Spatial and temporal roles of SARS-CoV PLpro -A snapshot.

SARS-CoV and SARS-CoV-2 encode four structural and accessory proteins (spike, envelope, membrane and nucleocapsid proteins) and two polyproteins (pp1a and pp1ab). The polyproteins are further cleaved by 3C-like cysteine protease (3CLpro ) and papain-like protease (PLpro ) into 16 nonstructural proteins (nsps). PLpro is released from nsp3 through autocleavage, and then it cleaves the sites between nsp1/2, between nsp2/3 and between nsp3/4 with recognition motif of LXGG, and the sites in the C-terminus of ubiquitin and of protein interferon-stimulated gene 15 (ISG15) with recognition motif of RLRGG. Alone or together with SARS unique domain (SUD), PLpro can stabilize an E3 ubiquitin ligase, the ring-finger, and CHY zinc-finger domain-containing 1 (RCHY1), through domain interaction, and thus, promote RCHY1 to ubiquitinate its target proteins including p53. However, a dilemma appears in terms of PLpro roles. On the one hand, the ubiquitination of p53 is good for SARS-CoV because the ubiquitinated p53 cannot inhibit SARS-CoV replication. On the other hand, the ubiquitination of NF-κB inhibitor (IκBα), TNF receptor-associated factors (TRAFs), and stimulator of interferon gene (STING), and the ISGylation of targeted proteins are bad for SARS-CoV because these ubiquitination and ISGylation initiate the innate immune response and antiviral state. This mini-review analyzes the dilemma and provides a snapshot on how the viral PLpro smartly manages its roles to avoid its simultaneously contradictory actions, which could shed lights on possible strategies to deal with SARS-CoV-2 infections.

INTRODUCTION

Being the largest positive‐sense RNA virus, severe acute respiratory syndrome coronavirus (SARS‐CoV) encodes four structural and accessory proteins: spike, envelope, membrane, and nucleocapsid proteins, and two polyproteins: pp1a and pp1ab, which are further cleaved by two proteases into 16 nonstructural proteins (nsps). , These two proteases, 3C‐like cysteine protease (3CLpro) and papain‐like protease (PLpro), come from nsp5 and nsp3, respectively. Of 16 nsps, nsp3 is the longest one with 1922 amino acid residues, and can be further cleaved into functional components with identified domains (top of Figure 1).

FIGURE 1

Roles of SARS‐CoV PLpro. The green dotted frames labeled with bad effects on SARS‐CoV are the scenarios, which are unlikely to happen when SARS‐CoV prevails

Initially, the domains in nsp3 of SARS‐CoV were predicted and identified as follows: N‐terminal glutamic acid rich acidic‐domain (Ac), X, , PLpro, and Y. , Thereafter, severe acute respiratory syndrome (SARS) unique domain (SUD) was added into the list of domains in nsp3. Eventually, the arrangement of domains from N‐terminus include (a) the first ubiquitin‐like domain (Ubl‐1), (b) a Ac, (c) ADP‐ribose‐1”‐phosphatase (ADRP) domain (X domain), (d) SUD, (e) the second ubiquitin‐like domain at the N‐terminus (Ubl‐2), (f) nucleic acid binding domain (NAB), (g) the marker domain (G2M), (h) a metal‐binding region (ZF) domain formed by four predicted transmembrane domains (TM1‐TM4), and (i) Y domain. ,

The release of 3CLpro and PLpro from pp1a/pp1ab is done through autocleavage at flanking sites on each protease, , , , , and hereafter, they perform the role of transcleavage (top of Figure 1). As ORF1b can encode polyprotein only if a ribosomal frameshift from ORF1a into ORF1b occurs, , presumably there will be a single copy of PLpro and 3CLpro when the frameshift does not occur but two copies of PLpro and 3CLpro when the frameshift occurs within a replication cycle of SARS‐CoV because the range from 18% to 30% was found for the frameshift frequency, of which less than 1% can synthesize pp1ab.

It is generally considered that PLpro has two roles: (a) the role of protease, by which it cleaves three sites between nsp1/2 (glycine‐180/alanine‐181), between nsp2/3 (glycine‐818/alanine‐819) and between nsp3/4 (glycine‐2740/lysine‐2741) (top of Figure 1), and (b) the role of deubiquitinating (DUB) and deISGylating enzyme, by which it deconjugates ubiquitin (the downward arrow in left part of Figure 1) and ubiquitin‐like (UBL) protein interferon‐stimulated gene 15 (ISG15) from their substrates (the downward arrow in right‐middle part of Figure 1).

Here, the functionality of PLpro is still not very clear: does a PLpro do several tasks throughout its lifetime or does a PLpro do only a single task during its lifetime? This question comes from the fact that PLpro domain alone is enough to cleave nsp1/2 and nsp2/3 sites with different efficiencies, but PLpro needs the downstream hydrophobic domain (amino acid residues 2207 to 2365) inserted into membranes in order to cleave nsp3/4 site in SARS‐CoV , (right top part of Figure 1). As PLpro is left from nsp3 through autocleavage, it must have a certain mechanism that PLpro associates with the downstream hydrophobic domain in order to cleave nsp3/4 site in SARS‐CoV. Since it is not clear whether or not PLpro is disassociated from the downstream hydrophobic domain, the open question is that two cleavage mechanisms may exist because the association of PLpro with the hydrophobic domain may change PLpro 3D structure.

Although the job of cleavaging nsp1/2, nsp2/3, and nsp3/4 by PLpro is needed for coronavirus RNA synthesis, another important role played by PLpro is directly related to its deubiquitinating and deISGylating functions because the deubiquitinating function is closely related to inhibiting the innate immune response to SARS‐CoV infection , , (the blue frame in left bottom part of Figure 1), and the deISGylating function is closely related to inhibiting the antiviral state (the violet dotted frame in middle bottom part of Figure 1). In fact, PLpro is structurally similar to the cellular deubiquitinase, herpesvirus‐associated ubiquitin‐specific protease (HAUSP)/USP7/USP14, , which is one of more than hundred human deubiquitinating enzymes (DUBs) encoded by human genome.

As a matter of fact, the deubiquitinating and deISGylating roles are essentially related to the proteolytic function of PLpro because both ubiquitin and ISG15 have the LXGG recognition motif at their C‐terminus, which are exactly as the same as the cleavage sites between nsp1/2, between nsp2/3 and between nsp3/4 (the blue dashed and violet dotted frames in lower part vs top part of Figure 1). These are very suggestive since the above mentioned literature indicates that the PLpro has multiple tasks to do in host cells.

It is generally considered that an enzyme usually works on a single type of substrate rather than several types of substrates simultaneously, thus, PLpro may have to make a choice as seen in most microorganisms. , In this mini‐review, we analyze the literature along this line of thought.

SWITCH BETWEEN TWO ROLES OF PLpro

X‐ray crystallography reveals that PLpro has both peptidase and isopeptidase activities, which correspond to the two roles of PLpro (top of Figure 1). It is likely that the PLpro peptidase function is exclusively directed to the recognition motif of LXGG between nsp1/2, between nsp2/3, and between nsp3/4. By contrast, an isopeptide bond is often associated with the ε‐amino group of lysine because an ubiquitin has seven lysines. It turns out RLRGG as recognition subsites at C‐terminus of ubiquitin, which suggests a different mechanism for substrate recognition, , , , because the cleavage of RLRGG‐7‐amido‐4‐methylcoumarin (AMC) (k cat/K  =  5.5 × 103 M−1 s−1) is slower than ubiquitin‐AMC (k cat/K  =  7.5 × 104 M−1 s−1). , Surely, these are kinetic evidence that PLpro does not work equally on its substrates, and its catalytic order is ISG15, ubiquitin, and then, nsp1/2, nsp2/3, and nsp3/4 in terms of efficiency and rapidity. , , When including chained ubiquitins, PLpro catalytic order is K48‐linked ubiquitin, ISG15, K63‐linked ubiquitin and ubiquitin.

Here, the recognition subsites with RLRGG may serve as the switch between peptidase and isopeptidase activities because RLRGG are located from R72 to G76 in ubiquitin and ISG15, while G76 would form an isopeptide bond with ε‐amino group of lysine side chains of another ubiquitin. Sometimes, C77 in ubiquitin can exist if a cysteine is attached, whose proteolysis could be inhibited by N‐ethylmaleimide (NEM) if we consider a similar case in 3CLpro.

Both ubiquitin and ISG15 play their roles in conjugating their target proteins through their C‐terminus. For example, K48‐linked ubiquitin conjugates NF‐κB inhibitor (IκBα) blocking the activation of the NF‐κB signaling pathway, , which initiates the immune response to viral infection , , , (the green dotted frame in left part of Figure 1). As another example, ISG15 conjugates its target proteins to establish the antiviral state (the green dotted frame in bottom middle part of Figure 1).

Overall, the cleavage of these conjugations by PLpro through an isopeptide bond with the recognition subsites of RLRGG would inhibit a series of activities against SARS‐CoV. Indeed, the current treatment coronavirus disease 2019 (COVID‐19) heavily depends on the patient immune system. For this reason, SARS‐CoV PLpro is considered as a strong antagonist against the ubiquitin‐dependent cellular responses to viral infection.

An interesting question is what form PLpro adopts when it cleaves the isopeptide bond with the recognition subsites RLRGG? This is not only because PLpro has two forms, one for the cleavage of nsp1/2 and nsp2/3, another for the cleavage of nsp3/4, which requires the attachment of PLpro to membrane, , but also because ubiquitin and ISG15 are in cytosol.

It showed that pp1a and pp1ab were difficult to be detected in vivo, and this difficulty is most probably due to the fact that pp1a and pp1ab were cotranslationally and autocatalytically processed into intermediates and nonstructural proteins, although the number and the origin of intermediates need to be determined. These findings implicate that the accumulation of pp1a and pp1ab is unlikely because various domains in nsps (nsp3 in our case) release immediately after the replication of SARS‐CoV.

THE THIRD ROLE OF PLpro

Theoretically, the conjugation of ubiquitin to its substrates requires the sequential three enzymes, E1 (ubiquitin‐activating enzymes) with consumption of ATP, E2 (ubiquitin‐conjugating enzymes) and E3 (ubiquitin ligases). Very likely, the chained ubiquitins, such as K48‐linked ubiquitin and K63‐linked ubiquitin, also undergo E1‐E2‐E3 system (the green leftward arrow in Figure 1) as ISGs including ISG15 do so (the green downward arrow in Figure 1). Consequently, the second role of PLpro as isopeptidase works on the products of E1‐E2‐E3 system (the green dotted frames in middle left and lower parts of Figure 1).

The ring‐finger and CHY zinc‐finger domain‐containing 1 (RCHY1) , , and mouse double minute 2 homolog (MDM2) are two E3 ubiquitin ligases. , , RCHY1 acts on the central region of p53, and then, ubiquitinates p53 (the green rightward arrow in Figure 1), and thus, promotes p53 degradation independently of MDM2 (the violet dotted frame in right part of Figure 1). Similarly, MDM2 not only can induce polyubiquitination and degradation of p53 at high levels, but can also cause the monoubiquitination and nuclear export of p53 at low levels, resulting in inhibition of apoptosis and innate immune signaling. At this point, it is not clear whether the chained ubiquitins and ISG15 interact with RCHY1 or MDM2, or both, or other E3 ligases such as ECS E3 ligase complex (Elongin B/C‐Cul2/5‐SOCS‐box protein ubiquitin ligase complex).

It seems counter‐intuitive for the common concept on the ubiquitin system because the ubiquitin system in principle should play positive roles to conjugate useless or harmful proteins, , but here ubiquitin system conjugates active and helpful proteins. Since p53 inhibits the SARS‐CoV replication and other virus infections, , p53 plays a positive role in antiviral infection (the green dotted frame in lower right part of Figure 1), for which E3 ubiquitin ligases should not ubiquitinate p53 leading to its degradation.

PLpro can stabilize RCHY1, which has a half‐life of 3.5 hours, and even more strongly stabilizes RCHY1 with SUD together through the interaction between residues 95‐144 of RCHY1 and 389‐652 of SUD (SUD‐NM) (upper middle part of Figure 1). As PLpro is autoproteolytic cleaved from nsp3, a question raised here is which protease cleaves SUD from nsp3 owing to the fact that SUD does not have the recognition subsites when looking at the terminus of their sequences (accession number: 2W2G_A and 2W2G_B)? Although the autocleavage mechanism is not fully understood, , , SUD is unlikely to be released from nsp3 through the autocleavage mechanism. Another mechanism, which needs to explore, is how PLpro is fused with SUD in vivo since PLpro should be released by autocleavage without SUD?

Evidently, a dilemma appears in consideration of PLpro role on E3 stabilization (central part of Figure 1). On the one side, the stabilization of E3 promotes the ubiquitination of p53 (the violet frame in middle right part of Figure 1), and thus, minimizes the role of p53 in inhibiting SARS‐CoV replication in host cells. On the other side, the stabilization of E3 promotes the ubiquitination of IκBα by K48‐linked ubiquitin, the ubiquitination of TNF receptor‐associated factors (TRAFs) and stimulator of interferon genes (STING) by K63‐linked ubiquitin, and the ISGylation of targeted proteins by ISG15, and thus, initiates the innate immune response to SARS‐CoV infection and antiviral state (the green frames in left and middle‐bottom parts of Figure 1). This dilemma is that the ubiquitination of p53 is useful and helpful against SARS‐CoV infection, whereas the ubiquitination of IκBα, TRAFs, and STING, and the ISGylation of targeted proteins are bad against SARS‐CoV infection. In short, the role of PLpro/PLpro+SUD on stabilizing E3 is to promote ubiquitination and ISGylation, which consequently plays dual effects: promoting intercellular reaction to SARS‐CoV (innate immune system response and antiviral state), but promoting SARS‐CoV replication within host cells (virus replication).

Furthermore, the role of PLpro is to cleave the ubiquitinated IκBα, TRAFs, and STING, and the ISGylated targeted proteins through the recognition subsites with RLRGG. The deubiquitination of IκBα, TRAFs, and STING, and the deISGylation of targeted proteins are good for SARS‐CoV because these actions inhibit the innate immune response and antiviral state (the blue frame in left part and violet dotted frame in middle‐bottom part of Figure 1). However, PLpro does not deubiquitinate the ubiquitinated p53. If such deubiquitination occurs, it will be bad for SARS‐CoV (the violet frame in right‐bottom part of Figure 1) because p53 inhibits SARS‐CoV replication. Although PLpro has the role of deubiquitination, it is not the case in this experiment. In short, the final consequence of second dilemma in effect is inhibiting intercellular reaction to SARS‐CoV, but inhibiting SARS‐CoV replication within host cells.

Collectively, PLpro has two important recognition functions, one for the recognition of LXGG in pp1a/pp1ab and of RLRGG in C‐terminus of ubiquitin and of ISG15, and the other for the recognition of a ring‐finger domain in E3. The latter enhances the functions of E3, which bring about counter effects on viral infection and on host‐cell reactions.

STRATEGY OF PLpro

The above analyzed dilemma reveals that PLpro may play contradictory roles in SARS‐CoV infection. It is very unlikely that the viruses evolve and conserve the codes whose products are harmful to themselves. Therefore, we need to figure out what strategy PLpro adapted to bypass this dilemma. If we arrange the roles of PLpro along the time course, we can find that these roles are well managed by SARS‐CoV (Figure 2).

FIGURE 2

Proposed strategy of PLpro to bypass its contradictory roles in infected host cells according to its kinetic preference

SARS‐CoV PLpro has its choice to cleave the chained ubiquitins, ISG15, ubiquitins (if any of them exists), and the sites between nps1/2 and between nps2/3 because the PLpro should kinetically prefer to recognition subsites RLRGG than LXGG. Without knowing the kinetics of association of PLpro with its downstream hydrophobic domain for cleaving nsp3/4, this kinetic preference dictates viral PLpro to cleave chained ubiquitins, ISG15, ubiquitins before cleaving nps1/2, nps2/3, and nsp3/4 (Event 2 vs Event 4 in Figure 2). This strategy suggests that SARS‐CoV at first inhibits the innate immune response and the initiation of viral state in infected host cells, and then, begins its replication (Event 3 vs Event 4 in Figure 2). The release of SUD is likely to be simultaneously with the accomplishment of cleaving nsp1/2, nsp2/3, and nsp3/4 by PLpro (Event 5 in Figure 2). Thereafter, PLpro and SUD work alone or together to stabilize E3 (Event 6 in Figure 2), and finally, proceed to p53 ubiquitination and degradation (Events 7 and 8 in Figure 2). Hence, PLpro accomplishes its series of tasks for viral infection.

Accordingly, there may theoretically be several interesting points to explore what happens after the release of PLpro from nsp3. Could it be possible to apply the artificial peptides of RLRGG as an aptamer to the recognition subsites in the host cells in order to derail the cleavage of nps1/2, nps2/3, and nsp3/4 to prevent the replication of SARS‐CoV? Also, it is not clear whether PLpro fusion with SUD is reversible or not. If it would not be so, then mimic of SUD interactive domain with PLpro could be a way to move out PLpro from its active duty.

Although the above discussed mechanisms were obtained from the studies on SARS‐CoV, mainly in vitro studies, it is highly likely that the knowledge on SARS‐CoV is equally applicable to severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) because the domains in nsp3 were predicted to be conserved across all CoV. Therefore, the development of antiviral drugs targeting PLpro is an important strategy. , , ,

Nevertheless, this mini‐review is exclusively directed to the roles of PLpro according to the current knowledge although many other players are also involved in the deubiquitinase activity, for example, the endoribonuclease from nsp15.

In conclusion, this review demonstrates that PLpro evolves a very efficient and clever strategy to suppress the host's response to SARS‐CoV infection, which sheds lights on understanding the roles of PLpro on SARS‐CoV‐2 infection as well as on possible strategies to deal with SARS‐CoV2 infections.

CONSENT FOR PUBLICATION

Both authors read and approved the manuscript.

COMPETING INTEREST

None of the authors has any competing interests in the manuscript.

AUTHOR CONTRIBUTIONS

G. Wu designed this review and wrote the first draft; S.M. Yan visualized the data, and both edited and finalized this review.