT(ϩ) or (Ϫ) denotes the presence or absence of essential hypertension. 5 min. The cells were then incubated with anti-GFP polyclonal antibody (CLONTECH) together with the monoclonal antibody against the 78-kDa glucose-regulated protein (BiP) (Accurate Chemicals), which resides in the endoplasmic reticulum lumen, in 0.5% bovine serum albumin, 0.05% Tween 20, PBS for 1 h at room temperature. After washing with PBS, the cells were incubated with a fluorescent-conjugated goat anti-rabbit IgG and a rhodamine-conjugated goat anti-mouse IgG. Fluorescence microscopy was performed using an Olympus microscope IX-70 (Olympus Co., Tokyo, Japan) with fluorescein isothiocyanate filter sets.
The cells and conditioned media from transfected COS-1 cells were subjected to Western blot analysis. Samples were loaded on SDS-PAGE gel and electrotransferred onto Hybond-C (Amersham Pharmacia Biotech). Membranes were blocked for 2 h in TBST (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.02% Tween 20) containing 3% BSA and then incubated with anti-GFP polyclonal antibody at room temperature for 2 h. Membranes were washed with TBST and incubated with antirabbit IgG-horseradish peroxidase (Amersham Pharmacia Biotech) at room temperature for 2 h. After washing three times with TBST, antibody binding was visualized by ECL (Amersham Pharmacia Biotech).

Analysis of RϪ30P Mutation in Human Angiotensinogen
Signal Peptide-While screening patients with essential hypertension, we identified a molecular variant, a G to A nucleotide substitution in exon 2 of AGT, which resulted in the replacement of arginine by proline residue at the Ϫ30 amino acid position (RϪ30P) in the angiotensinogen signal peptide (Fig.  1A). We extended the molecular screening in the proband's family, of Caucasian origin, and four subjects among the seven individuals analyzed were identified as heterozygote for the mutation (Fig. 1B). Heterozygous individuals (two were hypertensive and two were normotensive) had lower plasma angiotensinogen concentrations compared with the three normotensive non-carriers (1563 ng of ANG I/ml (range 1129 -1941) versus 1892 ng of ANG I/ml (range 1603-2072)) ( Fig. 1B).
Sequences encoding wild type and RϪ30P angiotensinogen were cloned and transiently transfected into COS-1 cells, which do not express measurable levels of angiotensinogen. Biosynthesis and secretion of recombinant angiotensinogen in COS-1 cells were examined by pulse-chase experiments. Radiolabeled angiotensinogen was immunoprecipitated from the cell lysate and media with anti-human angiotensinogen antibody, analyzed by SDS-PAGE gels, and quantified using a PhosphorImager (Fig. 2, A and B). Pulse-chase experiments showed that the total amount of wild type angiotensinogen (cell ϩ medium) was unchanged during a 4-h chase. The half-life of RϪ30P mutant protein was considerably shorter, and the secreted amount of RϪ30P mutant was significantly lower than the amount of wild type (Fig. 2, A and B). Portions of the immunoprecipitates were treated with endoglycosidase H, N-glycosidase F, or O-glycosidase (Fig. 2C). Intracellular angiotensinogen treated with endoglycosidase H showed a reduced heterogeneity in both wild type and RϪ30P mutant. On the other hand, secreted forms from wild type and mutant angiotensinogen were endoglycosidase H-resistant. Treatment of both cellular and secreted forms of angiotensinogen with Oglycosidase did not significantly change the molecular mass of the protein, suggesting that glycosylation of angiotensinogen is predominantly an N-linked glycosylation as demonstrated (17). We could not detect qualitative differences in the glycosylation between wild type and mutant protein.
To assess the effect of the RϪ30P mutation on the processing of angiotensinogen in the secretory pathway in more detail, we analyzed in vitro transcription, translation, and translocation using canine pancreatic microsomal membranes. As shown in Fig. 2D, by adding microsomal membrane, the in vitro product of angiotensinogen was processed to fuzzy heterogeneous bands showing retarded mobility in SDS-PAGE. The greater require-ment of microsomal membrane fraction for the RϪ30P mutation compared with wild type suggested reduced efficiency of translocation in the endoplasmic reticulum.
Two In-phase AUGs Are Used as Initiation Codon in Vitro-The sequence of the 5Ј end of human angiotensinogen mRNA revealed the existence of two AUG codons in the same reading frame (Fig. 3A). The first AUG codon was located 39 nucleotides from the cap site. Based on Kozak's initiation rule (18), an initiation consensus sequence "(G/A)NNATGG" can be recognized in the sequence around the second AUG but not in the sequence around the first AUG.
The regions containing signal sequences of angiotensinogen gene were compared between human, mouse, and rat (Fig. 3B). The signal peptides are very conserved between these species; the regions spanning residue Ϫ24 to Ϫ1 particularly are highly homologous. Mouse angiotensinogen contains two in-phase FIG. 2. Biochemical analysis of the R؊30P mutation. A, pulsechase experiments for wild type and RϪ30P mutant angiotensinogen were performed as described under "Materials and Methods." Immunoprecipitated and radiolabeled angiotensinogen was analyzed on SDS-PAGE followed by autoradiography. B, the autoradiogram shown in A was quantified on a PhosphorImager. Time courses of total (intracellular and secreted forms) and secreted amount of angiotensinogen are shown. The experiments were performed at least three times, and the data are expressed as mean Ϯ S.E. at each time point. C, analysis of angiotensinogen glycosylation. Aliquots of the protein immunoprecipitated during the pulse-chase experiments were subjected to endoglycosidase H, N-glycosidase F, or O-glycosidase treatment. D, processing of wild type and RϪ30P mutant angiotensinogen in vitro. In vitro transcription and translation were performed for each construct in the absence (0 eq) or presence (1, 2, 4, and 6 eq) of canine pancreatic microsomal membrane followed by the gel analysis.
AUGs, as does the human gene, but the position of the first mouse AUG is different from the human sequence. The signal peptide of rat angiotensinogen has only one AUG initiation codon.
To determine if the first AUG or second AUG could function as the translation initiation site of human angiotensinogen, three constructs that contain modified AUGs were generated. Each of the constructs, (i) wild type-ANG/pBS, (ii) ATC2-ANG/ pBS (the second ATG codon was changed to an ATC), and (iii) ATC1-ANG/pBS (the first ATG codon was changed to an ATC), was subjected to in vitro transcription and translation experiments (Fig. 4A). In some cases, the experiments were performed in the presence of canine microsomal membranes as a substitute for endoplasmic reticulum to study the protein processing.
As shown in Fig. 4B, the two forms of native angiotensinogen, ANG-1 and ANG-2, were produced from wild type/pBS in the absence of microsomal membrane due to the alternative initiation of translation at either one of the two in-frame AUG codons. Mutagenesis of the first or second ATG motifs abolished the synthesis of ATC1-ANG (ANG-2) and ATC2-ANG (ANG-1), respectively (Fig. 4B). PhosphorImager analysis showed that the ratio of ANG-1 to ANG-2 translated was approximately 5:1 when wild type-ANG/pBS was expressed. By adding microsomal membrane, four heterogeneous products (gp1, gp2, gp3, and gp4), which migrated more slowly than premature angiotensinogen on SDS-PAGE gels, were observed in all the constructs analyzed (Fig. 4B). These four products were predicted to be the glycosylated isoforms of translocated angiotensinogen (translocated ANG), in which the signal peptide was already cleaved by signal peptidase. This was confirmed by treatment with endoglycosidase H that could remove N-linked glycosylation. After the removal of N-linked sugar chain, the translational products from ATG-modified constructs showed the same major band (Fig. 4C).
Processing and Secretion of Angiotensinogen in COS-1 Cells-Typically, signal peptides exhibit one or more positively charged amino acids near their amino terminus, followed by a continuous stretch of several hydrophobic residues. If translation of angiotensinogen starts from the second AUG, then the translated molecule (ANG-2) lacks the positively charged amino acids, which are considered to be important for normal processing. By using an in vitro transcription, translation, and translocation system, ANG-2 was found to be translocated normally, and N-linked glycosylation occurred with the same efficiency as ANG-1. To ascertain the in vivo processing of all three transcripts, wild type-ANG/pMT2, ATC2-ANG/pMT2, or ATC1-ANG/pMT2 were transfected into COS-1 cells, and biosynthesis and secretion of each recombinant angiotensinogen was examined by pulse-chase experiments (Fig. 5A). Immunoprecipitated angiotensinogen from cells and media was separated on SDS-PAGE gels and quantified with a PhosphorImager. Although there was no major difference in half-life or secretion amount between wild type and transcript containing only the first AUG signal, the transcript containing only the second AUG codon led to quantitative reduction in both halflife and amount of angiotensinogen protein (Fig. 5B). Two Forms of Human Angiotensinogen Signal Peptide Alter the Subcellular Localization of EGFP-To examine further the function of either form of the signal peptide, recombinant an-giotensinogen signal peptide/EGFP chimeras were generated (Fig. 6A). COS-1 cells were transiently transfected with angiotensinogen signal peptide/EGFP plasmids as follows: wild type- FIG. 4. In vitro transcription, translation, and translocation experiments on human angiotensinogen. A, constructs for in vitro transcription, translation, and translocation experiments. The cDNA of human angiotensinogen was placed under the control of the T7 promoter in pBluescript SK(ϩ). Alterations of the first or second ATG codons were generated by site-directed mutagenesis of wild type-angiotensinogen (ANG)/pBS. The sequences of the 5Ј ends of the inserts are shown. In ATC1-ANG/pBS and ATC2-ANG/pBS, the first and second ATG codons are replaced by ATC, respectively. B, autoradiogram of [ 35 S]methionine-labeled angiotensinogen derived from wild type-ANG/pBS, ATC2-ANG/pBS, and ATC1-ANG/pBS. Transcription and translation were performed in the TNT T7 quick-coupled transcription/translation system in the absence or presence of canine pancreatic microsomal membranes. ATC1-ANG and ATC2-ANG lead to the formation of two distinct forms of nascent angiotensinogen. In the presence of canine microsomal membranes (6 U) four additional products (gp1, gp2, gp3, and gp4) are detected. C, glycosylation of wild type and mutant angiotensinogen. Wild type-ANG/pBS, ATC2-ANG/pBS and ATC1-ANG/pBS were transcribed and translated in the TNT T7 quick-coupled transcription/translation system in the presence of canine pancreatic microsomal membranes. Aliquots of the translation products were immunoprecipitated and treated with endoglycosidase H, followed by an SDS-PAGE gel.
FIG. 5. Pulse-chase analysis of angiotensinogen with or without alterations of either AUG codon. Pulsechase experiments were performed as described under "Materials and Methods." Radiolabeled angiotensinogen was immunoprecipitated and run on SDS-PAGE gel followed by autoradiography (A), and the quantification of the immunoprecipitated angiotensinogen signal was analyzed on a PhosphorImager (B). Time course of total (intracellular and secreted forms) and secreted amount of angiotensinogen are shown. The experiments were repeated at least three times, and the data are expressed as mean Ϯ S.E. at each time point. ANG/pEGFP, ATC2-ANG/pEGFP, and ATC1-ANG/pEGFP (Fig.  6A), and the subcellular localization of EGFP was detected by immunocytochemistry using anti-GFP antibody. The fluorescence distribution patterns of EGFP in COS-1 cells transfected with wild type-ANG/pEGFP, ATC2-ANG/pEGFP, and ATC1-ANG/pEGFP were a fine reticular network all over the cytoplasm and around the nucleus. This was suggestive of an endoplasmic reticulum staining pattern, and the localization of the signal was almost identical to that of BiP, which is localized in the endoplasmic reticulum (Fig. 6B, b-d) (19). In contrast, EGFP without signal peptide, as predicted, was not specifically localized to any subcellular compartment in transfected cells (Fig. 6B, a).
Western blot analysis was performed to detect EGFP secretion. COS-1 cells transfected with the chimeric constructs secreted EGFP, whereas COS-1 cells transfected with pEGFP did not secrete EGFP into the medium (Fig. 6C). These data together with the results on full-length angiotensinogen demonstrated that both of the amino-terminal peptides of 1st and 2nd ANG are functional as signal peptide as to drive the protein to the secretory pathway.

DISCUSSION
The RϪ30P Mutation Leads to Decreased Secretion of Angiotensinogen-Supporting evidence is accumulating that molecular variants of angiotensinogen constitute genetic susceptibility of human essential hypertension (1)(2)(3). The underlying mechanism mediating disease liability remains unknown. Rare variants in the gene may help in identifying the functional role of the protein in the pathogenesis of the disease. One such mutation at the renin cleavage site, L10F, was first identified in pre-eclamptic patients and caused altered kinetics in the reaction with renin (15). Here, we have identified a natural mutation at the Ϫ30 amino acid position of human angiotensinogen signal peptide, in which an arginine is replaced by a proline (RϪ30P). In the family studied, the presence of the RϪ30P mutation in the heterozygous state was associated with a modest decrease of plasma angiotensinogen concentration compared with control subjects in the family (Fig. 1B). In light of the small number of family members available for study, the issue of a possible relationship between mutation and essential FIG. 6. Transient expression of EGFP with or without fusion with the human angiotensinogen signal peptide/EGFP. A, chimeric constructs between human angiotensinogen signal peptide and EGFP. The coding region of human angiotensinogen signal peptide was added at the amino terminus of EGFP cDNA (wild type-angiotensinogen (ANG)/pEGFP). Site-directed mutagenesis was carried out to obtain ATC1-ANG/ pEGFP and ATC2-ANG/pEGFP. B, immunolocalization of GFP after transfection. COS-1 cells were transfected with pEGFP vector (a), wild type-ANG/pEGFP (b), ATC2-ANG/pEGFP (c), or ATC1-ANG/pEGFP (d). EGFP was visualized by fluorescence. As a marker of endoplasmic reticulum, BiP stained with monoclonal anti-BiP antibody is shown in boxed insets. C, Western blot analysis of GFP after transfection. COS-1 cells were transfected with GFP vectors as in B, and the cell lysate and medium were then harvested for protein blotting. GFP was incubated with a specific polyclonal antibody, followed by visualization using an ECL system. hypertension could not be addressed.
The RϪ30P mutation occurs in a cluster of basic residues at the amino terminus of the signal peptide. This substitution could affect the translocation process of human angiotensinogen, as the positively charged amino-terminal region of the signal peptide has been proposed to have an important role at an initial step of translocation (20,21). Pulse-chase experiments revealed that the RϪ30P mutation led to reduced secretion of angiotensinogen by reducing the net stability of the protein. Thus far, other signal peptide mutations have been reported that were mainly cleavage site mutations that could directly alter protein processing (22). Few cases have indicated that natural mutations in the hydrophobic core of the signal peptide have a direct correlation with defective secretion or associated pathological states such as familial isolated hypoparathyroidism due to prepro-parathyroid hormone gene mutation (23) and Crigler-Najjar type II due to bilirubin UDPglucuronosyltransferase gene mutation (24). In these cases, mutations in the hydrophobic core of the signal sequence resulted in defective signal peptide cleavage, leading to little or no secretion. In the case of the RϪ30P mutation, signal peptide cleavage did occur, as shown in Fig. 2, which might explain the relatively mild secretion defect.
Two Forms of Native Human Angiotensinogen Are Generated from Distinct AUG Initiation Codons-The sequence at the 5Ј end of human angiotensinogen mRNA revealed two AUG codons in the same reading frame. In the present study, we show that two forms of human angiotensinogen are translated by the alternative use of either one of two in-frame AUG initiation codons in both cell-free and cellular systems. Eukaryotic mRNAs generally adhere to the first AUG rule: in most cases the AUG codon nearest the 5Ј end is the unique site for initiation of translation (18,25). Two escape mechanisms from the first AUG rule have been considered (26). First, reinitiation at a downstream AUG codon may be possible when the 5Ј proximal AUG codon is followed shortly by a termination codon (27). Second, a "leaky scanning" mechanism has been reported. When the first AUG codon resides close to the cap site (within 12 nucleotides) or when the first AUG has an unfavorable context such as when the critical purine at position Ϫ3 or guanine at position ϩ4 is absent, initiation could occur from the second AUG codon (12, 18, 28 -32). In human angiotensinogen where the first AUG site contains a partial Kozak's consensus initiation sequence, position Ϫ3 is purine and ϩ4 is C, whereas the second AUG site has a perfect Kozak's consensus initiation sequence. Leaky scanning might permit the synthesis of two forms of angiotensinogen, but we could not confirm the presence of two forms of angiotensinogen in vivo because the signal peptide is cleaved from the secretory protein while it is still growing on the ribosome.
Two Functional Forms of Human Angiotensinogen Aminoterminal Signal Peptides-Protein targeting signals have been shown to be important both in directing proteins to and maintaining them in the target organelles or cell membrane. Typically, the signal peptide of a secretory protein has one or more positively charged amino acids near its amino terminus, followed by a continuous stretch of several hydrophobic residues (5,6). The deduced amino acid sequence of human angiotensinogen shows that the longer form of the signal peptide has three positively charged amino acids, Arg-Lys-Arg, followed by a continuous stretch of hydrophobic residues, whereas the shorter form lacks the positively charged amino acids in the amino terminus (Fig. 1B). The lack of basic amino acids could affect the translocation of human angiotensinogen, as the positively charged amino-terminal region of the signal peptide has been proposed to have an important role in the initial step of translocation (20,21). The 5Ј end mRNA sequence and the deduced amino acid sequence of human (7,8), mouse (33,34), and rat (35,36) angiotensinogen are compared in Fig. 1B. Human and mouse angiotensinogen have two AUG codons in the same reading frame, whereas the signal sequence of rat angiotensinogen contained only one. The region between residue Ϫ24 and Ϫ1 was highly conserved between human, mouse, and rat. Thus, it may be that the sequence following the second AUG in the human gene contains all the information necessary for secretion of angiotensinogen. Our studies show that whereas both forms of the angiotensinogen signal peptide can promote appropriate translocation both in vitro and in vivo, the additional segment present in the longer form may lead to increased efficiency of the process. That the RϪ30P mutation affects the net secretion rate of the protein suggests that the cluster of basic residues at the amino terminus of AGT mRNA may indeed improve the efficiency of translocation of the nascent protein. Our data, however, do not support the hypothesis that this cluster induces distinct subcellular sorting of the protein.