Direct targeting of host microtubule and actin cytoskeletons by a chlamydial pathogenic effector protein

Chlamydia pneumoniae (Cpn) is an important obligate intracellular bacterial pathogen, which infects lung epithelial cells, causing human respiratory diseases. It is responsible for 10% of community-acquired pneumoniae and 5% of bronchitis, pharyngitis and sinusitis cases (Grayston et al., 1995). Moreover, Cpn is associated with chronic obstructive pulmonary disease, asthma, atherosclerotic cardiovascular diseases and lung cancer (Campbell and Hahn, 2020). All chlamydial and chlamydia-related species undergo a very specific, biphasic infection cycle initiated by the infectious, extracellular and non-dividing elementary bodies (EBs), which invade host cells, and the intracellular metabolically active reticulate bodies (RBs), which replicate inside membrane vacuoles termed inclusions (Stelzner et al., 2023; Bayramova et al., 2018).

For obligate intracellular pathogens, host cell entry is of utmost importance. To ensure internalization Chlamydiae use a complex repertoire of bacterial adhesin – host cell receptor interactions as well as effector proteins translocated by the type-III secretion system (T3SS) (Nans et al., 2015). Cpn adhesins on the EB cell surface, such as OmcB, Pmp proteins and LipP, interact with host cell plasma membrane surface structures and these interactions are essential for chlamydial internalization (Wuppermann et al., 2008; Mölleken et al., 2013; Galle et al., 2019; Becker and Hegemann, 2014). Equally important for EB entry are the Cpn T3SS-mediated effector proteins such as SemC and SemD, which upon secretion into the host cell cytosol, bind and reshape the host cell plasma membrane underneath the adhered EB and recruit endocytotic proteins to support EB uptake (Hänsch et al., 2020; Spona et al., 2023).

A major target for soluble secreted effector proteins, such as the conserved TarP family, is the host cell cytoskeleton, which is composed of F-actin, microtubules (MTs), intermediate filaments and septins (Hohmann and Dehghani, 2019; Goodson and Jonasson, 2018; Svitkina, 2018; Herrmann and Aebi, 2016; Woods and Gladfelter, 2021). MTs are the most dynamic feature of the cytoskeleton of the cell, consisting of highly dynamic polar filaments that switch between phases of polymerization and depolymerization. These alterations of MT dynamics are controlled by the availability of α/β-tubulin dimers and by the association of a vast array of MT-associated proteins (MAPs) that modulate MT dynamics, especially from the MT plus-end (Meier et al., 2024). Additionally, the composition of a MT filament in terms of its different tubulin isotypes alters its dynamic parameter as does the presence of an array of post-translational modifications (McKenna et al., 2023).

Thus, the MTs even within one cell can have very different orientations and dynamics from each other (Stone et al., 2008; Tanenbaum et al., 2014; Feng et al., 2019), which can be altered quickly in time and space depending on need.

All cytoskeletal structures form a complex interacting meshwork involved in processes such as endocytosis, cell division, intracellular transport, motility, force transmission, reaction to external forces, adhesion and preservation, and adaptation of cell shape. The actin–MT crosstalk is particularly important for many of these processes (Dogterom and Koenderink, 2019). Importantly, the continued extensive crosstalk between these two cytoskeletons for basic biological functions has come into focus. An array of common regulators has been identified that co-modulate these cytoskeletons in time and space, pointing to an integrated system instead of the two separate actin and MT components of the cytoskeleton (Dogterom and Koenderink, 2019). This fact is not usually taken into consideration when the modulation of the actin cytoskeleton by bacterial pathogens is studied. In fact, typical protocols used for synchronous early chlamydia infection utilize spinning and 4°C incubation, which destroy the MT cytoskeleton. Thus, although the modulation of the host actin cytoskeleton by bacterial pathogens is well studied, the connection to the MT cytoskeleton is missing as is a thorough analysis of the MT impact on bacterial infection (Stevens et al., 2006; Colonne et al., 2016; Haglund and Welch, 2011).

All four cytoskeletal structures have been found to be associated with a C. trachomatis (Ctr) infection (Clifton et al., 2004; Volceanov et al., 2014; Al-Zeer et al., 2014; Kumar and Valdivia, 2008), but an understanding of their molecular function and role during infection is fragmentary, although the properties of specific chlamydial proteins in modulating MT dynamics have been analyzed extensively (Campanacci et al., 2019). It is known that early during infection, Ctr traffics along MTs to the MT-organizing center (MTOC; which is close to the nucleus) of the host cell to establish its intracellular niche, the inclusion (Grieshaber et al., 2003; Mital et al., 2015). Here the inclusion initiates and maintains a close association with the MTOC (also known as the centrosome) (Luis et al., 2023). A Ctr infection leads to MT dependent mitotic alterations (Grieshaber et al., 2006; Knowlton et al., 2011). Recently, it was shown that the Ctr-specific T3-secreted soluble effector protein CteG interacts with a key structural component of centrosomes and the absence of CteG impairs chlamydia's ability to replicate efficiently in primary cervical cells and in a murine model of Ctr infection (Steiert et al., 2023).

The important role of the MT cytoskeleton during Cpn infection is just beginning to emerge. Recently, we carried out a functional screen of 116 Cpn proteins and identified 13 Cpn effector proteins that modulate MT dynamics in different manners (Wevers et al., 2023). Six of the 13 proteins have a putative homolog in Ctr, seven have homologs in other Chlamydia and five proteins are unique to Cpn. Interestingly, one of the identified Cpn modulators altering the MT cytoskeleton is CPn0065, which has a Ctr homolog named IncM. A recent publication uncovers the role of IncM in MT modulation, suggesting that the hijacking of the MT cytoskeleton during a chlamydial infection might be a common trait (Luis et al., 2023).

The earliest known chlamydial effector, Ctr TarP is T3SS-translocated into the host cell cytosol within the first 5 min post infection, where it is directly involved in actin polymerization. Its proline-rich domain allows TarP oligomerization, which results in high G-actin concentrations mediated by the TarP actin-binding domain (ABD), facilitating direct F-actin polymerization (Jewett et al., 2010). This is indirectly supported by TarP binding to the actin adaptor protein vinculin, which also promotes F-actin recruitment (Thwaites et al., 2015). TarP homologs can be found in all species within the family Chlamydiaceae (Lutter et al., 2010). The Cpn TarP family member CPn0572 and Ctr TarP share the same actin nucleation activity in vitro and show a similar actin-recruiting activity early in infection in vivo (Jewett et al., 2010; Zrieq et al., 2017). In fact, both Ctr TarP and CPn0572 bind F-actin in vitro and in vivo (Jiwani et al., 2013; Zrieq et al., 2017). In addition, from previous publications on CPn0572, it is known that CPn0572 is present in EBs as mRNA (Mäurer et al., 2007) and as protein in EB lysates (Jewett et al., 2010; Zrieq et al., 2017), and is detected via antibody 15 min after infection colocalizing with actin and EBs (Zrieq et al., 2017). Moreover, CPn0572 also exhibits actin-related activities unknown for TarP. The CPn0572 C-terminal part stabilizes host F-actin by displacing members of the F-actin-severing factor cofilin family (Zrieq et al., 2017) and the N-terminal part of CPn0572 appears to contain a domain that inhibits the ability of the ABD to colocalize with actin (Braun et al., 2019).

These differences probably reflect the dissimilar requirements for the infection of the diverging anatomical sites: pulmonary epithelia versus eye and genital tract epithelia.

However, the most surprising difference between both proteins became apparent in their localization profile upon ectopic expression in human epithelial cells – a TarP–GFP fusion protein exclusively localized to distinct actin-containing aggregates, whereas a GFP–CPn0572 fusion protein colocalized with smaller actin aggregates and with filamentous structures emanating from them showing a continuous colocalization with distinct actin fibers. Remarkably, CPn0572 also localized to filamentous structures that were devoid of actin (Zrieq et al., 2017). The nature of these CPn0572-positive fibers and their relationship to other cellular structures remained obscure (Zrieq et al., 2017; Braun et al., 2019).

In this paper, we identified a new and unique function for Cpn TarP members. We found that CPn0572 can directly bind to MTs in vitro and in vivo and modulate MTs upon ectopic expression and during infection. The finding that CPn0572 can bind to either the MT and/or the actin cytoskeleton, and that the loss of the MT cytoskeleton alters its actin modulating properties suggests that CPn0572 might be able to alter the MT–actin crosstalk during early infection.

Previously, we have shown that ectopically expressed GFP–CPn0572 colocalized predominantly with the actin cytoskeleton of HEK293T cells as has been demonstrated for other members of the chlamydial TarP family (Braun et al., 2019; Zrieq et al., 2017; Clifton et al., 2004) (Fig. 1A). However, we found that in some cells CPn0572 also localized to non-actin microfilaments (Zrieq et al., 2017). To determine whether the non-actin localization might in fact be association with the MT cytoskeleton, HEp-2 cells were transfected with a plasmid expressing full-length GFP–CPn0572 (Fig. 1A) and fixed cells were stained with Rhodamine–phalloidin (actin) and an anti-α-tubulin antibody. Indeed, in 11% of cells, the chlamydial protein colocalized with both the MT and actin cytoskeleton (Fig. 1B and merged images in Fig. 1C; quantification in Fig. 1F). GFP–CPn0572 exclusive colocalization with the actin cytoskeleton (Fig. 1D) was observed in 81% cells, whereas exclusive colocalization with the MT cytoskeleton was scored for ∼8% of HEp-2 cells. (Fig. 1E; quantification in Fig. 1F). We conclude that CPn0572 can associate with both dynamic cytoskeletons of an interphase mammalian cell, although it is unclear what determines the preferred association with both or an individual cytoskeletal structure.

To identify the CPn0572 MT-binding region, we generated a number of GFP-tagged CPn0572 truncated variants and assayed their ability to associate with the actin and/or MT cytoskeleton in HEp-2 cells. The N-terminal variant CPn0572^1–536 showed exclusive actin colocalization, indicating that the MT-binding region must be present in the C-terminal part of the protein (Fig. 1G). C-terminal variants of this region showed that the absence of the G-actin- and F-actin-binding domains increased the number of cells where the variant protein showed exclusive colocalization with MTs (Fig. 1A,G). In particular, in 86% of cells, the variant GFP–CPn0572^595-755 localized with the MT cytoskeleton. The remaining transfected cells showed a non-specific cytoplasmic staining but no actin staining. Other C-terminal CPn0572 variants also showed no actin colocalization but a reduced MT colocalization frequency in comparison to CPn0572^595-755 although they were expressed in higher amounts (Fig. 1G; Fig. S1A,B). Thus, the MT-binding region (MBR) of CPn0572 is confined to the C-terminal 595–755 amino acids (Fig. 1H).

To determine whether the chlamydial protein associated exclusively with interphase MTs or was also able to colocalize with other MT structures, such as a mitotic spindle, human osteosarcoma U2OS cells were transfected with a plasmid-encoding GFP–CPn0572^595-755 for 18 h. U2OS cells are frequently used for mitotic analysis and the chosen CPn0572 variant was used as it showed a predominant association with interphase MTs. Indeed GFP–CPn0572^595-755 colocalized with spindles during all stages of mitosis (Fig. 1I; Fig. S2). To determine whether full length GFP–CPn0572 was also able to associate with the spindle, we expressed this protein in U2OS mitotic cells and also found that the protein associated with spindle MTs but was also able to localize to the cortical actin network adjacent to the plasma membrane (Fig. 1J).

We next determined whether association with the MT cytoskeleton is a conserved feature of other CPn0572/TarP family members. A comparison of 23 CPn0572 amino acid sequences from different Cpn isolates (Table S1) revealed that 20 out of 23 were identical to the CPn0572 sequence found in the GiD strain used in this study (Fig. S3A). The animal-derived Cpn isolate DC9 had a number of amino acid alterations, whereas the animal-derived isolates B21 and LPCoLN showed extensive differences. These include single amino acid (aa) changes, a large insertion of 211 amino acids, and a 24-amino-acid insertion in the region from aa 595 to aa 755, which we identified as the CPn0572 MBR from Cpn GiD (Fig. 2B; Fig. S3A). Thus, it was interesting to determine whether the insertion affected MT binding of this region. Furthermore, we selected Ctr (L2) TarP and C. psittaci (6BC) TarP family members for analysis, as these belong to different clades to Cpn (Fig. 2A). The selected CPn0572/TarP members were aligned to identify amino acid regions with homology to CPn0572^595-755 from strain GiD. A region of high similarity identified in GiD, LPCoLN and 6BC (red letters in Fig. S3B) was utilized to design N-terminal GFP fusions of these C-terminal domains (Fig. 2B, marked in blue). In the case of Ctr L2, no clear homology was found. Nevertheless, we tested a GFP–TarP^868-1005 fusion protein for its ability to bind MTs (Fig. 2B). Expression of these four variants in HEp-2 cells and subsequent microscopy analysis showed that the Cpn LPCoLN variant was also able to colocalize with MTs (Fig. 2C), whereas the C-terminal constructs derived from Ctr and C. psittaci family members did not (Fig. 2C). The latter two constructs instead colocalized with vinculin (Fig. S3C). Thus, the ability of TarP members to associate with the host MT cytoskeleton is a feature conserved in Cpn members.

In the above analysis, we noticed that the MT cytoskeleton in CPn0572-expressing cells was often aberrant, for example, showing abnormally thick interphase MT bundle structures, which were not observed in control cells expressing only GFP. To better characterize this phenotype, we analyzed interphase MTs in HEp-2 cells transfected with GFP–CPn0572^595-755 for various times prior to fixation (Fig. 3A). Interestingly, MT staining revealed a correlation between the amount of GFP–CPn0572^595-755 present and the number of abnormal, thicker MTs (Fig. 3A–C). The diameter of the MT bundles increased from ∼0.26 µm (control) to ∼0.51 µm 18 h after transfection (Fig. 3D). Thus, expression of GFP–CPn0572^595-755 results in abnormally thick interphase MTs, as has been observed upon overexpression of a MAP (Fong et al., 2013).

Next, we assayed whether the mitotic spindle was also altered upon expression of CPn0572^595-755 by transfecting U2OS cells with a plasmid encoding this CPn0572 variant for 18 h. We found that, in contrast to U2OS control cells expressing only GFP, 83% of the spindles were abnormal (Fig. 3F, quantified in Fig. 3E). A number of abnormal spindle phenotypes was scored, the most prominent, had nearly 40% of mitotic cells with elongated, irregular anaphase spindle structures with misaligned chromosomes, indicating a defect in bipolar chromosome–spindle attachment (Fig. 3F).

Thus, CPn0572^595-755 associates with interphase and mitotic MT structures, which leads to aberrant MTs with an altered function.

TarP family members have been classified as important chlamydial effector proteins that modulate the host actin cytoskeleton. With our discovery that Cpn members also colocalize with and modulate the MT cytoskeleton, it was of interest to determine whether this property affected chlamydial infection. Furthermore, we wanted to analyze the importance of the actin versus the MT-modulating effect by assessing the impact on infections of full-length CPn0572, which mainly associates with cellular actin structures but can also colocalize with MTs, and CPn0572^595-755, which associates with MTs only. We also expressed the N-terminal variant CPn0572^1-536 that only binds actin.

GFP-tagged full-length CPn0572, the N-terminal variant CPn0572^1-536 and the C-terminal variant CPn0572^595-755 were expressed in HEp-2 cells for 18 h prior to Cpn infection and the number of internalized EBs was scored after 2 h and the number of inclusions was scored after 30 h. Fig. 4A is a diagrammatic representation of the experimental set-up; Fig. 4B shows the CPn0572 variants used. Although expression of CPn0572 or CPn0572^595-755 had no visible effect on HEp-2 cells, the expression of CPn0572^1-536 in HEp-2 led to cells with typical apoptotic features after 48 h, demonstrating that CPn0572^1-536 expression was toxic. Thus, for this variant, we could only determine the number of internalized EBs 2 h post infection.

The number of internalized EBs in either CPn0572 and CPn0572^595-755 expressing cells was similar but significantly lower than those in control cells or CPn0572^1-536 expressing cells (Fig. 4C). Thus, the modulation of the MT cytoskeleton by ectopic expression of this protein affects invasion efficiency of the pathogen. In addition, the number of infected GFP-, GFP–CPn0572- or GFP–CPn0572^595-755-expressing HEp-2 cells was comparable (Fig. 4E), but the number of chlamydial inclusions per cell was reduced in the presence of GFP–CPn0572 or GFP–CPn0572^595-755 (Fig. 4D,F).

As ectopic expression of CPn0572 in epithelial cells affects MT structures and Cpn internalization, we next tested whether chlamydial expression of CPn0572 during infection would also modulate MT structures. As generation of a CPn0572 antibody useable for advanced immunofluorescence analysis was unsuccessful, we chose plasmid-borne expression of a tagged CPn0572 version in Ctr for the analysis. In contrast to Cpn, the plasmid expression system in Ctr is well established. A plasmid carrying full-length CPn0572 with a C-terminal FLAG tag under control of the constitutive IncD promoter (Agaisse and Derré, 2013; Belland et al., 2003) was transformed in Ctr L2 (Fig. 5A). In general, the developmental cycle was slower when CPn0572 was present. Thus, HEp-2 cells infected with CPn0572–FLAG-expressing Ctr transformants expressed CPn0572–FLAG for 48 h post infection (hpi) (Fig. 5B). To determine whether the protein is translocated into the host-cell cytoplasm early in infection, infected HEp-2 cells were fixed at 15 min post infection and analyzed by immunofluorescence microscopy. DNA-stained Ctr EBs were found to lie close to, or to partially overlap with, CPn0572–FLAG signals indicative of CPn0572 secretion (Fig. 5C). At 24 hpi, development of the CPn0572-expressing inclusion was affected, with the inclusion size reduced by more than 50% compared to untransformed control inclusions (Fig. 5D). Analysis of the subcellular localization of the CPn0572–FLAG fusion protein by immunofluorescence confocal microscopy at 48 hpi revealed a dense cytosolic localization exhibiting numerous filamentous structures, whereas the untransformed control cells only showed background cytosolic signals (Fig. 5E). Interestingly, a detailed look at focal planes lying between the inclusion membrane and the plasma membrane identified filamentous-like CPn0572 structures that were not associated with actin (Fig. S4) but rather with MT structures (Fig. 5E; Fig. S4), a phenotype not seen in Ctr control cells (Fig. 5E, see zoom and details of the areas denoted by the white boxes for FLAG and α-tubulin signals). Observing focal planes located between the inclusion membrane and the plasma membrane allowed us to detect clear individual MTs adjacent to the inclusion and to test colocalization; however, because these focal planes do not transverse the inclusion lumen, the clear ring-shaped MT cage observed during chlamydial infection (Al-Zeer et al., 2014; Meier et al., 2023; Haines et al., 2021; Dumoux et al., 2015) is not clearly detected here. Moreover, MT structures associated with CPn0572–FLAG signals were often aberrant compared to control MT structures. Specifically, the diameter of MT filaments in the vicinity of the inclusion in control cells was 0.21 µm, whereas MT structures in cells infected with CPn0572-expressing Ctr were thicker with 0.278 µm distal of the inclusions and 0.416 µm in proximity to the inclusion (Fig. 5F). When we infected the colon-derived epithelial Caco-2 cells with CPn0572-expressing Ctr cells, at 48 hpi CPn0572 filaments and MT filaments were associated and often overlapping, whereas the inclusion lumen was not stained (Fig. 5G), supporting the data obtained for infected HEp-2 cells.

These data show that CPn0572 can be expressed and secreted by Ctr and that this strain has a delayed inclusion development. The protein associates with invading EBs and, at 48 hpi, with MT structures, which are thickest near inclusions in comparison to MT structures proximal to inclusions in control cells.

CPn0572 can associate with and lead to the formation of thicker MTs when expressed ectopically in human cells and during an infection by Ctr expressing CPn0572. In addition, the exclusively MT-binding variant CPn0572^595-755 modulates MTs in a dose-dependent manner. Thus, it was of interest to determine whether CPn0572 could bind MTs directly and to analyze the nature of the MT modulation. To this aim, we purified recombinant bacterially expressed GST and GST–CPn0572 and either incubated it with Taxol-stabilized MTs or with buffer alone. After high-speed centrifugation GST and GST–CPn0572 were found to be predominantly in the supernatant in the absence of MTs (Fig. 6A, four rightmost lanes). In contrast, in the presence of MTs, GST–CPn0572 was exclusively present in the pellet fraction, whereas GST and a faint contaminant band in the GST–CPn0572 preparation remained in the supernatant (Fig. 6A, left and central lanes). We next tested whether purified CPn0572^595-755 was also able to interact with MTs directly. After high-speed centrifugation CPn0572^595-755 was found predominantly in the pellet fraction only in the presence of MTs (Fig. 6B). We next tested the strength of the CPn0572^595-755–MT interaction by determining their apparent equilibrium dissociation constant (K_d). We found that CPn0572^595-755 bound MTs with an apparent K_d of 70 nM (95% confidence interval: 26.8 to 113.2 nM) (Fig. 6C). Thus, we conclude that CPn0572 directly binds to MTs.

We next tested the effect of CPn0572^595-755 on MT stability by incubating GST or GST–CPn0572^595-755 with Taxol-stabilized MTs on ice. Following a 10 min ice incubation and high-speed centrifugation, depolymerized tubulin subunits were recovered preferentially in the supernatant of MTs alone or MTs incubated with GST (Fig. 6D, three left most lanes). However, MTs incubated with GST–CPn0572^595-755 were partially protected from ice depolymerization, as evident by their efficient recovery in the pellet fraction (Fig. 6D, three right most lanes).

To determine the impact of CPn0572 on MT characteristics in vivo, we expressed either GFP or GFP–CPn0572^595-755 in HEp-2 cells and incubated them on ice to induce MT depolymerization (Delphin et al., 2012) prior to immunostaining. In control GFP-expressing cells, the MT fibers present prior to ice incubation were barely detectable after 15 min and undetectable after 60 min on ice (Fig. 7A). In contrast, in GFP–CPn0572^595-755-expressing cells, short MT filament-like structures were still detectable even after 60 min on ice (Fig. 7B). In the latter case the MTs appeared fractured and displayed a punctate pattern that colocalized partially with GFP-CPn0572^595-755 signals (Fig. 7B, enlargements at the right).

Similarly, in a drug treatment experiment, MTs in GFP–CPn572^595-755-expressing HEp-2 cells displayed strongly increased resistance to the MT-depolymerization drug nocodazole (Movie 1). Thus, CPn0572^595-755 has the capacity to stabilize MTs in human cells in vivo.

MTs are diversified through a number of posttranslational modifications, which regulate MT properties and are read by cellular effectors, such as molecular motors and MAPs, to provide spatial and temporal specificity to MTs in cells (McKenna et al., 2023). Thus, we wondered whether the stabilization of MTs caused by CPn0572 would be reflected in its acetylation pattern, which has been described as making MTs more resistant to mechanical deformation (Portran et al., 2017). Indeed, in epithelial cells expressing GFP–CPn0572^595-755 almost all MTs associated with CPn0572 signals, which were thick and abnormally shaped, were acetylated, whereas in GFP-expressing control cells, acetylated MTs are less thick and fewer in numbers (Fig. 7C). Band intensity comparison revealed that in CPn0572-expressing human cells the amount of acetylated α-tubulin was increased 3-fold compared to that seen in control cells (Fig. 7D,E). These data imply that the binding of CPn0572 to MT results in their stabilization and in turn in an increased acetylation.

Although colocalization of full-length GFP–CPn0572 with MTs was only found in a minority of cells, we still wanted to test whether this association was sufficient for MT cold protection. Unlike GFP–CPn0572^595-755, full-length GFP–CPn0572 was unable to protect MTs after 60 min ice exposure (Fig. S5A). Surprisingly however, we noted a new actin phenotype which was the appearance of actin aggregates that colocalized with GFP–CPn0572 (Fig. S5A–D). We assume that the cells showing actin aggregation are those cells that before ice treatment had CPn0572 on MTs. Considering that depleting cells of MTs has an impact on the actin cytoskeleton of CPn0572-expressing cells, we conclude MTs have the capacity to play an important role in the regulation of CPn0572-mediated actin modulation.

The cytoskeleton of animal cells is a highly complex, adaptable structure required for a wide variety of distinct cellular processes. These range from cargo transport, endocytosis, force transmission and cell division to the processing and subsequent adaptation to extrinsic signals. To carry out these multiple tasks, the actin, MT and intermediate filament networks need to be highly adaptable and communicate with each other. Given the multiple roles of the eukaryotic cytoskeleton, it is unsurprising that the subversion of the infected cells cytoskeleton has evolved in many pathogens and is used for different tasks during the pathogen life cycle (Li et al., 2023; Hartland et al., 2022). For example, Listeria and Shigella use the host actin cytoskeleton for their uptake at the beginning of the infection cycle but also at the end of the cycle to spread from cell to cell (Hartland et al., 2022; Dowd et al., 2021).

The hijacking of the dynamic actin scaffold of the host is also crucial for the pathogenicity of the obligate intracellular bacteria Ctr and Cpn, both of which represent a worldwide major public health burden (Jordan et al., 2020; Campbell and Hahn, 2020). Here, members of the chlamydial TarP family are known early secreted effector proteins, which nucleate and regulate actin dynamics (Jewett et al., 2006; Caven and Carabeo, 2019). Moreover, TarP signaling activates host signaling proteins such as Rac1, phosphoinositide 3-kinase (PI3K) and the WAVE2 complex, in addition to recruiting the actin effectors formin and Arp2/3 (Romero et al., 2020; Romero and Carabeo, 2022).

In this work, we define an additional and unique role for Cpn TarP member CPn0572, beyond its well-established role as an actin nucleator, namely the association with and modulation of the MT cytoskeleton. Several eukaryotic proteins have been identified with the ability to physically crosslink F-actin and MT filaments, among them tau and MAP2 (Cabrales Fontela et al., 2017). However, only formins and adenomatous-polyposis-coli (APC) not only bind MTs and actin, but also have the intrinsic ability to polymerize actin filaments (Juanes et al., 2017; Henty-Ridilla et al., 2016). For instance, the formin mDia1 has the ability to polymerize actin from MT ends. Given that CPn0572 is able to polymerize actin (Jewett et al., 2010) and binds MTs, it is conceivable that this chlamydial effector protein functions in a similar way. Moreover, as these properties are not only found in the human Cpn isolate GiD but also in the CPn0572 homolog from the koala isolate LPCoLN, which is phylogenetically basal to the human isolates (Myers et al., 2009), we propose that the dual function of these proteins points to an evolutionarily conserved mechanism.

CPn0572 colocalizes to MTs in human interphase and mitotic cells and binds directly to MTs in vitro. Thus, CPn0572 fulfills the characteristics of a MAP. The apparent dissociation constant of CPn0572^595-755 to MTs of 70 nM is similar to that of tau (Goode et al., 1997; Gigant et al., 2014) and EB1 (Berrueta et al., 1998; Tirnauer et al., 2002).

MTs are highly dynamic structures with the ability to change the MT polymerization–depolymerization cycle according to the needs of the cell. As ectopic expression of CPn0572 in human cells protects MTs from cold- and drug-induced depolymerization in vivo and the presence of the C-terminal variant CPn0572^595-755 leads to increased resistance to cold-induced MT depolymerization in vitro, we propose that CPn0572 has a MT stabilization function. How this is achieved on a molecular level still needs to be determined; however, the finding that the presence of CPn0572^595-755 increases the amount of acetylated MTs in the cell might suggest that changing MT properties could involve tubulin post-translational modification (Janke and Montagnac, 2017).

To our knowledge, CPn0572 is the only bacterial protein identified that directly binds to and regulates both the actin and MT cytoskeletal systems, suggesting that this chlamydial protein might provide a functional crosstalk between the two cytoskeletal systems to modulate specific cellular process(es) (Dogterom and Koenderink, 2019).

What then is the role of CPn0572 in chlamydial infection? CPn0572 is expressed late in the Cpn infection cycle so to be present in the infectious Cpn EBs, and is detected 15 min post infection on actin and EBs (Jewett et al., 2010; Zrieq et al., 2017; Hänsch et al., 2020). Plasmid-borne expression of CPn0572 in Ctr shows that the secreted protein can associate with and modulate MTs in the vicinity of the chlamydial inclusion. The Ctr system used is artificial but is a very valuable tool to assay the connection between secreted CPn0572 and MTs. We do not at present know whether the actin-modulating and MT-modulating functions of CPn0572 occur in concert during early infection. However, the two functions appear to be intertwined as shown by the appearance of actin aggregates in cells expressing CPn0572 and a depolymerized MT cytoskeleton.

DNA sequences encoding CPn0572, CPn0572 deletion variants, or other members of the TarP family were amplified by PCR from genomic DNA [Cpn GiD, Cpn LPCoLN, Ctr L2 or C. psittaci 6BC; Jantos et al., 1997; Jan Rupp (Department of Infectious Diseases and Microbiology, University of Lübeck, Lübeck, Germany); ATCC no. VR-902B; and Li et al., 2022, respectively] or a pre-existing plasmid (full-length CPn0572, Braun et al., 2019) using custom-synthesized oligonucleotide primers (40-nucleotide homology to expression vector and 20-nucleotide homology to the respective gene) (Merck KGaA, Germany). Amplified DNA sequences were cloned into appropriate plasmids (pEGFP_2xFYVE; #140047 Addgene) with deletion of 2x_FYVE and addition of CEN6-ARSH4-TRP1 cassette from pYD1 (Thermo Fisher Scientific, Waltham, MA, USA), resulting in pAE67; or modified pGEX-5X-1 (GE Healthcare, Chicago, IL) harboring a CEN6-ARSH4-TRP1 where the factor Xa has been replaced with a TEV protease recognition site) using in vivo homologous recombination in Saccharomyces cerevisiae CEN.PK2 (as described previously; Moelleken and Hegemann, 2008)), amplified in Escherichia coli XL-1 blue (Stratagene) and the insertion verified by sequencing (Microsynth Seqlab GmbH, Germany). For CPn0572 expression in Ctr, plasmid p2TK2-SW2 IncDProm-RSGFP-IncDTerm (Agaisse and Derré, 2013) harboring an IncD promoter-RSGFP-IncD terminator expression cassette was used as backbone. For cloning by homologous recombination in S. cerevisiae, the CEN6-ARSH4-TRP1 cassette from pYD1 (Thermo Scientific, Waltham, MA, USA) was added upstream to the IncD promoter (yielding pKM209) and RSGFP was replaced by a 3× FLAG tag (resulting in pKM255). Finally, the DNA sequence encoding cpn0572 was cloned N-terminally to the 3× FLAG tag (pMH16).

Final constructs used were: CPn0572: bp 1–2265 (Braun et al., 2019); CPn0572^1-536: bp 1–1608 (Braun et al., 2019); CPn0572^478-755: bp 1432–2265 (Braun et al., 2019); CPn0572^536-755: bp 1606–2265 (Braun et al., 2019); CPn0572^595-755: bp 1783–2265; CPn0572^615-755: bp 1843–2265; CPn0572^635-755: bp 1903–2265; CPn0572^655-755: bp 1963–2265; CPn0572^806-990 (CPn LPCoLN): bp 2416-2970; TarP^712-870 (C. psittaci): bp 2134–2610; TarP^868-1005 (Ctr): bp 2602–3015.

Human epithelial HEp-2 (ATCC; CCL-23, contamination free) and U2OS (ATCC; HTB-96, contamination free) cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific) supplemented with 10% fetal calf serum (FCS), non-essential amino-acids (Thermo Fisher Scientific) and MEM vitamins (Thermo Fisher Scientific) at 37°C and 6% CO₂. For transfection, a 70% confluent monolayer was grown in 24-well plates (Sarstedt AG & Co. KG, Nümbrecht, Germany) on coverslips following medium exchange with fresh medium without FCS. After 18 h of transfection using TurboFect (Thermo Fisher Scientific), cells were fixed with 3% paraformaldehyde (Fisher Chemicals) in HBSS for 10 min at room temperature (RT) and washed three times with HBSS (Thermo Fisher Scientific). For permeabilization, 2% saponin (Merck KGaA, Germany) in PBS was used for 20 min at RT. Staining of MTs was performed with anti-α-tubulin antibody (Origene Technologies, Inc., Rockville, MD, USA; #BM753S; 1:150), acetylated MTs with anti-acetylated-α-tubulin antibody (Thermo Fisher Scientific; #32-2700; 1:100), vinculin with anti-vinculin antibody (Merck KGaA; #V9264; 1:100) and actin with Rhodamine–phalloidin as recommended by manufacturer (Thermo Fisher Scientific; #R415; 0.5 µl of 400× stock solution in 200 µl of PBS-Saponin solution for each coverslip), all detected with fluorescently conjugated secondary antibodies (Thermo Fisher Scientific). To visualize DNA DAPI (Merck KGaA; 1:500) was used. For western blot analysis, proteins were verified by using anti-GFP antibody (Thermo Fisher Scientific; #MA5-15256; 1:2500), anti-α-tubulin antibody (Merck KGaA; #T6199; 1:1000), anti-acetylated-tubulin antibody (Merck KGaA; #T7451; 1:20,000), anti-GAPDH antibody (Thermo Fisher Scientific; #MA5-15738; 1:1000).

Images were acquired using an inverse Nikon TiE Live Cell Confocal C2plus with 100× TIRF objective and a C2 SH C2 scanner. All images are displayed as maximum intensity projections, except when noted. Analysis of images and measurements were generated with Nikon Element software and ImageJ 1.47v (National Institute of Health, Bethesda MD, USA).

GST and GST–CPn0572^595-755 were expressed in E. coli Rosetta (DE3) cells (Merck KGaA) and purified using glutathione agarose beads (Merck KGaA) according to manufacturer's instructions and dialyzed overnight in buffer A1 (50 mM Hepes-HCl pH 7.4, 150 mM NaCl and 20% glycerol; w/v). The purity and integrity of fusion proteins were analyzed by SDS-PAGE and Coomassie Blue staining.

MT polymerization was undertaken according to the manufacturer's instruction (Cytoskeleton Inc.). MT binding assays were carried out as follows: 25 µM of Taxol-stabilized MTs (equivalent to 4 µM tubulin dimers) in 80 mM PIPES pH 7.0, 2 mM MgCl₂, 0.5 mM EGTA, 72 µM GTP, 0.4% glycerol (w/v) and 18 µM Taxol were incubated with 25 µl of 4.6 µM GST or GST–CPn0572^595-755 in Buffer A1 for 30 min at RT. Mixtures were then layered on top of a 100 µl glycerol cushion (80 mM PIPES pH 7.0, 1 mM EGTA, 60% glycerol and 20 µM Taxol) and centrifuged at 100,000 g at RT for 45 min. The supernatant was carefully collected before the extraction of the glycerol cushion. Pellets were resuspended and loaded next to their corresponding supernatant fraction in a 12% SDS gel, followed by Coomassie Blue staining. For MT stabilization assays, MT polymerization according to manufacturer's instruction was carried out followed by addition of 4.4 µM GST or GST–CPn0572^595-755 in a final volume of 32 µl. Mixtures were incubated for 25 min at 35°C and then incubated on ice for 10 min. Surviving MTs were stabilized by addition of 28 µl ice-cold taxol-stabilization buffer (80 mM PIPES pH 7.0, 2 mM MgCl₂, 0.5 mM EGTA and 20 µM Taxol), layered on top of a 100 µl-glycerol cushion, centrifuged at 100,000 g at RT for 45 min and processed as described above.

For cold-induced MT-depolymerization, HEp-2 cells were incubated on ice for 0, 15 or 60 min after transfection. MTs were stained with anti-α-tubulin antibody (Origene Technologies, Inc.; #BM753S; 1:150) and DNA was visualized with DAPI (Merck KGaA, Darmstadt, Germany; 1:500). For nocodazole treatment of HEp-2 cells, a 70% confluent monolayer of HEp-2 cells was grown in a µ-Dish with a glass bottom (ibidi GmbH, Gräfelfing, Germany) for life cell imaging. After 18 h of transfection, cells were washed twice with imaging buffer containing glucose following addition of SiR-tubulin (Spirochrome, Switzerland; #SC002; 1 µM) for MT visualization. Cells were incubated for 30 min at 37°C. DNA was visualized by using Hoechst 33342 (Thermo Fisher Scientific; 1:500). For MT de-polymerization, stained cells were treated with nocodazole (Merck KGaA; 10 µg/ml).

Cpn GiD cells (Jantos et al., 1997) was propagated in HEp-2 cells. EBs were purified in a 30% gastrographin solution (Bayer Vital GmbH, Leverkusen, Germany) and stored in SPG buffer (220 mM sucrose, 3.8 mM KH₂PO₄, 10.8 mM Na₂HPO₄, 4.9 mM L-glutamine) at −80°C. Transfected HEp-2 cells were infected at a given multiplicity of infection (MOI) with Cpn GiD by adding purified EBs suspended in DMEM followed by centrifugation at 950 g at 30°C for 1 h. The medium was replaced by fresh medium containing 12 µg/ml cycloheximide and cells were incubated at 37°C for 30 h. Paraformaldehyde- and methanol-fixed samples were used for microscopy. Inclusions were visualized with the Cpn inclusion membrane protein antibody anti-CPn0147 (generated in our lab; 1:50). For determination of internalized EBs, infected cells were fixed with 3% paraformaldehyde after 120 min post infection and uninternalized EBs were stained with a rabbit anti-Cpn GiD antibody (generated in our lab; 1:40) prior to cell permeabilization and DNA was stained with DAPI (Merck KGaA; 1:500).

A crude stock of Ctr L2 was centrifuged for 20 min at 21,885 g at 4°C. Resuspension of the pellet in CaCl₂ buffer (10 mM Tris-HCl pH 7.4, 50 mM CaCl₂) was followed by addition of 5 µg plasmid DNA and incubation for 30 min at RT. The suspension was placed on confluent HEp-2 cells and filled up with fresh medium. After spinning for 60 min at 950 g at 37°C the medium was replaced fresh medium containing 12 µg/ml cycloheximide (Merck KGaA; #C4859) and incubated over night at 37°C. The next day the medium was replaced by medium with cycloheximide and 1 µl/ml penicillin G (Merck KGaA; #13752) for selection. After incubation for 24 h at 37°C, wells were scraped, sonicated for 30 s and centrifuged for 15 min at 870 g at 37°C. The supernatant was split up to flasks with confluent HEp-2 cells and centrifuged for 60 min at 950 g at 37°C. After spinning, the medium was changed with fresh medium with cycloheximide and penicillin G and incubated for 48 h at 37°C. Cells were scraped, sonicated and centrifuged (15 min, 870 g, 37°C). The supernatant was centrifuged for 20 min at 21,885 g at 4°C. Then the pellet was resuspended in SPG buffer (220 mM sucrose, 3.8 mM KH₂PO₄, 10.8 mM Na₂HPO₄ and 4.9 mM L-glutamine) in an ultrasonic bath and frozen at −80°C. For protein verification via western blot an anti-CPn0572 antibody (generated in our lab; 1:50), anti-FLAG antibody (Thermo Fisher Scientific, Waltham, MA, USA; PA1-984B; 1:150) or anti-DnaK antibody (from Dr Svend Birkelund, Department of Health Science and Technology, Medical Microbiology and Immunology, Aalborg University, 9220 Aalborg, Denmark; 1:50) was used.

HEp-2 cells were infected by addition of control or plasmid transformed Ctr L2 and centrifugation for 60 min at 950 g at 37°C. The medium was replaced with fresh medium containing 12 µg/ml cycloheximide and cells incubated at 37°C for 15 min. Cells were fixed with methanol, washed two times with HBSS (Thermo Fisher Scientific) and blocked at RT for 10 min with PBS containing 0.1% Triton X-100 and 2% BSA. Cells infected for 48 h were pre-extracted with 80 mM PIPES pH 6.8, 1 mM MgCl₂, 5 mM EGTA and 0.5% Triton X-100 for 30 s, fixed by addition of 0.5% glutaraldehyde for 10 min and quenched with 0.1% NaBH₄ for 7 min. Cells were the washed with PBS, and blocked as described above. Anti-FLAG (mouse) antibody (Merck KGaA; #F1804; 1:500) or anti-FLAG (rabbit) antibody (Thermo Fisher Scientific; PA1-984B; 1:500) was incubated overnight at 4°C in blocking solution. Anti-α-tubulin antibody (Origene Technologies, Inc., Rockville, MD, USA; #BM753S; 1:150) was used for MT visualization. DAPI (Merck KGaA, Darmstadt, Germany; 1:500) was used to visualize DNA.

The multiple protein sequence alignment and phylogenetic tree of TarP homologues were generated using the Clustal Omega Multiple Sequence Alignment program (EMBL-EBI) selecting the MBED-LIKE CLUSTERING GUIDE-TREE and MBED-LIKE CLUSTERIN ITERATION.

For data analysis and statistical significance, Excel (Microsoft Corporation, USA) or Prism (GraphPad Software Inc., USA) was used. A two-tailed unpaired Student's t-test was used to determine statistical significance. Sample sizes in experiments were enough to ensure adequate detection of effect size. For figure creation Canvas™ 15 Software (ACD Systems of America, Inc.) was used. No AI tools were used.

We thank Alexandra Laska for initial work on CPn0572 variants, Roland Piekorz for the U2OS cell line and Sebastian Hänsch at the Center for Advanced Imaging for his help with microscopy (all at Heinrich-Heine-University, Düsseldorf, Germany). We also thank Dr Svend Birkelund (Department of Health Science and Technology, Medical Microbiology and Immunology, Aalborg University, 9220 Aalborg, Denmark) for the anti-DnaK antibody.