Inhibition of Hepadnavirus Reverse Transcriptase-ɛ RNA Interaction by Porphyrin Compounds

Abstract

The hepatitis B virus (HBV) reverse transcriptase (RT) plays a multitude of fundamental roles in the viral life cycle and is the key target in the development of anti-HBV chemotherapy. We report here that the endogenous small molecule iron protoporphyrin IX (hemin) and several related porphyrin compounds potently blocked a critical RT interaction with the viral RNA packaging signal/origin of replication, called ɛ. As RT-ɛ interaction is essential for the initiation of viral reverse transcription, which is primed by RT itself (protein priming), the porphyrin compounds dramatically suppressed the protein-priming reaction. Further studies demonstrated that these compounds could target the unique N-terminal domain of the RT protein, the so-called terminal protein. Hemin and related porphyrin compounds thus represent a novel class of agents that can block HBV RT functions through a mechanism and target that are completely distinct from those of existing anti-HBV drugs.

Hepatitis B virus (HBV) belongs to the family Hepadnaviridae, a group of small, hepatotropic DNA viruses that also includes related animal viruses, such as the duck HBV (DHBV) and the woodchuck hepatitis virus (31). Chronic HBV infection remains a major public health problem worldwide, with over 350 million chronic HBV carriers who are at serious risk of developing liver cirrhosis and hepatocellular carcinoma (5, 7, 12). At present, the immunomodulatory and antiviral cytokine alpha interferon and several nucleoside analogs represent two distinct classes of therapies for chronic HBV infections (14). Unfortunately, alpha interferon therapy induces a sustained antiviral response in only 20 to 30% of patients and is fraught with adverse reactions. In the short term, nucleoside analogs exhibit a potent inhibitory effect on HBV DNA synthesis. However, long-term treatment with nucleoside analogs is required to maintain viral suppression, which leads to the development of drug-resistant HBV variants (13, 26). Therefore, the current inventory of therapeutics against HBV infections is inadequate, and novel anti-HBV therapy needs to be developed.

All hepadnaviruses replicate the DNA genome through an RNA intermediate (the pregenomic RNA [pgRNA]) by reverse transcription (31, 33). The virally encoded reverse transcriptase (RT) protein carries the two essential enzymatic activities required for the conversion of pgRNA to the DNA copy, i.e., DNA polymerase activity and RNase H activity (18). In contrast to retroviral RTs, the hepadnavirus RT initiates viral DNA synthesis via a novel protein-priming mechanism, whereby the RT protein itself serves a dual role as a protein primer and as the polymerase. As an obligate template for the protein-priming action, a short (ca. 60-nucleotide) viral RNA structure termed ɛ, located at the 5′ end of pgRNA, is specifically recognized by the RT protein to form a stable ribonucleoprotein (RNP) complex (28, 38). Using an internal bulge located on the ɛ stem-loop structure as a template and a specific tyrosine residue of the RT protein as a primer, RT synthesizes a 3- to 4-nucleotide-long DNA oligomer, which becomes covalently attached to RT through the tyrosine residue (24, 25, 37, 40, 41). Following the synthesis of this short DNA oligomer, the nascent DNA-RT complex is translocated from ɛ to the 3′ end of the pgRNA (a minus-strand template switch) to continue DNA synthesis (27, 34, 36). In addition, the formation of the same RNP complex also serves to initiate nucleocapsid assembly, leading to the incorporation of both the RT protein and pgRNA into the subviral core particles, where viral reverse transcription takes place (3, 22, 28). Thus, the ɛ RNA serves both as the pgRNA packaging signal and the origin of reverse transcription, and furthermore, packaging of RT and that of pgRNA are mutually dependent.

The unique ability of the hepadnavirus RT to carry out specific RNA recognition and protein priming is reflected in its structural organization, which is similar to yet distinct from that of conventional RTs (4, 8, 18, 29). The N-terminal domain (terminal protein [TP]) is conserved among all hepadnaviruses but absent from all other known RTs or any other known proteins. It is within the TP domain that the primer tyrosine residue (for protein priming) is located. In contrast, the central RT domain and the C-terminal RNase H domain share sequence homologies with conventional RTs. A highly variable spacer, or tether, domain, which does not appear to have any essential functions, links the TP and RT domains. Both the TP and RT domains are required for ɛ binding and protein priming (15, 16, 25, 28, 30, 38). Furthermore, a host chaperone complex consisting of heat shock protein 90 (Hsp90) and its cochaperones associates with RT and facilitates the establishment and maintenance of an RT conformation that is competent for ɛ binding (15, 17, 19-21). Extensive truncations, however, led to the identification of a mini-DHBV RT protein (MiniRT2) that can carry out ɛ binding and protein priming independently of the host chaperone function (39).

Although RNP formation and protein priming represent excellent targets for developing anti-HBV agents, currently no anti-HBV drugs can block RT-ɛ interaction and few inhibit protein priming. The few nucleoside analogs that do inhibit protein priming act simply as competitive inhibitors of the initiating nucleotide (32). By using an in vitro DHBV protein-priming assay, we previously noted the existence of endogenous cellular factors in the rabbit reticulocyte lysate (RRL) that could dramatically inhibit protein priming (39). Ongoing efforts to isolate the putative cellular inhibitors led to the identification of iron protoporphyrin IX (hemin), a critical component of hemoglobin, and several related porphyrin compounds, as potent inhibitors of protein priming. In contrast to nucleoside analogs, which inhibit HBV DNA synthesis (mostly after protein priming) by targeting the polymerase active site in the central RT domain, hemin and related compounds blocked the essential RT-ɛ interaction and targeted the N-terminal TP domain.

MATERIALS AND METHODS

Chemical and reagents.

Hemin was obtained from Sigma, and the stock solution was prepared as described previously (9). Briefly, hemin was dissolved in 0.2 N NaOH, neutralized to pH 8.0 using 1 N HCl, and stored at −80°C until use. Alternatively, hemin was directly solubilized in dimethyl sulfoxide (Sigma). Similar results were obtained when hemin was solubilized by either method. Protoporphyrin IX, protoporphyrin IX disodium, biliverdin, porphine, and meso-tetraphenylporphine were purchased from Frontier Scientific (Utah) and dissolved in dimethyl sulfoxide. Delta-aminolevulinic acid and porphobilinogen were also from Frontier and were dissolved in distilled water. Human hemoglobin was purchased from Sigma and RRL from Promega.

Plasmids.

pcDNA-MiniRT2 and pGEX-MiniRT2 express a truncated DHBV RT (MiniRT2) and the glutathione-S-transferase (GST)-MiniRT2 fusion, respectively (15). pSP64pA-DP expresses the full-length DHBV RT downstream of the SP6 promoter in the pSP64pA vector. pGEX-TP and pGEX-RT express the DHBV TP (residues 75 to 220) and DHBV RT (residues 349 to 575) sequences, respectively, as GST fusions (GST-TP and GST-RT, respectively) in the vector pGEX-KT (15). pGEX-HTPRT/Dra expresses a truncated HBV RT-GST fusion, GST-HTPRT/Dra, as described previously (16, 17).

Protein and RNA expression and purification.

GST-MiniRT2, GST-TP, GST-RT, and GST-HTPRT/Dra were expressed in BL21-CodonPlus-RIL cells and purified using the glutathione resin, as previously described (15, 17). Expression of the full-length DHBV RT and the truncated DHBV MiniRT2, from pSP64pA-DP and pcDNA-MiniRT2, respectively, using in vitro transcription and translation in the RRL, was performed as previously described (15, 19). DHBV and HBV ɛ RNAs were produced by using the Megascript kit (Ambion) as described previously (17, 19).

In vitro protein priming and ɛ RNA binding.

The protein-priming assay was carried out as described previously (15, 39) with minor modifications. Briefly, the following components were added to a 10-μl reaction mixture: GST-MiniRT2 (ca. 0.1 μM) or 5 μl of the RRL translation reaction mixture expressing the full-length DHBV RT or MiniRT2, ɛ RNA (1 μM), TMnNK buffer (20 mM Tris-HCl, pH 8.0, 1 mM MnCl₂, 15 mM NaCl, 20 mM KCl), 4 mM dithiothreitol, and 0.5 μl of [α-³²P]dGTP (3,000 Ci/mmole; 10 mCi/ml). NP-40 (0.2% [vol/vol]) was added to stimulate protein priming by GST-MiniRT2 (39). To test for inhibition of priming, the RT protein was incubated with the indicated compounds for 20 min at room temperature prior to addition of the ɛ RNA, unless otherwise indicated. The protein-priming reaction was then conducted at 30°C for 2 h. The reaction products were resolved on sodium dodecyl sulfate-12% polyacrylamide gels and quantified by phosphorimaging. In vitro RT and ɛ RNA binding, in the presence or absence of the different compounds, was measured by the RNA gel mobility shift assay, as described previously (16, 17).

RESULTS

Hemin exhibited a potent inhibitory effect on the DHBV RT priming activity.

In our attempt to identify the putative cellular inhibitors in RRL (39), we subjected the lysate to biochemical fractionation using fast protein liquid chromatography. We observed that protein priming by the DHBV RT could be dramatically inhibited by fractions rich in hemoglobin (data not shown). Furthermore, DHBV RT priming could be inhibited by a commercial source of purified human hemoglobin. Since hemoglobin consists of two α- and two β-globin subunits, as well as four molecules of heme (which is the same as hemin, except that heme contains Fe²⁺ and hemin contains Fe³⁺), we extracted heme from the purified hemoglobin (11) and found that the isolated heme could also dramatically inhibit protein priming (data not shown). To directly test the effect of heme (or hemin) on protein priming, we then obtained hemin (Fig. 1) (9) and added it to the in vitro protein-priming reactions. Hemin indeed dramatically inhibited the protein-priming activity of the purified (from bacteria) DHBV RT fusion protein GST-MiniRT2, with a 50% inhibitory concentration (IC₅₀) of 1.1 ± 0.3 μM (Fig. 2). We also tested the effect of hemin on the full-length DHBV RT, as well as the truncated MiniRT2 protein, expressed in RRL and found that protein priming by these two DHBV RT proteins was also inhibited by hemin, with IC₅₀s of 21 ± 1.9 μM and 19 ± 1.6 μM, respectively (Fig. 3), indicating that both the full-length and the truncated DHBV RT proteins showed similar sensitivities to hemin inhibition. Interestingly, the IC₅₀ of hemin against the purified GST-MiniRT2 was 20-fold lower than that against the MiniRT2 (or full-length RT) translated in RRL. However, when the purified GST-MiniRT2 protein was mixed with RRL and used in the protein-priming assay, it was inhibited only at much higher concentrations of hemin, similar to the untagged MiniRT2 translated in RRL (data not shown). Most likely, the higher apparent IC₅₀ of hemin against protein priming as tested in the complex RRL compared with that obtained using purified RT proteins was due to the fact that RRL contains numerous proteins and other factors, some of which might bind hemin and thus decrease the effective (free) hemin concentration available to inhibit protein priming. This also likely explains why the DHBV RT expressed in RRL can efficiently bind to ɛ and carry out protein priming (Fig. 3) (37) despite the presence of endogenous hemin at a concentration (ca. 20 μM; Promega) that would have inhibited RT-ɛ interaction in the purified system. In any case, these results clearly showed that hemin was a potent inhibitor of protein-priming activity by the DHBV RT.

FIG. 1.

Chemical structures of compounds tested for the inhibition of hepadnavirus RT and their activities. Iron protoporphyrin IX (hemin), protoporphyrin IX (PPP-IX), protoporphyrin IX disodium (PPP-IX-Na), and biliverdin showed potent inhibition of RT activity. In contrast, delta-aminolevulinic acid (ALA), porphobilinogen (PPB), and porphine (PP), as well as meso-tetraphenylporphine (MTPP), showed little inhibitory effect on RT.

FIG. 2.

Inhibition of DHBV GST-MiniRT2 protein priming by porphyrin compounds. (A) GST-MiniRT2, purified from bacteria, was assayed for protein-priming activity in vitro in the presence of iron protoporphyrin IX (hemin) (lanes 2 to 5), protoporphyrin IX (PPP-IX) (lanes 6 to 9), protoporphyrin IX disodium (PPP-IX-Na) (lanes 10 to 13), biliverdin (lanes 14 to17), porphine (PP) (lane 19), meso-tetraphenylporphine (MTPP) (lane 20), delta-aminolevulinic acid (ALA) (lane 21), and porphobilinogen (PPB) (lane 22) at the concentrations indicated. C, control reactions without any compound (lanes 1 and 18). The ³²P-labeled GST-MiniRT2, as a result of protein priming, was resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and detected by autoradiography. (B) The protein-priming activities of GST-MiniRT2 in the presence of different concentrations of the indicated compounds are expressed as percentages of the untreated control activity. All data points represent the means and standard deviations from four independent experiments.

FIG. 3.

Inhibition of full-length DHBV RT and MiniRT2 expressed in RRL. The full-length DHBV RT and MiniRT2 were translated in RRL and assayed for protein-priming activity without (lane 1) or with (lanes 2 to 4) hemin. Protein-priming activity was detected as described in the legend to Fig. 2. The ³²P-labeled full-length (FL) RT and MiniRT2 are indicated. The protein-priming activities of the full-length RT and MiniRT2 in the presence of hemin are expressed as percentages of that of the untreated control.

Hemin-related porphyrin compounds also exhibited a potent inhibitory effect on DHBV RT priming activity.

We then asked whether iron or other structural components of hemin were required for the inhibition of RT priming activity. A number of hemin-related porphyrin compounds were tested, including the hemin analogs (protoporphyrin IX, protoporphyrin IX disodium, biliverdin, porphine, and meso-tetraphenylporphine), as well as two hemin precursors (delta-aminolevulinic acid and porphobilinogen) (Fig. 1). The protein-priming assay was performed as described above with various concentrations of the indicated compounds. The results showed that the GST-MiniRT2-priming activity was inhibited in a dose-dependent manner by protoporphyrin IX, protoporphyrin IX disodium, and biliverdin (Fig. 2). The IC₅₀ values were 1.1 ± 0.2 μM for protoporphyrin IX, 1.1 ± 0.3 μM for protoporphyrin IX disodium, and 4.1 ± 1.8 μM for biliverdin, which were similar to, or (in the case of biliverdin) slightly higher than, that for hemin. In contrast, porphine and meso-tetraphenylporphine showed little to no inhibitory activity even at 100 μM, nor did delta-aminolevulinic acid or porphobilinogen (Fig. 2). Taken together, these results demonstrated that hemin and the three closely related analogs, protoporphyrin IX, protoporphyrin IX disodium, and biliverdin, showed potent inhibitory effects against DHBV RT priming.

Hemin and related porphyrin compounds blocked RNP complex formation between DHBV RT and ɛ RNA.

To gain insight into the mechanism of inhibition of protein priming by hemin and related compounds, we asked whether these same compounds could inhibit RT-ɛ RNP complex formation, as RNP formation is a prerequisite for protein priming. Using an in vitro RNA binding assay, we found that hemin, protoporphyrin IX, protoporphyrin IX disodium, and biliverdin could all inhibit MiniRT2-ɛ RNA interaction in a dose-dependent fashion (Fig. 4). In contrast, no significant inhibition of RNA binding by porphine, meso-tetraphenylporphine, delta-aminolevulinic acid, or porphobilinogen was observed (data not shown). Thus, the inhibitory effects of the various compounds on RT-ɛ interaction were in good agreement with those on protein priming (Fig. 2), indicating that the inhibition of DHBV RT priming activity by these compounds was mainly, if not exclusively, due to the blockade of RNP complex formation. To further confirm that the inhibition of RT priming was due to the inhibition of RT-ɛ interaction, we reversed the order of addition of hemin and the ɛ RNA by preincubating MiniRT2 with ɛ first to form the RNP complex before adding hemin. As expected, the priming activity of the preformed RNP complex was inhibited to a much smaller degree by hemin, thus supporting the notion that hemin inhibited RT priming mainly by blocking RT-ɛ RNA interaction (Fig. 5). The residual sensitivity of the preformed RNP complex to hemin inhibition suggests that hemin may be able to weakly inhibit the RT DNA synthesis activity, in addition to blocking RT-ɛ interaction (see Discussion below).

FIG. 4.

Inhibition of DHBV RT-ɛ complex formation by porphyrin compounds. GST-MiniRT2 was used in a gel mobility shift assay in vitro to measure its binding activity to the ³²P-labeled DHBV ɛ RNA in the presence of iron protoporphyrin IX (hemin) (lanes 2 to 4), protoporphyrin IX (PPP-IX) (lanes 5 to 7), protoporphyrin IX disodium (PPP-IX-Na) (lanes 8 to 10), or biliverdin (lanes 11 to 13). C, untreated control (lane 1). The ³²P-labeled free DHBV ɛ and GST-MiniRT2-ɛ complex are indicated. The ɛ binding activities of GST-MiniRT2 in the presence of different concentrations of the indicated compounds are expressed as percentages of that of the untreated control.

FIG. 5.

Inhibition of DHBV GST-MiniRT2 priming before or after RNP complex formation. GST-MiniRT2 was incubated with hemin before the addition of the ɛ RNA (hemin/ɛ) (lanes 2 and 3) or with ɛ before the addition of hemin (ɛ/hemin) (lanes 5 and 6). Untreated controls are shown in lanes 1 and 4. Protein-priming activity is indicated at the bottom as a percentage of that of controls. The minus sign indicates that protein-priming activity was not detectable. The ³²P-labeled GST-MiniRT2 is indicated.

Hemin targeted the TP domain.

To gain insight into the mechanism of inhibition of RT-ɛ interaction by hemin, we then asked whether the inhibitory effect was mediated by hemin interacting with the RT protein or the ɛ RNA. Circumstantial evidence in support of hemin binding to the RT instead of ɛ was obtained when an increased concentration of RT protein, but not ɛ RNA, was able to partially rescue protein priming (Fig. 6A and data not shown). Based on these results, we hypothesized that the RT protein may contain one or more hemin binding sites, which may be responsible for mediating the inhibitory effect of hemin. If this is true, an isolated RT domain containing the putative hemin binding site(s) may be able to specifically compete with MiniRT2 for hemin binding and thus rescue the protein-priming activity of MiniRT2. Since the requirement for the N-terminal TP and central RT domain in ɛ RNA binding has been well characterized, we considered that one or both of these domains, particularly those sequences contained in MiniRT2, might harbor a hemin binding site. To test this hypothesis, different domains of the RT proteins were purified as GST fusion proteins and added to the priming reactions. MiniRT2 showed similar priming activities in all reactions without hemin, whether or not the TP or RT domain was added (Fig. 6B, odd-numbered lanes). In the presence of hemin (Fig. 6B, even-numbered lanes), MiniRT2 priming was still completely inhibited in the presence of GST alone, indicating that GST by itself could not bind to hemin. In contrast, protein-priming activity was restored almost completely in the presence of increasing concentrations of the TP domain, suggesting the existence of a hemin binding site in the TP domain. Addition of the isolated RT domain could also partially restore protein priming, although to a lesser degree than that achieved with the TP domain, suggesting that the RT domain may harbor another hemin binding site with a lower affinity than that of the TP site.

FIG. 6.

Rescue of protein priming by excess TP and RT domains but not by ɛ RNA. (A) The protein-priming activity of GST-MiniRT2 (0.05 μM) was measured in the presence (odd-numbered lanes) or absence (even-numbered lanes) of hemin (1 μM). As indicated, increasing amounts of ɛ RNA were added to the reaction mixtures. The ³²P-labeled GST-MiniRT2 is indicated. (B) The protein-priming activity of GST-MiniRT2 (0.05 μM) was measured in the presence (even-numbered lanes) or absence (odd-numbered lanes) of hemin (1 μM). As indicated, different amounts of GST (lanes 3 to 6), GST-TP (lanes 7 to 12), or GST-RT (lanes 13 to 18) were also added before hemin. C, control reactions without additional proteins (lanes 1 and 2). The ³²P-labeled GST-MiniRT2 is indicated. Protein-priming activity in the presence of hemin is expressed as a percentage of that without hemin under each condition.

Hemin and its analogs inhibited HBV RT-ɛ RNP complex formation.

Since the above-mentioned results were obtained using DHBV RT, we then asked whether hemin was able to inhibit HBV RT-ɛ RNA interaction, as measured by using the in vitro HBV RNA binding assay (16, 17). A purified HBV RT protein (GST-HTPRT/Dra) (17) was incubated with the radiolabeled HBV ɛ RNA, with or without different concentrations of hemin or its analogs. The results showed that HBV RT-ɛ interaction was inhibited in a dose-dependent manner by hemin (Fig. 7) and its analogs protoporphyrin IX, protoporphyrin IX disodium, and biliverdin (data not shown). In contrast, there was no significant inhibition by either porphine or meso-tetraphenylporphine (Fig. 7), delta-aminolevulinic acid, or porphobilinogen (data not shown). These results agreed very well with the inhibitory effects of these compounds on the DHBV RT-ɛ RNA interaction and thus indicated that hemin and its analogs could inhibit RT-ɛ RNA interaction in both HBV and DHBV.

FIG. 7.

Inhibition of HBV RT-ɛ complex formation by porphyrin compounds. GST-HTPRT/Dra, purified from bacteria, was measured for ɛ binding activity, using an in vitro gel mobility shift assay, in the absence of any treatment (control [C]) (lanes 1 and 4) or in the presence of hemin (lanes 2 and 3), porphine (PP) (lane 5), or meso-tetraphenylporphine (MTPP) (lane 6) at the concentrations indicated. The ³²P-labeled free HBV ɛ RNA and GST-HTPRT/Dra-ɛ complex are indicated. The ɛ binding activity is expressed as a percentage of that of untreated controls.

DISCUSSION

The HBV RT protein plays a multitude of critical functions in viral assembly and replication and has been the focus of intense efforts to develop anti-HBV chemotherapy. The only function of RT that has been successfully targeted so far is its DNA synthesis activity, which is inhibited by several nucleoside analog drugs directed at the polymerase active site. The phenylpropenamide derivatives AT-61 and AT-130 have been reported to inhibit HBV pgRNA encapsidation, but the exact target, either viral or host, remains unknown (10, 23). Here, we report that the essential RT interaction with the viral RNA signal, ɛ, which is critical for protein-primed initiation of viral reverse transcription, as well as the packaging of RT and ɛ into replication-competent nucleocapsids, could be inhibited dramatically by the endogenous small molecule hemin and several related porphyrin compounds. Furthermore, in contrast to nucleoside analogs, hemin and its analogs could target the N-terminal TP domain, a region conserved in all hepadnaviruses but absent from any other known proteins.

Hemin is a cyclic compound formed by the linkage of four pyrrole rings with a central iron ion (Fig. 1). We excluded the requirement for iron by demonstrating that two iron-free hemin analogs, protoporphyrin IX and protoporphyrin IX disodium, showed inhibitory activities similar to that of hemin. On the other hand, biliverdin, which is structurally similar to protoporphyrin but with an open porphine ring, was approximately fourfold less active, suggesting that the closed porphine ring may contribute to the potency of inhibition. Moreover, delta-aminolevulinic acid, a metabolic precursor of hemin with a simple linear structure similar to the side chain modification of hemin, and porphobilinogen, another precursor of hemin with only a single pyrrole ring, showed little inhibitory activity, indicating that the complex tetrapyrrole (porphine) ring may contribute to the strong inhibitory activity (Fig. 1). On the other hand, porphine, a compound with the tetrapyrrole ring alone without any side chain modification, was ineffective, demonstrating that side chain modifications, in addition to the porphine ring, were required for efficacy. Furthermore, meso-tetraphenylporphine, which contains the tetrapyrrole ring with four phenyl side chain groups, showed little inhibitory effect on RT-ɛ interaction, indicating that only certain specific side chain modifications could confer the inhibitory function. Further structure-function analyses may help to produce synthetic porphyrin-based compounds, particularly with certain specific side chain modifications, that will show even greater potency and, more importantly, specificity (see below) against HBV RT-ɛ interaction and thus can be developed as novel, effective anti-HBV drugs.

Several lines of evidence indicated that the inhibition of in vitro DHBV RT protein-priming activity by hemin and its analogs were mostly, if not exclusively, due to the blockade of RT-ɛ interaction. First, inhibition of RT-ɛ binding could be directly measured in vitro using the gel mobility shift assay, which showed that hemin and related compounds behaved almost identically in the protein-priming and RNA binding assays. The only other compound that is known to block hepadnavirus RT-ɛ binding is the antibiotic geldanamycin (17, 19, 21). In contrast to hemin, which directly targets the RT protein (see below), geldanamycin indirectly inhibits HBV and DHBV RT-ɛ interactions by blocking the function of the cellular chaperone protein Hsp90, which is required to establish and maintain an ɛ-binding-competent conformation of the RT proteins. Second, the inhibition of protein-priming activity could be partially reversed by allowing the ɛ RNA to bind to the RT protein prior to the addition of hemin. On the other hand, the residual sensitivity of the preformed DHBV RT-ɛ complex to hemin suggests that hemin may have a secondary activity that could block the DNA synthesis function of RT. This second activity was supported by the observation (L. Li and J. Hu, unpublished results) that hemin, at higher concentrations than that needed to block RT-ɛ interaction, could inhibit the HBV endogenous polymerase reaction, which measures the DNA strand elongation activity of RT (following RT-ɛ binding and protein priming) inside native HBV nucleocapsids, and was also able to inhibit the in vitro protein-priming activity of HBV RT protein purified from insect cells (24, 25), which may have the RNA template already bound (25).

In order to block RT-ɛ interaction, hemin could target the RT protein and/or the ɛ RNA. Our results showed that it is unlikely that hemin targets the ɛ RNA. Instead, hemin clearly could target the RT protein. Thus, protein priming could be partially rescued by the presence of excess RT but not excess RNA. Furthermore, the isolated TP domain and, to a lesser extent, the isolated RT domain could rescue protein priming from hemin inhibition by titrating out the compound so that it became unavailable to bind the RT protein. These results thus further indicate the existence of one or more hemin binding sites within the TP, as well as the RT, domain. It has been previously reported that hemin could also inhibit human immunodeficiency virus (HIV) RT (2) by potentially binding to residues 398 to 407 within the connection domain of HIV RT. Further mutagenesis studies showed that the tryptophan residues at positions 401 and 402 were important for hemin binding to the HIV RT (1). Future studies will be required to define the exact hemin binding site(s) in the hepadnavirus TP and RT domains.

Exactly how hemin and related porphyrin compounds block RT-ɛ interaction remains to be elucidated. By binding to the RT protein (via the TP, and potentially the RT, domain), these compounds could alter the RT conformation and render it incompetent for ɛ binding. Since the DHBV MiniRT2 protein, which can function independently of the Hsp90 chaperone complex (39), remains sensitive to the porphyrin compounds, the inhibition of RT-ɛ interaction by these compounds is apparently independent of the RT conformational maturation chaperoned by the cellular proteins. Alternatively, hemin may directly compete for the same binding site of the RT protein that is also involved in ɛ binding, although the lack of rescue of protein priming by excess ɛ RNA appears to argue against this possibility.

For hemin or its analogs to be developed as viable anti-HBV drugs, a major hurdle will have to be overcome, i.e., the specificity of hemin activity. As mentioned above, hemin has been shown to inhibit retroviral RT and other DNA polymerases (6, 35). We have also found that hemin and the related porphyrin compounds could inhibit the purified Moloney murine leukemia virus RT and the Klenow fragment of the Escherichia coli DNA polymerase I with a potency only slightly less than their inhibitory effects on the HBV and DHBV RT (Li and Hu, unpublished). This pleiotropic effect has hampered our efforts so far to demonstrate clearly the assumed anti-HBV activities of the porphyrin compounds in the cells. At concentrations that were tolerated by the cells, hemin and protoporphyrin IX could indeed partially block HBV and DHBV DNA synthesis and RNA packaging (as predicted from their in vitro inhibitory effect on RT-ɛ interaction) in transfected hepatoma cells (Li and Hu, unpublished). However, the compounds also significantly decreased the expression levels of the viral capsid protein, which is also required for viral RNA packaging and DNA synthesis (31), making the interpretation of these results difficult. Additional structure-function analyses are clearly required to improve the targeting specificities of the porphyrin compounds while maintaining or enhancing their efficacies. The effects, if any, of the endogenous heme and related metabolites in the cells on natural HBV infection also remain to be explored.

However, the potent inhibitory effects of hemin and its analogs on HBV RT-ɛ interaction and their abilities to target the unique TP domain, as reported here, should encourage further development of this class of compounds as novel, effective anti-HBV agents. As the RT-ɛ interaction plays a critical dual role in the protein-primed initiation of viral reverse transcription, as well as nucleocapsid assembly, agents targeting this essential viral interaction should display high efficacy, as shown here for hemin and its analogs. Together with the previous discovery that the antibiotic geldanamycin can block HBV RT-ɛ interaction (indirectly by targeting the required host cofactor, Hsp90), the observations reported here should facilitate efforts to target RT-ɛ interaction, perhaps through combined action against both the RT protein itself and host cofactors. These agents should complement the current nucleoside analog drugs well, as they can target a distinct reaction (RT-ɛ interaction) and a different domain (TP) of the RT protein or a host protein and thus should synergize with existing nucleoside analog drugs and, furthermore, remain potent against drug-resistant HBV variants selected for by nucleoside analog therapy.