Multiple in Vitro Inhibition of HIV ‑ 1 Proteins by 2,6-Dipeptidyl-anthraquinone Conjugates Targeting the PBS RNA

: We recently reported a series of 2,6-dipeptidyl-anthraquinone conjugates (AQs) as Trans-Activation Response element (TAR) RNA-binding agents able to inhibit in vitro the HIV-1 nucleocapsid (NC) protein-mediated processes. Because NC is a highly adaptable nucleic acid chaperone assisting several crucial steps along reverse transcription, in this study we investigate the ability of AQs to interact with other virus-derived nucleic acid structures thus potentially inhibiting multiple NC functions. Focusing on the HIV-1 Primer Binding Site (PBS) RNA sequence, we demonstrate that properly substituted dipeptidyl-anthraquinone conjugates e ﬃ ciently inhibit the NC-mediated primer annealing in the low micromolar range. Similarly, we extended the analysis to the HIV-1 trans-activator of transcription (Tat) peptide, which has been recently shown to mimic the annealer functions of NC upon interacting with the same nucleic acid regulatory sequences. Our results highlight how RNA-targeting agents can act as multimode inhibitors of key viral proteins a ﬀ ecting their chaperone activity in reverse transcription processes.

A ntiretroviral therapy exploits a great variety of drugs that efficiently control HIV-1 infection, but the onset of resistance demands the identification of novel pharmacological targets that are less prone to mutate during chronic treatment. 1 Recently, a great effort has been made looking for strategies to inhibit the HIV-1 nucleocapsid protein (NC), a nucleic acid chaperone that catalyzes many steps of the viral replication process, supporting the activity of the reverse transcriptase (RT) and participating in the late step of viral particle maturation. 2,3 The strict conservation and multiple roles that NC plays throughout the entire viral replication cycles make this protein an attractive and promising target against HIV. 4 Even if different classes of molecules have been proposed as anti-NC agents, 5 drug-candidates interfering with NC functions are still missing. In the attempt to identify small molecules able to impair NC binding/annealing activities, we recently focused our work on dipeptidyl-anthraquinone conjugates (AQs), nucleic acid binding agents acting as threading intercalators able to recognize and stabilize the dynamic structures of nucleic acids which are substrate of NC, thus impairing the expected nucleic acid rearrangements catalyzed by the viral protein. 6−8 These AQs are characterized by a common anthraquinone scaffold and two side chains at opposite sides of the polycyclic aromatic nucleus: the peptidylside chains are built combining a first amino acid referred as "linker" (R1) and a second amino acid acting as a cationic terminal residue (R2), whose function is to facilitate the interaction with the nucleic acid phosphate backbone. We therefore designed a chemical library including four series of analogues identified by the R1 linkers (β-alanine, glycine, L-or D-alanine) and different R2 residues and reported their in vitro activity as potent inhibitors of the NC-mediated TAR/cTAR annealing, a suitable in vitro model of the minus-strand transfer event during the HIV-1 reverse transcription process. 6−8 A schematic representation of AQs chemical structure is shown in Figure 1A.
Interestingly, the structure−activity relationship (SAR) of these NC inhibitors allowed identification of positive hits targeting preferentially the Trans-Activation Response element (TAR), an essential RNA substrate of NC, and we demonstrated that TAR binding is a key event in their molecular mechanisms of NC inhibition. 6,8,9 The studies highlighted the crucial role of the highly basic nature of the terminal (R2) residues as well as the flexibility and stereochemistry of the R1 linker connecting them to the aromatic nucleus to achieve the desired binding and stabilization of the RNA substrate. 6,8 Moreover, these TAR-binding agents displayed the ability to hamper specific interactions of TAR with a peptide mimicking the HIV-1 trans-activator of transcription (Tat) factor, 8,9 opening the possibility to consider AQs as multitarget agents acting against different viral proteins through specific interaction with the common TAR RNA substrate. RNA-binding small molecules are an important and highly challenging area of therapeutic research. 10 Our findings prompted us to further investigate selected 2,6-dipeptidyl anthraquinones as RNA-specific ligands.  Since many RNA regulatory sequences in the HIV-1 genome are locked up into stable secondary structure requiring the NC chaperone activity, we propose that AQs may target several of these nucleic acid structures, potentially inhibiting NC at multiple levels. To test this hypothesis, we moved our attention to the Primer Binding Site (PBS) RNA sequence. In the early steps of reverse transcription, the primer tRNA Lys3 must be annealed to the complementary PBS sequence of the viral RNA genome. This RNA−RNA annealing allows the minus-strand synthesis, a step not occurring in the absence of NC. 12 Annealing involves a stretch of 18 nucleotides in the tRNA Lys3 3′-acceptor stem, with the same sequence as (−)PBS RNA, forming an RNA-duplex with PBS, (+)PBS RNA, that is selectively used as primer by the HIV-1 RT. 12 The stable secondary structure of (+)PBS RNA, similarly to TAR RNA, is the target of the chaperone activity of NC. 13 In this study, we therefore selected from our chemical library of AQs a cluster of 2,6-dipeptidyl-anthraquinone conjugates ( Figure 1B) to explore their putative ability to inhibit NC-mediated (+)/(−)PBS RNA annealing in vitro. Interestingly, the placement of tRNA Lys3 primer onto viral RNA is also promoted by the HIV-1 trans-activator of transcription (Tat) protein. 14,15 This protein is thought to support the annealing activities of NC during reverse transcription, thus earning itself the definition of nucleic acids annealer. 16,17 For this reason, we developed an assay relying on the annealer activity of both NC and Tat. We named it "RNA annealer-mediated primer annealing" (RAMPA) assay, since it allows us to employ either NC or Tat to mediate the annealing of the (+)PBS RNA construct with its complementary sequence (−)PBS RNA, leading to the formation of the annealed RNA-duplex (dsRNA) as shown in Figure 2. In RAMPA assay, the analysis of the (+)/(−)PBS RNA annealing reaction can be investigated by native polyacrylamide gel electrophoresis (PAGE).
We employed the RAMPA assay to assess the ability of AQs to interfere with the annealer-mediated (+)/(−)PBS RNA annealing in vitro. Preliminarily, we optimized the RAMPA assay with NC. Analysis of the (+)PBS RNA secondary structure by UV-absorbance and circular dichroism melting/ annealing experiments, reported in Figure S1 (Supporting Information), confirmed the predicted secondary structure of the RNA construct consistent with the 4-base pair stem reported in Figure 2. In the absence of the annealer proteins, spontaneous RNA-duplex formation occurred rather slowly with less than 40% of RNA-duplex formed after incubation at room temperature up to 60 min ( Figure S2, Supporting Information), indicating that (+)PBS and (−)PBS RNA are kinetically stable into monomeric structures under the condition used. The (+)/(−)PBS RNA annealing reaction was instead extremely accelerated by the presence of the fulllength NC protein. A representative RAMPA assay is reported in Figure 3, showing that complete RNA−RNA annealing was achieved after 10 min in the presence of 3 μM NC, i.e. at 12 nucleotide (nt):NC molar ratio. As an additional control experiment, we performed RAMPA assay in the presence of APO-NC (a zinc-less version of NC prepared as detailed in the Supporting Information). The results are reported in Figure S3 (Supporting Information) and showed that the APO-NC possesses a similar annealer activity compared to the zincbound protein. These results confirmed that even in the absence of folded Zn fingers the APO-NC mutant preserves an efficient annealing activity, while the role of properly folded NC Zn fingers is critical to guide the extended duplex nucleation through loop−loop interactions. 18−21 Then, we set up the RAMPA assay to assess the Tatmediated (+)/(−)PBS RNA annealing. We employed the Tat(48−57) peptide (Figure 2), corresponding to the essential Tat basic domain responsible for the nucleic acid chaperone activity. 14 Similarly to NC, Tat(48−57) was able to dramatically speed up the reaction with a complete annealing of the two RNAs achieved at 6 μM peptide concentration, i.e. Six nt:Tat(48−57) molar ratio, indicating that the basic peptide is about 2-fold less efficient compared to NC as nucleic acid annealer in our RAMPA assay ( Figure S4, Supporting Information).
Having defined a suitable platform to analyze in vitro either the NC-or Tat-mediated (+)/(−)PBS RNA annealing step, we proceeded to evaluate the inhibitory activity of the selected 2,6-dipeptidyl-anthraquinone conjugates. RAMPA assay was therefore performed with all selected compounds to evaluate their inhibitory activity on either NC-or Tat-mediated annealing of (+)/(−)PBS RNA. A representative RAMPA assay conducted in the presence of NC and compound 4a is shown in Figure 4. The gradual disappearing of the RNAduplex band clearly indicates inhibition of the annealermediated annealing in the presence of increasing concentration of compound. Experiments for all compounds are reported in Figures S5 and S6 (Supporting Information).
Quantification of the band corresponding to the annealerinduced RNA-duplex was assessed to calculate the IC 50 as a measure of the ability of each compound to slow down the annealing activity of either NC or Tat(48−57) under the selected experimental conditions. Results are reported in Table  1.
The obtained results allowed for the direct comparison of the inhibitory activity of each compound toward the two different viral proteins. When NC was examined, AQs with 11atom liner-aliphatic side chain and two ionizable groups (1b, 2a, and 4a) resulted in the most active inhibitors, with 4a as the best among all the series (IC 50 value: 12.5 ± 1.5 μM, Table  1). Slightly lower inhibitory activity was observed for 10-atom linear side chain compounds (1c, 2b, 3b, and 4b); the activity drastically decreased by further shortening the side chain length as observed with compounds 4c and 4d. Similarly, inhibition decreased with excessive lengthening of the side chain (12-atom length, 1a). Side chain chirality also played a major role in the NC-inhibition as highlighted by the conjugates bearing the D-or L-alanine linker. Comparing the activity of the D-and L-stereoisomers having the same terminal residue (3a to 4a and 3b to 4b), the non-natural linker configuration (D-Ala) enhances the inhibition of NC activity.
Interestingly, AQs displayed the same behavior in inhibiting the NC-mediated TAR/cTAR annealing step. 8 When RAMPA assay was performed employing Tat(48−57) as annealer, the activity profile of the tested AQs mirrored the one observed for NC (Table 1). These results highlighted side chain length and linker configuration as the major structural determinants affecting the AQs inhibitory activity of both the NC-and Tat-mediated (+)/(−)PBS RNA-annealing. Hence, AQs displayed the same potency scale when testing NC and Tat RNA-annealing properties, suggesting that their mechanism of action is based on their RNA-directed targeting.
Unambiguous evidence of AQs direct binding to the (+)PBS and (−)PBS RNA constructs was assessed by employing electrospray ionization mass spectrometry (ESI-MS) under nondenaturing conditions. 22 Figure S8 (Supporting Information). All compounds were able to bind both RNA   6,7,9 Quantification of the different species peaks reveals the accurate partitioning between free and bound forms in solution, which was used to calculate the relative binding affinity of the various compounds for the different RNA substrates, employing the percentage of bound substrate observed in each spectrum. 6 A graphical representation of the obtained results is reported in Figure S9. Most compounds showed higher affinity for (+)PBS RNA rather than for the (−)PBS RNA sequence. This can be related to a folding structure of (+)PBS that is more favorable for drug binding and to a higher content in purine bases that can offer stronger stacking interactions with the anthraquinone scaffold of AQs. When looking at the structural requirements of AQs that maximize binding to the RNAs, a striking correlation was observed with the SAR obtained by RAMPA assay. The correlation between the inhibitory activity of NC-mediated (+)/(−)PBS RNA annealing and AQs binding affinity for (+)PBS construct is plotted in Figure 5.
Compounds with 11-atom side chain length (1b, 2a, and 4a) proved to be the best binders, followed by shorter analogues (compounds 1c, 2b, and 4b). Affinity was strongly impaired by further shortening the cationic side chains to 9-and 8-atoms (compounds 4c and 4d), by L-conformation of the peptide linker (e.g., compound 3a) and by bulky terminal residues such as those in compounds 4e and 3f. Interestingly, we previously reported analogous SAR when testing the same AQs in the NC-mediated TAR/cTAR annealing process: compounds 1b, 2a, and 4a proved to be the most potent NC inhibitors of our chemical library with comparable activity in the micromolar range, while compound 3f was inactive. 6−8 Also in the TAR system, we observed a strict correlation between NC inhibition and binding affinity to TAR construct. 6,8 These observations demonstrate that RNA binding is a key event in AQs molecular mechanisms of proteins inhibition.
Further investigations were performed to identify the location of the AQs binding sites on the (+)PBS RNA construct. Compound 4a was selected because it resulted to be the most potent NC and Tat inhibitor and the strongest (+)PBS binder to perform a footprinting experiment by MS. The effects of 4a complexation on the gas-phase fragmentation of the (+)PBS structure were investigated by comparing the results provided by tandem MS (MS/MS) of free and bound RNA construct. The free (+)PBS substrate produced the typical fragmentation pattern expected from this type of nucleic acid, 25,26 as shown in Figure S10 (Supporting Information). Complete sequence coverage obtained under the selected experimental conditions offered no evidence of possible fragmentation inhibition that could be ascribed to base-pairing between the stem complementary strands. Collision induced activation of the 1:1 complex of 4a with (+)PBS induced fragmentation of the oligonucleotide backbone ( Figure 6A). A distinctive gap in the fragmentation pattern of (+)PBS corresponding to nucleotides C7−C12 ( Figure 6B) indicates a protective effect on the 5′-side of the hairpin loop structure by the compound (Figure 6C). These results show that 4a preferentially binds to dynamic regions in the RNA structure such as the apical loop of the (+)PBS structure, which can better accommodate both stacking of the anthraquinone scaffold with base pairs and electrostatic interaction between the cationic terminal residues with the phosphate backbone.
In conclusion, this study led to the identification of a promising set of 2,6-dipeptidyl-anthraquinone conjugates able to inhibit NC-and Tat-mediated primer annealing in the low micromolar range. By exploring their SAR, we demonstrated that AQs' mechanism of action resides in their binding with good affinity to the (+)PBS and (−)PBS RNA sequences which are shared substrates of both NC and Tat proteins. Hence, the selected AQs revealed their dual activity opening the possibility to consider them as multitarget inhibitors. In addition, the most potent AQs in the PBS system resulted to be highly active as NC inhibitors also in the previously analyzed TAR system, 6−8 highlighting their putative ability to impair NC activities at multiple levels. These findings indicate that AQs potentially exert a pleiotropic antiviral activity by targeting important RNA regulatory sequences while inhibiting the activity of multiple virus chaperones. NC and Tat are multifunctional proteins involved in crucial steps of the HIV-1 life cycle, both acting through interaction with viral RNA domains. Based on this, we strongly believe that agents capable of interfering with those critical steps are expected to block HIV-1 replication at multiple stages. This multimode mechanism of action represents a valuable approach that may help prevent the onset of drug resistance by targeting multiple and conserved structures within the HIV-1 genome. Although further investigations are needed, additional studies are underway to evaluate the AQs' interactions with other viral RNA domains that possess dynamic structure similar to TAR and PBS RNA, such as the stem loops in the psi-RNA region, Figure 5. Correlation between the inhibitory activity of NC-mediated (+)/(−)PBS RNA annealing and AQs binding affinity for (+)PBS construct. The relative activity in the x-axis was calculated from the data reported in Table 1 by normalizing the IC 50 values of compounds to the value of the best inhibitor compound 4a, obtained by the RAMPA assay conducted with NC. The relative affinity in the y-axis was calculated from the data reported in Figure S9