Fibril-mediated oligomerization of pilin-derived protein nanotubes
© Petrov et al.; licensee BioMed Central Ltd. 2013
Received: 3 May 2013
Accepted: 1 July 2013
Published: 5 July 2013
Skip to main content
© Petrov et al.; licensee BioMed Central Ltd. 2013
Received: 3 May 2013
Accepted: 1 July 2013
Published: 5 July 2013
Self-assembling protein nanotubes (PNTs) are an intriguing alternative to carbon nanotubes for applications in bionanotechnology, in part due to greater inherent biocompatibility. The type IV pilus of the gram negative bacteria Pseudomonas aeruginosa is a protein-based fibre composed of a single subunit, the type IV pilin. Engineered pilin monomers from P. aeruginosa strain K122-4 (ΔK122) have been shown to oligomerize into PNTs both in solution and at surfaces. In order to fully exploit PNTs in bionanotechonological settings, an in-depth understanding of their assembly, physical characteristics and robustness, both in solution and when constrained to surfaces, is required.
This study details the effectiveness of multiple initiators of ΔK122-derived PNT oligomerization and characterize the formation of PNTs in solution. The optimal initiator for the oligomerization of ΔK122 in solution was observed to be 2-methyl-2,4-pentanediol (MPD). Conversely, larger PEG molecules do not trigger oligomerization. Multi-angle light scattering analysis indicates that the pilin protein exists in a monomer-dimer equilibrium in solution, and that an intermediate species forms within three hours that then coalesces over time into high molecular weight PNTs. Transmission Electron Microscopic analysis was used to observe the formation of oligomerized ΔK122 fibrils prior to assembly into full-length PNTs.
The oligomerization of ΔK122 pilin derived PNTs is a fibril mediated process. The optimal trigger for PNT oligomerization in solution is MPD, and the observation that PEGs do not induce oligomerization may enable the oligomerization of pilin-derived PNTs on PEG-functionalized surfaces for implantable bionanodevices.
The development of peptide and protein-based nanotubes as biologically accepted nanosystems have several advantages when compared to their inorganic counterparts such as carbon nanotubes (CNTs), which are significantly more cytotoxic and present biocompatibility issues [1–11]. Peptide and protein based nanotubes can be assembled utilizing both template and non-template assembly mechanisms under milder conditions (ambient temperature, physiological pH), and provide a readily customizable system via modern protein engineering methods [12, 13]. In addition, studies have shown that wormlike, filamentous nanoparticles are better than spherical ones at avoiding immune responses allowing for longer circulation times due to the difficulty of macrophages have adjusting tertiary and/or quaternary structure to engulf such elongated particles [14–16]. Therefore, peptide and protein-based nanotubes will likely have applications as drug-delivery vehicles as their relatively large inner cavity and high surface areas would enable them to transport drug molecules, nucleic acids or antigens to targeted cell surface.
Several recent studies have examined the applicability of nanotubes from peptides [17–20], proteins [21–29], and viruses [30–38]. For instance, the mutation of the pIII and pVII coat proteins of the M13 phage enabled the modified phage to scaffold metal oxides [34, 35]. The resultant protein-metal hybrid bionanowires demonstrated significant initial and reversible storage capacity [35, 38], suggesting the utility of these nanocomposites for power generating applications. Another system ex-amined for protein nanotube (PNT) development is based on the bacterial flagella, where the flagellin protein FliC has been modified to contain a thioredoxin domain . The resultant FliC-thioredoxin chimera was shown to form PNTs on surfaces  as well as enable metal nanowire synthesis . These studies highlight the applicability of using bio-inspired PNTs for various applications in nanoelectronics and as biosensors.
In addition to protein-based nanostructures derived from viral coat proteins and flagella, PNTs have been shown to assemble from an engineered form of the type IV pilin [23, 24], the monomeric unit of the type IV pilus (T4P) of many gram-negative bacteria including Pseudomonas aeruginosa. Opportunistic infections by P. aeruginosa are a significant cause morbidity and mortality in individuals with compromised immune systems (e.g. burn victims  and cystic fibrosis patients ), with infections being initiated through interaction of T4P with cellular receptors [42–47]. In addition to cellular adherence, T4P are involved in a number of functions including surface adherence [48, 49], twitching motility [47, 50–54], DNA uptake [55–57], and biofilm formation [47, 58–60]. T4P are robust structures assembled and disassembled via a membrane-spanning complex whose architecture is evolutionarily related to a type II secretion system [47, 50, 60]. P. aeruginosa T4P have also been demonstrated to retract at rates of 0.5-1 μm s-1 (~1500 subunits s-1)  generating forces exceeding 100 pN . The T4P has an outer diameter of approximately 6–8 nm and can reach lengths up to tens of microns [44, 46, 47, 50, 62–65]. T4P are polymers of the type IV pilin, and cryo-EM [66–68] and fibre diffraction  analyses of T4P have demonstrated that T4P exhibit a three-start helical assembly of pilin monomers [44, 62]. The type IV pilin monomer is comprised of a four-stranded antiparallel β-sheet wrapped around a hydrophobic α-helix connected by a variable loop region [66, 70–76]. Surface adherence and cell-host adhesion is mediated by a C-terminal loop known as the D-region, which is disulfide-bound in most pilins [66, 68, 70–74, 76], although the FimA pilin of Dichelobacter nodosus displays a conserved structure without the disulfide bond . The observation that truncated pilins from P. aeruginosa strain K122-4 (ΔK122) could form PNTs morphologically similar to T4P in the presence of a hydrophobe (C11-SH), both in solution and when the hydrophobe was surface constrained [23, 24], presents an interesting avenue for the development of bionano applications that target the T4P, for example pilus-specific biosensors.
Several studies highlight the potential applications of PNTs including targeted drug delivery systems, tissue-engineering scaffolds and biosensing devices [35, 38, 77–83]. However, reports characterizing the assembly and properties of PNTs generated from full-length proteins in solution or at surfaces are more limited; it is in this light that we undertook the characterization of the oligomerization of pilin-derived PNTs in solution. Pilin-derived PNTs may have an advantage of being a more biologically accepted nanosystem when compared to their CNT counterparts. However in order to fully exploit PNTs for application development, a detailed understanding of their assembly and physical characteristics in solution and when surface-constrained is required. In the current study, we examine the assembly of ΔK122-derived PNTs in solution, monitoring PNT oligomerization through liquid chromatography, multi-angle light scattering and negatively stained transmission electron microscopic methods. We identify an optimal trigger molecule, 2-methyl-2,4-pentanediol (MPD), characterize pilin oligomerization in solution, and discuss the assembly of ΔK122-derived PNTs through intermediate pilin fibrils.
The identification that monomeric pilins from P. aeru-ginosa oligomerized into PNTs [12, 13, 23] suggests that these structures could be adapted for a variety of applications. Previous studies, employing a polyclonal antibody that recognizes the C-terminal region of the pilin from multiple strains of P. aeruginosa[84–86], have shown that the structure of and receptor binding properties of ΔK122 are unaffected upon oligomerization into PNTs [23, 48, 55]. Furthermore, the observation that pilin-derived PNTs can assemble both in solution  and at surfaces [24, 49] suggests that these structures could be adapted for applications such as biosensors and in bionanoelectronics while retaining several functional features associated with the native pilus itself.
The most effective trigger compound for the oligomerization of ΔK122 PNTs was found to be 2-methyl-2,4-pentanediol (MPD) (Figure 1G). MPD had been expected to behave similarly to 1-propanol in its ability to trigger PNT oligomerization. This is because the presence of the second hydroxyl group in MPD increases its hydrophilicity despite containing 2 more methylene groups than propanol. This increased hydrophilicity enables MPD to more effectively interact with proteins in solution and is therefore often used in protein crystallization . In fact, the PNT peak associated with the incubation of ΔK122 with MPD is larger than any of the other samples, both in peak height as well as its ratio to the monomeric ΔK122-4 peak (Figure 1). This indicates that the added solubility due to the second hydroxyl of MPD adds sufficient hydrophilicity for a more favorable interaction with the ΔK122 monomer in order to initiate PNT oligomerization. In contrast to MPD, exposure of the protein to polyethylene glycol (PEG) 3350 (Figure 1H) or PEG 8000 (Figure 1I) shows very little PNT oligomerization. PEGs are also frequently used in protein crystallization [88, 89], and have been shown to reduce non-specific adsorption of proteins to implantable devices . These data are important for the future development of PNT-containing nanodevices where the bio-nonfouling nature of PEG additives as surface coatings is exploited for increasing implantable device lifetimes in the body . The lack of PNT oligomerization with PEGs may allow the pre-functionalization of surfaces with PEGs and exposed hydrophobes for site-localized PNT oligomerization; we are currently examining this possibility.
SEC-MALS analysis of ΔK122 oligomerization
M w (kDa)
26.5 ± 1.2
183.6 ± 17.4
256.7 ± 37.7
520.7 ± 146.3
M w /M n
1.11 ± 0.07
1.27 ± 0.22
1.71 ± 0.46
1.16 ± 0.49
R h (nm)
4.32 ± 0.11
4.20 ± 0.13
3.9 ± 0.1
4.1 ± 0.1
Dt (x10 -9 cm 2 s -1 )
5.82 ± 0.15
6.10 ± 0.22
6.6 ± 0.2
6.15 ± 0.2
In order to observe the species identified by SEC-MALS analysis, aliquots of SEC-separated ΔK122 pilin (15 mg·mL-1) after 24 hours incubation with MPD corresponding to Peaks 1/2 and 3/4 were negatively stained with 4% uranyl acetate, visualized with TEM (Figure 3B, C), and compared to that PNTs oligomerized using the original C11-SH hydrophobe as the inducer [23, 87]. TEM analysis of a Peak 1/2 aliquot (Figure 3B) show pilin fibrils (highlighted with arrows) interspersed among a general aggregation of the ΔK122 monomers/dimers. The components of a Peak 3/4 aliquot shows the presence of full PNTs (Figure 3C), consistent with previous observations of pilin-derived PNTs in solution  and at surfaces . MALS analysis of this Peak 3/4 aliquot was challenging due to the high protein concentration (15 mg·mL-1) resulting in signal overload at the detector. However a MALS-determined Mw of 51,610 ± 4,900 kDa was observed for the species in this SEC-separated aliquot. This Mw corresponds to a structure that is microns in length containing ~4040 ΔK122 monomers, a structure which is observed in the TEM analysis of the Peak 3/4 aliquot (Figure 3C). Pilin-derived PNTs were also observed to further bundle into larger structures where PNT bundles ranging in width from ~25-65 nm (Figure 3C) to greater than 250 nm in cross-section . Given the predicted outer diameter of ~6 nm for native T4P [44, 46, 62, 63, 66, 67, 69] and/or K122-derived PNTs [12, 23, 24], the observed structures would correspond to bundles of ~4-11 PNTs.
The development of protein-based nanotubes for biologically based nanosystems is receiving increased interest due to their richness in structural diversity, adaptability through protein engineering approaches and inherent biocompatibility. The adaptation of the T4P as protein nanotubes through engineered type IV pilin monomers has shown distinct promise in that these structures can assemble both in solution and at surfaces in a template independent fashion [23, 24, 49, 87, 95]. In the current report, we have shown that the ΔK122 pilin is in a monomer-dimer equilibrium in solution, and that oligomerization of the pilin can be induced from short alkyl chains in solution, however optimal oligomerization is achieved when MPD is used as an initiator (Figure 1). Upon addition of the MPD initiator to the ΔK122 solution, the protein forms fibrils that then assemble into full length PNTs (Figures 2, 3), although the exact assembly mechanism is at this time unclear. Research in our group is on going to further characterize ΔK122-derived PNT assembly (both in solution and at surfaces), understand the structural and mechanistic requirements of PNT oligomerization and fibril stabilization, and to develop these structures for applications as bionanowires and biosensors.
The truncated form of the monomeric type IV pilin from P. aeruginosa strain K122-4 [pilA (Δ1-28); ΔK122] was expressed and purified as previously reported [23, 24, 71, 94]. The ΔK122 pilin was purified as an MBP-fusion construct and isolated from the MBP tag via cation exchange chromatography using a linear gradient of 0–1 M NaCl in 10 mM Tris (pH 7.4) following trypsin digestion of the MBP-ΔK122 fusion protein (500:1 protein:trypsin ratio; 10 min on ice). Freshly purified, monomeric ΔK122 pilin was concentrated to 15 mg·mL-1 or diluted to 1 mg·mL-1 as appropriate and used for all oligomerization experiments. All experiments conducted in this study were reviewed and approved by the York University's Biological Safety Committee, the institutional body responsible for oversight of such research.
The initially reported solution for initiating oligomerization was composed of 1.1 M 1-undecanethiol (C11-SH) in methanol containing 1 mM EDTA, 1 mM dithiothreitol (DTT) at pH 6.4 . Studies of PNT oligomerization from surfaces suggest that the initiation of PNT assembly could be achieved with smaller chain alkylthiols, and/or mixtures thereof [24, 87]. Characterization of an optimal trigger molecule in solution was conducted by incubating 15 mg·mL-1 ΔK122 with buffer (10 mM Tris, 300 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 7.4) alone, or with buffer plus methanol (3.2 M), 1-propanol (1.7 M), 1-undecanethiol (C11-SH, 0.6 M), 1-tetradecanethiol (C14-SH, 0.5 M), 2,4-methylpentanediol (MPD, 1.0 M), polyethylene glycol 3350 (PEG 3350, 6.5% (w/v)) or polyethylene glycol 8000 (PEG 8000, 6.5% (w/v)) (Figure 1). PNT oligomerization was initiated through the addition of the trigger solution to ΔK122 in a 10:1 (v/v) protein to hydrophobe ratio, and the ΔK122-trigger solution was incubated at room temperature with nutation for 96 hours. PNT oligomerization was monitored using size exclusion chromatography (SEC) on a G50 Sephadex column (separation range 1.5 kDa - 30 kDa, standardized with blue Dextran 2000, V0/VTotal = 17.23/49.9l mL) on an Akta Purifier (GE Healthcare) at a flow rate of 1 mL·min-1.
The oligomerization of 1 mg·mL-1 ΔK122 triggered with MPD in buffer was analyzed using SEC and multi-angle light scattering (MALS) using an Akta Purifier 10 (GE Healthcare) connected in-line to a Dawn Heleos II and Optilab T-rEX light scattering system (Wyatt Technology) (Figure 2, 3). Analysis of 100 μL protein samples was performed at a flow rate of 0.5 mL min-1 in SEC buffer (10 mM Tris, pH 7.4) on a silica-based column (Wyatt Technology, 10 nm pore size, separation range 100 Da - 100 kDa, VTotal = 10.71 mL). After chromatographic separation, the column eluate traveled to the MALS flow cell where light scattering (658 nm laser light source) of the separated species was monitored by 15 angularly separated static light scattering detectors and a quasi-elastic light scattering (QELS) detector at a collection angle of 100.2° (Figure 3A; Table 1). Hydrodynamic radii (Rh) and diffusion coefficients (Dt) were calculated from an autocorrelation function using the accompanying Astra 6 software package (Table 1).
Transmission electron microscopy (TEM) of SEC-MALS separated PNTs was conducted in the Department of Biology’s Core Imaging Facility at York University, and samples for TEM analysis were prepared as follows. Ten microlitre aliquots of PNT solutions were dispensed onto plastic-coated nickel mesh grids and allowed to dry in air for 10 min; any remaining liquid was carefully removed by blotting with filter paper. Samples were stained with a 4% aqueous uranyl acetate, which was added to the grid and allowed to incubate for 10 minutes at room temperature, following which excess stain was removed by blotting with filter paper. Samples were imaged using a Philips 210 Transmission Electron Microscope operating at an accelerating voltage of 60 kV, and images visualized using the ImageJ software package .
The truncated pilin from P. aeruginosa strain K122-4
Size exclusion chromatography
multi-angle light scattering
Transmission electron microscopy
This research was supported by grants to GFA from the Natural Sciences & Engineering Council of Canada (NSERC), the Canadian Foundation for Innovation (CFI) and York University. AP and SL acknowledge financial support from the Ontario Graduate Scholarship Program, NSERC and York University. The authors acknowledge technical assistance by Ms. Karen Rethoret (Dept. of Biology, York University) for TEM studies.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.