DOI: 10.1128/JVI.01862-08
Crystal structures of two coronavirus ADP-ribose-1 ″-monophosphatases and their complexes with ADP-ribose: a systematic structural analysis of the viral ADRP …
The coronaviruses are a large family of plus-strand RNA viruses that cause a wide variety of
diseases both in humans and in other organisms. The coronaviruses are composed of three
main lineages and have a complex organization of nonstructural proteins (nsp9s). In the
coronavirus, nsp3 resides a domain with the macroH2A-like fold and ADP-ribose-1"-
monophosphatase (ADRP) activity, which is proposed to play a regulatory role in the
replication process. However, the significance of this domain for the coronaviruses is st
diseases both in humans and in other organisms. The coronaviruses are composed of three
main lineages and have a complex organization of nonstructural proteins (nsp9s). In the
coronavirus, nsp3 resides a domain with the macroH2A-like fold and ADP-ribose-1"-
monophosphatase (ADRP) activity, which is proposed to play a regulatory role in the
replication process. However, the significance of this domain for the coronaviruses is st
Article
The coronaviruses are a large family of plus-strand RNA viruses that cause a wide variety of diseases both in humans and in other organisms. The coronaviruses are composed of three main lineages and have a complex organization of nonstructural proteins (nsp's). In the coronavirus, nsp3 resides a domain with the macroH2A-like fold and ADP-ribose-1"-monophosphatase (ADRP) activity, which is proposed to play a regulatory role in the replication process. However, the significance of this domain for the coronaviruses is still poorly understood due to the lack of structural information from different lineages. We have determined the crystal structures of two viral ADRP domains, from the group I human coronavirus 229E and the group III avian infectious bronchitis virus, as well as their respective complexes with ADP-ribose. The structures were individually solved to elucidate the structural similarities and differences of the ADRP domains among various coronavirus species. The active-site residues responsible for mediating ADRP activity were found to be highly conserved in terms of both sequence alignment and structural superposition, whereas the substrate binding pocket exhibited variations in structure but not in sequence. Together with data from a previous analysis of the ADRP domain from the group II severe acute respiratory syndrome coronavirus and from other related functional studies of ADRP domains, a systematic structural analysis of the coronavirus ADRP domains was realized for the first time to provide a structural basis for the function of this domain in the coronavirus replication process.The coronaviruses are positive-strand RNA viruses with the largest known genome sizes and the most complex replication mechanisms. After generations of evolution, the coronaviruses that have been characterized to date produce a striking number of virus-encoded nonstructural proteins (nsp's) which assemble into a large membrane-bound complex to perform the rapid viral replication process (23, 30, 35, 46). Current understanding of the coronavirus genome suggests that a single large replicase gene encodes all the proteins involved in the process. This gene contains two open reading frames (ORFs) (designated ORF1a and ORF1b) and is transcribed into two polyproteins, pp1a (from ORF1a) and pp1ab (from ORF1a and ORF1b) (46). The synthesis of the ORF1b-encoded part in the latter polyprotein requires a −1 ribosomal frameshift upon translation of the viral mRNA (8, 9). In order to produce functional nsp's, the two polyproteins are cleaved by two virus-encoded proteases, the main protease (Mpro or 3CLpro) and the papain-like protease (PLpro), to produce up to 16 nsp's (nsp1 to nsp16), the final product of this intricate process (46, 48). Among these nsp's, nsp3 is the largest and possesses a variety of putative domains that are conserved among coronaviruses. These domains have been shown to harbor diverse enzymatic activities, including a domain with ADP-ribose-1"-monophosphatase (ADRP) activity (14, 37, 46, 47). As structural and functional evidence accumulates, it would appear that the enzymatic activities harbored by the viral nsp's are essential for the coronavirus to achieve its highly coordinated replication process (4, 5, 7, 17, 19, 20, 41, 42, 45).The ADRP domain of nsp3 is proposed to belong to the macroH2A-like family, which is characterized by the possession of a structural module called the “macro domain” with high-affinity ADP-ribose (and, in some cases, poly-ADP-ribose [PAR]) binding (21). The macroH2A-like family is named after the nonhistone macro domain of the histone macroH2A, a prototype of this family (28). Noticeably, their recognition of ADP-ribose and its derivative in animal cells has been demonstrated to be associated with many key physiological processes including ADP ribosylation, an important posttranslational protein modification involved in DNA damage repair, transcription regulation, chromatin remodeling, and so on (1, 21, 33). The coronaviruses characterized to date all possess the ADRP domain as part of nsp3, yet very few other viruses are known to contain this module. Only rubella virus, alphaviruses, and hepatitis E virus have been shown to possess an ADRP domain to date (37). Given the ubiquity and functional significance of the macroH2A-like family of proteins, it would seem that viral ADRP domains may play an essential role in the replication of coronavirus or other viruses containing such a module. How this domain is involved in the complicated viral replication process or why it exists exclusively in such a limited range of virus families remains unclear. Until now, there has been no clear evidence to suggest any specific interactions between the viral ADRP domains and biological pathways in the host cells. Moreover, a reverse genetics study recently revealed that mutations in the active site of the viral ADRP domain resulted in no significant effects on virus replication when viral transcription levels were assayed in cell culture. Hence, it has been suggested that this domain may be involved in the regulation of viral replication rather than in the process itself (31).In yeast (Saccharomyces cerevisiae) and plant cells, proteins with the macroH2A-like fold have been shown to involve in the tRNA splicing pathway by acting as an ADRP (22, 25, 36). Further studies from both structural and functional perspectives have confirmed that the ADRP domains in coronaviruses, including severe acute respiratory syndrome coronavirus (SARS-CoV), human coronavirus 229E (HCoV-229E), and transmissible gastroenteritis virus, also possess this enzymatic activity with high specificity. Although this may point toward a potential function of viral ADRP domains in regulating the metabolism of ADP-ribose derivatives, the poor turnover numbers in enzymatic assays (from 5 to 20 min−1 for the three positive-strand RNA viruses reported) indicate an insufficiency in metabolite processing and argue against this hypothesis (12, 25, 31, 32, 34, 37). Another possibility is that viral ADRP domains could serve as PAR-recognizing modules and may interact with host proteins to regulate cellular responses to viral infection. Such processes may include a counteraction of apoptosis-signaling pathways induced by viral entry and the subsequent transcription of the viral RNA genome (16). In support of this hypothesis, a recent structural and functional study on the SARS-CoV ADRP domain demonstrated the mechanism of substrate binding and showed that viral ADRP domains have a high affinity for PAR (12). However, the question of how and why coronaviruses uniquely evolved this domain as part of their replication complex remains a mystery. Thus far, no studies have been conducted that could provide a comprehensive understanding of the significance of the conserved sequence of the ADRP domains among coronavirus and how this conservation is related to their three-dimensional structural features and corresponding functions in the viral replication process.Here we report the crystal structures of two coronavirus nsp3 ADRP domains from avian infectious bronchitis virus (IBV) and HCoV-229E to 1.8-Å and 2.1-Å resolutions, respectively, along with those of their corresponding ADP-ribose complexes. These structures reveal a novel dimerization state in IBV, and, more significantly, observable variations in the structural organization of the substrate binding pocket, despite their conserved amino acid sequence. This is the first structure-based comparison of viral ADRP domains involving three distinct structures, from HCoV-229E, SARS-CoV, and IBV, which are related to each of the three main coronavirus lineages currently identified (38). Subsequent analysis of the structural and functional differences of viral ADRP domains found in the three coronavirus groups demonstrates a highly conserved active site among the coronavirus ADRP domains, from both sequence and structural perspectives. Thus, our work provides the first systematic study of how these highly conserved amino acid sequences translated into three-dimensional structural features that direct the function of this domain in the coronavirus life cycle. Collectively, these results could provide insights into the potential role of the viral ADRP domain in the coronavirus replication process and host-virus interaction and in the evolution of coronavirus nsp's. Additionally, our study may shed new light on the structurally based design of new antiviral drugs targeting the active site harbored in viral ADRP domains, an approach which has been demonstrated in previous reports concerning coronavirus main protease (42-44).MATERIALS AND METHODS
Protein expression and purification.The sequences encoding the nsp3 ADRP domains from IBV (isolate M41, residues 1005 to 1178 of the polyprotein) and HCoV-229E (residues 1269 to 1436 of the polyprotein) were cloned from virus cDNA libraries by PCR. The two sequences were both inserted between the BamHI and XhoI sites of the pGEX-6p-1 plasmid (GE Healthcare). The forward and reverse PCR primers used for amplification were IBV-nsp3-ADRP-F (5′-CGGGATCCGTTAAACCAGCTACATGTGA-3′), IBV-nsp3-ADRP-R (5′-CCGCTCGAGTTACTTACAAGTTGCATCGAAAT-3′), 229E-nsp3-ADRP-F (5′-CGCGGATCCAAAGAGAAGTTGAACGCCT-3′), and 229E-nsp3-ADRP-R (5′-CCGCTCGAGTTACACTAAACCAGACACAA-3′). The resulting plasmids with the two inserted sequences were transformed into Escherichia coli BL21(DE3) cells as glutathione S-transferase (GST) fusion proteins IBV-nsp3-ADRP-GST and 229E-nsp3-ADRP-GST and purified using glutathione affinity chromatography. The GST tag was removed by PreScission protease (GE Healthcare), leading to five additional residues (GPLGS) at the N terminus for both proteins. The proteins were further purified by cation-exchange chromatography using a Resource S column (GE Healthcare) with elution buffer containing 20 mM MES (morpholineethanesulfonic acid) (pH 6.0), 1 M NaCl and by size exclusion chromatography using a Superdex 75 column (GE Healthcare) in 20 mM MES (pH 6.0), 150 mM NaCl. The protein was finally concentrated to 25 mg·ml−1 before crystallization.Protein crystallization.The nsp3 ADRP domains from IBV and HCoV-229E were both crystallized..FIG. 1.
Three-dimensional structures of the viral ADRP domains from IBV and HCoV-229E. (A) Overall structure of IBV ADRP domain in one asymmetric unit. Molecule A (Mol A; red) and Mol B (blue) form a homodimer. (B) Subunit of the IBV ADRP domain (Mol A). Secondary structures (helices, strands, and loops) are colored from blue (N terminus) to red (C terminus) in a rainbow fashion; α-helices are numbered from α1 to α6, and β-strands are numbered from β1 to β6. (C) Subunit of the HCoV-229E ADRP domain. Secondary-structure elements are colored in the same way as for IBV; α-helices are numbered from α1 to α6, and β-strands are numbered from β1 to β7.The nsp3 ADRP domain from HCoV-229E was cloned and expressed in the same manner. The coded protein contains amino acid residues 1269 to 1436 of pp1a, which are renumbered 1 to 168 hereinafter for convenience. The crystal structure was determined using the same SAD method from a Se-Met derivative diffracting to 2.1-Å resolution, as described in Materials and Methods. In the HCoV-229E crystal, the nsp3 ADRP domain exists as a single molecule in the asymmetric unit with dimensions of approximately 35 by 40 by 45 Å3. After final refinement, electron densities for the five leading residues left from the tag and Val168 at the C terminus were not observed. The final refinement statistics are also shown in Table 1.The monomer fold.In the crystal of the full-length IBV nsp3 ADRP domain, each subunit is comprised of six α-helices and six β-strands (Fig. 1B). As typically observed for the macroH2A-like fold, the six β-strands assume an almost parallel three-dimensional arrangement in the order of β1-β6-β5-β2-β4-β3 to form a central six-stranded β-sheet (21). The last strand on one side of the sheet, namely, the β3 strand, is uniquely antiparallel to the rest. The surrounding six α-helices have a sandwich-like topology and form a three-layered α/β/α motif with the central β-sheet, with three on one side of the sheet, namely, α1, α2, and α3, and the other three on the other side. In the HCoV-229E nsp3 ADRP domain crystal, despite the same α/β/α three-layer overall arrangement, the monomer has an additional β-strand at the N terminus compared with its counterpart from IBV (Fig. 1C). This β-strand and the other six β-strands constitute the central β-sheet in the order β1-β2-β7-β6-β3-β5-β4. The first and last strands are antiparallel to the rest. The overall topology of the HCoV-229E nsp3 ADRP domain is thus similar to that of the equivalent domain from SARS-CoV, which has been demonstrated in previous reports (34)... these results from the structure-based comparison, in combination with previous reports on the SARS-CoV nsp3 ADRP domain, unambiguously demonstrate that the viral nsp3 ADRP domain in all three main lineages of coronavirus belongs to the canonical macroH2A-like fold family (34)... In both cases, the ADP-ribose adopts a curved shape as it binds into the pocket. The adenine moiety fits into the hydrophobic cavity formed by residues Leu21, Ala40, Val51, Pro127, Ile133, and Phe159 of the IBV ADRP domain and by residues Val20, Leu46, Pro120, Ile126, Phe150, and Tyr152 of the HCoV-229E ADRP domain. A series of hydrogen bonds are also involved in the binding of ADP-ribose. The N6 atom of the adenine ring makes three hydrogen bonds with surrounding water molecules, through which it interacts with Asp20 in IBV or with the equivalent Asp19 in HCoV-229E (Fig. 4A). The equivalent residue in the SARS-CoV ADRP domain is Asp23, which has also been demonstrated to be involved in hydrogen bonding with the adenosine moiety from previous structural reports (12). This residue has been revealed to be critical for the binding specificity of the ADRP domain by a study on AF1521, a macro domain from Archaeoglobus fulgidus (2). Structure-based sequence alignment of the viral ADRP domain also shows that this residue is highly conserved among the three main coronavirus lineages (Fig. 5). Collectively, these facts indicate that Asp20 in the IBV ADRP domain is indeed conserved in terms of both amino acid sequence and structural interactions, confirming its role in conveying the substrate specificity to the viral ADRP domain. The first ribose moiety and the two phosphate groups make strong hydrogen bonds with the main chain of surrounding residues. This complicated set of residues includes Gly49, Val51, Ala52, Ser130, Gly132, Ile133, and Phe134 in the IBV ADRP domain and Gly44, Leu46, Ala47, Ser123, Gly125, Ile126, and Phe127 in the HCoV-229E ADRP domain. Surprisingly, although these residues are involved only in the binding of the ADP moiety, all of them are highly conserved in sequence among different coronavirus species (Fig. 5).FIG. 4FIG. 5.
Structure-based sequence alignment of the viral ADRP domains from all three main coronavirus lineages. Shown are the following: HCoV-229E (group Ib, DDBJ/EMBL/GenBank accession number P0C6U2); feline infectious peritonitis virus (FIPV; group Ia, DDBJ/EMBL/GenBank accession number Q98VG9); HCoV-NL63 (group Ib, DDBJ/EMBL/GenBank accession number P0C6X5); HCoV-OC43 (group IIa, DDBJ/EMBL/GenBank accession number P0C6X6); SARS-CoV (group IIb, DDBJ/EMBL/GenBank accession number P0C6X7); bat coronavirus HKU5 (BCoV_HKU5; group IIc, DDBJ/EMBL/GenBank accession number P0C6W4); bat coronavirus HKU9 (BCoV_HKU9; group IId, DDBJ/EMBL/GenBank accession number P0C6W5); coronavirus SW1 (CoV_SW1; group III, DDBJ/EMBL/GenBank accession number YP_001876435); and IBV (group III, DDBJ/EMBL/GenBank accession number P0C6V5). Secondary structures of the HCoV-229E ADRP domain (above) and the IBV ADRP domain (below) are indicated in the aligned sequence. Residue numbers of ADRP domain from HCoV-229E are indicated by black dots above the HCoV-229E sequence (one dot corresponding to 10 residues). The residues located in the active site of the ADRP domain, namely, Asn37, His42, Gly43, Gly44, and Phe127 (numbering from HCoV-229E), are labeled by blue arrows. The sequence alignment was generated using MUSCLE (11) and presented using ESPript (15).The terminal ribose, which harbors the site of cleavage in the catalytic hydrolysis reaction, interacts with Asn42, His47, Gly49, and Phe134 in the IBV ADRP domain through a complex hydrogen-bonding network (Fig. 4A). Noticeably, a water molecule serves as an intermediate bridge between the cleavage site on the terminal ribose and the catalytically significant residues, i.e., Asn42 and His47. This indicates that Asn42 and His47 may be responsible for the catalytic activity of the ADRP domain through which ADPR-1"-P is converted into ADP-ribose. This result is consistent with previous structural data obtained from the yeast ADRP domain, in which it was shown to employ similar residues to achieve its catalytic activity (22). Additional biochemical studies on the viral ADRP domain also demonstrated that when the residues in the SARS-CoV ADRP domain corresponding to Asn42, His47, Gly49, and Phe134 in IBV are mutated, the ADRP domain will lose most of its catalytic activity (12).
Inga kommentarer:
Skicka en kommentar