A targeting sequence for dense secretory granules resides in the active renin protein moiety of human preprorenin.

A Targeting Sequence for Dense Secretory Granules Resides in the Active Renin Protein Moiety of Human Preprorenin

William N. Chu, John D. Baxter, and Timothy L. Reudelhuber Department of Medicine Metabolic Research Unit and the Graduate Program in Endocrinology University of California San Francisco, California 94143

Human renin plays an important role in blood pressure homeostasis and is secreted in a regulated manner from the juxtaglomerular apparatus of the kidney in response to various physiological stimuli. Many aspects of the regulated release of renin (including accurate processing of prorenin to renin, subcellular targeting of renin to dense secretory granules, and regulated release of active renin) can be reproduced in mouse pituitary AtT-20 cells transfected with a human preprorenin expression vector. Using protein engineering, we have attempted to define the roles of various structures in prorenin that affect its production and trafficking to dense core secretory granules, resulting in its activation and regulated secretion. Replacement of the native signal peptide of human preprorenin with that of a constitutively secreted protein (immunoglobulin M) had no apparent effect on either the constitutive secretion of prorenin or the regulated secretion of active renin in transfected AtT-20 cells. Removal of the pro segment resulted in a marked reduction in total renin secretion, but did not prevent renin from entering the regulated secretory pathway. Single or combined mutations in the two glycosylation sites of human renin did not prevent its regulated secretion; however, the complete elimination of glycosylation resulted in a significant increase in the ratio of renin/prorenin secreted by the transfected cells. Thus, these results suggest that 1) at least one of the sequences that target human renin to dense secretory granules lies within the protein moiety of active renin; 2) the presence of the pro segment is important for efficient prorenin and renin production; and 3) glycosylation can quantitatively affect the proportion of active renin secreted. (Molecular Endocrinology 4: 1905-1913, 1990)

sin-l (Al). Al is subsequently converted by angiotensinconverting enzyme to the octapeptide angiotensin-ll (All), a potent vasoconstrictor and regulator of salt balance. The limiting reagent in the circulating reninangiotensin system appears to be renin released from the kidney (reviewed in Ref. 1), although components of this system present in various other organs and tissues may also direct local generation of All (reviewed in Ref. 2). Inappropriate activity of the renin-angiotensin system is implicated in the pathogenesis of certain cardiovascular disorders, including hypertension, congestive heart failure, and chronic renal dysfunction (1-5). The primary translation^ product of renin mRNA, preprorenin, is cotranslationally processed to the inactive zymogen, prorenin. The pre-, or signal, peptide sequence of 23 amino acids is cleaved off to yield prorenin (6, 7), which is predominantly inactive. The pro segment contains 43 amino acids and, by analogy to the related enzyme pepsinogen (8), probably occupies a cleft in the enzyme harboring its active site and also wraps around the surface of the protein. Cleavage of the pro segment irreversibly activates renin and is likely to cause the enzyme to undergo a major structural change, as is the case with pepsinogen (8). It is not known whether the pro segment plays some other role in the initial folding or stability of prorenin or its secretion. Since the 43-amino acid pro segment emerges biosynthetically into the lumen of the endoplasmic reticulum before the mature renin peptide sequence, it is possible that the propeptide domain folds independently first and then provides a scaffold around which the rest of the molecule subsequently folds. If this were the case, renin devoid of its pro region would be incapable of folding into its active conformation. A portion of the prorenin produced in the kidney is sorted to secretory granules, where it is processed, and the active renin is stored and released in response to secretagogues such as through /3-adrenergic stimuli (1). Recombinant human and mouse submaxillary gland (SMG) prorenins are handled in a similar fashion by transfected mouse pituitary AtT-20 cells, which process

INTRODUCTION Renin is an aspartyl protease which cleaves circulating angiotensinogen to release the decapeptide angioten0888-8809/90/1905-1913S02.00/0 Molecular Endocrinology Copyright © 1990 by The Endocrine Society

1905

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 21 November 2015. at 12:47 For personal use only. No other uses without permission. . All rights reserved.

VoUNo. 12

MOL ENDO-1990 1906

endogenous POMC to ACTH (7-10). By contrast, human prorenin expressed in cultured Chinese hamster ovary (CHO) cells, which lack such a regulated secretory pathway secrete exclusively prorenin in a constitutive manner (11). These results suggest that prorenin contains a signal, which is recognized by the cellular machinery (across tissue and species barriers), that directs it to dense secretory granules. While subcellular sorting signals have been characterized for various proteins residing in nuclear, mitochondrial, peroxisomal, glycosomal, and lysosomal subcellular compartments (12-16), the molecular signals that sort prorenin or any other protein to dense core secretory granules are not presently known. It appears, however, that some structure within the protein must dictate their sorting, since not all proteins expressed in a cell containing a regulated secretory pathway are secreted in a regulated manner (17). In addition, these signals appear to be dominant, since fusion of a protein secreted by the regulated pathway to a protein secreted by the constitutive pathway results in regulated secretion of the hybrid protein (18). Dense core secretory granule sorting signals could conceivably be in the form of some type of posttranslational modification or some as yet undiscerned feature of the tertiary structure of the protein. Although the physical segregation of proteins to be secreted by either the regulated or constitutive pathways has been reported to occur as late as the frans-Golgi network (19), it is not known how early the commitment to this pathway of secretion is made (e.g. it could occur as early as translocation into the endoplasmic reticulum). In addition, while the signals for dense secretory granule sorting appear to be recognized by components of the secretory machinery across tissue and species lines (20), the precise location of such sorting signals may vary among proteins destined for the regulated pathway. This latter point is illustrated in the case of prosomatostatin, where the dense core secretory granule sorting signal appears to be in the pro segment of the hormone (21), whereas removal of the pro segment of either chymotrypsin (22) or the internal C-peptide of insulin (23) does not prevent regulated secretion of the recombinant proteins. Two consensus sites for N-linked glycosylation (Asnx-Thr) have been identified in human renin, located 5 and 75 amino acids to the carboxy side of the prorenin to renin cleavage site. While the glycosylation status of a number of proteins profoundly affects their secretion (reviewed in Ref. 24), it is clear that prorenin can be secreted from heterologous cultured cells transfected with expression vectors encoding glycosylation site mutated human prorenins (25) or the nonglycosylated mouse SMG prorenin (10). In addition, AtT-20 cells transfected with a human preprorenin expression vector are still competent for secretion of prorenin in the presence of tunicamycin, an inhibitor of N-linked glycosylation (26). In this latter study it was also proposed that the lack of glycosylation might decrease the activation of prorenin. Thus, more studies are needed to

define the role of N-linked glycosylation in the intracellular trafficking and processing of human prorenin. In the present report we describe the use of protein engineering to investigate the molecular signals contained in human preprorenin that direct its sorting to dense core secretory granules, the role of the pro segment in prorenin production, and the role of glycosylation in the intracellular trafficking and release of prorenin and renin. The results suggest that a dense core secretory granule sorting signal is contained within the body of the active renin molecule, that neither the native signal peptide nor the pro segment is essential for granule sorting, deletion of the pro segment decreases prorenin production, and that glycosylation quantitatively influences the pattern of prorenin and renin secretion.

RESULTS Secretion of Native Prorenin and Renin after Transient Transfection into AtT-20 Cells The human preprorenin expression vector pRhRHOO (see Materials and Methods and Fig. 1 A) directed synthesis of the protein at sufficiently high levels in transfected cells that analysis could be carried out using transient transfection. The results of these experiments are reported in two ways. Table 1 lists the mean of the absolute levels of secretion of the various prorenins and renins from transfected AtT-20 cells; however, since these levels varied from experiment to experiment due to factors such as the efficiency of transfection, the secretagogue responses of the various recombinant proteins are depicted in Fig. 2 as the relative secretion compared within each individual experiment to the level of prorenin secreted by the transfected cells in the absence of secretagogue. The results obtained by transient transfection of the expression vector encoding the native human preprorenin into AtT-20 cells (Table 1 and Fig. 2) confirm our earlier findings using a stably transfected expression vector (7). In the absence of secretagogue, the transfected cells secrete predominantly prorenin during the 4-h assay period (Table 1 and Fig. 2, NATIVE, - FORSKOLIN); treatment of transfected cells with forskolin, which stimulates an increase in intracellular cAMP levels and a release of storage granules, causes a rapid and selective stimulation of the release of active renin (Table 1 and Fig. 2, NATIVE, + FORSKOLIN). Role of the Signal Peptide Sequence in Regulated Secretion of Human Renin To assess the role of the signal peptide sequence in renin subcellular sorting, the native renin signal peptide was replaced with the corresponding signal peptide from immunoglobulin M (IgM) heavy chain (30), a protein that is constitutively secreted in AtT-20 cells (17) to form the expression vector /*Ren (Fig. 1B; see Materials


Determinants of Regulated Secretion in Human Renin

1907

A.

CLYCOSVLATION SITU MUTATIONS

s ~J

Cly Aia GCC

human prtprorcniB cDNA

B.

Cl,

ccc

6 Tht

74 III!

73 Am

74 Cl>

Str Thr TCC ACC Ml! Ser Cly CAC TCC CCA

Clttia

4

hRaili •!(>•) ptplldf

PRO-SECMENT DELETION

|ci

Leu Cl CTCCAC

ACC

TTT

GCC

I

2

i

CTC

AC0

CTT

ItU ugnat ptpude

Fig. 1. A, Schematic Representation of the Human Preprorenin Expression Cassette of the Vector pRhR1100 The arrow represents the transcription start site in the Rous sarcoma virus long terminal repeat (RSV-LTR). STOP signifies the translations stop codon (TGA) and (A)n represents the location of the polyadenylation signals. B-D, Description of the mutations introduced into the human prorenin cDNA. Nucleotides and resulting encoded amino acids that have been altered are shown in italics. The enclosed C residue in D represents a conservative nucleotide difference between the cDNA used in this study and that reported by Imai et al. (41). See text for additional details.

Table 1. Absolute Secretion Levels of Prorenin and Renin in AtT20 Cells Transfected with the Various Prorenin Expression Vectors Forskolin

Native

n Prorenin Renin Prorenin + Renin Prorenin Renin Prorenin + Renin

8 202.5 ± 1 1 . 8 55.5 ± 4.0 258 ± 13.3

+

ji-Ren

ASN5-SER

ASN75-SER

ASN5.75-SER

8 83.4 ± 17.6 83.0 ±14.8 166.3 ± 31.7

Pro-Del

12 ND 28.6 ± 6.0 28.6 ± 6.0

153.5 ± 11.8 52.5 ± 7.2

123.8 ± 6 . 5 28.7 ± 5.9

217.5 ±9.8

206 ±18.5 159.6 ±10.3

152.5 ±7.1

110.6 ± 14.7 55.4 ± 6.5 166 ± 20.3

133 ±8.5

129.4 ±16.9

97.5 ±16.9

ND

157.6 ± 8 375.1 ± 16.7

146.4 ±14.3 306 ± 23.8

120.5 ± 6.0 253.5 ±11.2

161.5 ± 17.9 291 ± 34.6

258.4 ± 48.6 355.8 ± 64.6

36.5 ± 23.3 36.5 ± 23.3

All values are in nanograms of angiotensin-l generated per ml culture supernatant/h (see Materials and Methods) and are calculated as the mean ± SEM. n, Number of independent transfections analyzed; ND, not determined.

and Methods). Transfection of uRen into AtT-20 cells resulted in the release of both prorenin and renin. Further, the total level of prorenin and renin produced (306 ng/ml-h; n = 8; Table 1) was comparable to that for the native protein (375 ng/ml-h; n = 8; Table 1), implying that the signal peptide switch did not have a major influence on prorenin secretion. Further, the IgM signal peptide replacement had no effect on the ability of forskolin to stimulate the release of active renin compared to the native preprorenin (Fig. 2, compare NATIVE and uRen); i.e. forskolin preferentially stimulates the release of active renin (Fig. 2, compare solid bars - and +). This result suggests that the protein sequences directing human preprorenin to secretory granules in AtT-20 cells do not reside within the native preprorenin signal peptide. Role of the Prosegment and the Body of Renin in Regulated Secretion of Human Renin To test the role of the pro segment in prorenin production and in directing human preprorenin to secretory

granules, a deletion of the pro segment was made by site-directed mutagenesis to generate prerenin (Fig. 1D, Pro-Del). Transfection of AtT-20 cells with this recombinant construction led, as expected, to the exclusive secretion of active renin (Table 1). Forskolin treatment of the transfected cells resulted in a relative increase in active renin release of 1.37 ± 0.08-fold (Fig. 2, ProDel). By comparison, total secretion of the native protein (prorenin and active renin) was increased 1.47 ± 0 . 1 1 fold by forskolin treatment (Fig. 2, NATIVE). Thus, these are the expected results if human prorenin devoid of its pro segment is being released via the regulated pathway in response to secretagogue treatment as well as being released simultaneously by the constitutive pathway. To further substantiate this conclusion, cell lines were established which were stably transfected with expression vectors for either native or pro segment-deleted prorenin, and the kinetics of prorenin/renin secretion were further characterized by pulse-chase. Confirming our previous results (7) and consistent with activity measurements (Fig. 2 and Table 1), AtT-


MOL ENDO-1990 1908

Vol4No. 12

FOLD INDUCTION 4.00 3.50 3.00 2.50 V~A PRORENIN

RENIN

2.00 1.50 1.00 0.50 0.00 FORSKOLIN

—

+

NATIVE

—

-f-

M Ren

Asn75>-Ser

Asn5,75^Ser

Pro-Del.

Fig. 2. Effect of Secretagogue on Secretion of Native or Mutated Prorenin and Renin from Transfected AtT-20 Cells • , Active renin; H, prorenin measured in culture supernatants incubated for 4 h in the absence of secretagogue (-) and for a subsequent 4 h in the presence of 10 \m forskolin (+). Data are recalculated from the individual experiments used to derive the data in Table 1. Levels of secretion are represented as the mean of 6-12 independent transfections ± SEM, normalized to the level of prorenin secreted in the same experiment in the absence of secretagogue (set at 1.0), except in the case of Pro-Del, where secretion is normalized to the level of active renin released in the secretagogue-unstimulated state.

20 cells transfected with an expression vector encoding native preprorenin secrete predominantly prorenin in the overnight labeling medium and the first 3-h chase period (Fig. 3, Native). However, by the second 3-h chase period, most of the labeled prorenin has been chased out of the constitutive pathway (Fig. 3, Native, Medium - ) , and administration of forskolin increases the amount of labeled renin released, but not the amount of labeled prorenin released (Fig. 3, Native, Medium +). Examination of the pattern of immunoprecipitable prorenin/renin remaining in the cells reflects the expected selective decrease in intracellular renin levels in response to secretagogue treatment (Fig. 3, Native, Cells). Thus, pulse-chase analysis is capable of discriminating between the rapidly turning over pool of constitutively secreted prorenin and the stored pool of active renin. A similar analysis of a pool of cells stably transfected with the Pro-Del expression vector reveals that the release of pulse-labeled active renin into the culture medium is consistently stimulated in three independent experiments by treatment of the cells with forskolin (Fig. 3, Prosegment deletion, Medium), substantiating the conclusion that human renin is capable of being targeted to the regulated secretory pathway in the absence of the pro segment. Of additional note is that the total renin produced by the Pro-Del vector was much lower than that with the vector that expressed the native preprorenin in both

the transiently transfected (Table 1) and stably transfected (Fig. 3; note different times of exposure) cell lines. This was also true when the Pro-Del expression was compared to that directed by the vectors expressing the other mutated preprorenins (Table 1). These results may imply that either the pro segment DNA or protein sequences are important in some other way for optimal renin production. Effect of Glycosylation Site Mutation on Renin Secretion To test for the role of the glycosyl residues in regulated secretion of renin, the two glycosylation sites in preprorenin were altered, either individually or as a pair, by using site-directed mutagenesis (Fig. 1C; see Materials and Methods), and the various mutated expression vectors were transfected into AtT-20 cells and analyzed for regulated secretion of active renin. The results, presented in Table 1 and Fig. 2, reveal the following. First, all of the glycosylation site mutated preprorenins are efficiently secreted and processed to active renin, as evidenced by the presence of both prorenin and active renin in the supernatants of the transfected cultures (Table 1, Asn5-Ser, Asn75-Ser, Asn5.75-Ser). Second, the glycosylation site mutated renins are targeted to the regulated secretory pathway, as evidenced by the selective secretagogue-dependent stimulation of



± Forskohn treatment Overnight labelling

I Native

#1. Prosegment deletion

#2.

#3.

3 hr chase

II

r r *"*" "* %* -%»

Medium [

| -

Cells 11

+

|

-

+

~ZL T

-II -21

• •

Forskolin •*- Prorenin •*- Renin

• - Renin

• - Renin

• - Renin

Fig. 3. Pulse-Chase Analysis of Prorenin/Renin Secretion from Stably Transfected AtT-20 Cells AtT-20 cells that had been stably transfected with expression vectors encoding either native or pro segment-deleted prorenin (Pro-Del) were plated at 50% confluence in parallel wells of a 12-well dish. After 24 h, the medium was replaced with 250 n\ methionine-free medium containing 125 fiC\ [35S] methionine (labeling medium). After incubation of the cells for 16 h, the labeling medium was collected and replaced with fresh medium lacking radiolabel. This medium was collected after 3 h (3-h chase) and replaced with 250 n\ fresh medium either containing (+) or lacking (-) 10 /IM forskolin. After an additional 3 h, this medium was collected, and the cells were lysed in 250 p\ detergent mix (7). All of the samples were then immunoprecipitated with monospecific antibody against human prorenin, and the immunoprecipitated products were separated by SDS-PAGE. Note that the order of the pairs remains consistent throughout each set {i.e. the left sample in each pair represents supernatant from the same well in each panel of the corresponding set). Prosegment deletion 1-3 represents three separate experiments carried out with a single pool of cells stably transfected with Pro-Del. The positions of migration of prorenin and renin were determined by coelectrophoresis of previously characterized standards. Autoradiographic exposure time was 48 h for the set of native samples and 2 weeks for all of the pro segment deletion samples.

the active forms (Table 1, compare - and + Forskolin). Third, the complete removal of the glycosylation sites leads to a significant increase in the ratio of renin to prorenin released in both the absence and presence of secretagogue (Table 1, compare Asn5.75-Ser, - and + forskolin, to Native, - and + forskolin, respectively). A similar trend was observed with the single glycosylation mutants, although the results were not statistically significant. By contrast, transfection of the double glycosylation mutant into CHO cells, which lack detectable renin-processing activity and the ability to secrete renin in a regulated manner (11), led to constitutive secretion exclusively of prorenin whose accumulation in the medium was linear for a period of up to 96 h and whose susceptibility to trypsin was indistinguishable from that of the native prorenin (data not shown). Taken together, these results suggest that the renin activity secreted constitutively from AtT-20 cells transfected with the glycosylation mutations is not simply due to spurious activation of prorenin by destabilization of the pro segment in a mechanism similar to that responsible

1909

for acid activation (31), resulting in either spontaneous renin activity or increased susceptibility to activation in the culture supernatants. Furthermore, we have found no difference in the intracellular/extracellular ratio between the deglycosylated and native prorenin in transfected CHO cells (DeNoto, F. M., and T. L. Reudelhuber, unpublished observations), suggesting that the apparent increase in renin activity in the supernatant of AtT-20 cells transfected with the deglycosylated human prorenins is not due to retention of prorenin by the cells. Thus, whereas glycosylation of human renin is not critical for its processing, secretion, or sorting to secretory granules, it does appear to quantitatively affect the proportion of active renin secreted in both the unstimulated and secretagogue-stimulated transfected AtT-20 cells.

DISCUSSION

We have sought to define the roles of various structures on human preprorenin that affect its production and intracellular trafficking in mouse pituitary AtT-20 cells. To do this, we first established that a human preprorenin expression vector could express enough prorenin and renin after transient transfection of AtT-20 cells to obtain reliable measurements. The results with a vector expressing the human preprorenin sequence were similar to those we (7) and others (9) obtained with stably transfected cells. Preprorenin was processed to active renin by these cells, and active renin release was selectively increased by a secretagogue, indicating that renin was correctly targeted to the regulated secretory pathway. By contrast, transiently transfected CHO cells released only prorenin and did not increase their release of the protein after secretagogue treatment. It was first necessary to exclude a role for the pre sequence of preprorenin in targeting renin to the regulated secretory pathway, both because a role for this sequence has not been generally excluded and to interpret subsequent results. When the pre sequence of preprorenin was replaced by the signal peptide sequence of the IgM heavy chain that is released constitutively, renin was released in a manner identical to the native sequences. Thus, the signal peptide directs neither the processing of prorenin to renin nor the targeting of the protein to the regulated secretory pathway. To assess the role of the pro segment in targeting of renin to the regulated secretory pathway, a gene was constructed in which the pro segment was deleted and the signal peptide sequences were linked directly to the sequence encoding renin. Both the pattern of secretion of active renin by AtT-20 cells transiently transfected with the Pro-Del expression vector (Fig. 2 and Table 1) and the pattern of secretion of pulse-labeled active renin in AtT-20 cells stably transfected with Pro-Del (Fig. 3) were consistent with the ability of renin to enter the regulated secretory pathway in the absence of the pro segment. However, because cells transfected with the


MOL ENDO-1990 1910

Pro-Del mutant have the capacity to secrete renin by both the constitutive and regulated pathways, we cannot at present accurately quantitate the efficiency of sorting of pro segment-deleted prorenin to the dense core secretory granules, and therefore, we cannot rule out that removal of the pro segment decreases the efficiency of this process. Nevertheless, it is clear from these results that neither the presence nor the removal of the propeptide appear to be essential for targeting of prorenin to dense core secretory granules. This latter point is reinforced by the recent report that expression of a noncleavable mutant of human prorenin in AtT-20 cells leads to regulated' release of prorenin (27). The findings that the signal peptide sequence of preprorenin does not direct renin to the regulated secretory pathway and that a protein containing only renin amino acids is directed to this pathway suggest that the body of renin contains the sequences that direct renin to the regulated secretory pathway. Further, these sequences are probably not ones that are affected by removal of the pro segment from the active site and its wrapping around the body of renin. In the absence of obvious linear sequence homology between the various proteins secreted by the regulated secretory pathway, it seems likely that more complicated signals, such as surface patch structures, may be recognized by the cell. We are currently investigating this possibility by scanning deletion analysis of the active renin moiety. Although the pro segment-deleted renin was correctly targeted to the regulated secretory pathway, the total renin production by the vector expressing this protein in both transiently (Table 1) and stably (Fig. 3) transfected AtT-20 cells was only 10-15% of the total (renin plus prorenin) produced by cells expressing the native preprorenin. While it has been reported that pro segment-deleted human prorenin can be efficiently expressed in subclones of stably transfected mouse myeloma cells (28), a more recent report suggests that removal of the pro segment-coding sequences has a detrimental effect on the expression of renin in transfected CHO cells that could not be explained by a difference in expression at the RNA level (29). Taken together, these results imply that the pro segment is in some way important for efficient prorenin production. Further studies will be needed to determine the mechanisms of this decreased expression. This could involve influences on the folding or stability of the translational product or its secretion. It is conceivable that the pro segment of renin could be critical for correct folding of the protein, since the pro segment occupies part of the surface of renin and has a dramatic effect on its conformation. Of particular interest is the finding that elimination of the glycosylation sites in human preprorenin led to an increase in both the constitutive and secretagoguestimulated secretion of active renin vs. prorenin in transfected AtT-20 cells. Paul et al. (26) treated renin-secreting AtT-20 cells with tunicamycin to inhibit glycosylation and found an increase in the ratio of secreted to cellular renin. They concluded that the intracellular

Vol4No. 12

transit time for prorenin in treated cells was greatly reduced and proposed that activation of prorenin would consequently be reduced in the absence of glycosylation, a result that is not reflected in the current study. While we have no direct explanation for the difference in these results, it could suggest an effect of tunicamycin on proteins other than prorenin, which would affect the overall processing ability of the transfected cells. It has recently been reported that human renin acquires phosphomannose residues that direct its targeting to lysosomes in injected Xenopus oocytes and, to a much lesser extent, in mammalian L-cells (32). In mammalian cells, much of the renin targeted to lysosomes appears to be degraded (32). One possible explanation for the relative increase in the secretion of active renin in the cells transfected with the glycosylation site mutated renins could be that in the absence of glycosylation, less of the renin is diverted to lysosomes, resulting in a greater proportion of the protein being targeted to the dense secretory granules. Such an explanation would predict that removal of the glycosylation sites would lead to an increase in the absolute levels or total protein (prorenin plus renin) secreted in response to secretagogue. Although the absolute secretion levels are subject to variations in transfection efficiency (see Materials and Methods), we saw no such trend in multiple experiments (Table 1, compare prorenin + renin in Native and Asn5.75-Ser). Alternatively, removal of the glycosylation sites may allow the AtT-20 cell to more efficiently process prorenin to renin and to secrete more processed active renin via both the constitutive and regulated pathways. Pratt et al. (33) have reported that mouse SMG renin (a product of the Ren-2 gene which is not present in all mouse strains), which is not glycosylated, displays a rapid removal of the pro segment concomitant with the constitutive secretion of the processed protein. In addition, transfection of AtT-20 cells with an expression vector encoding mouse SMG preprorenin results in a pattern of renin and prorenin secretion that is strikingly similar to that seen with the Asn5.75-Ser mutation (compare Fig. 2, Asn5.75-Ser to Ref. 10, Fig. 5). While in the present experiments the total amount of prorenin and renin secreted did not seem to increase with removal of the glycosylation sites, the increase in secretagoguestimulated release of renin seen with this mutant was balanced by a corresponding decrease in the absolute level of prorenin secretion (Table 1, compare + Forskolin, Native, and Asn5.75-Ser). One possible explanation for these results is that a greater proportion of the glycosylation site mutated prorenin is being processed to active renin. These findings are consistent with the proposal that one role of the glycosylation sites may be to protect renin from activation by proteases before encountering the acidic environment of the maturing secretory granule. Although we did not find the nonglycosylated prorenin to be either more sensitive to trypsin digestion or less stable in culture supernatants than the native protein, our results differ somewhat in this last



regard with those reported by Hori et al. (25). While we do not yet understand the reason for this discrepancy, it could be due to differences in either the cell types transfected in the two studies (CHO vs. COS), the mutations introduced in the prorenin (Ser vs. Ala), or such variables as the collection or incubation times. Regardless of the stability of the mutated prorenin in the culture supernatants, removal of the glycosylation sites might still make the prorenin more sensitive to cleavage by an intracellular processing enzyme, perhaps distinct from the enzyme that normally removes the pro segment. Neutral processing activity has been described for pancreatic islet cells, which process proinsulin to insulin (34). Finally, removal of the glycosyl residues may increase the efficiency of targeting of prorenin to dense core secretory granules. This possibility is consistent with the apparent decrease in constitutively secretory prorenin as the amount of active renin secreted by the regulated pathway increases (Table 1, compare Native to Asn5.75-Ser). Such an effect might be seen if the glycosyl residues partially occlude the targeting signal of prorenin. In light of the proposed heterogeneity of glycosylation of mammalian renins (35), altered pathways of subcellular sorting or sensitivity to endogenous prohormone-processing enzymes might provide an additional mechanism by which the physiological regulation of renin activity could be accomplished.

MATERIALS AND METHODS Recombinant Plasmid Construction All plasmid constructions were carried out using standard recombinant DNA techniques, as described by Maniatis et al. (36). The human preprorenin expression vector pRhRHOO (Fig. 1A) was constructed as follows. The human renin cDNA was excised from the expression vector pHR14 (11) by digestion with restriction enzymes A/col (cuts immediately preceding the initiating methionine) and Sacl (cuts between the termination codon and the polyadenylation signal). Unique Hind\\\ (5'end) and SamHI (3'-end) sites were created flanking the cDNA sequences by linker addition. This fragment was then inserted down-stream of the promoter/enhancer of the Rous sarcoma virus long terminal repeat (RSV-LTR) (37) in plasmid pUC9 and is fused at its 3' end to the naturally occurring SamHI site in the second exon on the rabbit (9-globin gene (38). The resulting vector (Fig. 1 A) contains a potent promoter/enhancer which is competent in a broad variety of cell types (37), driving the expression of the human preprorenin-encoding cDNA. The RNA transcript contains the second intron and polyadenylation signal of the rabbit /3-globin gene for increased RNA stability (38), although the protein translation product ends with the natural humanrenin termination codon (Fig. 1A, Stop). Site-directed mutations in the human renin cDNA were carried out by oligonucleotide mutagenesis by a modification of the method of Zoller and Smith (39), using a mutagenesis kit from Amersham (Arlington Heights, IL) according to the manufacturer's instructions. Substitution of the native preprorenin signal peptide with the signal peptide of IgM heavy chain (30) was carried out by first introducing a unique Pst\ site into the preprorenin cDNA just 3' to the signal peptide cleavage site (Fig. 1B) using the following oligonucleotide (the engineered Pst\ site is italicized): CTCCCGACAC7GC4GACCACCTTTAA.

1911

A unique H/ndlll site was introduced by linker addition at the 5' end of the IgM cDNA to make it compatible with the pRhRHOO expression vector. It was then possible to make use of a naturally occurring Pst\ site just 3' to the IgM signal cleavage site to switch the analogous signal peptide encoding H/ndlll-Psfl fragments of the two engineered cDNAs (Fig. 1B). Glycosylation site mutations were introduced as described in Fig. 1C, using the following oligonucleotides: CAGTTGGCTCCACCACCTCC (Asn5-Ser) and CTACAAGCACTCCGGAACAGAA (Asn75-Ser). Deletion of the pro segment was carried out by "looping out" the native pro segment encoding cDNA (Fig. 1D) using the following oligonucleotide: GTACCTTTGGTCTGACACTTG. Mutated cDNAs were inserted in place of the native sequences of pRhRHOO using the unique SamHI and H/ndlll sites (Fig. 1A). All recombinant plasmid constructions were verified by DNA sequencing using the Sequenase kit (U.S. Biochemical Corp., Cleveland, OH). Maintenance and Transfection of Cell Cultures Mouse pituitary AtT-20 cells and CHO cells were maintained in monolayer culture in Dulbecco's Modified Eagle's Medium (DMEM) H21 supplemented with 10% fetal calf serum (FCS) and were grown at 37 C in an atmosphere of 12% CO 2 -88% ambient air (AtT-20 cells) or 5% CO 2 -95% ambient air (CHO cells). For transient transfection analysis, cells were replated 4-16 h before transfection at approximately 50% confluence in individual wells of a six-well culture dish (Costar, Cambridge, MA) and were transfected with 1.5 ng of the appropriate pRhRHO expression vector using the calcium phosphate method (40). After 20-24 h, the culture medium was removed, and the cells were shocked for 3 min with 15% glycerol in DMEM, washed twice with DMEM containing 10% FCS, and incubated for an additional 24 h in the same medium before prorenin/renin secretion was assayed. Medium was again removed 40-48 h after transfection, cells were washed twice with 1 ml normal medium (DMEM H21 plus 10% FCS), and 1 ml fresh prewarmed medium was placed on the cells. After 4 h, this medium was removed and used as the secretagogueunstimulated control (-). The cultures were again washed and incubated with fresh medium containing 10 HM forskolin for 4 h. This medium was then collected and used as the secretagogue-stimulated samples (+). Use of tandem assays of single cultures as opposed to assaying parallel cultures reduced variability but did not qualitatively change the results obtained. Stable transfection of AtT-20 cells with both the native and Pro-Del prorenin expression vectors, pulse-labeling analyses, immunoprecipitation of culture supernatants, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of immunoprecipitated proteins were carried out as previously described (7), except that G418-resistant pools of transfected cells were used for pulse-labeling analysis. After the second 3-h chase period, intracellular prorenin/renin was immunoprecipitated by lysing the cell monolayer in a detergent mix containing 0.4% sodium deoxycholate, 1 % Nonidet P-40, 0.1% sodium dodecyl sulfate, 10 mM EDTA, and 10 mM Tris (pH 7.5). All of the culture supernatants were immunoprecipitated using 2 n\ monospecific antihuman prorenin antibody and 5 n\ wet packed volume protein-A-Sepharose CL4B (Sigma Chemical Co., St. Louis, MO). The washed beads were lyophilized to dryness before resuspension in SDS-PAGE sample buffer. Active renin (direct measurements) and total renin (measurement after treating supernatants with 150 ^g/ml trypsin for 1 h at room temperature) levels were determined using the Al generation assay kit (Dade-Baxter Travenol, Cambridge, MA) and sheep angiotensinogen (from nephrectomized sheep plasma). Prorenin levels were calculated as the difference in these two values. All assays were performed on 90-^1 aliquots of 4-h cultured supernatants within 12 h of collection in the absence of freezing or exposure to low pH in order to avoid


MOL ENDO-1990 1912

spurious activation of prorenin (31). When samples required dilution to keep the generated Al within the linear range of the assay, dilution was carried out using tissue culture medium containing 10% FCS (from the identical lot of FCS used in all of the experiments described herein). The amount of trypsin used to activate prorenin (150 fig/m\) was twice the concentration required to obtain maximal activity under these assay conditions and did not result in any detectable decrease in activity due to degradation of the generated renin. Under these conditions, prorenin secreted by CHO cells (which lack a processing enzyme) routinely demonstrated less than 5% Algenerating activity. Results presented in Table 1 and Fig. 2 are the mean of six to eight independent transfections ± SE. Since the absolute level of expression varied from one experiment to another (see Table 1) due to variations in transfection efficiency, the levels of secretion shown in Fig. 2 have been normalized to the constitutive level of prorenin secretion in each experiment (set at 1.0), except for Pro-Del, where the unstimulated active renin level was set at 1.0. Statistical analysis was carried out using the Wilcoxon multiple range test.

Acknowledgments We would like to thank Drs. Brian West and Rudolph Grosschedl for their generous gift of plasmids, Mai Phuong Hoang for expert technical assistance, Susan Corke and Vivianne Jodoin for typing the manuscript, and Dr. Oscar Carretero for his critical reading of the manuscript.

Received March 5,1990. Revision received September 13, 1990. Accepted September 17, 1990. Address requests for reprints to: Dr. Timothy L. Reudelhuber, Clinical Research Institute of Montreal, 110 Avenue des Pins Ouest, Montreal, Quebec, H2W 1R7 Canada. This work was supported by NIH Grant HL-35706, MRC of Canada Grant MA-11179, and a Grant-in-Aid from the American Heart Association, California Affiliate.

REFERENCES 1. Baxter JD, Perloff D, Hsueh W, Biglieri EG 1987 The endocrinology of hypertension. In: Felig P, Baxter JD, Broadus AE, Frohman LH (eds) Endocrinology and Metabolism. McGraw-Hill, New York, pp 693-788 2. Campbell DJ 1987 Circulating and tissue angiotensin system. J Clin Invest 79:1-6 3. Dzau VJ 1987 Renin-angiotensin system and renal circulation in clinical congestive heart failure. Kidney Int 20:S203-S209 4. Ledingham JGG 1987 Effects of angiotensin II and of angiotensin converting enzyme inhibition in chronic renal failure. Kidney Int 20:S112-S116 5. Riegger GAJ 1987 Experimental models of heart failure implications for renin and the kidney. Kidney Int 20:S210S212 6. Murakami K, Hirose S, Miyazaki H, Imai MS, Hori H, Hayashi T, Kakeyama R, Ohkubo H, Nakanishi S 1984 Complementary DNA sequences of renin. Hypertension 6:195-1100 7. Fritz LC, Haidar MA, Arfsten AE, Schilling JW, Carilli C, Shine J, Baxter JD, Reudelhuber TL 1987 Human renin is correctly processed and targeted to the regulated secretory pathway in mouse pituitary AtT-20 cells. J Biol Chem 262:12409-12412 8. James MNG, Sielecki AR 1986 Molecular structure of an aspartic proteinase zymogen, porcine pepsinogen, at 1.8 resolution. Nature 319:33-38 9. Pratt RE, Flynn JA, Hobart PM, Paul M, Dzau VJ 1988

Vol4No. 12

Different secretory pathways of renin from mouse cells transfected with the human renin gene. J Biol Chem 263:3137-3141 10. Ladenheim RG, Seidah N, Lutfalla G, Rougeon F 1989 Stable and transient expression of mouse submaxillary gland renin cDNA in AtT20 cells: proteolytic processing and secretory pathways. FEBS Lett 245:70-74 11. Fritz LC, Arfsten AE, Dzau VJ, Atlas SA, Baxter JD, Fiddes JC, Shine J, Cofer CL, Kushner P, Ponte PA 1986 Characterization of human prorenin expressed in mammalian cells from cloned cDNA. Proc Natl Acad Sci USA 83:4114-4118 12. Dingwall C, Laskey RA 1986 Protein import into the cell nucleus. Annu Rev Cell Biol 2:367-390 13. van Loon AP, Schatz G 1987 Transport of proteins to the mitochondrial intermembrane space: the sorting domain of the cytochrome d presequence is a stop-transfer sequence specific for the mitochondrial inner membrane. EMBO J 6:2441-2448 14. Small GM, Szabo LJ, Lazarow PB 1988 Acyl-CoA oxidase contains two targeting sequences each of which can mediate protein import into peroxisomes. EMBO J 7:1167-1173 15. Swinkels BW, Evers R, Borst P1988 The topogenic signal of the glycosomal (microbody) phosphoglycerate kinase of Crithidia fasciculata resides in a carboxyl-terminal extension. EMBO J 7:1159-1165 16. Mellman I, Kornfeld S 1989 The biogenesis of lysozomes. Annu Rev Cell Biol 5:483-525 17. Matsuuchi L, Buckley KM, Lowe AW, Kelly RB 1988 Targeting of secretory vesicles to cytoplasmic domains in AtT-20 and PC-12 cells. J Cell Biol 106:239-251 18. Moore HPH, Kelly RB 1986 Re-routing of a secretory protein by fusion with human growth hormone sequences. Nature 321:443-446 19. Orci L, Ravazzola M, Amherdt M, Perrelet A, Powell SK, Quinn DL, Moore HP 1987 The trans-most cisternae of the Golgi complex: a compartment for sorting of secretory and plasma membrane proteins. Cell 51:1039-1051 20. Chung KN, Walter P, Aponte GW, Moore HP 1989 Molecular sorting in the secretory pathway. Science 243:192197 21. Stoller TJ, Shields D 1989 The propeptide of preprosomatostatin mediates intracellular transport and secretion of alpha-globin from mammalian cells. J Cell Biol 108:1647-1655 22. Burgess TL, Craik CS, Matsuuchi L, Kelly RB 1987 In vitro mutagenesis of trypsinogen: role of the amino terminus in intracellular protein targeting to secretory granules. J Cell Biol 105:659-668 23. Powell SK, Orci L, Craik CS, Moore HP 1988 Efficient targeting to storage granules of human proinsulins with altered propeptide domain. J Cell Biol 106:1843-1851 24. Lodish, H 1988 Transport of secretory and membrane glycoproteins from the rough endoplasmic reticulum to the golgi. J Biol Chem 263:2107-2110 25. Hori H, Yoshino T, Ishizuka Y, Yamauchi T, Murakami K 1988 Role of N-linked oligosaccharides attached to human renin expressed in COS cells. FEBS Lett 232:391394 26. Paul M, Nakamura N, Pratt R, Dzau VJ 1988 Glycosylation influences intracellular transit time and secretion rate of human prorenin in tranfected cells. J Hypertension [Suppl] 6:S487-S489 27. Nakayama K, Nagahama M, Kim W-S, Hatsuzawa K, Hashiba K, Murakami M 1989 Prorenin is sorted into the regulated pathway independent of its processing to renin in mouse pituitary AtT-20 cells. FEBS Lett 257:89-92 28. Harrison TM, Chidgey MAJ, Brammar WJ, Adams GJ 1989 The pro-peptide is not necessary for active renin secretion from transfected mammalian cells. Proteins 5:259-265 29. Nagahama M, Nakayama K, Hori H, Murakami K 1989 Expression of a deletion mutant of the pro segment of



30.

31.

32.

33.

34.

35.

human prorenin in Chinese hamster ovary cells. FEBS Lett 259:202-204 Bothwell ALM, Paskind M, Reth M, Imanishi-Kari T, Rajewsky K, Baltimore D 1981 Heavy chain variable region contribution to the NP6 family of antibodies: somatic mutation evidence in a 2a variable region. Cell 24:625627 Derkx FHM, Schalekamp MPA, Schalekamp MADH 1987 Two-step prorenin to renin conversion. J Biol Chem 262:2472-2477 Faust PL, Chirgwin JM, Kornfeld S 1987 Renin, a secretory glycoprotein, acquires phosphomannosyl residues. J Cell Biol 105:1947-1955 Pratt RE, Ouelette AJ, Dzau VJ 1983 Biosynthesis of renin: multiplicity of active and intermediate forms. Proc Natl Acad Sci USA 80:6809-6813 Davidson HW, Rhodes CJ, Hutton JC1988 Intraorganellar calcium and pH control proinsulin cleavage in the pancreatic beta cell via two distinct site-specific endopeptidases. Nature 333:93-96 Shier DN, Malvin RL 1985 Differential secretion and removal of multiple renin forms. Am J Physiol 249:R79-R84

1913

36. Maniatis T, Fritsch EF, Sambrook J 1982 Molecular Cloning-A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor 37. Gorman CM, Merlino GT, Willingham MC, Pastan I, Howard BH 1982 The Rous sarcoma virus long terminal repeat is a strong promoter when introduced into a variety of eukaryotic cells by DNA-mediated transfection. Proc Natl Acad Sci USA 79:6777-6781 38. Kabnick KS, Housman DE 1988 Determinants that contribute to cytoplasmic stability of human c-fos and betaglobin mRNAs are located at several sites in each mRNA. Mol Cell Biol 8:3244-3250 39. Zoller MJ, Smith M 1984 Oligonucleotide-directed mutagenesis: a simple method using two oligonucleotide primers and a single-stranded DNA template. DNA 3:479488 40. Graham F, Van der Eb A 1973 A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52:456-457 41. Imai T, Miyazaki H, Hirose S, Hori S, Hori H, Hyashi T, Kageyama R, Ohkubo H, Nakanishi S, Murakami K 1983 Cloning and sequence analysis of cDNA for human renin precursor. Proc Natl Acad Sci USA 80:7405-7409


The Human Mitochondrial DEAD-Box Protein DDX28 Resides in RNA Granules and Functions in Mitoribosome Assembly.

Protein mobility within secretory granules.

Isolation of dense granules from human platelets.

Transfer of a dense granule protein of Plasmodium falciparum to the membrane of ring stages and isolation of dense granules.

Biogenesis of secretory granules.

Nucleotide sequence of a human oncogene active in tumors of secretory epithelium.

Menin immunoreactivity in secretory granules of human pancreatic islet cells.

Characterization of a monoclonal antibody specific for human active renin.

Anchorage of secretion-competent dense granules on the plasma membrane of bovine platelets in the absence of secretory stimulation.

Chromogranin A: posttranslational modifications in secretory granules.

Aggregations of dense granules in mitochondria of active pulmonary lymphatic endothelial cells.

Human Eosinophil Leukocytes Express Protein Disulfide Isomerase in Secretory Granules and Vesicles: Ultrastructural Studies.

Mast cell secretory granules: armed for battle.

Lumenal protein within secretory granules affects fusion pore expansion.

Bacterial mimetics of endocrine secretory granules as immobilized in vivo depots for functional protein drugs.

Sarcocystis muris (Apicomplexa): a thiol protease from the dense granules.

Proteoglycans in secretory granules of NK cells.

Low molecular weight renin as a storage form in renin granules of the dog.

Measurement of human active renin heterogeneity.

The nature of breast dense core granules: chromogranin reactivity.

Functional ryanodine receptors in the membranes of neurohypophysial secretory granules.

Regional differences in statistical geometry of endothelial dense granules in human extremity veins.

Arrested exocytosis of atrial secretory granules.

Decreased numbers of dense granules in fetal and neonatal platelets.