Large nanodiscs going viral

Abstract

Covalently circularized and DNA-corralled nanodisc technologies have enabled engineering of large sized-bilayer nanodiscs up to 90 nm. These large nanodiscs have the potential to extend the applicability of nanodisc technology from studying small and medium-sized membrane proteins to acting as surrogate membranes to investigate functional and structural aspects of viral entry. Here, we discuss the recent technical developments leading to construction of large circularized and DNA-corralled nanodiscs and examine their application in viral entry.

Introduction

Phospholipid bilayer nanodiscs have attracted great interest from structural biologists as a native-like membrane mimetic for studying membrane proteins. Nanodiscs are detergent-free lipid bilayer models, which enable the study of membrane proteins in a physiologically relevant environment [1–3]. A conventional nanodisc is composed of a phospholipid bilayer patch (~ 8–16 nm in diameter) encircled by two copies of amphipathic helical proteins termed membrane scaffold proteins (MSPs) (Figure 1). MSPs are engineered forms of apolipoprotein A-1, which is the major component of high-density lipoprotein. The hydrophobic residues of MSP interact with the lipid acyl chain; the hydrophilic face is located at the outside surface to allow the nanodisc to stay in solution. Any kind of membrane forming lipid can be used to form nanodiscs and the diameter of the captured lipid patch is determined by the length of the scaffold protein[2,3]. Nanodiscs have been used successfully to stabilize many membrane proteins and render them soluble in aqueous buffers, facilitating high-resolution structural determination in lipid bilayer using cryoEM and NMR. Examples of single-particle cryo-EM structures include the tetrameric TRPV1 ion channel (330 KDa) [4]. The TRPV1 structure revealed the mechanism for channel activation by bioactive lipids. Moreover, it showed how certain phospholipid interactions enhance binding of a spider toxin to the channel. Another remarkable cryo-EM structure is that of the ryanodine receptor RyR1 (2.3 MDa) [5]. The structure identified the calcium binding domain and revealed how the calcium allosterically regulates gating of the channel. There are many more examples of cryo-EM structures of membrane proteins in nanodiscs. These examples include the lipid exporter ABCB4[6], MsbA transporter [7], Transient receptor potential channel subfamily A member 1 (TRPA1)[8], a voltage-activated potassium channel[9], CorA magnesium channel[10]. In addition to facilitating the structure determination of a single membrane protein, nanodiscs have been used to assemble and enable the cryo-EM structure of membrane protein complexes such as the ribosome-SecYE complex[11].

Figure 1.

Traditional nanodisc self-assembly process. Membrane scaffold protein (MSP), lipid and target membrane protein solubilized in detergent assemble nanodisc with incorporated membrane protein upon detergent removal.

Nanodiscs have been also used in both solution and solid-state NMR to reveal critical details of small membrane protein structure and function [12,13]. There are many other applications for nanodisc technology which are covered by several excellent reviews [14–16].

Further engineering efforts are expanding the nanodisc size and stability thus extending the potential applications. Our group and others have developed methods to covalently link the N- and C- termini of membrane scaffold protein variants. As a result of the covalent circularization, we and others have produced nanodiscs with a large range of discrete sizes up to 50 nm [17–19]. The 50 nm nanodisc has been successfully used in studying the dynamics of trans-SNARE complexes and investigate the number of SNAREs needed to drive fusion [20]. Moreover, the 50 nm nanodisc has been used to study poliovirus entry [17]. Other efforts to expand the nanodisc size beyond 50 nm include using a DNA origami scaffold to help construct and stabilize nanodiscs within the DNA cavity. These latter efforts led to the production of nanodiscs up to 90 nm, which were employed to investigate poliovirus interactions with its receptor embedded in bilayer[21]. Here, we give a concise overview of the design and applications of these large nanodiscs in studying viral entry.

Engineering large nanodiscs

1. Covalent protein circularization

The covalent circularization of proteins results in improved thermal stability [22–24] and proteolytic resistance [24–28]. MSP represents an attractive target for covalent circularization because its N- and C- termini are in close proximity [29,30]. There are several established methods for protein circularization. These methods include the use of intein-fusion proteins [31,32], which permit the covalent circularization through expressed-protein ligation or protein trans-splicing. Alternatively, circularized proteins can be produced by using sortase enzymes [33], butelase enzyme [34] or chemical ligations [35]. Combining protein circularization with MSPs enables the creation of large (> 16 nm) and more stable nanodiscs of well-defined circular and polygonal shapes.

Exploiting the potential advantages of circularized proteins, our group engineered recombinant versions of MSP that can be circularized using sortase and we were able to assemble very stable and homogenous nanodiscs of varying sizes up to 50 nm [17]. We have engineered four different variants of circularizable scaffold proteins: NW9, NW11, NW30, and NW50 which assemble ~ 8.5, 11, 15, 50 nm nanodiscs respectively. The NW constructs contain a TEV protease-cleavable N-terminal His tag followed by a single glycine, and a C-terminal sortase-cleavable His tag (Figure 2b). The presence of these two sites ensures covalent linkage between the N and C termini of NWs while still preserving the function to form nanodiscs. Following the publication of the sortase-mediated circularization of nanodiscs, several other groups worked on optimizing the circularization protocol to improve the yield and minimize byproducts. Yusuf et al [19] used detergent during the circularization reaction in order to improve the yield and minimize the high molecular weight oligomers. Another group aimed at continued improvement of the circularized nanodisc protocol and optimized the original MSP sequence to incorporate extra negatively charged residues. With these modifications , they were able to improve the solubility of MSP and perform the purification and circularization reaction in the absence of any detergents [36]. Furthermore, a recent study demonstrated the production of the circularized MSP by using the in vivo split intein ligation in E. coli (Figure 2a). This split intein method has been shown to yield circularized MSP nanodiscs of varying sizes from 8 nm to 26 nm [18].

Figure 2. Different strategy that have been successfully used for production of large nanodiscs.

(a) Design of the Npu DnaE split-intein MSP fusion construct and mechanism of intein splicing in vivo, leading to the generation of circularized MSP. (b) A general outline of the constructs for making covalently circularized nanodiscs using sortase. Glycine (G) and LPXTG at the N and C termini respectively are necessary for the linkage by sortase. (c) DNA-corralled reconstitution strategy. Addition of detergent and lipids to small nanodisc- decorated barrels followed by dialysis results in the construction of a large bilayer nanodisc within the DNA cavity. Each DNA barrel is decorated with ssDNA overhang as handles to hybridize with ssDNA chemically conjugated small nanodisc.

2. DNA-Corralled nanodisc

Recently, we reported a method of constructing a large nanodisc inside the cavity of DNA corrals[21]. Each DNA corral recruits a number of ~11 nm diameter nanodiscs that are assembled by noncircularized, oligonucleotide-functionalized MSPs, and directs their reconstitution into a single large nanodisc (Figure 2c). We employed two different sized barrels: 90 and 60 nm outer diameter to reconstitute ~70 or ~45 nm DNA-corralled nanodiscs, respectively. The DNA corral offers a number of useful features including the ability to assemble large nanodiscs while circumventing the need for sortase circularization and production of larger MSP variants. In addition, DNA corral acts as a bumper case to prevent nanodisc aggregation and facilitates control over stoichiometry, geometry, and orientation of inserted proteins through tethering to the corral. The enclosed nanodiscs are relatively stable and tolerant of a broad range of pH levels and divalent ion concentrations. This DNA-corralled nanodisc system was used to reconstitute two membrane-protein clusters and to study poliovirus entry[21]. Because of the unprecedented possibilities of controlling the stoichiometry and orientation of inserted proteins through tethering to the DNA corral, DNA-corralled nanodiscs have great potential in studying the stoichiometry of viral receptor engagement. Moreover, they can be utilized to incorporate multimeric receptor complexes and study the impact of these complexes on fusion and penetration dynamics as well as subsequent viral uncoating. A similar approach to DNA-corralled nanodisc was reported by Iric et al [37] to prepare DNA-encircled bilayer (DEB) structures. The DEB structures are made of multiple copies of an alkylated oligonucleotide hybridized to a single-stranded minicircle. The presence of the alkyl modification enables interactions between negatively charged hydrophilic DNA and lipids.

Virus entry

1. Imaging of membrane interaction

Viruses interact with, modulate and penetrate the cellular membranes during cell entry and exit. In order to ensure successful viral replication, the entry of viruses into host cells requires the disruption of the membrane without compromising cell integrity. For enveloped viruses, transfer of viral genome is achieved by fusion of viral and cellular membranes[38]. This fusion step is facilitated by the glycoproteins on the surface of enveloped viruses. On the other hand, most of the non-enveloped viruses transiently destabilize the host membranes using amphipathic or hydrophobic capsid peptides[39]. Many questions regarding the mechanisms of membrane penetration, genome translocation and disassembly of non-enveloped virus remain unanswered. Thus far, elucidating these mechanisms have been challenging, due in part to technical difficulties relating to direct visualization of viral gene delivery and size heterogeneity of liposomes, which are commonly used membrane models for nonenveloped viral entry studies. Previous studies using liposomes have resulted in several low-resolution structural models of cell entry and have shed light on the mechanisms of the early stages of viral entry. Strauss and colleagues have shown that poliovirus 135s particles form a 50 Å elongated extension or “umbilical connector”, attaching the poliovirus to the membrane to allow the passage of the viral genome and initiating the infection[40]. Kumar and colleagues reported that the protein shell of membrane-bound human rhinovirus 2 intimately interacts with the membrane; they did not observe any umbilical connector [41]. Another study reported a cryo-EM structure for a complex of a small nanodisc (10 nm) containing coxsackie-adenovirus receptors with human pathogen coxsackie virus B3 (CVB3)[42]. The structure was resolved to 7.8 Å without or to 3.9 Ǻ with imposed icosahedral symmetry and revealed local dynamics in capsid structures. The authors identified an extension of electron density near the nanodisc membrane and formation of a channel on the surface of the virus capsid. Even though the use of small nanodiscs only minimally interfered with capsid imaging and triggered local structural rearrangement, it failed to trigger RNA release.

Unlike small nanodiscs, large covalently circularized nanodiscs successfully triggered RNA release when incubated with poliovirus [17]. This could be attributed to the fact that poliovirus can accommodate multiple copies of viral receptors and has enough surface for the RNA-translocation complex during viral uncoating. The quality of the cryo-EM data we collected on the poliovirus–nanodisc complexes (Figure 3) was substantially better than the data collected using the poliovirus-liposome complexes. The cryo-EM images in figure 3a–c show poliovirus bound to 50 nm cNDs decorated with CD155 ectodomain receptors, formation of a putative pore/channel and ejection of RNA across the membrane. Similarly, using 60 nm DNA-corralled nanodiscs functionalized with CD155 ectodomain, we were successful in initiating the receptor mediated uncoating of poliovirus and observed the early steps of virus attachment to the bilayer and pore formations in the nanodisc (Figure 4).

Figure 3.

Poliovirus caught in the act. (a) Cryo-EM image of poliovirus interacting with 50 nm covalently circularized nazanodiscs containing CD155 receptor. (b) Tilted views of the nanodisc were observed in ice. (c) Cryo-EM image of individual viral particle tethered to nanodisc containing CD155. (d) Schematic of pore formation and RNA release from poliovirus capsid across the bilayer nanodisc.

Figure 4.

Poliovirus interactions with DNA corralled nanodiscs containing CD155 receptor. (a) Negative-stain TEM images of 60 nm DNA-corralled nanodiscs containing CD155 ectodomain. (b) TEM images of individual Negative-stain TEM images of DCND (60 nm outer diameter) containing CD155 ectodomain interacting with poliovirus. Some of the nanodiscs were partially released from their DNA scaffolds after binding the virus. (c) TEM images of individual viral particles showing the bending of the bilayer and the creation of a pore in nanodisc by the poliovirus. This figure is adapted from reference [21^••].

Poliovirus particles have been shown to induce the formation of channels or pores in planar membranes. This was demonstrated by electrophysiology experiments [43] and genetic studies [44]. Danthi at al. have demonstrated that mutations in the VP4 caspid protein alter the ability of poliovirus to form channels and release RNA during infection [31]. This has led to the hypothesis that the insertion of VP4 into membranes enables the formation of pores or channels in host membranes that allow the translocation of the viral genome into the cytoplasm of the cell [45].

2. Role of host lipids in viral entry

Lipids play a critical role in viral entry and replication. Several studies have shown that the alterations of membrane lipid composition can block viral release and entry. For example, a study by Snyder and colleagues shows that the phosphatidylethanolamine lipid cooperates with the reovirus membrane penetration peptide to facilitate viral particle uncoating and enables the structural transition of intermediate subviral particles [46]. Another study demonstrated that the infectivity of bovine rotavirus is directly influenced by lysobisphosphatidic acid during the ESCRT-mediated entry [47]. Similarly, lysobisphosphatidic acid cooperates with the membrane penetration peptide, VP5, of the bluetongue virus to facilitate viral uncoating [48]. On the other hand, Ceramides have been shown to inhibit viral entry, likely due to self segregating into ceramide-rich microdomains[49].

In addition to certain lipids, membrane structures such as lipid rafts have been shown to facilitate the entry of many viruses [50,51]. The diameter of lipid rafts are on the order of 10–50 nm [52] and many viral receptors and/or co-receptors are often localized in the rafts [53].

Host proteins, other than the specific receptor(s) for any given virus, have been shown to play critical roles in membrane penetration by nonenveloped viruses. For example SV40 and human BK viruses, members of the polyomavirus family, have been shown to use EMC1 and SGTA chaperones in addition to the heat shock protein Hsp105 to enable their release from the ER membrane and transportation into the cytosol [54,55]. Another study using a genome-wide haploid genetic screen, has identified the lipid-modifying enzyme PLA2G16 as a host factor that functions early during picornaviruses infection, enabling genome delivery into the cytoplasm [56]. Large nanodiscs can be used to create a functional replication of membrane systems, including duplication of lipid compositions, fluidity and potentially membrane curvature, as well as lipid rafts and asymmetry. Moreover, they can provide a large membrane area to allow the co-incorporation of the specific receptor(s) along with other host protein factors important for the entry and replication of any given virus. Research has only begun to tap the broad potential applications of the large designed nanodiscs.

Conclusions and future directions

This review is an effort to shed light on the recent technical developments leading to the construction of large circularized protein and DNA-corralled nanodiscs and discuss some of the viral entry applications. These nanodiscs are sufficiently large to accommodate multiple copies of any given viral receptor and provide enough surface area to act as a surrogate membrane for the genome-translocation complexes during viral uncoating. They can be used to mimic specific membrane features including lipid compositions and curvature. Although extensive research over the past four decades has resulted in identification of many membrane-penetrating peptides and many models to explain how genome of non-enveloped viruses is released across the bilayer, questions regarding the mechanisms of membrane penetration and the influence of host factors remain unanswered. The receptor-decorated large nanodisc model system represents a very attractive tool, in that it can allow virus structures to be determined in the context of membranes and has already begun to reveal some exciting information about the nature of pore/channel that is triggered by non-enveloped viruses.

Despite rapid progress in advancing nanodisc technology, there are areas that warrant further exploration. Currently it is difficult to create asymmetric nanodiscs whose lipid composition is different on their two faces. Biological membranes are asymmetric with respect to lipid distributions across the bilayer and this asymmetry gives rise to a nonzero intrinsic potential difference between the two sides of the membrane in the absence of ions and has important functional consequences. Therefore, reconstructing this asymmetry in nanodisc system will provide invaluable tool that can be widely used to evaluate the role of this parameter in membrane protein function.