Agrobacterium tumefaciens has two polyphosphate (polyP) kinases, one of which (PPK1AT) is responsible for the formation of polyP granules, while the other (PPK2AT) is used for replenishing the NTP pools by using polyP as a phosphate donor to phosphorylate nucleoside diphosphates. Fusions of eYFP with PPK2AT or of the polyP granule-associated phosin PptA from Ralstonia eutropha always co-localized with polyP granules in A. tumefaciens and allowed the tracking of polyP granules in time-lapse microscopy experiments without the necessity to label the cells with the toxic dye DAPI. Fusions of PPK1AT with mCherry formed fluorescent signals often attached to, but not completely co-localizing with, polyP granules in wild-type cells. Time-lapse microscopy revealed that polyP granules in about one-third of a cell population migrated from the old pole to the new cell pole shortly before or during cell division. Many cells de novo formed a second (nonmigrating) polyP granule at the opposite cell pole before cell division was completed, resulting in two daughter cells each having a polyP granule at the old pole after septum formation. Migration of polyP granules was disordered in mitomycin C-treated or in PopZ-depleted cells, suggesting that polyP granules can associate with DNA or with other molecules that are segregated during the cell cycle.

© 2022 The Author(s). Published by S. Karger AG, Basel

Introduction

The formation of polyphosphate (polyP) granules (also designated as volutin granules) in microorganisms has been known for more than a hundred years [Meyer, 1904; Harold, 1966]. In prokaryotes, polyP is formed as globular inclusions with diameters mostly in the range from 50 to ≈200 nm (recently reviewed in [Jendrossek, 2020]). PolyP granules are reservoirs of phosphorous and (divalent) cations but can have several other functions, e.g., resistance to various forms of stresses (for reviews see [Kornberg et al., 1999; Rao et al., 2009]). Recent studies in Ralstonia eutropha (also Cupriavidus necator) showed that several proteins, including polyP kinases (PPKs) [Tumlirsch et al., 2015], phosins (proteins with CHAD-motifs such as PptA or PptB) [Tumlirsch and Jendrossek, 2017], and others, are specifically attached to polyP granules in vivo, suggesting that polyP granules are supramolecular complexes for which the designation as polyphosphatosomes has been proposed [Jendrossek, 2020]. Only little attention has been paid, however, to the question whether the localization of polyP granules (in prokaryotes) is random or is a controlled process. In cyanobacteria, Synechococcus elongatus, for example, a specific positioning system (McdA and McdB) has been recently discovered that determines the equidistant positioning of carboxysomes in the nucleoid region [MacCready et al., 2018; MacCready et al., 2020]. In Magnetospirillum gryphiswaldense, the positioning of the magnetosome chain in the middle of the spirillum-shaped cells depends on a magnetosome-specific cytoskeleton (MamK and MamJ) including an active repositioning of the split magnetosome chain system after cell division [Toro-Nahuelpan et al., 2016]. The finding that the positioning of several organelle-like structures is actively controlled suggests that the positioning of polyP granules (polyphosphatosomes) might be also controlled by the cell. In a few publications, polyP granules were found located in or associated with the nucleoid region of Myxococcus xanthus [Voelz et al., 1966], Pseudomonas aeruginosa [Takade et al., 1991; Racki et al., 2017], Caulobacter crescentus [Henry and Crosson, 2013], R. eutropha [Tumlirsch et al., 2015], or S. elongatus [Murata et al., 2016], but specific target molecules or target structures, which polyP granules might interact with or bind to, have not been identified.

In contrast to the specific localization of polyP granules in the nucleoid region, polyP granules in some bacterial species are specifically localized in the cell pole region: Corynebacterium glutamicum, for example, forms two polyP granules, one at each cell pole [Pallerla et al., 2005]. A similar observation was made in Agrobacterium tumefaciens, a species with a polar growth mode [Brown et al., 2012]. Most A. tumefaciens cells formed one polyP granule that was generally located near one of the cell poles [Seufferheld et al., 2003; Frank and Jendrossek, 2020]. Due to the impact of the polar growth mode on cell polarity in A. tumefaciens, we wondered whether the polyP granule is located at the growth pole (GP) or at the old cell pole (OP) and what happens if the GP switches to an OP after cell division. To our surprise, we observed that the polyP granules in some cells migrated from the OP to the upcoming new GP of the cell during the process of cell growth and cell division. To our knowledge, a subcellular microscopically trackable migration of organelle-like supramolecular complexes has been so far described only for the magnetosome chain, which is repositioned from the new cell poles to mid-cell after cell division in M. gryphiswaldense [Toro-Nahuelpan et al., 2016].

Results and DiscussionPPK2AT Is Always Attached to polyP Granules

Plasmid-derived expression of an eyfp-ppk2AT fusion (or mCherry-ppk2AT) in A. tumefaciens resulted in the formation of cell pole-localized fluorescent foci. Cells had either one, two, or – rarely – three eYFP-PPK2AT (or mCherry-PPK2AT) foci. Staining of the cells with DAPI and imaging at a DAPI-polyP-specific wavelength revealed that 70%, 29%, and 0.8% of the cells had one, two, or three polyP granules, respectively, that always co-localized with eYFP-PPK2AT (or mCherry-PPK2AT) foci (online suppl. Fig. s1a, b, f; for all online suppl. material, see www.karger.com/doi/10.1159/000521970). We never detected a DAPI-polyP signal without an eYFP-PPK2AT focus. This confirmed our previous observation that PPK2AT of A. tumefaciens is a polyP-associated protein and that expression of eyfp-ppk2AT is a suitable tool to monitor the formation and the localization of polyP granules. The same result was obtained when a fusion of the eyfp gene with the gene coding for the CHAD domain-containing phosin protein PptA of R. eutropha [Tumlirsch and Jendrossek, 2017] was expressed in A. tumefaciens (online suppl. Fig. s1c). The eYFP-PptARE fusion always co-localized with polyP granules and confirmed the in vitro binding of phosins to polyP [Lorenzo-Orts et al., 2019; Werten et al., 2019] also in vivo. Thus, the monitoring of eYFP-PPK2AT or eYFP-PptARE allowed us to follow the polyP granule localization in living cells in time-lapse experiments by avoiding staining of the cells with the toxic dye DAPI.

No eYFP-PPK2AT foci (or mCherry-PPK2AT foci or eYFP-PptARE) were detected when eyfp-ppk2AT (or mCherry-ppk2AT or eyfp-pptARE) was expressed in a polyP-free A. tumefaciens background (∆ppk1AT∆ppk2AT mutant). Instead, fluorescence of eYFP-PPK2AT (or mCherry-PPK2AT or eYFP-PptARE) was dispersed throughout the cell and staining of these cells with DAPI confirmed the absence of DAPI-polyP-specific foci (online suppl. Fig. s1d, e). This result is in agreement with our previous observation that PPK2AT does not catalyze the formation of polyP in vivo in A. tumefaciens and suggests that PPK2AT of A. tumefaciens is important for replenishing NTP pools at the expense of previously accumulated polyP [Frank et al., 2020].

Cells with only a single polyP granule generally had this polyP granule located at the OP as confirmed by co-expression of an eyfp-popZ fusion in combination with DAPI-polyP staining. PopZ is a pole-organizing protein well-known in alpha-proteobacteria [Bowman et al., 2008; Ebersbach et al., 2008; Grangeon et al., 2015; Pfeiffer et al., 2019], and PopZAT-eYFP in growing cells is located at the GP. Consequently, PopZAT-eYFP foci and DAPI-polyP foci were located at opposite cell poles (online suppl. Fig. s2). Cells with more than one polyP granule never had two granules at/near the same cell pole. Instead, most cells with two polyP granules had one polyP granule at each cell pole, and only occasionally we detected cells with one polyP granule at a cell pole and a second polyP granule somewhere between the cell poles (online suppl. Fig. s3). Figure 1 shows a graphic representation of the distance from the OP of 257 eYFP-PPK2AT foci representing 257 polyP granules in relation to cell length. It is evident that short cells (<3.5 µm) mainly had only one polyP granule that was located near the OP. By contrast, longer cells with >3.5 µm length had either one or two polyP granules. In the majority of the cells with two polyP granules, these polyP granules were located at each of the cell poles (83%), and only a minor fraction (17%) had one polyP granule between the cell poles. Since larger cells are in a later stage of the cell cycle compared to short cells, the frequent detection of two polyP granules in long cells suggests that new polyP granules are formed at the GP, thus explaining the general localization of one polyP granule at the OP in freshly divided (short) cells.

Fig. 1.

Dependence of polyP granule localization on cell length. The positions of polyP granules were determined in FM1-43-stained A. tumefaciens cells harboring pBBR1MCS2-eyfp-ppk2AT and were plotted against cell lengths. FM1-43 preferentially stains the OP [Zupan et al., 2013].

PPK1AT Is Closely Associated with Formed polyP Granules

A. tumefaciens mutants with deletion of ppk1AT or of both ppk1AT and ppk2AT (but not with a deletion of ppk2AT only) are deficient in polyP granule formation. Therefore, only PPK1AT is essential and responsible for polyP formation in A. tumefaciens. To monitor the subcellular localization of PPK1AT, we expressed fusions of mCherry with the ppk1AT gene in A. tumefaciens, which generally led to single fluorescence foci. When ppk1AT-mCherry was expressed in a wild-type background, the fluorescence signals were almost exclusively located at the GP (Fig. 2a, black dots), suggesting that this is the location from where polyP granules originate. Remarkably, an inactive form of PPK1AT, in which the active site (H450) had been replaced by an alanine [PPK1AT(H450A)-mCherry variant], localized mainly (≈89%) at or near the OP and was detected only in about 11% of all cases at the GP (Fig. 2a, green dots). In the absence of polyP, i.e., in a ∆ppk1AT, ∆ppk2AT background, PPK1(H450A)-mCherry foci were detected not only at the poles but also throughout the cells (Fig. 2b, orange dots). This result means that foci formation of PPK1AT(H450A)-mCherry is an intrinsic property of the enzyme and does not depend on the presence of polyP granules. This is in contrast to PPK2AT, which is soluble in a polyP-deficient (∆ppk1AT) background. Apparently, the correct localization of PPK1AT requires a catalytically active enzyme and/or the presence of polyP granules. A similar observation of a mis-localized PPK1 has been previously described for an inactive Venus-PPK1(H434A) variant in C. crescentus [Henry and Crosson, 2013]. The detection of about 11% of correctly (at the GP) localized PPK1AT(H450A)-mCherry foci in a wild-type background might be explained by hetero-oligomer formation of active PPK1AT-mCherry monomers with inactive PPK1AT(H450A)-mCherry monomers. The reason why PPK1AT-mCherry foci in the ∆ppk1AT, ∆ppk2AT background were detected mainly (75%) at the OP and only rarely (24%) at the GP (Fig. 2b, blue dots) remains unclear.

Fig. 2.

Localization of PPK1AT-mCherry variants and polyP granules. The positions of wild-type PPK1AT-mCherry foci and inactive PPK1AT(H450A)-mCherry foci relative to the OP in A. tumefaciens backgrounds as indicated are shown in (a) and (b), respectively. The distances of polyP granules detected in the form of eYFP-PPK2AT foci (c) or DAPI-polyP foci (d) relative to PPK1AT-mCherry foci at the GP are shown in (c) and (d). The position of the highest fluorescence intensity of the elongated PPK1AT-mCherry foci or PPK1AT(H450A)-mCherry foci were taken for determination of intracellular distances. All strains were grown in lysogeny broth medium.

We assume that PPK1AT-mCherry proteins assemble to higher molecular weight complexes (foci) as it has been shown for PPK1 of C. crescentus [Henry and Crosson, 2013]. Presumably, the growing polyP chain is released from PPK1AT-mCherry and aggregates to globular polyP granules. Only in the stage of ongoing polyP synthesis, the polyP granule is closely associated with PPK1AT-mCherry. Once a polyP granule has been formed, it can now bind other proteins, such as PPK2AT, phosins, and others, and at this stage, PPK1AT-mCherry can detach from the polyP granule. This “aging phenomenon” would explain why a certain fraction of polyP granules is always located in close association with PPK1AT-mCherry foci and another fraction of polyP granules is free of associated PPK1AT-mCherry. A similar aging phenomenon has been described for carbonosomes (PHB granules) in R. eutropha: for this species, a detachment of eYFP-labeled PHB synthase (PhaC) as well as of its activating protein, PhaM, was observed in time-lapse experiments [Bresan and Jendrossek, 2017].

Dependent on whether ppk1AT-mCherry was expressed from a plasmid or from the native genomic locus, the fluorescence signals were either intense, relatively long (up to 0.5 µm) and slightly curved (online suppl. Fig. s4a), or represented globular foci of relatively low fluorescence intensity, respectively (online suppl. Fig. s4b). We assume that a gene dosis effect is leading to an overexpression of ppk1-mcherry in the plasmid system and causes the elongated and intense fluorescence signals in comparison to globular foci of low intensity by genomic ppk1AT-mCherry expression from the own promoter. The polyP granules were generally located in close proximity to the PPK1AT-mCherry signals but did not co-localize with them or were located at the opposite cell pole, the OP, independent of PPK1AT-mCherry. Interestingly, in those cells that harbored two polyP granules, only one of the polyP granules was associated with a PPK1AT-mCherry signal at the GP, while the other was “free” of PPK1AT-mCherry and was located at the OP. Figure 2c, d shows a statistical analysis of the distances of polyP granules from the position of the PPK1AT-mCherry signals. It is evident that about half of the polyP granules were closely located to the GP-localized PPK1AT-mCherry signal, while the other half of the polyP granules were free of a PPK1AT-mCherry signal and were located near the opposite cell pole, the OP. PolyP granules were visualized either by co-expression of eyfp-ppk2AT (Fig. 2c) or by DAPI staining (Fig. 2d).

The close association of the PPK1AT-mCherry signals with polyP granules was investigated at higher resolution by imaging the cells with three-dimensional structured illumination microscopy (3D-SIM). Unfortunately, the fluorescence intensity of the genome-integrated ppk1AT-mcherry gene product was too low to conduct 3D-SIM. We, therefore, show the results of plasmid-derived overexpressed ppk1AT-mcherry. As shown in Figure 3, simultaneous co-expression of ppk1AT-mCherry and eyfp-ppk2AT again resulted in fluorescence signals that were close to each other but did not overlap: eYFP-PPK2AT formed spherical, halo-like signals (in x-y direction) around unstained (“black”) regions resembling the core of polyP granules. This result suggests that eYFP-PPK2AT molecules form a protein layer on the surface of polyP granules but are unlikely to reside inside of the granules. PPK1AT-mCherry produced nonspherical but longish and bent fluorescence signals suggesting that PPK1AT is attached to one side of the eYFP-PPK2AT-foci.

Fig. 3.

Three-dimensional structured illumination microscopy (3D-SIM) of A. tumefaciens. A. tumefaciens wild-type cells harboring pBBR1MCS2-ppk1AT-mCherry and pTrc200-Plac-eyfp-ppk2AT are shown. Orthogonal cross-sections shown to the right relate to the positions marked in yellow boxes.

PolyP Granules Can Migrate from the Old Pole to the Upcoming New Pole during Cell Cycle

A. tumefaciens cells grow asymmetrically at the GP and, after cell division, the former GP becomes an OP, while the new poles of the daughter cells become the new GPs [Grangeon et al., 2015]. To find out when and where new polyP granules are formed, A. tumefaciens cells expressing eyfp-ppk2AT (to monitor the position of polyP granules without the use of toxic DAPI) were grown in liquid culture (lysogeny broth [LB]) and on LB-agar pads in time-lapse experiments. The presence of the plasmid harboring the eyfp-ppk2AT fusion had no detectable impact on growth of the cells (online suppl. Fig. s5). Most cells of an exponentially growing culture had one eYFP-PPK2AT focus that co-localized with DAPI-stained polyP granules as described above (online suppl. Fig. s1b). The same observation was made for cells at the beginning of a time-lapse experiment (Fig. 4, movie 1, https://bit.ly/37fet4o): the two adjacent cells both harbored one eYFP-PPK2AT focus each representing one polyP granule. Analysis of the cells shown in the next time frames indicated that the eYFP-PPK2AT foci of both cells were located at the OPs as evident from growth at the opposite cell ends (GPs). Remarkably, after ≈90 min of growth, the eYFP-PPK2AT focus, i.e., the polyP granule, of one cell (the lower cell in Fig. 4, movie 1) began to migrate to the direction of the GP. Shortly after the onset of the migration of the polyP granule/eYFP-PPK2AT focus, cell invagination at mid-cell became detectable and indicated that the cell was in the process of cell division. Roughly at the time when the migrating eYFP-PPKAT focus had reached the (former) mid-cell position, cell division was completed. The eYFP-PPK2AT focus was now located at the new pole (GP) of one daughter cell. In addition, two new (weakly fluorescent) eYFP-PPK2AT foci appeared shortly before septum formation was completed: one at the GP and one at the OP. After cell division, the upper daughter cell had one eYFP-PPK2AT focus at the former GP (now OP) and the lower daughter cell had two eYFP-PPK2AT foci, one at each cell pole. A similar process was observed in the other cell (the upper cell in Fig. 4, movie 1, https://bit.ly/37fet4o), in which the process of polyP granule/eYFP-PPK2AT movement and cell division started ≈15 min later. A statistical analysis of cells (n = 54) showed that the polyP granules of only approximately one-third of the cells (29%) showed a directional migration, while the polyP granules of the other cells (71%) did not substantially move or were formed de novo. Similar findings were obtained in several repetitions of these experiments.

Fig. 4.

Migration of polyP granules in A. tumefaciens. Snapshots of a time-lapse experiment of A. tumefaciens harboring pBBR1MCS2-eyfp-ppk2AT at times as indicated are shown. Migration of eYFP-PPK2AT foci (polyP granules) and de novo formation of new eYFP-PPK2AT foci are indicated by white and red arrows, respectively.

PPK1AT of A. tumefaciens Is Involved in polyP Migration

PPK2AT of A. tumefaciens is a soluble protein distributed in the cytoplasm in the absence of polyP granules as revealed by a homogeneous fluorescence of eYFP-PPK2AT in a ∆ppk1 ∆ppk2 background (online suppl. Fig. s1d). PPK2AT is, therefore, unlikely to be involved in or responsible for the movement or subcellular positioning of polyP granules. To test whether the migration of polyP granules could depend on PPK1AT, we followed the localization of PPK1AT-mCherry foci in A. tumefaciens by time-lapse microscopy. As shown in Figure 5a and movie 2 (https://bit.ly/37fvIlS), PPK1AT-mCherry foci were able to migrate from the OP to the GP. To test whether the migration of PPK1AT depended on the presence of polyP granules, we followed the migration of the inactive PPK1(H450A)AT-mCherry variant in a polyP-free background (∆ppk1AT, ∆ppk2ATA. tumefaciens). Interestingly, PPK1(H450A)AT-mCherry was also able to migrate. However, the orientation of migration was not restricted to the direction from the OP to the GP, but also the reverse migration of PPK1(H450A)AT-mCherry foci from the GP to the OP was detected (Fig. 5b and movie 3, https://bit.ly/2V5RBlp). We conclude that a migration of PPK1(H450A)AT-mCherry foci between the cell poles does not depend on the presence of polyP granules. It seems that an active form of PPK1AT (PPK1AT-mCherry) or a wild-type background is necessary for the correct direction of migration from the OP to the GP. This finding is in agreement with the more or less randomly localized PPK1(H450A)AT-mCherry foci in a polyP-deficient background (Fig. 2b, orange dots) and the nonrandom localization of active PPK1AT in close association with polyP at the GP (Fig. 2a, black dots). Our result suggests that the presence of polyP is not necessary for migration of PPK1AT, but that polyP or other not yet identified molecules are necessary for a correct direction of the migration. The finding that only about one-third of polyP granules is migrating might be explained by a detachment of PPK1AT from polyP granules (“aging” phenomenon, see above): if (only) PPK1AT is responsible for migration, as a consequence, a PPK1AT-released polyP granule cannot migrate anymore. The reason, however, why the unidirectional migration from the OP to the GP requires an active form of PPK1AT remains unclear. Perhaps, the polyP-associated CHAD motif containing phosin PptA (Atu4492) or another, not yet identified, polyP granule-attached protein is involved in the coordination of the migration process. We have no evidence (blast analysis) of the presence of related positioning systems for organelle-like structures, such as the McdAB-related carboxysome positioning system in cyanobacteria [MacCready et al., 2018], the PhaM- or PhaF-related proteins in carbonosome-forming bacteria [Galan et al., 2011; Pfeiffer et al., 2011], or the actin-related cytoskeleton-like proteins (MamK/MamJ) in magnetotactic bacteria [Uebe and Schüler, 2016], that might be present in A. tumefaciens and could be involved in positioning of polyP granules.

Fig. 5.

Migration of PPK1AT-mCherry foci in A. tumefaciens. Snapshots of a time-lapse experiment of wild-type A. tumefaciens harboring pBBR1MCS2-ppk1AT-mCherry (a) and of ∆ppk1AT∆ppk2ATA. tumefaciens harboring pBBR1MCS2-ppk1(H450A)AT-mCherry (b) at times as indicated are shown.

Migration of PPK1a of R. eutropha in A. tumefaciens

To test whether the movement of polyP-associated PPK1AT in A. tumefaciens is a specific feature of PPK1AT, we expressed eyfp-ppk1a of R. eutropha, a beta-proteobacterial species in a polyP-free background (∆ppk1AT, ∆ppk2AT) of A. tumefaciens and performed time-lapse studies. PPK1a of R. eutropha is responsible for polyP formation in R. eutropha; however, polyP granules in R. eutropha are located in the nucleoid region and do not migrate during the cell cycle [Beeby et al., 2012; Tumlirsch et al., 2015]. The expression of eyfp-ppk1aRE in ∆ppk1AT, ∆ppk2ATA. tumefaciens restored the formation of polyP granules, but eYFP-PPK1aRE did not colocalize with polyP and formed foci distinct from DAPI-polyP foci as shown in online supplementary Figure s6. Migration of eYFP-PPK1aRE foci was clearly detectable in time-lapse experiments (movie 4, https://bit.ly/3ii0Jft). eYFP-PPK1aRE foci migrated to the GP of the cell. Immediately after cell division, the eYFP-PPK1aRE foci turned back, now moving to the new GP of the daughter cell. Remarkably, the migration of eYFP-PPK1aRE foci appeared to be more rapid than the migration of polyP granules with attached eYFP-PPK2AT. These results indicated that the signal to migrate or the signal to bind to a moving molecule is also present in PPK1aRE of R. eutropha and does not depend on a co-localized polyP granule. This is somehow unexpected since polyP granules and/or eYFP-PPK1aRE did not show an active movement in R. eutropha [unpubl. observation]. We assume that PPK1AT in A. tumefaciens interacts with and/or binds to an unknown mobile molecule and that this interaction also functions with PPK1aRE of R. eutropha. The primary amino acid sequences of PPK1AT and PPK1aRE share an identity/similarity of 41/58%, respectively. The reason why movement of eYFP-PPK1aRE was more rapid in comparison to PPK1AT might be the early release of the PPK1aRE enzyme from a freshly formed polyP granule in contrast to a much longer association of PPK1AT with polyP. The complex of PPK1AT with polyP has a much higher mass and higher diameter and this might explain the slower movement of eYFP-PPK1AT foci compared to PPK1aRE.

Movement of polyP Granules in the Presence of Mitomycin C

What could be the driving force for the PPK1AT movement? The amino acid sequences of PPK1AT and PPK1aRE do not contain known motifs for nucleotide binding such as Walker A/B motifs that could be responsible for providing the energy for migration. We assume that PPK1AT is not a migrating protein per se but is able to bind to another protein or molecule that is moving during the cell cycle and, thus, could explain the movement of PPK1AT and/or PPK1AT with associated polyP. The genome of each cell must replicate and segregate during cell division (for recent reviews on DNA segregation systems see [Gogou et al., 2021; Hu et al., 2021]), and this process includes an active movement of the daughter chromosome(s) to the upcoming daughter cell. We, therefore, hypothesized that PPK1AT could interact with and bind to DNA, thus explaining the observed movement shortly before cell division. To test this hypothesis, we followed the movement of polyP granules that were “stained” with eYFP-PPK2AT in A. tumefaciens cells in the presence of mitomycin C (movie 5, https://bit.ly/3C64jAZ). Mitomycin C interacts with DNA and covalently connects the two DNA strand. As a consequence, DNA replication is blocked and, in turn, the cells stop cell division, leading to the formation of dramatically elongated and often branched cells. Protein biosynthesis and other metabolic reactions are not affected by mitomycin C. In some of our experiments, the A. tumefaciens circular chromosome was labeled via expression of a parBAT-mcherry fusion. When such cells were grown in the presence of 2 µg/mL mitomycin C, the originally rod-shaped short cells turned into long and branched filaments that could still elongate but did not divide anymore. As shown in movie 6 (https://bit.ly/3fjbJaA), the large, branched cell harbored two visible ParBAT-mCherry foci that did not, however, migrate during the experiment. They remained more or less at the same position in the middle of the long filament. Analogous experiments in the absence of mitomycin C clearly showed the formation of a ParBAT-mCherry focus that divided into two foci, one of which rapidly migrated to the other cell pole before cell division, similar to what has been shown by others previously [Robalino-Espinosa et al., 2020]. This indicated that the concentration of 2 µg/mL of mitomycin C is suited to inhibit cell division and segregation of the circular chromosome but is below lethal levels. When the position of polyP-attached eYFP-PPK2AT foci granules was followed, a back and forth movement of the two eYFP-PPK2AT foci at each end of the approximately 14-µm-long filament by up to 2–4 µm towards the center of the filament and backwards was observed. Apparently, some movement or oscillation of polyP granules was still possible, but unidirectional migration was clearly disturbed. Since the ParBAT-mCherry foci did not migrate in the presence of mitomycin C, it is unlikely that polyP granules are attached to the circular chromosome in the neighborhood of parS-bound ParB-mCherry. We cannot, however, exclude that PPK1AT is bound elsewhere, to the linear chromosome or to one of the two plasmids of A. tumefaciens.

Directed Movement of polyP Depends on Pole Integrity

Due to the polar growth mode of A. tumefaciens, the direction of growth alternates after each cell division [Brown et al., 2012]. As a consequence, the GP of a cell turns into an OP after cell division. This implies that the two poles are differently organized. Indeed, several GP- and OP-specific proteins have been identified in A. tumefaciens and in other α-proteobacteria in the last decade. PopZ and PodJ, for example, are specific for the GP and the OP, respectively [Grangeon et al., 2015]. They do not migrate but appear and/or disappear at the respective poles dependent on the cell cycle. Cells without PodJ or without PopZ are disordered in segregation of the circular and the linear chromosome [Robalino-Espinosa et al., 2020]. To test whether mutants impaired in cell pole organization are also impaired in polyP granule migration we followed movement of eYFP-PPK2AT-marked polyP granules in PopZ-depleted cells of A. tumefaciens. As shown in Figure 6 and movie 7 (https://bit.ly/37cmgjl), depletion of PopZ resulted in the formation of branched cells as previously described [Grangeon et al., 2015; Grangeon et al., 2017; Howell et al., 2017]. In these cells, the eYFP-PPK2AT-marked polyP granules did not show a directed migration. Instead, the polyP granules seemed to be located to the area of the cell poles and some of them showed a random oscillation around its position. Apparently, a correctly organized GP is necessary for directed migration of polyP granules.

Fig. 6.

Migration of polyP granules in PopZ-depleted A. tumefaciens. Snapshots of a time-lapse experiment of PopZ-depleted A. tumefaciens cells harboring pBBR1MCS2-eyfp-ppk2AT at times as indicated are shown. The black arrowhead points to an oscillating eYFP-PPK2AT focus.

It is known that segregation of the circular and the linear chromosomes is disordered in PopZ-depleted cells and often results in the formation of mini-cells with absent or incomplete genomes [Grangeon et al., 2017]. We also observed the formation of mini-cells in PopZ-depleted cells. Interestingly, polyP granules in these mini-cells showed a rapid back and forth oscillation between the cell poles of the mini-cells. Further studies including labeling of each genomic element are necessary to obtain deeper insights into the background of PPK1AT/polyP movement. Furthermore, the determination of the polyP proteome of A. tumefaciens is necessary to identify other, so far unknown, polyP-associated proteins that might be involved in migration of polyP granules. Interestingly, a similar approach in P. aeruginosa very recently led to the identification of AlgP as a novel polyP-associated protein that is responsible for equidistant spacing of multiple polyP granules in P. aeruginosa [Chawla et al., 2021].

Materials and MethodsBacterial Strains, Plasmids and Culture Conditions

Bacterial strains and plasmids used in this study are listed in Table 1. A. tumefaciens C58 strains were cultivated in LB (88 rpm agitation)-medium supplemented with the appropriate antibiotics at 30°C or 28°C. The A. tumefaciens ∆popZ depletion strain was pre-cultured in the presence of IPTG to induce expression of the popZ gene. For depletion experiments, the pre-culture was centrifuged and the cell pellet was resuspended in fresh LB-medium without IPTG.

Table 1.

Strains and plasmids used in this study

Construction of Gene Fusions and of Gene Deletions

The PCR-assisted construction of in frame gene fusions of ppk genes with mCherry or eyfp and transformation of A. tumefaciens cells was performed as previously described [Lenz and Friedrich, 1998; Frank and Jendrossek, 2020]. For construction of gene deletions, the sacB-containing pLO3 suicide vector was used [Frank and Jendrossek, 2020]. An analogous procedure was applied for the formation of genomic gene fusions and exchange of the parental gene by the gene fusion. Each constructed plasmid/strain was verified by DNA sequencing of the total region of the modified DNA.

Microscopic Methods

Fluorescence microscopy was done as described in detail elsewhere [Hildenbrand et al., 2019; Frank and Jendrossek, 2020]. 3-μL portions of cells were immobilized on agarose pads (1% (wt/vol) in phosphate buffered saline) and covered with a coverslip. For time-lapse experiments, 3 μL of the respective cell suspension were added to the bottom surface of an LB-Agar pads (1% (wt/vol) in LB-medium) and – after drying for 15–30 s – were put into a nunc glass base dish (12 mm). The borders of the agar pad were sealed with ≈150 µL of liquid (50°C) LB-agar. Drying of the agar pad during the time-lapse experiment was reduced by adding water-soaked filter paper into the nunc glass base dish. Care was taken that the filter paper did not touch the agar pad. The closed dishes were incubated at 28°C or 30°C in an incubation chamber and imaged every 4–5 min for up to 12 h.

Structured Illumination Microscopy (3D-SIM)

3D-SIM (striped illumination at 3 angles and 5 phases) was performed using an Eclipse Ti2-E N-SIM E microscope (Nikon) equipped with a CFI SR Apo TIRF AC 100×H NA1.49 oil objective lens, LU-N3-SIM laser unit (Nikon) (488/561/640 nm wavelength lasers; emission filters: 525/50, 605/70, 700/75 nm), and an Orca Flash4.0 LT Plus sCMOS camera (Hamamatsu). Sample preparation employing precision cover glasses (0.17 mm thickness, No. 1.5H, Marienfeld) and calibration of the motorized objective correction collar and SIM grating focus using fluorescent beads with a diameter of 100 nm (T-7279 TetraSpeck microspheres) were performed at 24°C as described previously [Pfeiffer et al., 2019]. To correct for color shift between channels during 3D-SIM image reconstruction, the system was calibrated using T-7279 TetraSpeck microspheres and the N-SIM color registration. 3D-SIM imaging was performed using exposure times in the range of 20–80 ms at 40–60% laser power. z-series were acquired at a total thickness of 1.50–2.64 µm with 50–120 nm z-step spacing. Image reconstruction was performed in NIS-Elements 5.11 (Nikon) using either the “slice reconstruction” algorithm (illumination modulation contrast: auto; high resolution noise suppression: 0.1; blur suppression: 0.01) or the “stack reconstruction” algorithm for z-stack data (illumination modulation contrast: auto; high resolution noise suppression: 0.1).

Evaluation of Microscopical Pictures

Images were processed using Fiji/ImageJ software [Schindelin et al., 2012] in combination with ObjectJ [Vischer et al., 2015] to determine cell length and position of fluorescent foci (scale 15.5px/μm).

Acknowledgements

We gratefully acknowledge the support by P. Brown/University of Missouri, Columbia, MO, USA) for providing the A. tumefaciens ∆popZ depletion strain and P. Zambrysky (University of California, Berkeley, CA, USA) for advice in time-lapse experiments. We also thank F. Narberhaus (University of Bochum, Bochum, Germany) for support in A. tumefaciens techniques and helpful discussions.

Statement of Ethics

Ethical approval was not required for this study (no experiments with humans or animals).

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This work was supported by a grant of the Deutsche Forschungs­gemeinschaft (DFG) to D.J. and RTG1708.

Author Contributions

C.F. performed most of the experiments, prepared the figures, and wrote parts of the manuscript. D.P. performed 3D-SIM experiments. M.A. supported the project by advice and assistance on how to manipulate A. tumefaciens. D.J. designed the study and wrote most parts of the manuscript. All authors evaluated the data and read the manuscript.

Data Availability Statement

All data generated or analyzed during this study are included in this article and in its online supplementary material files. Links to electronic files of time-lapse movies are embedded in the text and are additionally provided in the online supplementary material files.

This article is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC). Usage and distribution for commercial purposes requires written permission. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.