Flavivirus proteases: The viral Achilles heel to prevent future pandemics

Abstract

Flaviviruses are important human pathogens and include dengue (DENV), West Nile (WNV), Yellow fever virus (YFV), Japanese encephalitis (JEV) and Zika virus (ZIKV). DENV, transmitted by mosquitoes, causes diseases ranging in severity from mild dengue fever with non-specific flu-like symptoms to fatal dengue hemorrhagic fever and dengue shock syndrome. DENV infections are caused by four serotypes, DENV1-4, which interact differently with antibodies in blood serum. The incidence of DENV infection has increased dramatically in recent decades and the CDC estimates 400 million dengue infections occur each year, resulting in ∼25,000 deaths mostly among children and elderly people. Similarly, ZIKV infections are caused by infected mosquito bites to humans, can be transmitted sexually and through blood transfusions. If a pregnant woman is infected, the virus can cross the placental barrier and can spread to her fetus, causing severe brain malformations in the child including microcephaly and other birth defects. It is noteworthy that the neurological manifestations of ZIKV were also observed in DENV endemic regions, suggesting that pre-existing antibody response to DENV could augment ZIKV infection. WNV, previously unknown in the US (and known to cause only mild disease in Middle East), first arrived in New York city in 1999 (NY99) and spread throughout the US and Canada by Culex mosquitoes and birds. WNV is now endemic in North America. Thus, emerging and re-emerging flaviviruses are significant threat to human health. However, vaccines are available for only a limited number of flaviviruses, and antiviral therapies are not available for any flavivirus. Hence, there is an urgent need to develop therapeutics that interfere with essential enzymatic steps, such as protease in the flavivirus lifecycle as these viruses possess significant threat to future pandemics. In this review, we focus on our E. coli expression of NS2B hydrophilic domain (NS2BH) covalently linked to NS3 protease domain (NS3Pro) in their natural context which is processed by the combined action of both subunits of the NS2B-NS3Pro precursor. Biochemical activities of the viral protease such as solubility and autoproteolysis of NS2BH-NS3Pro linkage depended on the C-terminal portion of NS2BH linked to the NS3Pro domain. Since 2008, we also focus on the use of the recombinant protease in high throughput screens and characterization of small molecular compounds identified in these screens.

Fig. 1.. Flavivirus NS2B-NS3 protease.

A. Flavivirus polyprotein topology and predicted transmembrane domains. Viral protease NS3 cleaves polyprotein from the cytoplasmic side. B. Arrangement of flavivirus DENV2 NS2B-NS3 protease. NS3 protease uses NS2B, a transmembrane (TM) protein with four helices (cyan), as a cofactor. The conserved hydrophilic region of NS2B (NS2BH; 49–96 aa), linked to NS3Pro (yellow), is required for protease activity. The NS2BH-NS3Pro construct used in the HTS is described below. C. Crystal structure of ZIKV protease (PDB code: 5GPI; (Zhang et al., 2016). The protease consists of NS2B peptide (49 aa, pink) and NS3 protease domain (yellow). The catalytic triad, H51, D75 and S135 are shown.

Fig. 2.. Alignment of NS2B-NS3pro of DENV1, −2, −3, and −4.

The NS2B is 130aa and NS3Pro domain shown is 180aa after QR (the N-terminal 10 residues of NS3Pro are underlined). The sequences in red represent conserved hydrophilic domain essential for protease activity. The amino acid residues in bold of NS2B upstream of the C-terminus of NS2B are hydrophobic and contribute to insolubility in E, coli expression of DENV2 NS2BH (Yusof et al., 2000).

Fig 3.. WNV protease inhibitors with 8-OHQ scaffold.

IC₅₀ was determined by protease assay using DENV2 and WNV NS2BH-NS3Pro. The positions of substitutions in comparison to compound 1 are indicated by an arrow.

Fig 4.. Hit compounds identified by DENV2 NS3 protease HTS.

A. Chemical structures of hit compounds A-H. B. IC₅₀ values of the compounds A-H against DENV1,2,3,4 and WNV NS3 protease. C. Ki, EC₅₀, CC₅₀, and selective index for compound A-H. Ki was measured by in vitro protease assay and EC₅₀ was measured by plaque and replicon assays.

Fig 5.. Substrate binding site of flavivirus protease.

A. Substrate binding site of DENV NS2B-NS3 protease (boxed) is shown in the same orientation as Fig 1C. The substrate-binding site residues that differ in DENV1–4 protease (B) are colored in red. The active site residues are shown in blue. C. DENV2 protease inhibitors identified by virtual screen.

Fig 6.. Hits identified from ZIKV protease HTS and viral infection assays in neural stem cells (NSC).

The 15 compounds can be divided into 9 classes of chemicals, A-I. The IC₅₀ values were determined using ZIKV protease and EC₅₀ determined in ZIKV infection assay in NSC.