Design, high-level expression, purification and characterization of soluble fragments of the hepatitis C virus NS3 RNA helicase suitable for NMR-based drug discovery methods and mechanistic studies.

RNA helicases represent a family of enzymes that unwind double-stranded (ds) RNA in a nucleoside triphosphate (NTP)-dependent fashion and which are required in all aspects of cellular RNA metabolism and processing. The hepatitis C virus (HCV) non-structural 3 (NS3) protein possesses a serine protease activity in the N-terminal one-third, whereas RNA-stimulated NTPase and helicase activities reside in the C-terminal portion of the 631 amino acid residue bifunctional enzyme. The HCV NS3 RNA helicase is of key importance in the life cycle of HCV, which makes it a target for the development of therapeutics. However, neither the precise mechanism nor the substrate structure has been defined for this enzyme. For nuclear magnetic resonance (NMR)-based drug discovery methods and for mechanistic studies we engineered, prepared and characterized various truncated constructs of the 451-residue HCV NS3 RNA helicase. Our goal was to produce smaller fragments of the enzyme, which would be amenable to solution NMR techniques while retaining their native NTP and/or nucleic acid binding sites. Solution conditions were optimized to obtain high-quality heteronuclear NMR spectra of nitrogen-15 isotope-labeled constructs, which are typical of well-folded monomeric proteins. Moreover, NMR binding studies and functional data directly support the correct folding of these fragments.

Probing the relationship between RNA-stimulated ATPase and helicase activities of HCV NS3 using 2'-O-methyl RNA substrates.

The hepatitis C virus (HCV) NS3 protein contains an amino terminal protease (NS3 aa. 1-180) and a carboxyl terminal RNA helicase (NS3 aa. 181-631). NS3 functions as a heterodimer of NS3 and NS4A (NS3/4A). NS3 helicase, a nucleic acid stimulated ATPase, can unwind RNA, DNA, and RNA:DNA duplexes, provided that at least one strand of the duplex contains a single-stranded 3' overhang (this strand of the duplex is referred to as the 3' strand). We have used 2'-O-methyl RNA (MeRNA) substrates to study the mechanism of NS3 helicase activity and to probe the relationship between its helicase and RNA-stimulated ATPase activities. NS3/4A did not unwind double-stranded (ds) MeRNA. NS3/4A unwinds hybrid RNA:MeRNA duplex containing MeRNA as the 5' strand but not hybrid duplex containing MeRNA as the 3' strand. The helicase activity of NS3/4A was 50% inhibited by 40 nM single-stranded (ss) RNA but only 35% inhibited by 320 nM ss MeRNA. Double-stranded RNA was 17 times as effective as double-stranded MeRNA in inhibiting NS3/4A helicase activity, while the apparent affinity of NS3/4A for ds MeRNA differed from ds RNA by only 2.4-fold. However ss MeRNA stimulated NS3/4A ATPase activity similar to ss RNA. These results indicate that the helicase mechanism involves 3' to 5' procession of the NS3 helicase along the 3' strand and only weak association of the enzyme with the displaced 5' strand. Further, our findings show that maximum stimulation of NS3 ATPase activity by ss nucleic acid is not directly related to procession of the helicase along the 3' strand.

Structure of the hepatitis C virus RNA helicase domain.

Helicases are nucleotide triphosphate (NTP)-dependent enzymes responsible for unwinding duplex DNA and RNA during genomic replication. The 2.1 A resolution structure of the HCV helicase from the positive-stranded RNA hepatitis C virus reveals a molecule with distinct NTPase and RNA binding domains. The structure supports a mechanism of helicase activity involving initial recognition of the requisite 3' single-stranded region on the nucleic acid substrate by a conserved arginine-rich sequence on the RNA binding domain. Comparison of crystallographically independent molecules shows that rotation of the RNA binding domain involves conformational changes within a conserved TATPP sequence and untwisting of an extended antiparallel beta-sheet. Location of the TATPP sequence at the end of an NTPase domain beta-strand structurally homologous to the 'switch region' of many NTP-dependent enzymes offers the possibility that domain rotation is coupled to NTP hydrolysis in the helicase catalytic cycle.

Isolation and purification of a biologically active human platelet-derived growth factor BB expressed in Escherichia coli.

Human platelet-derived growth factor (PDGF) was expressed in Escherichia coli from a high-level cytoplasmic expression vector. A cDNA fragment encoding the mature form of the human PDGF B chain (hPDGF-B) was cloned into a plasmid under transcriptional control of the inducible E. coli Tac promoter. Expression of hPDGF-B from the final construct, pTacBIq, is regulated by the lactose repressor (LacIq). Upon induction, a polypeptide of approximately 14 kDa that had the same molecular mass and immunoreactivity as authentic hPDGF-B was produced. The production of recombinant hPDGF-B was significantly increased in an E. coli strain (CAG629) defective in expression of the lon protease. Expression of hPDGF-B in the CAG629 strain accounted for approximately 1% of total cell protein. In this system, hPDGF-B is expressed as an insoluble, intracellular protein and can readily be obtained in a partially purified form after differential centrifugation. Amino acid sequence determination of the purified protein has verified that the amino-terminal portion of the recombinant PDGF is correct. After renaturation into dimers, the purified recombinant hPDGF is fully functional in assays for receptor binding and mitogenesis.

Immunological evidence that inactive renin is prorenin.

Antibody raised to a synthetic dodecapeptide, corresponding to the C-terminal portion of the human renin pro-segment, was tested for its ability to recognize highly purified human inactive or active (mature) renins; immune complexes were detected by precipitation with protein A-Sepharose. Serial antibody dilutions caused identical binding of renal or plasma inactive renins but failed to bind active renin. In contrast, antibody to active renin recognized both active and inactive forms. Reversible acid activation of inactive renin enhanced its binding to the anti-prorenin antibody, whereas irreversible trypsin activation significantly reduced binding. Binding was abolished following prolonged exposure to trypsin or if inactive renin was acidified prior to trypsin treatment. These results indicate that inactive renin shares immunochemical determinants with prorenin; they suggest that acidification alters the conformation of the pro-segment and that trypsin can convert the molecule both to fully mature renin and to intermediate form(s).

Characterization of inactive renin ("prorenin") from renin-secreting tumors of nonrenal origin. Similarity to inactive renin from kidney and normal plasma.

Inactive renin comprises well over half the total renin in normal human plasma. There is a direct relationship between active and inactive renin levels in normal and hypertensive populations, but the proportion of inactive renin varies inversely with the active renin level; as much as 98% of plasma renin is inactive in patients with low renin, whereas the proportion is consistently lower (usually 20-60%) in high-renin states. Two hypertensive patients with proven renin-secreting carcinomas of non-renal origin (pancreas and ovary) had high plasma active renin (119 and 138 ng/h per ml) and the highest inactive renin levels we have ever observed (5,200 and 14,300 ng/h per ml; normal range 3-50). The proportion of inactive renin (98-99%) far exceeded that found in other patients with high active renin levels. A third hypertensive patient with a probable renin-secreting ovarian carcinoma exhibited a similar pattern. Inactive renins isolated from plasma and tumors of these patients were biochemically similar to semipurified inactive renins from normal plasma or cadaver kidney. All were bound by Cibacron Blue-agarose, were not retained by pepstatin-Sepharose, and had greater apparent molecular weights (Mr) than the corresponding active forms. Plasma and tumor inactive renins from the three patients were similar in size (Mr 52,000-54,000), whereas normal plasma inactive renin had a slightly larger Mr than that from kidney (56,000 vs. 50,000). Inactive renin from each source was activated irreversibly by trypsin and reversibly by dialysis to pH 3.3 at 4 degrees C; the reversal process followed the kinetics of a first-order reaction in each instance. The trypsin-activated inactive renins were all identical to semipurified active renal renin in terms of pH optimum (pH 5.5-6.0) and kinetics with homologous angiotensinogen (Michaelis constants, 0.8-1.3 microM) and inhibition by pepstatin or by serial dilutions of renin-specific antibody. These results indicate that a markedly elevated plasma inactive renin level distinguishes patients with ectopic renin production from other high-renin hypertensive states. The co-production of inactive and active renin by extrarenal neoplasms provides strong presumptive evidence that inactive renin is a biosynthetic precursor of active renin. The unusually high proportion of inactive renin in plasma and tumor extracts from such patients is consistent with ineffective precursor processing by neoplastic tissue, suggesting that if activation of "prorenin" is involved in the normal regulation of active renin levels it more likely occurs in the tissue of origin (e.g., kidney) than in the circulation.

Biochemical similarity of partially purified inactive renins from human plasma and kidney.

Inactive renin was partially purified from 4.5 liters of human plasma (502-fold, specific activity 0.8 X 10(-3) Goldblatt units/mg protein) and from 207 g renal cortex (103-fold, 52 X 10(-3) Goldblatt units/mg). In contrast to active renin, inactive renin from each source bound to Cibacron blue-agarose and was unable to bind to pepstatin-Sepharose. Both plasma and renal inactive renin had weaker affinity for anion-exchange resins than the active form, both bound to concanavalin A-Sepharose and were eluted with carbohydrate, and both bound tightly to hydrophobic gels. Each substance could be isolated in a completely inactive form during small-scale pilot studies, but "spontaneous" activation did occur, to a limited degree, during large-scale purification; this was possibly due to a plasma serine protease that fractionated with inactive renin during the initial purification steps. Both plasma and renal inactive renin were activated irreversibly by trypsin. Following activation, each substance lost it ability to bind to Cibacron blue-agarose. Each could be activated fully by acidification at 4 degrees C, but this activation was reversed during subsequent incubation at higher temperature and pH. There was no evidence of acid protease activity in either preparation. Activated inactive renin from both plasma and kidney were identical to partially-purified active renal renin in terms of pH optimum (pH 5.5-6.0) and reaction kinetics (Km 0.8-1.3 microM) with homologous angiotensinogen, noncompetitive inhibition by pepstatin (ki 2.5-3.5 microM), and an identical inhibition profile by monospecific antirenin antibodies. These results suggest that inactive renin from plasma and kidney may be the same substance and that their activated forms are similar to the endogenously produced active enzyme, consistent with the possibility that inactive renin is a precursor of circulating active renin.

Apparent molecular size difference between plasma and renal inactive renins.

Partially-purified inactive renins from human plasma and kidney seem to be identical in most respects except for apparent molecular size. To evaluate this difference, we determined apparent molecular weights by gel filtration with internal radiolabeled standards, using trypsin activation in the presence of benzamidine and albumin to provide reproducible detection. While there was a suggestion of a shoulder in the 50,000-dalton region of the plasma inactive renin peak, the major form (56,000) was consistently larger than that of renal inactive renin (50,000), confirming our previous observation. Since both renal and plasma inactive renins appear to be glycoproteins, based on their ability to bind to concanavalin A-Sepharose, it is possible that differences in carbohydrate composition might contribute to this discrepancy in gel filtration behavior. The striking similarity of these substances in all other respects, including inhibition of the activated forms by monospecific antirenin antibodies, makes it unlikely that they differ in primary structure.

Quantitation of inactive renin in human and dog plasma: techniques for activation.

For human samples quantitation of inactive renin can be carried out by incubation with trypsin under defined conditions, followed by RIA of the activated renin. For dog samples we were unable to obtain evidence for the presence of inactive renin in the plasma by using trypsin, acid or cold to activate. Increases in angiotensin generation did occur with trypsin and acid but they both changed renin substrate such that the rate of angiotensin generation by exogenous renin was increased at pH 7.4, but not at pH 5.7; also following trypsin or acid treatment angiotensin I was cleaved from renin substrate by a plasma acid protease that normally does not cleave renin substrate in plasma. Therefore, for dog samples, it is important to demonstrate that an increase in the rate of angiotensin generation is indeed due to activation of inactive renin and not to changes in pH optimum of renin with angiotensinogen or to the effect of another enzyme.

Detection and isolation of inactive, large molecular weight renin in human kidney and plasma.

Estimation of apparent molecular weight (mw) of inactive renin by gel filtration of human plasma was found to be inaccurate when "acid activation" or "cryoactivation" was used for detection; recoveries were only 5% to 20%. Trypsin activation produced greater recoveries, but the apparent elution volume of inactive renin varied with the concentration of trypsin used; the presence of trypsin inhibitors increased trypsin requirements to 100 to 200 micrograms/ml in the 60,000 dalton region, while low protein concentration in the 50,000 dalton region resulted in destruction of renin by as little as 10 microgram/ml trypsin. A composite trypsin-activated inactive renin peak corresponded to a mw of 56,000 +/- 1500 daltons (104% to 120% recovery), while active plasma renin was 48,000 +/- 2000 daltons. When this prorenin-like substance was isolated by affinity chromatography, it was found to be completely inactive. It was also nearly free of trypsin inhibitors, so that a single trypsin concentration correctly identified and confirmed the elution characteristics of inactive renin peak following gel filtration. The apparent mw of trypsin-activated inactive renin was slightly lower (52,000 daltons) than that of inactive renin. Human renal cortex was also found to contain a trypsin-activable form of renin. Like plasma inactive renin, it could be isolated by chromatography on Cibacron blue-agarose (Affi-Gel blue). It was found to be completely inactive following passage over a pepstatin affinity column. This inactive renal renin, as well as a similar substance in perfusate of normal human kidney, had a mw of 49,500 +/- 1000, while active renal renin was 39,500 +/- 500. Trypsin-activated inactive renal renin had a mw of 46,500 +/- 500; its pH optimum was identical with that of active renal renin, and it no longer bound to Cibacron blue-agarose. We conclude that both human plasma and kidney contain an inactive, prorenin-like substance that can be detected reliably by trypsin activation. There appear to be slight differences in the apparent mw of plasma renins and kidney renin, but the similarity of other characteristics suggests that the inactive, prorenin-like substances in renal cortex, renal perfusate, and plasma may be one and the same substance.

An inactive, prorenin-like substance in human kidney and plasma.

1. Plasma prorenin (inactive renin), which accounts for about 70% of the total renin in human plasma, was almost completely separated from active renin by affinity chromatography on Cibacron blue F3G-A-agarose. The slight residual renin activity present in the prorenin peak can be removed on concanavalin A-Sepharose, demonstrating that prorenin is completely inactive. 2. The renin activity of both human renal cortical extract and renal perfusate increased after incubation with trypsin. This trypsin-activable renin accounted for 15 and 40% of the total renin in extract and perfusate respectively. 3. Trypsin-activable renin from both renal extract and renal perfusate was, like plasma prorenin, almost completely separated from active renin on Cibacron blue F3G-A-agarose. After additional chromatographic steps, the trypsin-activable renin from renal cortical extract was found to be completely inactive. 4. We conclude that human kidney contains, and is able to release, a trypsin-activable renin that resembles plasma prorenin. It may differ from many of the 60 000 molecular-weight forms of renin previously identified in renal extracts, since these possess considerable intrinsic renin activity and probably represent a complex of renin with a binding protein.