Reconstitution of Spiroplasma swimming by expressing two bacterial actins in synthetic minimal bacterium

Motility is one of the most important features of life, but its evolutionary origin is still unknown. Here, we focus on Spiroplasma, commensal or parasitic bacteria. They swim by the helicity switching of a ribbon-like cytoskeleton composed of six proteins each evolved from a nucleosidase and a bacterial actin called MreB. We expressed these proteins in a synthetic minimal bacterium, JCVI-syn3B whose genome was computer-designed and chemically synthesized. The synthetic bacterium showed swimming motility with the features common with Spiroplasma swimming. Moreover, some combinations of two proteins showed helical cell shape and swimming, suggesting that the swimming was originated from differentiation and coupling of the bacterial actin. One-Sentence Summary Expression of two bacterial actins gave cell helicity and swimming to a synthetic minimal cell.

Motility is observed in many forms of life, and is arguably one of the major determinants for survival. If we focus on the force-generating units of cell motility, all cell motilities reported so far can be classified into 18 mechanisms (1). Generally, the direct evolutional ancestor of the individual mechanisms cannot be identified, probably because many of these have been in existence for a long time. However, it is possible to discuss the origins and their evolution. Cell 5 motility is thought to originate from the movements of housekeeping proteins after they were amplified and transmitted to cell outside. However, this process has not been experimentally demonstrated. class Mollicutes are parasitic or commensal bacteria, characterized by a small genome (2,3). Interestingly, there are three unique motility mechanisms in class Mollicutes (4)(5)(6). It is likely that when phylum Firmicutes evolved to stop peptidoglycan synthesis in order to 10 escape the host's immune system, they also stopped flagellar motility, which depends on the peptidoglycan layer, and then acquired unique motility (1,5). In one of the three types of motilities, Spiroplasma swimming, they push water backwards by switching the handedness of their helicity (4,(7)(8)(9). These schemes are completely different from those of the spirochete, a group of bacteria whose cells are also helical. The helical shape of Spiroplasma is likely 15 determined by a ribbon-like cytoskeleton, which is consist of mainly fibrils evolved from nucleosidases (10)(11)(12) and five classes of Spiroplasma MreBs evolved from MreB, the bacterial actin (12-15). Here, we call Spiroplasma MreBs as SMreBs, because they are distantly related to MreBs found in walled-bacteria (13,16,17). The helicity of ribbon is determined by fibril protein, but the mechanism of helicity switching is unknown. 20 A synthetic bacterium JCVI-syn3.0B (syn3B in short), was established by J. Craig Venter Institute (JCVI) in 2016, as a combination of a cell of Mycoplasma capricolum and a genome designed based on Mycoplasma mycoides. Both Mycoplasma species belong to Spiroplasma clade, one of four Mollicutes clades. It has a fast growth rate beneficial for genome manipulation, a roughly spherical morphology, and no motility (18,19). In this study, we 25 reconstituted Spiroplasma swimming in syn3B by adding seven genes and identified the minimal gene set for Spiroplasma cell helicity and swimming.

Results
Reconstitution of Spiroplasma swimming in syn3B. We focus on Spiroplasma eriocheiris, an 30 actively swimming pathogen of crustaceans (14). Seven genes which are likely related to swimming are encoded in four loci in the genome, comprising fibril, five classes of SMreB, and a non annotated conserved gene. We assembled these genes into a 8.4 kb DNA fragment and incorporated it into the syn3B genome, using the Cre/loxP system (Fig. 1A, Fig. S1, Table S1) (20,21). An active promoter of syn3B, Ptuf was inserted at the upstream of the gene cluster. 35 Surprisingly, under optical microscopy 48% of the syn3B cells showed filamentous cell shape and active movements, presumably accompanied by force generation, and moreover 13% had helical shape and swimming motility (Fig. 1B, Movie S1). Then, we named this construct as syn3Bsw. The width and pitch of the cell helices analyzed with light microscopy were slightly different from those of Spiroplasma cells (Fig. 1C). If we focus on cells that are partially bound 40 to the glass, we can see that a free part of cell was rotating with some reversals (Fig. 1D, Movie S2), meaning that helicity switching causes the helix rotation in syn3Bsw, like Spiroplasma swimming. Next, we analyzed the helices and their handedness of cell images in each frame of swimming video (Fig. 1E, F). The handedness of the cell helix was different depending on the axial position, and the helicity changed with time. Then, we measured the movement and 45 rotation speed of the helix from the part where the helix appeared to move along the cell axis smoothly. The helix movement and rotation speeds were 8.2 ± 3.7 µm/s and 11.6 ± 4.8 /s (n=10) for syn3sw, not significantly different from 8.8 ± 2.8 µm/s and 12.0 ± 3.6 /s (n=16) for Spiroplasma. In the cryo-electron microscopy (EM) image of syn3Bsw cells, fi1aments running along the axis were observed in the inner part of the curvature, like Spiroplasma cells (Fig. S2). The filaments recovered from syn3Bsw cells showed periodicity and width, that were similar to 5 fibril filament from Spiroplasma cells (Fig. 1G, Fig. S3) (10). In addition, electrophoretic and mass spectrometric analyses of cell lysates showed that fibril and all SMreBs were expressed in syn3sw cells (Fig. S4, Table S2). These results show that the expression of Spiroplasma proteins inside syn3Bsw cells resulted in the formation of internal filaments that reconstituted helical shape, helicity switching, and swimming. 10 Differences in swimming between syn3Bsw and Spiroplasma. The speeds of helix movement and rotation were not significantly different between syn3Bsw and Spiroplasma (Fig. 1F). However, the trajectory of the cells over 10 s showed that syn3Bsw could not travel long distances, unlike Spiroplasma (Fig. 1H). The reason can be seen in the time course of helicity 15 switching, showing little continuity in the rotation that hampers long distance traveling (Fig. 1F). This may be caused by lack of cooperativity in the helicity switching generating the helix rotation. EM images of syn3Bsw cells did not show the tapered pole including an inner architecture called "dumbbell", unlike Spiroplasma cells ( Fig. 1I) (14), suggesting that the tapered pole made by unknown proteins have some roles for continuous helicity switching of the 20 ribbon.

Role of component proteins.
To examine the role of each protein, we made and analyzed constructs in which each protein was not expressed (Fig. 2, A and B, Movie S3). In order not to affect gene expression by the change in the DNA and RNA structures, we introduced nonsense 25 mutations to one of the 8th-22nd codons of each structural gene (Fig. S1). We confirmed by electrophoresis that the target proteins were no longer expressed in the mutant cells (Fig. S5). No significant differences from syn3Bsw were observed in cell structures and behaviors for five of the six constructs ( Fig. 2A). However, in the construct missing SMreB5, the helix width was 0.64 ± 0.13 µm, significantly larger than that of syn3Bsw, in half of the filamentous cells, and 30 the cells moved but did not swim. The distinctive features by the lack of SMreB5 are consistent with a previous observation that Spiroplasma citri lost helicity and swimming by the lack of SMreB5 (13). These results suggest that the seven proteins have redundant roles for helix formation and swimming.
We then examined syn3B constructs expressing each protein (Fig. 2C, Movie S4). The cells 35 expressing only fibril protein formed helical cell shape with a pitch of 0.72 ± 0.08 µm and a width of 1.0 ± 0.10 µm, which is wider than Spiroplasma cells. The pitch of the helix is in good agreement with the number from isolated fibrils, which is consistent with the fact that fibril is the major component of the ribbon (10,11). The cells expressing SMreB2 formed filamentous morphology, and some of them formed helices with variety of pitches as 0.66 ± 0.12µm. The 40 cells expressing only SMreBs1, 3, or 4, did not show difference in cell shape from the original syn3B.
Expressing pair of SMreBs. Next, we analyzed the shapes and behaviors of cells expressing ten combinations of SMreB protein pairs (Fig. 3A, Fig. S1). As five classes of SMreB can be divided into three groups from amino acid sequence: 5-2, 4-1, and 3 (13,16,17), here we will discuss the results based on this classification. In the pairs of SMreBs selected from each of the 5-2 and 4-1 groups, surprisingly, the cells in the 5-1 and 5-4 combinations showed helix formation and movements, and some cells showed swimming like syn3Bsw, with occurrence frequencies comparative to syn3Bsw (Fig. 3, B and C, Movie S5). The cells of 2-1 showed filamentous morphology and movements. The cells of 2-4 combination showed filamentous morphology but 5 basically immotile. However, a few in several hundred cells showed movements. In the combinations of one in 5-2 or 4-1 groups paired with 3, cells of 3-2 formed right-handed helix (Fig. 3D, Movie S6). In the combinations of 3-1, 3-4 and 3-5, the cells did not show differences from the original syn3B. In the combinations in the same group, 5-2 and 4-1 cells were filamentous, and 4-1 cells rarely formed a short, right-handed helix. 10 In the construct of 5-4, we fused a fluorescent protein mCherry into SMreB 5 and 4 at a position suggested by previous studies (Fig. 3E, Fig. S1, Movie S7)(22). The cells expressing SMreB5 fused with mCherry showed a helical cell shape and swimming as observed in the 5-4 cells. Fluorescence was observed throughout a cell, suggesting that SMreB5 filaments were formed along the entire cell axis. Also this result showed that the mCherry fusion did not interfere the 15 functions of SMeB5. The 5-4 cells with mCherry fusion to SMreB4 did not show helicity basically. Even helical cells found in hundreds of cells did not show any movements. To clarify the roles of fibril, a major component of ribbon structure, we analyzed cells expressing fibril additionally to SMreBs 4 and 5 (Fig. 3F, Movie S8). The differences between presence and absence of fibril protein were subtle in analyses conducted in this study. 20

Discussion
MreB belonging to actin superfamily, forms a short antiparallel double-strand filament, based on ATP energy (23,24). It has the ability sensing the curvature of the peripheral structures and serves to guide the bacterial peptidoglycan synthase to the positions required for the synthesis 25 (25). Isolated SMreBs also form fibers similar to those of MreB (13,26). Our results indicate that the helix formation and force generation of Spiroplasma occur by the interaction between different SMreBs. The mechanism can be explained as follows (Fig. 4). Protofilaments made of proteins belonging to either SMreB 5 and 2 or 4 and 1 group are aligned along the cell axis, and are bound together. If the unit length in each protofilament is different, some curvature is 30 induced in the double strand, resulting in helix formation. If these protofilaments undergo a local length change at different timing using ATP energy, the curvature changes like a Bi-metallic strip resulting in helicity switching (4). The length change may be related to polymerization and depolymerization in terms of the change in axial distance between subunits. Remarkably, the difference in amino acid sequence between SMreBs 5-2, and 4-1 in S. eriocheiris is only less 35 than 34% (16). This small number of differences suggest that the ancestors of SMreB may have acquired stability, helicity, and switching after accidental acquisition of different properties. In other words, it may represent the moment when a small structural change in a housekeeping protein is amplified by an accidental accumulation of mutations, leading to motility. The reason for the existence of as many as five SMreBs, even though two proteins are capable of acquiring 40 helicity and force generation, is unclear. It may be advantageous for efficient and robust swimming, possibly in different environments, or for chemotaxis. The participation of fibril can be explained in a similar way.
Here, we used JCVI-syn3B as an experimental platform (18,19). Since the genes of synthetic bacteria are derived from organisms related to Spiroplasma, it remains possible that factors derived from synthetic bacteria, such as proteins and membrane structures, are essential for the helix formation and swimming. Then, it is still a future challenge to elucidate the mechanism of Spiroplasma swimming completely. Nevertheless, the results of this study demonstrate that syn3B is a good system to study cell evolution. 5
Plasmid construction. The Spiroplasma genome was isolated as previously described (27). The plasmid used to transform JCVI-syn3B to obtain syn3Bsw (pSeW001) were constructed as follows (Fig. S1). Focused Spiroplasma DNA regions, puroR gene, and vector fragment were 15 amplified from the Spiroplasma genome DNA and pSD079 DNA (21) as five PCR products, using primer sets listed in Table S1. The DNA fragments were assembled by In-Fusion ® HD Cloning Kit (Takara Bio Inc. Kusatsu, Japan). pSeW002 was constructed by replacing the upstream region of the 1st gene, fibril in pSeW001 to Ptuf fragment (promoter from the EF-Tu gene) amplified from pSD079. pSeW102, pSeW202, pSeW302, pSeW402, pSeW502, pSeW602, 20 and pSeW702 were modified to introduce nonsense mutation in individual genes. The plasmids to construct other strains were modified from pSeW005, which was constructed by the process stated above, using pSeW002 as the PCR template. Ptuf or Pspi (Spiralin promoter from Spiroplasma) (21) were inserted at the 5' end of the 1st orfs. All DNA fragments were verified for DNA sequence. 25 Protein analyses. Profiling and identification of protein in cells were performed as previously described (14,28,29).
Phase-contrast and fluorescence microscopy. The cultured cells of Spiroplasma and syn3B 30 were observed in 0.5 × SP4 medium diluted by PBS, containing 0.5% methylcellulose and 0.5 mg/mL BSA. The cell density was adjusted by centrifugation at 11, 000 × g for 10 min, followed by suspension with the diluted medium. The cell suspension was inserted into a tunnel slide (14,27,30,31) and observed by an inverted microscope IX71 (Olympus, Tokyo, Japan) equipped by UPlanSApo 100× 1.4 NA Ph3 and a CMOS (complementary metal-oxide-semiconductor) 35 camera, DMK33UX174 (The Imaging Source Asia Co., Ltd. Taipei, Taiwan). The videos were analyzed by ImageJ ver.1.53f51 (Fiji) using plugins, MTrackJ and EGT (Empirical Gradient Threshold), and a color foot printing macro (32).

Electron microscopy.
To observe the intact cells, cultured cells were collected by the 40 centrifugation, suspended to be 10-fold density of the original in the medium, and fixed using 0.5% glutaraldehyde for 5 min at 25°C. After quenching by 500 mM Tris-HCl pH7.5, the cells were collected by the centrifugation, washed and suspended in PBS to be 40-fold density of the original. The cell suspension was placed on a carbon-coated grid for 5 min, removed, rinsed by PBS three times, and then stained with 2% phosphotungstic acid for 60 s. To observe the internal structure, the cell suspension was treated with PBS including 0.1 mg/mL DNase, 1 mM MgCl2 and 1mM PMSF (phenylmethylsulfonyl fluoride) for 10 min at 4°C, and centrifuged at 200,000 × g for 30 min at 4°C. The pellet was suspended in PBS to be 160-fold density of the original, placed on the EM grid for 2 min, and stained with 2% phosphotungstic acid for 60 s. The images were acquired using a JEM1010 EM (JEOL, Akishima, Japan) equipped with a FastScan-F214(T) CCD (charged-coupled device) camera (TVIPS, Gauting, Germany). For cryoEM, the 5 cultured cells were collected and suspended to be 10-fold density of the original and frozen as described previously (33). The images were captured by Talos F200C EM (Thermo Fisher Scientific, Waltham, USA) equipped by 4k × 4k Ceta CMOS camera (Thermo). The images were analyzed using the ImageJ.     4. Schematic for origin and mechanism of Spiroplasma swimming. The swimming mechanism may be acquired through four steps as presented by arrows.
Step 1: The MreB protein derived from walled bacteria differentiated into two classes with different characters through accumulated mutations. Association of heterogeneous protofilaments allowed stable 5 filament formation.
Step 2: Small differences in length generated curvature, resulting in helicity of the heterogenous filament.
Step 3: Change in length caused by ATP energy induces change in curvature, causing helicity switching. The early stage of swimming was acquired.
Step 4: The acquired swimming was refined to be equipped by five classes of SMreBs, fibril, dumbbell structure and so on. Corresponding cell morphology and behaviors are presented in the bottom.