- Email: [email protected]
Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion
Journal of Bioscience and Bioengineering VOL. xxx No. xxx, xxx, xxxx www.elsevier.com/locate/jbiosc
Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion Sota Aoki,1 Shogo Yoshimoto,1 Masahito Ishikawa,1 Dirk Linke,2 Andrei Lupas,3 and Katsutoshi Hori1, * Department of Biomolecular Engineering, Graduate School of Engineering, Nagoya University, Nagoya, Aichi 464-8603, Japan,1 Department of Biosciences, University of Oslo, 0316 Oslo, Norway,2 and Department of Protein Evolution, Max Planck Institute for Developmental Biology, 72076 Tübingen, Germany3 Received 21 August 2019; accepted 30 September 2019 Available online xxx
AtaA, a trimeric autotransporter adhesin from Acinetobacter sp. Tol 5, exhibits nonspecific, high adhesiveness to abiotic surfaces. For identification of the functional domains of AtaA, precise design of domain-deletion mutants is necessary so as not to cause undesirable structural distortion. Here, we designed and constructed three types of AtaA mutants from which the same domain, FGG1, was deleted. The first mutant was designed to preserve the periodicity of hydrophobic residues in the coiled-coil segments sandwiching the deleted region. After the deletion, the protein was properly displayed on the cell surface and had the same adhesive function as the wild type. Transmission electron microscopy (TEM) imaging and circular dichroism (CD) spectroscopy showed that its isolated passenger domain had the same fiber structure as in the AtaA wild type. In contrast, a mutant designed to disturb the coiled-coil periodicity at the deletion site failed to reach the cell surface. Although secretion occurred for the mutant designed with a flexible connector between the coiled coils, the cells exhibited a decrease in adhesiveness. Furthermore, TEM imaging of the mutant fibers showed bending at the fiber tip and changes in their CD spectrum indicated a decrease in secondary structure content. Thus, we succeeded to natively display the huge homotrimeric fiber structure of AtaA on the cell surface after precise deletion of a domain, maintaining the proper folding state and adhesive function by preserving its coiled-coil periodicity. This strategy enables us to construct various domain-deletion mutants of AtaA without structural distortion for complete functional mapping. Ó 2019, The Society for Biotechnology, Japan. All rights reserved. [Key words: Trimeric autotransporter adhesin; Coiled-coil; Protein structure; In-frame deletion; Multidomain protein; Acinetobacter]
In-frame deletion of domains is commonly used to identify the functional site or epitope of large and multi-domain proteins, because it enables efficient mapping of regions larger than a point mutation (1e6). However, the deletion of an internal domain often causes undesirable structural distortion and steric conflicts in the protein, and can lead to misfolding and functional inactivation of the constructed mutant. Therefore, great care and precise design are necessary when constructing domain-deletion mutants of large and complex proteins. Previously, we discovered AtaA, a member of the trimeric autotransporter adhesin (TAA) family, in the Acinetobacter sp. Tol 5, which exhibits noteworthy autoagglutination and high adhesiveness to various abiotic surfaces from hydrophobic plastics to hydrophilic glass and metals (7e9). TAAs are outer membrane (OM) proteins of gram-negative bacteria, and polypeptide chains of a TAA form a homotrimeric structure with an N-terminal passenger domain (PSD) corresponding to its biological functions and a Cterminal transmembrane domain transporting and anchoring the PSD onto the OM (10,11). The AtaA polypeptide is composed of 3630 amino acids and is thus one of the largest TAAs. AtaA’s long PSD of 225 nm in length consists of mosaically arranged multiple domain repeats including two Ylheads, 15 GINs, 12 Trp-rings, ten GANGs,
* Corresponding author. Tel: þ81 52 789 3339; fax: þ81 52 789 3218. E-mail address: [email protected] (K. Hori).
ten DALLs, five FGGs, and several neck domains, which often appear in TAAs (9,12). While AtaA shares structural features with other TAAs, which usually bind bacterial cells to specific biotic molecules, AtaA is unique in terms of its nonspecific adhesiveness to various abiotic surfaces. This adhesive feature of AtaA can be conferred to other non-adhesive gram-negative bacteria by transformation with the ataA gene (9). We invented a new method for immobilizing bacterial cells utilizing AtaA and demonstrated its advantages; large numbers of bacterial cells expressing AtaA can be quickly immobilized onto various material supports, the immobilized cells can be reversibly detached by rinsing with deionized water (dH2O) (13) or by adding casein hydrolyzates (14), and the immobilized cells can be efficiently used for bioproduction (15e17). However, little is known about the adhesion mechanism and functional regions of AtaA. Domains of TAAs are connected by trimeric coiled coils (10,12,18,19). Coiled coils are helical bundles with periodically occurring hydrophobic residues, which form a hydrophobic core along the central axis of the bundle. Their most prevalent periodicity is of 7 residues over 2 helical turns (the so-called heptad repeat), but TAAs also have coiled coils with other periodicities such as 11, 15, and 19 residues (10,20). The periodicities of heptad and hendecad (11 residues) repeats are denoted as abcdefg and abcdefghijk, in which positions a and d in heptad repeats and a, d, e, and h in hendecad repeats come with hydrophobic residues (21,22). It is considered that domains between two coiled coils can be deleted
1389-1723/$ e see front matter Ó 2019, The Society for Biotechnology, Japan. All rights reserved. https://doi.org/10.1016/j.jbiosc.2019.09.022
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
2
AOKI ET AL.
J. BIOSCI. BIOENG.,
FIG. 1. Design of domain-deletion mutants of AtaA. (A) Schematic representations of AtaA. The dotted box indicates the deleted region in the mutants. The colored regions represent annotated domains as indicated in the box. (B) Structure of an FGG domain, FGG5 in AtaA (PDB ID: 3WPA). The monomeric chain is shown in magenta and two symmetry-related molecules are shown in gray. FGG domains have common structure consisting of a b-meander sandwiched between two coiled coils although their detail structures are varied in different FGGs. (C) Schematic figures of heptad and hendecad coiled coils. Left, the periodicity of heptad repeats is denoted as abcdefg. Right, the periodicity of hendecad repeats is denoted as abcdefghijk. Residues at the positions a and d in heptad repeats and a, d, e, and h in hendecad repeats, which are shown in red, contribute to interhelical hydrophobic interactions. (D) Amino acid sequences around the deleted region. The deleted residues in each mutant are shown as dashes. The periodicities of hendecad and heptad repeats in the coiled coils are shown above the amino acid sequence. Positions forming a hydrophobic core of coiled coils are shown in red.
without adverse effect on the structure by designing a resulting construct to preserve the periodicity of hydrophobic residues in the coiled coils after the deletion. Numerous domain-deletion mutants of TAAs have been designed and constructed with this strategy, and their production, cell surface display, and functions have been evaluated (12,23,24). However, their native folding state has never been directly confirmed because there was no effective method for analyzing the structure or morphology of TAAs after secretion onto the OM. Recently, we developed a method to isolate the PSD of AtaA by genetically introducing a recognition site for human rhinovirus 3C (HRV 3C) protease (25). Specific cleavage by the protease isolates AtaA’s PSD nanofibers from the cell surface, which enables structural and morphological analyses of isolated AtaA’s PSD in the native molecular state. In the current study, we succeeded to obtain the proper fiber structure of AtaA with correct folding and biological function after the deletion of a domain, by preserving the periodicity of flanking coiled-coil segments. Upon deviation from the periodicity, maintenance of the biological function failed due to disturbance of the proper folding of the protein.
DNA manipulation Plasmids and primers used in this study are listed in Supplementary Tables S2 and S3, respectively. The plasmid pDONR::ataA was digested with EcoRV at two sites in the original ataA and the linearized plasmid was re-circularized by self-ligation, generating pDONR::Ntemp that included a region encoding AtaA1e496. The procedures for the construction of pIFD-DFGG1A, B, and C and pIFD3C-DFGG1A and C are schematically shown in Figs. S1 and S2, respectively. Fragment-1 was amplified by PCR from pDONR::Ntemp using primers Ntemp-FGG_F/Ntemp-HiFi_R. Fragment-2A was amplified by inverse PCR from pAtaA using primers pAtaA_invPCR_F/pAtaA_invPCR2_R. Fragment-3 was excised from pDONR::ataA by digestion with EcoRV and XbaI. The three fragments were assembled using NEBuilder HiFi DNA Assembly master mix (New England BioLabs, Ipswich, MA, USA), generating pIFD-DFGG1A. To construct pIFD-DFGG1B and pIFD-DFGG1C, fragment-2B and fragment-2C were amplified by inverse PCR from pAtaA using the primer sets pAtaA_invPCR_F/pAtaA_invPCR3_R and pAtaA_invPCR_F/pAtaA_invPCR4_R, respectively, and then each fragment was assembled with fragment-1 and fragment-3 as described above. To construct pIFD3C-DFGG1A and pIFD3C-DFGG1C, fragment-3C was excised from pTA2::3CataA by digestion with EcoRV and XbaI, and assembled with fragment-1 and fragment-2A or fragment-2C as described above. Transformation of Tol 5 4140 (27) with these expression vectors was carried out by conjugal transfer from Escherichia coli S17-1 strain, as previously described (9). To produce a recombinant protein of the N-terminal head domain (Nhead) of AtaA as an antigen, the DNA fragment encoding AtaA59e325 was amplified from pDONR::ataA by PCR with the primer set AtaA59F/AtaA325R. The PCR product was digested with XbaI and BsaI, and inserted into the same site of pIBA-GCN4tri-His plasmid (28) by ligation. The constructed plasmid, pIBA-AtaA59e325-GCN4tri-His, was used for the transformation of E. coli BL21 Star(DE3). The protein produced by the E. coli transformant was used to generate the polyclonal anti-Nhead antiserum.
MATERIALS AND METHODS
Structural analysis The PSDs of AtaA wild type (WT), IFD-DFGG1A, and IFDDFGG1C were purified from the cell surface of Tol 5 4140 harboring p3CAtaA, pIFD3C-DFGG1A, and pIFD3C-DFGG1C as described previously (25). Bacterial cells
Bacterial strains and cell analysis The bacterial strains used in this study are listed in Supplementary Table S1. They were grown as described previously (9). Arabinose was added to the culture medium to induce the expression of the ataA gene and its domain-deletion mutant genes on plasmid vectors in the DataA mutant strain of Acinetobacter sp. Tol 5, 4140. The concentration of arabinose was varied from 0.3% to 0.5% to control the production levels of the proteins, which were examined by SDS-PAGE and Coomassie Brilliant Blue (CBB) staining as previously described (26). The cell surface display of the AtaA mutants was confirmed by flow cytometry and confocal laser scanning microscopy (CLSM) as previously described (26) with slight modification; anti-Nhead antiserum and anti-rabbit IgG antibody conjugated to AlexaFluor 488 (Cell Signaling Technology, Danvers, MA, USA) were used for the primary and secondary antibodies, respectively. Adherence assay (25) and autoagglutination assay (7) of bacterial cells were performed as described previously.
were treated with HRV 3C protease at 4 C for 2 days, and then centrifuged at 10,000g for 10 min. The supernatant was fractionated by precipitation in a 30% solution of the saturated concentration of ammonium sulfate. The precipitate was dissolved in dH2O and subjected to ultrafiltration with a 100 kDa molecular weight cut-off filter by centrifugation at 10,000g, and finally concentrated in a final volume of 50 mL of dH2O solution. The protein concentration was quantified by bicinchoninic acid (BCA) assay, and the purity was confirmed by SDS-PAGE followed by CBB staining. The purified PSDs were subjected to transmission electron microscopy (TEM) as described previously (25). To sum up, the purified PSD dissolved in 10 mM KCl solution was placed onto a copper grid with carbon substrate (UHR-C10; Okenshoji, Tokyo, Japan). After 10-min incubation, the grid was washed by dipping 20 times into 10 mM KCl solution and three times into dH2O. Negative staining of the proteins
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
VOL. xxx, xxxx
NATIVE DISPLAY OF TRIMERIC AUTOTRANSPORTER ADHESINS
on the grid was performed with the organo-tungstate compound Nano-W (Nanoprobes, NY, USA) according to a standard method. The specimen was observed using a TEM system (JEM-1400EX; JEOL, Tokyo, Japan) operated at 100 kV. Circular dichroism (CD) spectra were also recorded in dH2O as described previously (25) with slight modification; protein samples of 0.08 mg/ml dissolved in dH2O were used for the measurements.
RESULTS Molecular design of domain-deletion mutants of AtaA We designed three types of domain-deletion mutants of AtaA lacking FGG1, which follows upon the N-terminal Ylhead (Nhead) in AtaA (Fig. 1A). The FGG is a widespread domain of TAAs and is generally composed of a b-meander including a couple of b-strands, which is sandwiched between the N-terminal and Cterminal coiled coils (Fig. 1B). The periodicities of the N-terminal and C-terminal coiled coils are hendecad and heptad repeats, respectively (Fig. 1C). Accordingly, we designed IFD-DFGG1A lacking residues from leucine 326 (L326) to alanine 362 (A362), in which isoleucine 363 was expected to act as a substitute of L326 so as to maintain the periodicity of hydrophobic residues in the coiled-coil (Fig. 1D). For comparison, in addition to IFDDFGG1A, we also designed IFD-DFGG1B and IFD-DFGG1C, keeping the C-terminal residue of the deletion identical, namely, A362. IFD-DFGG1B was a mutant lacking residues from alanine 328 to A362, and the periodicity of hydrophobic residues in the coiled coil was disturbed by just the two extra residues remaining at the ligation joint. IFD-DFGG1C was a mutant lacking residues from glycine 335 to A362, in which a flexible sequence (GTGS) connected the coiled coils. The expression vectors for these three mutants (pIFD-DFGG1A, B, and C) were constructed and introduced into Tol 5 4140, a DataA mutant strain of Tol 5. Production and cell surface display of the domain-deletion mutant proteins To examine the production of the AtaA domain-deletion mutant proteins, the cell lysates of Tol 5 4140 expressing the mutant genes were separated by SDS-PAGE. The calculated molecular weights of mature proteins of AtaA WT, IFDDFGG1A, IFD-DFGG1B, and IFD-DFGG1C were 353.2, 349.7, 349.9, and 350.5 kDa, respectively. On the gel stained with CBB, each AtaA mutant polypeptide was detected with almost the same electrophoretic mobility as the WT protein, indicating that all of the mutant polypeptides were properly produced (Fig. 2A).
3
We next examined the cell surface display of the AtaA mutants. Tol 5 4140 cells expressing the mutant genes were stained with anti-Nhead antiserum and Alexa 488-conjugated anti-rabbit antibody, and then subjected to CLSM and flow cytometry. The cells expressing IFD-DFGG1A and IFD-DFGG1C exhibited similar fluorescence intensity as the cells expressing AtaA WT, and the fluorescence surrounded the cells (Fig. 2B and C). On the other hand, the cells expressing IFD-DFGG1B exhibited much weaker fluorescence intensity than them. These results demonstrated that IFD-DFGG1A and IFD-DFGG1C were displayed on the cell surface while the secretion of IFD-DFGG1B onto the OM failed. Functionality of the domain-deletion mutant proteins To evaluate the functionality of the domain-deletion mutants of AtaA, we performed adherence and autoagglutination assays. As shown in Fig. 3, cells expressing IFD-DFGG1A exhibited high adhesiveness to both polystyrene (PS) and glass surfaces and the autoagglutination ratio was similar to that of cells expressing AtaA WT. On the other hand, the cells expressing IFD-DFGG1B, whose cell surface display failed, showed greatly decreased adhesiveness to both material surfaces compared with the cells expressing AtaA WT and lost the autoagglutinating nature to the same level as DataA. Although IFD-DFGG1C seemed to be regularly secreted onto the cell surface, the cells expressing the mutant showed partial decreases in adhesiveness to PS and glass surfaces. The autoagglutinating nature of this mutant was slightly reduced. These results suggest that only the design of IFDDFGG1A successfully achieved deletion of the FGG1 domain without an adverse effect on the biological functions of AtaA. Morphological characteristics of the domain-deletion mutant fibers Next, we attempted to examine the morphological characteristics of the fiber structure of IFD-DFGG1A and IFDDFGG1C, which were displayed on the cell surface. For this purpose, a nucleotide sequence encoding human rhinovirus 3C (HRV 3C) protease recognition site and a tri-glycine linker was inserted into the FGG domain that was situated at the base of the PSD (FGG5), to produce pIFD3C-DFGG1A and pIFD3C-DFGG1C. These domaindeletion mutant AtaA fibers were displayed on the cell surface of Tol 5 4140 and chemically reaped from the cells with HRV 3C protease. Thus, the PSDs of these mutants were isolated by a previously established method (25). As shown in Fig. 4A, the PSDs of IFD-DFGG1A and IFD-DFGG1C were purified as with the PSD of AtaA WT and no other contaminated proteins were not
FIG. 2. Production and cell surface display of the wild type (WT) of AtaA or the domain-deletion mutants of AtaA on Tol 5-derived cells. (A) SDS-PAGE followed by CBB staining of whole-cell lysates of the Tol 5-derived cells. (B) Immunofluorescence microscopy of the Tol 5-derived cells using anti-Nhead antiserum and Alexa Fluor 488-conjugated anti-rabbit antibody. Scale bar: 2 mm. (C) Flow cytometry of the Tol 5-derived cells using anti-Nhead antiserum and Alexa Fluor 488-conjugated anti-rabbit antibody. For a control, DataA cells were also subjected to the same analyses ().
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
4
AOKI ET AL.
J. BIOSCI. BIOENG., negative band with a peak at 208 nm corresponding to a-helix than that of AtaA WT. This smaller negative band of IFD-DFGG1C PSD was independently reproduced with samples from different purification lots. This result suggests that the molecular structure of IFD-DFGG1C had a distortion that was not detected by TEM in addition to the bending at the tip. In contrast, the CD spectrum of the PSD of IFD-DFGG1A exhibited a negative band with a peak at 208 nm and almost overlapped with that of AtaA WT. This and TEM observation implied that IFD-DFGG1A had the proper folded structure just as AtaA WT.
DISCUSSION
FIG. 3. Functional analyses of Tol 5 derivatives expressing domain-deletion mutants of AtaA. (A) Adherence assay of Tol 5 derivatives expressing AtaA (WT) and the AtaA mutants. Bacterial cells adhering to polystyrene (PS) and glass plates were quantified by crystal violet staining. (B) Autoagglutination assay of Tol 5 derivatives expressing AtaA (WT) and the AtaA mutants. The autoagglutination ratio was quantified by a decrease in the optical density at 660 nm (OD660) of the cell suspension after settling for 3 h. As a control, DataA cells were also subjected to the same analyses (). Data are expressed as mean SEM (n ¼ 3).
detected, indicating that the purity was adequately high. Subsequently, the fiber morphology of these PSDs was observed by TEM. TEM images of the PSD fibers of IFD-DFGG1A with a globular tip of Nhead were similar to those of AtaA WT (Fig. 4B and C). The fiber lengths of the PSDs of AtaA WT and IFDDFGG1A, which were determined from the measurement of fifteen fibers’ images, were 224 12 nm (mean SD) and 225 13 nm, respectively. There was no significant difference in a fiber length between IFD-DFGG1A and WT (p ¼ 0.82, t-test). This was reasonable because the deleted region was too short to affect the fiber length. The PSD of IFD-DFGG1C also had fibrous morphology with the globular tip. However, the fiber length of the PSD of IFDDFGG1C decreased to 211 12 nm (p < 0.01, t-test) and some of the tips were markedly bent (Fig. 4D), implying that IFD-DFGG1C did not retain the native fiber morphology. Secondary structure of the domain-deletion mutants To confirm the folding states of IFD-DFGG1A and IFD-DFGG1C, CD spectra, which reflect the secondary structure of a protein and are widely used for this purpose, were measured. As shown in Fig. 5, the CD spectrum of the PSD of IFD-DFGG1C exhibited a smaller
For functional mapping, domain-deletion mutants are frequently constructed and examined for the preservation of their functions. In general, a domain would be recognized as a functional site if the function fails upon its deletion. However, the failure of a function can be caused not only by the deletion of its functional domain but also by the destruction or distortion of protein structure by the deletion. In this study, in two FGG1-deletion mutants, IFD-DFGG1B and IFD-DFGG1C, the functionality of AtaA was lost or reduced; the former mutant was deficient in the secretion of its PSD and the latter mutant showed a partial decrease in its adhesive function. However, it can be concluded that FGG1 is not a domain responsible for the adhesive function of AtaA because mutant cells expressing IFD-DFGG1A exhibited adhesiveness to abiotic surfaces and autoagglutinating property at the same level as cells expressing AtaA WT. Our results reaffirm the importance of the precise design of domain-deletion mutants in functional mapping to avoid disturbing the folding and structure formation of the mutant proteins. In the case of TAAs, the register of coiled-coil segments, which frequently exist in their PSDs, can be utilized for precise design (12,23,24). Our results regarding AtaA proved the appropriateness of this strategy in the functional mapping of a TAA and simultaneously showed that deviation from the periodicity could cause failure in the secretion onto the OM or partial disruption of the secondary structure of the protein together with abnormal fiber morphology. This strategy can be employed even when the crystal structure of a TAA of interest or its domain is unknown if the secondary structure and domain annotation can be deduced. TAAs are known to be secreted via the type Vc secretion pathway (29,30). In this pathway, the synthesized polypeptide chain is translocated into the periplasm by the Sec machinery, and then three polypeptide chains trimerize, forming a C-terminal b-barrel pore at the OM. Subsequently, the N-terminal PSD is transported onto the OM through the pore and folded into a fiber structure. Although the driving force for the transportation of the PSD through the OM remains an open research problem, the folding of the polypeptide chain on the OM is speculated to be the major energy source (30). Fig. 2 shows that the PSD of IFD-DFGG1B was produced adequately, but its secretion onto the OM failed. Since IFD-DFGG1B has a signal sequence of the Sec machinery, which transports polypeptides in an unfolded state, it must have passed through the inner membrane. However, the disruption of the coiled-coil periodicity should disturb folding of the PSD of this mutant, resulting in the lack of energy for its secretion. Therefore, at least a part of Nterminal PSD of the non-secreted IFD-DFGG1B polypeptide molecules should have accumulated in the periplasmic space. However, we do not know if their C-terminal transmembrane domain was inserted into the OM, which is the next challenge. Although IFD-DFGG1C seemed to be regularly secreted onto the cell surface, its adhesive function was partially decreased. In such a case, it is difficult to understand the reason for the functional deterioration and to clarify the involvement of the deleted domain in the protein function. In this study, we examined the fiber
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
VOL. xxx, xxxx
NATIVE DISPLAY OF TRIMERIC AUTOTRANSPORTER ADHESINS
5
FIG. 4. Purification and TEM observation of PSDs of the domain-deletion mutants of AtaA. (A) SDS-PAGE followed by CBB staining of the PSDs of AtaA WT, IFD-DFGG1A, and IFD-DFGG1C. An arrow indicates the bands corresponding to the PSDs. (BeD) TEM images of negatively stained PSDs of AtaA WT (B), IFD-DFGG1A (C), and IFD-DFGG1C (D). The image on the bottom of each panel shows a representative fiber of each sample. Dotted circles in panel D indicate the bending tips. Scale bar: 100 nm.
morphology and the secondary structure of the mutant protein. This is the first report confirming the folding state of domaindeletion mutants of TAAs isolated from the cell surface. This revealed that some of the PSD fibers of IFD-DFGG1C had a morphology with a bent tip (Fig. 4). Although TAAs typically have straight fiber morphologies, some TAAs, such as Aggregatibacter actinomycetemcomitans EmaA, Moraxella catarrhalis UspA1, and E. coli EibD, have been reported to bend naturally at specific sites in the PSD (31e33). In EmaA, a flexible region composed of glycine and serine was considered to be the bending site (34). As for IFDDFGG1C, because the two coiled coils sandwiching the excised region were reconnected by a flexible sequence, the bending of the tip of the PSD of IFD-DFGG1C was thought to be associated with the flexibility. Therefore, the tip of the fiber would randomly swing in liquid at the reconnected site, and some fibers could be bent and others could be in straight at one time. Indeed, the tips of some fibers, albeit not all of them, were observed to be bent. The CD spectrum revealed the decrease in the secondary structure of the PSD of IFD-DFGG1C. We do not know if the morphological change just at the reconnected site caused such a relatively large change in the CD spectrum. When we isolated PSDs, the recovery amount was always less from cells expressing pIFD3C-DFGG1C than cells expressing pIFD3C-DFGG1A or p3CAtaA (Table S4), suggesting the decrease in the accessibility of the protease to its recognition site in the former mutant cells. The protease recognition site was
FIG. 5. CD spectra of PSDs of the domain-deletion mutants of AtaA. Gray line, AtaA WT; orange line, IFD-DFGG1A; blue line, IFD-DFGG1C. The ellipticities of protein samples were measured at a concentration of 0.08 mg/mL. Each spectrum is the average of 10 scans.
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
6
AOKI ET AL.
J. BIOSCI. BIOENG.,
introduced into FGG5, which was situated at the base of the AtaA PSD in these mutants. Therefore, we cannot rule out the possibility that the folding of this domain was affected far away by the structural change at the deletion site of IFD3C-DFGG1C. In IFDDFGG1B, the local misfolding at the reconnected site might have disabled the entire assembly and folding of the protein, resulting in failure in its secretion. Also, in IFD-DFGG1C, disturbing proper folding around reconnected site by the flexible linker might affect the entire folding and structure formation of the protein. We think that mutant molecules with different degrees of misfolding were mixed rather than all of the molecules were uniformly misfolded. In the field of protein engineering, a flexible sequence is frequently used for the connection of domains as a linker (35,36). However, the current results remind us that flexible linkers are not always effective for the connection of protein domains. In conclusion, we succeeded in restoring the huge homotrimeric fiber structure of AtaA with proper folding and biological functions after deletion of a domain, utilizing its coiled-coil periodicity. This revealed that the deleted domain FGG1 is unnecessary for the adhesive function of AtaA. Preservation of the coiled-coil periodicity will enable us to design and construct various domain-deletion mutants of AtaA without structural distortion and lead us to complete functional mapping and molecular engineering of AtaA. Supplementary data to this article can be found online at https://doi.org/10.1016/j.jbiosc.2019.09.022. ACKNOWLEDGMENTS The authors wish to thank the Division for Medical Research Engineering, Graduate School of Medicine, Nagoya University, for the usage of TEM equipment. This work was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant Numbers JP17H01345 and JP18K14062). The authors declare no conflicts of interest. References 1. Hartford, O., McDevitt, D., and Foster, T. J.: Matrix-binding proteins of Staphylococcus aureus: functional analysis of mutant and hybrid molecules, Microbiology, 145, 2497e2505 (1999). 2. Fine, D. H., Kaplan, J. B., Furgang, D., Karched, M., Velliyagounder, K., and Yue, G.: Mapping the epithelial-cell-binding domain of the Aggregatibacter actinomycetemcomitans autotransporter adhesin Aae, Microbiology, 156, 3412e3420 (2010). 3. Corrigan, R. M., Rigby, D., Handley, P., and Foster, T. J.: The role of Staphylococcus aureus surface protein SasG in adherence and biofilm formation, Microbiology, 153, 2435e2446 (2007). 4. Kao, L., Lam, V., Waldman, M., Glassock, R. J., and Zhu, Q.: Identification of the immunodominant epitope region in phospholipase A2 receptor-mediating autoantibody binding in idiopathic membranous nephropathy, J. Am. Soc. Nephrol., 26, 291e301 (2015). 5. Sundstrom, M., White, R. L., de Parseval, A., Sastry, K. J., Morris, G., Grant, C. K., and Elder, J. H.: Mapping of the CXCR4 binding site within variable region 3 of the feline immunodeficiency virus surface glycoprotein, J. Virol., 82, 9134e9142 (2008). 6. Mangan, R. J., Stamper, L., Ohashi, T., Eudailey, J. A., Go, E. P., Jaeger, F. H., Itell, H. L., Watts, B. E., Fouda, G. G., Erickson, H. P., and other 3 authors: Determinants of Tenascin-C and HIV-1 envelope binding and neutralization, Mucosal Immunol., 12, 1004e1012 (2019). 7. Ishikawa, M., Shigemori, K., Suzuki, A., and Hori, K.: Evaluation of adhesiveness of Acinetobacter sp. Tol 5 to abiotic surfaces, J. Biosci. Bioeng., 113, 719e725 (2012). 8. Hori, K., Yamashita, S., Ishii, S., Kitagawa, M., Tanji, Y., and Unno, H.: Isolation, characterization and application to off-gas treatment of toluenedegrading bacteria, J. Chem. Eng. Jpn., 34, 1120e1126 (2001). 9. Ishikawa, M., Nakatani, H., and Hori, K.: AtaA, a new member of the trimeric autotransporter adhesins from Acinetobacter sp. Tol 5 mediating high adhesiveness to various abiotic surfaces, PLoS One, 7, e48830 (2012). 10. Bassler, J., Alvarez, B. H., Hartmann, M. D., and Lupas, A. N.: A domain dictionary of trimeric autotransporter adhesins, Int. J. Med. Microbiol., 305, 265e275 (2015).
11. Linke, D., Riess, T., Autenrieth, I. B., Lupas, A., and Kempf, V. A. J.: Trimeric autotransporter adhesins: variable structure, common function, Trends Microbiol., 14, 264e270 (2006). 12. Koiwai, K., Hartmann, M. D., Linke, D., Lupas, A. N., and Hori, K.: Structural basis for toughness and flexibility in the C-terminal passenger domain of an Acinetobacter trimeric autotransporter adhesin, J. Biol. Chem., 291, 3705e3724 (2016). 13. Yoshimoto, S., Ohara, Y., Nakatani, H., and Hori, K.: Reversible bacterial immobilization based on the salt-dependent adhesion of the bacterionanofiber protein AtaA, Microb. Cell Fact., 16, 123 (2017). 14. Ohara, Y., Yoshimoto, S., and Hori, K.: Control of AtaA-mediated bacterial immobilization by casein hydrolysates, J. Biosci. Bioeng., 128, 544e550 (2019). 15. Nakatani, H., Ding, N., Ohara, Y., and Hori, K.: Immobilization of Enterobacter aerogenes by a trimeric autotransporter adhesin, AtaA, and its application to biohydrogen production, Catalysts, 8, 159 (2018). 16. Hori, K., Ohara, Y., Ishikawa, M., and Nakatani, H.: Effectiveness of direct immobilization of bacterial cells onto material surfaces using the bacterionanofiber protein AtaA, Appl. Microbiol. Biotechnol., 99, 5025e5032 (2015). 17. Ishikawa, M., Shigemori, K., and Hori, K.: Application of the adhesive bacterionanofiber AtaA to a novel microbial immobilization method for the production of indigo as a model chemical, Biotechnol. Bioeng., 111, 16e24 (2014). 18. Szczesny, P. and Lupas, A.: Domain annotation of trimeric autotransporter adhesins d daTAA, Bioinformatics, 24, 1251e1256 (2008). 19. Hartmann, M. D., Grin, I., Dunin-Horkawicz, S., Deiss, S., Linke, D., Lupas, A. N., and Alvarez, B. H.: Complete fiber structures of complex trimeric autotransporter adhesins conserved in enterobacteria, Proc. Natl. Acad. Sci. USA, 109, 20907e20912 (2012). 20. Alvarez, B. H., Gruber, M., Ursinus, A., Dunin-Horkawicz, S., Lupas, A. N., and Zeth, K.: A transition from strong right-handed to canonical left-handed supercoiling in a conserved coiled-coil segment of trimeric autotransporter adhesins, J. Struct. Biol., 170, 236e245 (2010). 21. Lupas, A. N. and Gruber, M.: The structure of a-helical coiled coils, Adv. Protein Chem., 70, 37e78 (2005). 22. Lupas, A.: Coiled coils: new structures and new functions, Trends Biochem. Sci., 21, 375e382 (1996). 23. Biedzka-Sarek, M., Salmenlinna, S., Gruber, M., Lupas, A. N., Meri, S., and Skurnik, M.: Functional mapping of YadA- and ail-mediated binding of human factor H to Yersinia enterocolitica serotype O:3, Infect. Immun., 76, 5016e5027 (2008). 24. Kaiser, P. O., Linke, D., Schwarz, H., Leo, J. C., and Kempf, V. A. J.: Analysis of the BadA stalk from Bartonella henselae reveals domain-specific and domainoverlapping functions in the host cell infection process, Cell Microbiol., 14, 198e209 (2012). 25. Yoshimoto, S., Nakatani, H., Iwasaki, K., and Hori, K.: An Acinetobacter trimeric autotransporter adhesin reaped from cells exhibits its nonspecific stickiness via a highly stable 3D structure, Sci. Rep., 6, 28020 (2016). 26. Nakatani, H., Kanie, J., and Hori, K.: On-fiber display of a functional peptide at sites distant from the cell surface using a long bacterionanofiber of a trimeric autotransporter adhesin, Biotechnol. Bioeng., 116, 239e249 (2019). 27. Ishikawa, M. and Hori, K.: A new simple method for introducing an unmarked mutation into a large gene of non-competent Gram-negative bacteria by FLP/ FRT recombination, BMC Microbiol., 13, 86 (2013). 28. Alvarez, B. H., Hartmann, M. D., Albrecht, R., Lupas, A. N., Zeth, K., and Linke, D.: A new expression system for protein crystallization using trimeric coiled-coil adaptors, Protein Eng. Des. Sel., 21, 11e18 (2008). 29. Sikdar, R., Peterson, J. H., Anderson, D. E., and Bernstein, H. D.: Folding of a bacterial integral outer membrane protein is initiated in the periplasm, Nat. Commun., 8, 1309 (2017). 30. Leo, J. C., Grin, I., and Linke, D.: Type V secretion: mechanism(s) of autotransport through the bacterial outer membrane, Philos. Trans. R. Soc.B, 367, 1088e1101 (2012). 31. Leo, J. C., Lyskowski, A., Hattula, K., Hartmann, M. D., Schwarz, H., Butcher, S. J., Linke, D., Lupas, A. N., and Goldman, A.: The structure of E. coli IgG-binding protein D suggests a general model for bending and binding in trimeric autotransporter adhesins, Structure, 19, 1021e1030 (2011). 32. Agnew, C., Borodina, E., Zaccai, N. R., Conners, R., Burton, N. M., Vicary, J. A., Cole, D. K., Antognozzi, M., Virji, M., and Brady, R. L.: Correlation of in situ mechanosensitive responses of the Moraxella catarrhalis adhesin UspA1 with fibronectin and receptor CEACAM1 binding, Proc. Natl. Acad. Sci. USA, 108, 15174e15178 (2011). 33. Ruiz, T., Lenox, C., Radermacher, M., and Mintz, K. P.: Novel surface structures are associated with the adhesion of Actinobacillus actinomycetemcomitans to collagen, Infect. Immun., 74, 6163e6170 (2006). 34. Yu, C., Mintz, K. P., and Ruiz, T.: Investigation of the three-dimensional architecture of the collagen adhesin EmaA of Aggregatibacter actinomycetemcomitans by electron tomography, J. Bacteriol., 191, 6253e6261 (2009). 35. Chen, X., Zaro, J. L., and Shen, W. C.: Fusion protein linkers: property, design and functionality, Adv. Drug Deliv. Rev., 65, 1357e1369 (2013). 36. Reddy Chichili, V. P., Kumar, V., and Sivaraman, J.: Linkers in the structural biology of protein-protein interactions, Protein Sci., 22, 153e167 (2013).
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion Sota Aoki,1 Shogo Yoshimoto,1 Masahito Ishikawa,1 Dirk Linke,2 Andrei Lupas,3 and Katsutoshi Hori1, * Department of Biomolecular Engineering, Graduate School of Engineering, Nagoya University, Nagoya, Aichi 464-8603, Japan,1 Department of Biosciences, University of Oslo, 0316 Oslo, Norway,2 and Department of Protein Evolution, Max Planck Institute for Developmental Biology, 72076 Tübingen, Germany3 Received 21 August 2019; accepted 30 September 2019 Available online xxx
AtaA, a trimeric autotransporter adhesin from Acinetobacter sp. Tol 5, exhibits nonspecific, high adhesiveness to abiotic surfaces. For identification of the functional domains of AtaA, precise design of domain-deletion mutants is necessary so as not to cause undesirable structural distortion. Here, we designed and constructed three types of AtaA mutants from which the same domain, FGG1, was deleted. The first mutant was designed to preserve the periodicity of hydrophobic residues in the coiled-coil segments sandwiching the deleted region. After the deletion, the protein was properly displayed on the cell surface and had the same adhesive function as the wild type. Transmission electron microscopy (TEM) imaging and circular dichroism (CD) spectroscopy showed that its isolated passenger domain had the same fiber structure as in the AtaA wild type. In contrast, a mutant designed to disturb the coiled-coil periodicity at the deletion site failed to reach the cell surface. Although secretion occurred for the mutant designed with a flexible connector between the coiled coils, the cells exhibited a decrease in adhesiveness. Furthermore, TEM imaging of the mutant fibers showed bending at the fiber tip and changes in their CD spectrum indicated a decrease in secondary structure content. Thus, we succeeded to natively display the huge homotrimeric fiber structure of AtaA on the cell surface after precise deletion of a domain, maintaining the proper folding state and adhesive function by preserving its coiled-coil periodicity. This strategy enables us to construct various domain-deletion mutants of AtaA without structural distortion for complete functional mapping. Ó 2019, The Society for Biotechnology, Japan. All rights reserved. [Key words: Trimeric autotransporter adhesin; Coiled-coil; Protein structure; In-frame deletion; Multidomain protein; Acinetobacter]
In-frame deletion of domains is commonly used to identify the functional site or epitope of large and multi-domain proteins, because it enables efficient mapping of regions larger than a point mutation (1e6). However, the deletion of an internal domain often causes undesirable structural distortion and steric conflicts in the protein, and can lead to misfolding and functional inactivation of the constructed mutant. Therefore, great care and precise design are necessary when constructing domain-deletion mutants of large and complex proteins. Previously, we discovered AtaA, a member of the trimeric autotransporter adhesin (TAA) family, in the Acinetobacter sp. Tol 5, which exhibits noteworthy autoagglutination and high adhesiveness to various abiotic surfaces from hydrophobic plastics to hydrophilic glass and metals (7e9). TAAs are outer membrane (OM) proteins of gram-negative bacteria, and polypeptide chains of a TAA form a homotrimeric structure with an N-terminal passenger domain (PSD) corresponding to its biological functions and a Cterminal transmembrane domain transporting and anchoring the PSD onto the OM (10,11). The AtaA polypeptide is composed of 3630 amino acids and is thus one of the largest TAAs. AtaA’s long PSD of 225 nm in length consists of mosaically arranged multiple domain repeats including two Ylheads, 15 GINs, 12 Trp-rings, ten GANGs,
* Corresponding author. Tel: þ81 52 789 3339; fax: þ81 52 789 3218. E-mail address: [email protected] (K. Hori).
ten DALLs, five FGGs, and several neck domains, which often appear in TAAs (9,12). While AtaA shares structural features with other TAAs, which usually bind bacterial cells to specific biotic molecules, AtaA is unique in terms of its nonspecific adhesiveness to various abiotic surfaces. This adhesive feature of AtaA can be conferred to other non-adhesive gram-negative bacteria by transformation with the ataA gene (9). We invented a new method for immobilizing bacterial cells utilizing AtaA and demonstrated its advantages; large numbers of bacterial cells expressing AtaA can be quickly immobilized onto various material supports, the immobilized cells can be reversibly detached by rinsing with deionized water (dH2O) (13) or by adding casein hydrolyzates (14), and the immobilized cells can be efficiently used for bioproduction (15e17). However, little is known about the adhesion mechanism and functional regions of AtaA. Domains of TAAs are connected by trimeric coiled coils (10,12,18,19). Coiled coils are helical bundles with periodically occurring hydrophobic residues, which form a hydrophobic core along the central axis of the bundle. Their most prevalent periodicity is of 7 residues over 2 helical turns (the so-called heptad repeat), but TAAs also have coiled coils with other periodicities such as 11, 15, and 19 residues (10,20). The periodicities of heptad and hendecad (11 residues) repeats are denoted as abcdefg and abcdefghijk, in which positions a and d in heptad repeats and a, d, e, and h in hendecad repeats come with hydrophobic residues (21,22). It is considered that domains between two coiled coils can be deleted
1389-1723/$ e see front matter Ó 2019, The Society for Biotechnology, Japan. All rights reserved. https://doi.org/10.1016/j.jbiosc.2019.09.022
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
2
AOKI ET AL.
J. BIOSCI. BIOENG.,
FIG. 1. Design of domain-deletion mutants of AtaA. (A) Schematic representations of AtaA. The dotted box indicates the deleted region in the mutants. The colored regions represent annotated domains as indicated in the box. (B) Structure of an FGG domain, FGG5 in AtaA (PDB ID: 3WPA). The monomeric chain is shown in magenta and two symmetry-related molecules are shown in gray. FGG domains have common structure consisting of a b-meander sandwiched between two coiled coils although their detail structures are varied in different FGGs. (C) Schematic figures of heptad and hendecad coiled coils. Left, the periodicity of heptad repeats is denoted as abcdefg. Right, the periodicity of hendecad repeats is denoted as abcdefghijk. Residues at the positions a and d in heptad repeats and a, d, e, and h in hendecad repeats, which are shown in red, contribute to interhelical hydrophobic interactions. (D) Amino acid sequences around the deleted region. The deleted residues in each mutant are shown as dashes. The periodicities of hendecad and heptad repeats in the coiled coils are shown above the amino acid sequence. Positions forming a hydrophobic core of coiled coils are shown in red.
without adverse effect on the structure by designing a resulting construct to preserve the periodicity of hydrophobic residues in the coiled coils after the deletion. Numerous domain-deletion mutants of TAAs have been designed and constructed with this strategy, and their production, cell surface display, and functions have been evaluated (12,23,24). However, their native folding state has never been directly confirmed because there was no effective method for analyzing the structure or morphology of TAAs after secretion onto the OM. Recently, we developed a method to isolate the PSD of AtaA by genetically introducing a recognition site for human rhinovirus 3C (HRV 3C) protease (25). Specific cleavage by the protease isolates AtaA’s PSD nanofibers from the cell surface, which enables structural and morphological analyses of isolated AtaA’s PSD in the native molecular state. In the current study, we succeeded to obtain the proper fiber structure of AtaA with correct folding and biological function after the deletion of a domain, by preserving the periodicity of flanking coiled-coil segments. Upon deviation from the periodicity, maintenance of the biological function failed due to disturbance of the proper folding of the protein.
DNA manipulation Plasmids and primers used in this study are listed in Supplementary Tables S2 and S3, respectively. The plasmid pDONR::ataA was digested with EcoRV at two sites in the original ataA and the linearized plasmid was re-circularized by self-ligation, generating pDONR::Ntemp that included a region encoding AtaA1e496. The procedures for the construction of pIFD-DFGG1A, B, and C and pIFD3C-DFGG1A and C are schematically shown in Figs. S1 and S2, respectively. Fragment-1 was amplified by PCR from pDONR::Ntemp using primers Ntemp-FGG_F/Ntemp-HiFi_R. Fragment-2A was amplified by inverse PCR from pAtaA using primers pAtaA_invPCR_F/pAtaA_invPCR2_R. Fragment-3 was excised from pDONR::ataA by digestion with EcoRV and XbaI. The three fragments were assembled using NEBuilder HiFi DNA Assembly master mix (New England BioLabs, Ipswich, MA, USA), generating pIFD-DFGG1A. To construct pIFD-DFGG1B and pIFD-DFGG1C, fragment-2B and fragment-2C were amplified by inverse PCR from pAtaA using the primer sets pAtaA_invPCR_F/pAtaA_invPCR3_R and pAtaA_invPCR_F/pAtaA_invPCR4_R, respectively, and then each fragment was assembled with fragment-1 and fragment-3 as described above. To construct pIFD3C-DFGG1A and pIFD3C-DFGG1C, fragment-3C was excised from pTA2::3CataA by digestion with EcoRV and XbaI, and assembled with fragment-1 and fragment-2A or fragment-2C as described above. Transformation of Tol 5 4140 (27) with these expression vectors was carried out by conjugal transfer from Escherichia coli S17-1 strain, as previously described (9). To produce a recombinant protein of the N-terminal head domain (Nhead) of AtaA as an antigen, the DNA fragment encoding AtaA59e325 was amplified from pDONR::ataA by PCR with the primer set AtaA59F/AtaA325R. The PCR product was digested with XbaI and BsaI, and inserted into the same site of pIBA-GCN4tri-His plasmid (28) by ligation. The constructed plasmid, pIBA-AtaA59e325-GCN4tri-His, was used for the transformation of E. coli BL21 Star(DE3). The protein produced by the E. coli transformant was used to generate the polyclonal anti-Nhead antiserum.
MATERIALS AND METHODS
Structural analysis The PSDs of AtaA wild type (WT), IFD-DFGG1A, and IFDDFGG1C were purified from the cell surface of Tol 5 4140 harboring p3CAtaA, pIFD3C-DFGG1A, and pIFD3C-DFGG1C as described previously (25). Bacterial cells
Bacterial strains and cell analysis The bacterial strains used in this study are listed in Supplementary Table S1. They were grown as described previously (9). Arabinose was added to the culture medium to induce the expression of the ataA gene and its domain-deletion mutant genes on plasmid vectors in the DataA mutant strain of Acinetobacter sp. Tol 5, 4140. The concentration of arabinose was varied from 0.3% to 0.5% to control the production levels of the proteins, which were examined by SDS-PAGE and Coomassie Brilliant Blue (CBB) staining as previously described (26). The cell surface display of the AtaA mutants was confirmed by flow cytometry and confocal laser scanning microscopy (CLSM) as previously described (26) with slight modification; anti-Nhead antiserum and anti-rabbit IgG antibody conjugated to AlexaFluor 488 (Cell Signaling Technology, Danvers, MA, USA) were used for the primary and secondary antibodies, respectively. Adherence assay (25) and autoagglutination assay (7) of bacterial cells were performed as described previously.
were treated with HRV 3C protease at 4 C for 2 days, and then centrifuged at 10,000g for 10 min. The supernatant was fractionated by precipitation in a 30% solution of the saturated concentration of ammonium sulfate. The precipitate was dissolved in dH2O and subjected to ultrafiltration with a 100 kDa molecular weight cut-off filter by centrifugation at 10,000g, and finally concentrated in a final volume of 50 mL of dH2O solution. The protein concentration was quantified by bicinchoninic acid (BCA) assay, and the purity was confirmed by SDS-PAGE followed by CBB staining. The purified PSDs were subjected to transmission electron microscopy (TEM) as described previously (25). To sum up, the purified PSD dissolved in 10 mM KCl solution was placed onto a copper grid with carbon substrate (UHR-C10; Okenshoji, Tokyo, Japan). After 10-min incubation, the grid was washed by dipping 20 times into 10 mM KCl solution and three times into dH2O. Negative staining of the proteins
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
VOL. xxx, xxxx
NATIVE DISPLAY OF TRIMERIC AUTOTRANSPORTER ADHESINS
on the grid was performed with the organo-tungstate compound Nano-W (Nanoprobes, NY, USA) according to a standard method. The specimen was observed using a TEM system (JEM-1400EX; JEOL, Tokyo, Japan) operated at 100 kV. Circular dichroism (CD) spectra were also recorded in dH2O as described previously (25) with slight modification; protein samples of 0.08 mg/ml dissolved in dH2O were used for the measurements.
RESULTS Molecular design of domain-deletion mutants of AtaA We designed three types of domain-deletion mutants of AtaA lacking FGG1, which follows upon the N-terminal Ylhead (Nhead) in AtaA (Fig. 1A). The FGG is a widespread domain of TAAs and is generally composed of a b-meander including a couple of b-strands, which is sandwiched between the N-terminal and Cterminal coiled coils (Fig. 1B). The periodicities of the N-terminal and C-terminal coiled coils are hendecad and heptad repeats, respectively (Fig. 1C). Accordingly, we designed IFD-DFGG1A lacking residues from leucine 326 (L326) to alanine 362 (A362), in which isoleucine 363 was expected to act as a substitute of L326 so as to maintain the periodicity of hydrophobic residues in the coiled-coil (Fig. 1D). For comparison, in addition to IFDDFGG1A, we also designed IFD-DFGG1B and IFD-DFGG1C, keeping the C-terminal residue of the deletion identical, namely, A362. IFD-DFGG1B was a mutant lacking residues from alanine 328 to A362, and the periodicity of hydrophobic residues in the coiled coil was disturbed by just the two extra residues remaining at the ligation joint. IFD-DFGG1C was a mutant lacking residues from glycine 335 to A362, in which a flexible sequence (GTGS) connected the coiled coils. The expression vectors for these three mutants (pIFD-DFGG1A, B, and C) were constructed and introduced into Tol 5 4140, a DataA mutant strain of Tol 5. Production and cell surface display of the domain-deletion mutant proteins To examine the production of the AtaA domain-deletion mutant proteins, the cell lysates of Tol 5 4140 expressing the mutant genes were separated by SDS-PAGE. The calculated molecular weights of mature proteins of AtaA WT, IFDDFGG1A, IFD-DFGG1B, and IFD-DFGG1C were 353.2, 349.7, 349.9, and 350.5 kDa, respectively. On the gel stained with CBB, each AtaA mutant polypeptide was detected with almost the same electrophoretic mobility as the WT protein, indicating that all of the mutant polypeptides were properly produced (Fig. 2A).
3
We next examined the cell surface display of the AtaA mutants. Tol 5 4140 cells expressing the mutant genes were stained with anti-Nhead antiserum and Alexa 488-conjugated anti-rabbit antibody, and then subjected to CLSM and flow cytometry. The cells expressing IFD-DFGG1A and IFD-DFGG1C exhibited similar fluorescence intensity as the cells expressing AtaA WT, and the fluorescence surrounded the cells (Fig. 2B and C). On the other hand, the cells expressing IFD-DFGG1B exhibited much weaker fluorescence intensity than them. These results demonstrated that IFD-DFGG1A and IFD-DFGG1C were displayed on the cell surface while the secretion of IFD-DFGG1B onto the OM failed. Functionality of the domain-deletion mutant proteins To evaluate the functionality of the domain-deletion mutants of AtaA, we performed adherence and autoagglutination assays. As shown in Fig. 3, cells expressing IFD-DFGG1A exhibited high adhesiveness to both polystyrene (PS) and glass surfaces and the autoagglutination ratio was similar to that of cells expressing AtaA WT. On the other hand, the cells expressing IFD-DFGG1B, whose cell surface display failed, showed greatly decreased adhesiveness to both material surfaces compared with the cells expressing AtaA WT and lost the autoagglutinating nature to the same level as DataA. Although IFD-DFGG1C seemed to be regularly secreted onto the cell surface, the cells expressing the mutant showed partial decreases in adhesiveness to PS and glass surfaces. The autoagglutinating nature of this mutant was slightly reduced. These results suggest that only the design of IFDDFGG1A successfully achieved deletion of the FGG1 domain without an adverse effect on the biological functions of AtaA. Morphological characteristics of the domain-deletion mutant fibers Next, we attempted to examine the morphological characteristics of the fiber structure of IFD-DFGG1A and IFDDFGG1C, which were displayed on the cell surface. For this purpose, a nucleotide sequence encoding human rhinovirus 3C (HRV 3C) protease recognition site and a tri-glycine linker was inserted into the FGG domain that was situated at the base of the PSD (FGG5), to produce pIFD3C-DFGG1A and pIFD3C-DFGG1C. These domaindeletion mutant AtaA fibers were displayed on the cell surface of Tol 5 4140 and chemically reaped from the cells with HRV 3C protease. Thus, the PSDs of these mutants were isolated by a previously established method (25). As shown in Fig. 4A, the PSDs of IFD-DFGG1A and IFD-DFGG1C were purified as with the PSD of AtaA WT and no other contaminated proteins were not
FIG. 2. Production and cell surface display of the wild type (WT) of AtaA or the domain-deletion mutants of AtaA on Tol 5-derived cells. (A) SDS-PAGE followed by CBB staining of whole-cell lysates of the Tol 5-derived cells. (B) Immunofluorescence microscopy of the Tol 5-derived cells using anti-Nhead antiserum and Alexa Fluor 488-conjugated anti-rabbit antibody. Scale bar: 2 mm. (C) Flow cytometry of the Tol 5-derived cells using anti-Nhead antiserum and Alexa Fluor 488-conjugated anti-rabbit antibody. For a control, DataA cells were also subjected to the same analyses ().
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
4
AOKI ET AL.
J. BIOSCI. BIOENG., negative band with a peak at 208 nm corresponding to a-helix than that of AtaA WT. This smaller negative band of IFD-DFGG1C PSD was independently reproduced with samples from different purification lots. This result suggests that the molecular structure of IFD-DFGG1C had a distortion that was not detected by TEM in addition to the bending at the tip. In contrast, the CD spectrum of the PSD of IFD-DFGG1A exhibited a negative band with a peak at 208 nm and almost overlapped with that of AtaA WT. This and TEM observation implied that IFD-DFGG1A had the proper folded structure just as AtaA WT.
DISCUSSION
FIG. 3. Functional analyses of Tol 5 derivatives expressing domain-deletion mutants of AtaA. (A) Adherence assay of Tol 5 derivatives expressing AtaA (WT) and the AtaA mutants. Bacterial cells adhering to polystyrene (PS) and glass plates were quantified by crystal violet staining. (B) Autoagglutination assay of Tol 5 derivatives expressing AtaA (WT) and the AtaA mutants. The autoagglutination ratio was quantified by a decrease in the optical density at 660 nm (OD660) of the cell suspension after settling for 3 h. As a control, DataA cells were also subjected to the same analyses (). Data are expressed as mean SEM (n ¼ 3).
detected, indicating that the purity was adequately high. Subsequently, the fiber morphology of these PSDs was observed by TEM. TEM images of the PSD fibers of IFD-DFGG1A with a globular tip of Nhead were similar to those of AtaA WT (Fig. 4B and C). The fiber lengths of the PSDs of AtaA WT and IFDDFGG1A, which were determined from the measurement of fifteen fibers’ images, were 224 12 nm (mean SD) and 225 13 nm, respectively. There was no significant difference in a fiber length between IFD-DFGG1A and WT (p ¼ 0.82, t-test). This was reasonable because the deleted region was too short to affect the fiber length. The PSD of IFD-DFGG1C also had fibrous morphology with the globular tip. However, the fiber length of the PSD of IFDDFGG1C decreased to 211 12 nm (p < 0.01, t-test) and some of the tips were markedly bent (Fig. 4D), implying that IFD-DFGG1C did not retain the native fiber morphology. Secondary structure of the domain-deletion mutants To confirm the folding states of IFD-DFGG1A and IFD-DFGG1C, CD spectra, which reflect the secondary structure of a protein and are widely used for this purpose, were measured. As shown in Fig. 5, the CD spectrum of the PSD of IFD-DFGG1C exhibited a smaller
For functional mapping, domain-deletion mutants are frequently constructed and examined for the preservation of their functions. In general, a domain would be recognized as a functional site if the function fails upon its deletion. However, the failure of a function can be caused not only by the deletion of its functional domain but also by the destruction or distortion of protein structure by the deletion. In this study, in two FGG1-deletion mutants, IFD-DFGG1B and IFD-DFGG1C, the functionality of AtaA was lost or reduced; the former mutant was deficient in the secretion of its PSD and the latter mutant showed a partial decrease in its adhesive function. However, it can be concluded that FGG1 is not a domain responsible for the adhesive function of AtaA because mutant cells expressing IFD-DFGG1A exhibited adhesiveness to abiotic surfaces and autoagglutinating property at the same level as cells expressing AtaA WT. Our results reaffirm the importance of the precise design of domain-deletion mutants in functional mapping to avoid disturbing the folding and structure formation of the mutant proteins. In the case of TAAs, the register of coiled-coil segments, which frequently exist in their PSDs, can be utilized for precise design (12,23,24). Our results regarding AtaA proved the appropriateness of this strategy in the functional mapping of a TAA and simultaneously showed that deviation from the periodicity could cause failure in the secretion onto the OM or partial disruption of the secondary structure of the protein together with abnormal fiber morphology. This strategy can be employed even when the crystal structure of a TAA of interest or its domain is unknown if the secondary structure and domain annotation can be deduced. TAAs are known to be secreted via the type Vc secretion pathway (29,30). In this pathway, the synthesized polypeptide chain is translocated into the periplasm by the Sec machinery, and then three polypeptide chains trimerize, forming a C-terminal b-barrel pore at the OM. Subsequently, the N-terminal PSD is transported onto the OM through the pore and folded into a fiber structure. Although the driving force for the transportation of the PSD through the OM remains an open research problem, the folding of the polypeptide chain on the OM is speculated to be the major energy source (30). Fig. 2 shows that the PSD of IFD-DFGG1B was produced adequately, but its secretion onto the OM failed. Since IFD-DFGG1B has a signal sequence of the Sec machinery, which transports polypeptides in an unfolded state, it must have passed through the inner membrane. However, the disruption of the coiled-coil periodicity should disturb folding of the PSD of this mutant, resulting in the lack of energy for its secretion. Therefore, at least a part of Nterminal PSD of the non-secreted IFD-DFGG1B polypeptide molecules should have accumulated in the periplasmic space. However, we do not know if their C-terminal transmembrane domain was inserted into the OM, which is the next challenge. Although IFD-DFGG1C seemed to be regularly secreted onto the cell surface, its adhesive function was partially decreased. In such a case, it is difficult to understand the reason for the functional deterioration and to clarify the involvement of the deleted domain in the protein function. In this study, we examined the fiber
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
VOL. xxx, xxxx
NATIVE DISPLAY OF TRIMERIC AUTOTRANSPORTER ADHESINS
5
FIG. 4. Purification and TEM observation of PSDs of the domain-deletion mutants of AtaA. (A) SDS-PAGE followed by CBB staining of the PSDs of AtaA WT, IFD-DFGG1A, and IFD-DFGG1C. An arrow indicates the bands corresponding to the PSDs. (BeD) TEM images of negatively stained PSDs of AtaA WT (B), IFD-DFGG1A (C), and IFD-DFGG1C (D). The image on the bottom of each panel shows a representative fiber of each sample. Dotted circles in panel D indicate the bending tips. Scale bar: 100 nm.
morphology and the secondary structure of the mutant protein. This is the first report confirming the folding state of domaindeletion mutants of TAAs isolated from the cell surface. This revealed that some of the PSD fibers of IFD-DFGG1C had a morphology with a bent tip (Fig. 4). Although TAAs typically have straight fiber morphologies, some TAAs, such as Aggregatibacter actinomycetemcomitans EmaA, Moraxella catarrhalis UspA1, and E. coli EibD, have been reported to bend naturally at specific sites in the PSD (31e33). In EmaA, a flexible region composed of glycine and serine was considered to be the bending site (34). As for IFDDFGG1C, because the two coiled coils sandwiching the excised region were reconnected by a flexible sequence, the bending of the tip of the PSD of IFD-DFGG1C was thought to be associated with the flexibility. Therefore, the tip of the fiber would randomly swing in liquid at the reconnected site, and some fibers could be bent and others could be in straight at one time. Indeed, the tips of some fibers, albeit not all of them, were observed to be bent. The CD spectrum revealed the decrease in the secondary structure of the PSD of IFD-DFGG1C. We do not know if the morphological change just at the reconnected site caused such a relatively large change in the CD spectrum. When we isolated PSDs, the recovery amount was always less from cells expressing pIFD3C-DFGG1C than cells expressing pIFD3C-DFGG1A or p3CAtaA (Table S4), suggesting the decrease in the accessibility of the protease to its recognition site in the former mutant cells. The protease recognition site was
FIG. 5. CD spectra of PSDs of the domain-deletion mutants of AtaA. Gray line, AtaA WT; orange line, IFD-DFGG1A; blue line, IFD-DFGG1C. The ellipticities of protein samples were measured at a concentration of 0.08 mg/mL. Each spectrum is the average of 10 scans.
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022
6
AOKI ET AL.
J. BIOSCI. BIOENG.,
introduced into FGG5, which was situated at the base of the AtaA PSD in these mutants. Therefore, we cannot rule out the possibility that the folding of this domain was affected far away by the structural change at the deletion site of IFD3C-DFGG1C. In IFDDFGG1B, the local misfolding at the reconnected site might have disabled the entire assembly and folding of the protein, resulting in failure in its secretion. Also, in IFD-DFGG1C, disturbing proper folding around reconnected site by the flexible linker might affect the entire folding and structure formation of the protein. We think that mutant molecules with different degrees of misfolding were mixed rather than all of the molecules were uniformly misfolded. In the field of protein engineering, a flexible sequence is frequently used for the connection of domains as a linker (35,36). However, the current results remind us that flexible linkers are not always effective for the connection of protein domains. In conclusion, we succeeded in restoring the huge homotrimeric fiber structure of AtaA with proper folding and biological functions after deletion of a domain, utilizing its coiled-coil periodicity. This revealed that the deleted domain FGG1 is unnecessary for the adhesive function of AtaA. Preservation of the coiled-coil periodicity will enable us to design and construct various domain-deletion mutants of AtaA without structural distortion and lead us to complete functional mapping and molecular engineering of AtaA. Supplementary data to this article can be found online at https://doi.org/10.1016/j.jbiosc.2019.09.022. ACKNOWLEDGMENTS The authors wish to thank the Division for Medical Research Engineering, Graduate School of Medicine, Nagoya University, for the usage of TEM equipment. This work was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant Numbers JP17H01345 and JP18K14062). The authors declare no conflicts of interest. References 1. Hartford, O., McDevitt, D., and Foster, T. J.: Matrix-binding proteins of Staphylococcus aureus: functional analysis of mutant and hybrid molecules, Microbiology, 145, 2497e2505 (1999). 2. Fine, D. H., Kaplan, J. B., Furgang, D., Karched, M., Velliyagounder, K., and Yue, G.: Mapping the epithelial-cell-binding domain of the Aggregatibacter actinomycetemcomitans autotransporter adhesin Aae, Microbiology, 156, 3412e3420 (2010). 3. Corrigan, R. M., Rigby, D., Handley, P., and Foster, T. J.: The role of Staphylococcus aureus surface protein SasG in adherence and biofilm formation, Microbiology, 153, 2435e2446 (2007). 4. Kao, L., Lam, V., Waldman, M., Glassock, R. J., and Zhu, Q.: Identification of the immunodominant epitope region in phospholipase A2 receptor-mediating autoantibody binding in idiopathic membranous nephropathy, J. Am. Soc. Nephrol., 26, 291e301 (2015). 5. Sundstrom, M., White, R. L., de Parseval, A., Sastry, K. J., Morris, G., Grant, C. K., and Elder, J. H.: Mapping of the CXCR4 binding site within variable region 3 of the feline immunodeficiency virus surface glycoprotein, J. Virol., 82, 9134e9142 (2008). 6. Mangan, R. J., Stamper, L., Ohashi, T., Eudailey, J. A., Go, E. P., Jaeger, F. H., Itell, H. L., Watts, B. E., Fouda, G. G., Erickson, H. P., and other 3 authors: Determinants of Tenascin-C and HIV-1 envelope binding and neutralization, Mucosal Immunol., 12, 1004e1012 (2019). 7. Ishikawa, M., Shigemori, K., Suzuki, A., and Hori, K.: Evaluation of adhesiveness of Acinetobacter sp. Tol 5 to abiotic surfaces, J. Biosci. Bioeng., 113, 719e725 (2012). 8. Hori, K., Yamashita, S., Ishii, S., Kitagawa, M., Tanji, Y., and Unno, H.: Isolation, characterization and application to off-gas treatment of toluenedegrading bacteria, J. Chem. Eng. Jpn., 34, 1120e1126 (2001). 9. Ishikawa, M., Nakatani, H., and Hori, K.: AtaA, a new member of the trimeric autotransporter adhesins from Acinetobacter sp. Tol 5 mediating high adhesiveness to various abiotic surfaces, PLoS One, 7, e48830 (2012). 10. Bassler, J., Alvarez, B. H., Hartmann, M. D., and Lupas, A. N.: A domain dictionary of trimeric autotransporter adhesins, Int. J. Med. Microbiol., 305, 265e275 (2015).
11. Linke, D., Riess, T., Autenrieth, I. B., Lupas, A., and Kempf, V. A. J.: Trimeric autotransporter adhesins: variable structure, common function, Trends Microbiol., 14, 264e270 (2006). 12. Koiwai, K., Hartmann, M. D., Linke, D., Lupas, A. N., and Hori, K.: Structural basis for toughness and flexibility in the C-terminal passenger domain of an Acinetobacter trimeric autotransporter adhesin, J. Biol. Chem., 291, 3705e3724 (2016). 13. Yoshimoto, S., Ohara, Y., Nakatani, H., and Hori, K.: Reversible bacterial immobilization based on the salt-dependent adhesion of the bacterionanofiber protein AtaA, Microb. Cell Fact., 16, 123 (2017). 14. Ohara, Y., Yoshimoto, S., and Hori, K.: Control of AtaA-mediated bacterial immobilization by casein hydrolysates, J. Biosci. Bioeng., 128, 544e550 (2019). 15. Nakatani, H., Ding, N., Ohara, Y., and Hori, K.: Immobilization of Enterobacter aerogenes by a trimeric autotransporter adhesin, AtaA, and its application to biohydrogen production, Catalysts, 8, 159 (2018). 16. Hori, K., Ohara, Y., Ishikawa, M., and Nakatani, H.: Effectiveness of direct immobilization of bacterial cells onto material surfaces using the bacterionanofiber protein AtaA, Appl. Microbiol. Biotechnol., 99, 5025e5032 (2015). 17. Ishikawa, M., Shigemori, K., and Hori, K.: Application of the adhesive bacterionanofiber AtaA to a novel microbial immobilization method for the production of indigo as a model chemical, Biotechnol. Bioeng., 111, 16e24 (2014). 18. Szczesny, P. and Lupas, A.: Domain annotation of trimeric autotransporter adhesins d daTAA, Bioinformatics, 24, 1251e1256 (2008). 19. Hartmann, M. D., Grin, I., Dunin-Horkawicz, S., Deiss, S., Linke, D., Lupas, A. N., and Alvarez, B. H.: Complete fiber structures of complex trimeric autotransporter adhesins conserved in enterobacteria, Proc. Natl. Acad. Sci. USA, 109, 20907e20912 (2012). 20. Alvarez, B. H., Gruber, M., Ursinus, A., Dunin-Horkawicz, S., Lupas, A. N., and Zeth, K.: A transition from strong right-handed to canonical left-handed supercoiling in a conserved coiled-coil segment of trimeric autotransporter adhesins, J. Struct. Biol., 170, 236e245 (2010). 21. Lupas, A. N. and Gruber, M.: The structure of a-helical coiled coils, Adv. Protein Chem., 70, 37e78 (2005). 22. Lupas, A.: Coiled coils: new structures and new functions, Trends Biochem. Sci., 21, 375e382 (1996). 23. Biedzka-Sarek, M., Salmenlinna, S., Gruber, M., Lupas, A. N., Meri, S., and Skurnik, M.: Functional mapping of YadA- and ail-mediated binding of human factor H to Yersinia enterocolitica serotype O:3, Infect. Immun., 76, 5016e5027 (2008). 24. Kaiser, P. O., Linke, D., Schwarz, H., Leo, J. C., and Kempf, V. A. J.: Analysis of the BadA stalk from Bartonella henselae reveals domain-specific and domainoverlapping functions in the host cell infection process, Cell Microbiol., 14, 198e209 (2012). 25. Yoshimoto, S., Nakatani, H., Iwasaki, K., and Hori, K.: An Acinetobacter trimeric autotransporter adhesin reaped from cells exhibits its nonspecific stickiness via a highly stable 3D structure, Sci. Rep., 6, 28020 (2016). 26. Nakatani, H., Kanie, J., and Hori, K.: On-fiber display of a functional peptide at sites distant from the cell surface using a long bacterionanofiber of a trimeric autotransporter adhesin, Biotechnol. Bioeng., 116, 239e249 (2019). 27. Ishikawa, M. and Hori, K.: A new simple method for introducing an unmarked mutation into a large gene of non-competent Gram-negative bacteria by FLP/ FRT recombination, BMC Microbiol., 13, 86 (2013). 28. Alvarez, B. H., Hartmann, M. D., Albrecht, R., Lupas, A. N., Zeth, K., and Linke, D.: A new expression system for protein crystallization using trimeric coiled-coil adaptors, Protein Eng. Des. Sel., 21, 11e18 (2008). 29. Sikdar, R., Peterson, J. H., Anderson, D. E., and Bernstein, H. D.: Folding of a bacterial integral outer membrane protein is initiated in the periplasm, Nat. Commun., 8, 1309 (2017). 30. Leo, J. C., Grin, I., and Linke, D.: Type V secretion: mechanism(s) of autotransport through the bacterial outer membrane, Philos. Trans. R. Soc.B, 367, 1088e1101 (2012). 31. Leo, J. C., Lyskowski, A., Hattula, K., Hartmann, M. D., Schwarz, H., Butcher, S. J., Linke, D., Lupas, A. N., and Goldman, A.: The structure of E. coli IgG-binding protein D suggests a general model for bending and binding in trimeric autotransporter adhesins, Structure, 19, 1021e1030 (2011). 32. Agnew, C., Borodina, E., Zaccai, N. R., Conners, R., Burton, N. M., Vicary, J. A., Cole, D. K., Antognozzi, M., Virji, M., and Brady, R. L.: Correlation of in situ mechanosensitive responses of the Moraxella catarrhalis adhesin UspA1 with fibronectin and receptor CEACAM1 binding, Proc. Natl. Acad. Sci. USA, 108, 15174e15178 (2011). 33. Ruiz, T., Lenox, C., Radermacher, M., and Mintz, K. P.: Novel surface structures are associated with the adhesion of Actinobacillus actinomycetemcomitans to collagen, Infect. Immun., 74, 6163e6170 (2006). 34. Yu, C., Mintz, K. P., and Ruiz, T.: Investigation of the three-dimensional architecture of the collagen adhesin EmaA of Aggregatibacter actinomycetemcomitans by electron tomography, J. Bacteriol., 191, 6253e6261 (2009). 35. Chen, X., Zaro, J. L., and Shen, W. C.: Fusion protein linkers: property, design and functionality, Adv. Drug Deliv. Rev., 65, 1357e1369 (2013). 36. Reddy Chichili, V. P., Kumar, V., and Sivaraman, J.: Linkers in the structural biology of protein-protein interactions, Protein Sci., 22, 153e167 (2013).
Please cite this article as: Aoki, S et al., Native display of a huge homotrimeric protein fiber on the cell surface after precise domain deletion, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.09.022