The Taspase1/Myosin1f-axis regulates filopodia dynamics

Summary The unique threonine protease Tasp1 impacts not only ordered development and cell proliferation but also pathologies. However, its substrates and the underlying molecular mechanisms remain poorly understood. We demonstrate that the unconventional Myo1f is a Tasp1 substrate and unravel the physiological relevance of this proteolysis. We classify Myo1f as a nucleo-cytoplasmic shuttle protein, allowing its unhindered processing by nuclear Tasp1 and an association with chromatin. Moreover, we show that Myo1f induces filopodia resulting in increased cellular adhesion and migration. Importantly, filopodia formation was antagonized by Tasp1-mediated proteolysis, supported by an inverse correlation between Myo1f concentration and Tasp1 expression level. The Tasp1/Myo1f-axis might be relevant in human hematopoiesis as reduced Tasp1 expression coincided with increased Myo1f concentrations and filopodia in macrophages compared to monocytes and vice versa. In sum, we discovered Tasp1-mediated proteolysis of Myo1f as a mechanism to fine-tune filopodia formation, inter alia relevant for cells of the immune system.

Among the latter is the protein Myosin1f (Myo1f), a monomeric class I myosin, which belongs to the subgroup of unconventional myosins (Crozet et al., 1997) that do not form filaments unlike conventional myosin (Kalhammer and Bä hler, 2000). Myosins are actin-based, ATP-dependent motor proteins characterized by a tripartite structure: the middle 'neck' domain interconnects the conserved N-terminal motor domain and the highly divergent C-terminal 'tail' domain.
Class-I myosins are known to execute diverse biological functions. They are involved in membrane dynamics and actin organization, thus affecting cell migration, endo-, exo-and phagocytosis (Mermall et al., 1998;Maravillas-Montero and Santos-Argumedo, 2012). Myo1f is expressed predominantly in the spleen, mesenteric lymph nodes, thymus and lung, and in particular in certain cells of the mammalian immune system including natural killer cells, macrophages (M4), dendritic cells and neutrophils (Kim et al., 2006). Myo1f-deficient neutrophils show a severely impaired migration rate on fibronectin and Myo1f-deficient mice are more susceptible to infection (Kim et al., 2006). Furthermore, Myo1f is required for neutrophil migration in a 3D environment during acute inflammation (Salvermoser et al., 2018). However, the exact molecular function of Myo1f in immune cell motility and thus the innate immune response still remains to be elucidated.
Myosin-X, another unconventional myosin, was revealed to stimulate the formation and elongation of filopodia (Berg et al., 2000;Berg and Cheney, 2002;Zhang et al., 2004;Bohil et al., 2006;Tokuo et al., 2007). Filopodia are thin, finger-like protrusions of the plasma membrane that contain tightly packed parallel bundles of actin filaments (Small, 1988). They are typically less than 10 mm long but can extend up to 70 mm, depending on the cell type (Mogilner and Rubinstein, 2005;Schä fer et al., 2011).
Here, we demonstrate that Myo1f induces the formation of filopodia-like plasma membrane protrusions which lead to increased cell adhesion and migration properties. Moreover, we discovered that Myo1f is a bona fide Tasp1 substrate and we also provide valuable hints for the physiological relevance of this proteolysis during human hematopoiesis.

Myo1f is a bona fide Tasp1 substrate
As endogenous Myo1f is rarely detectable in non-immune cells, we first employed the full-length open reading frame of MYO1F, encoding a protein of 1098 amino acids and 125 kDa, as a GFP fusion for mammalian expression. As an important control, we also generated a non-cleavable Myo1f mutant, in which the Tasp1 consensus cleavage site was deleted by a 245 DG 246 to alanine mutation (Myo1f-245 AA 246 -GFP). A semi-in vitro Tasp1 substrate cleavage assay was established to investigate a putative Tasp1-mediated Myo1f cleavage. Briefly, 293T cells with neglectable endogenous Tasp1 expression were transfected with the respective Myo1f constructs. Then cell lysates were incubated with recombinant His-tagged Tasp1, resulting in a decreased amount of full-length Myo1f-GFP and the appearance of a lower migrating molecular species consistent with the calculated molecular weight (124 kDa) of the C-terminal Myo1f-GFP cleavage product ( Figure 1A). However, following prolonged incubation with recombinant Tasp1-His, both Myo1f-GFP species disappeared in a time-dependent manner, whereas the full-length Myo1f in the untreated samples remained stable over time. Thus, we repeated the assay in presence of a two-fold concentrated protease inhibitor mix, which indeed allowed the stabilization of at least the C-terminal Myo1f-fragment against proteolytic degradation over time ( Figure 1B). Of note, Tasp1 itself is not inhibited by common protease inhibitors (Hsieh et al., 2003a). Moreover, this setup enabled us to confirm the predicted cleavage site in Myo1f, as we could not detect any cleavage product of the Myo1f-245 AA 246 -GFP mutant, even after prolonged incubation with recombinant His-Tasp1 ( Figure 1B). We further set out to verify proteolytic cleavage in a more physiological environment. Therefore, we analyzed 293T cell lysates with ectopic co-expression of Myo1f-GFP or the cleavage-site-deficient mutant and Tasp1-HA in the presence or absence of 50 nM bortezomib, a specific inhibitor of the 26S proteasome (Adams et al., 1999). Again, we could detect the C-terminal Tasp1 cleavage product only for the wild-type protein, but not the mutant ( Figure 1D), further substantiating Myo1f as a bona fide Tasp1-substrate. Notably, in absence of the inhibitor, the C-terminal cleavage fragment was hardly detectable. As the cleavage is suggested to remove the ATP binding site pivotal for ATPase activity of Myo1f, the remaining Myo1f likely represents a non-functional motor protein destined for proteasomal degradation.
Myo1f is a nucleo-cytoplasmic shuttle protein However, it was still unclear how proteolytic processing and the accompanying removal of the ATP binding site on Myosin's head domain may affect the cytosolic and membrane-associated localization of Myo1f. Therefore, localization of full-length Myo1f-GFP and GFP-fusions of myc-tagged Myo1f fragments representing the cleavage products resulting from Tasp1-mediated proteolysis was investigated with confocal fluorescence microscopy ( Figure 2A). Here, the N-terminal Tasp1 cleavage fragment (myc-Myo1f-N-term-GFP, 55 kDa) exclusively showed nuclear accumulation, whereas the respective C-terminal cleavage construct (myc-Myo1f-C-term-GFP, 124 kDa), was solely localized in the cytoplasm ( Figure 2B, right panel). These results prompted us to screen the Myo1f protein sequence for the presence of intracellular transport signals, in particular nuclear localization (NLS) and nuclear export signals (NES). First, the web-based motif predictor tool ''cNLS mapper'' (Kosugi et al., 2009) predicted a bipartite NLS in the extreme N-terminus of Myo1f and at least 5 potential nuclear export mediating sequences in the C-terminal portion of the protein were identified by in silico analysis with ''NES Finder 0.2'' and ''LocNES'' (Xu et al., 2015) ( Figure S1). Thus, we decided to make use of the general nuclear export inhibitor Leptomycin B (LMB) irreversibly inhibiting the export receptor Exportin1/Crm1. Indeed, for full length Myo1f-GFP, which predominantly localizes to the plasma membrane and the cytoplasm in untreated cells, we could detect a partial, but robust nuclear iScience Article accumulation in response to treatment with 5 nM LMB for 3 h ( Figure 2B, left panel). This could be quantified in four independent experiments, in terms of a significant increase in the mean nuclear fluorescence intensity ( Figure 2C) as well as a clear shift in the respective frequency distribution ( Figure 2D). Moreover, the subcellular distribution pattern of Myo1f-GFP and its truncations could be confirmed by subcellular protein fractionation. Likewise, LMB treatment results in an enrichment of Myo1f-GFP in the nuclear protein extracts ( Figure S2).
The potential of Myo1f to access the nucleus might be directly linked to its cleavage by the predominantly nuclear/nucleolar Tasp1, enabling optimal access of the protease to its substrates within the same cellular compartment. Moreover, its characterization as a nucleo-cytoplasmic shuttle protein places Myo1f in good company with further equally classified unconventional myosins, such as Myosin1C (Nevzorov et al., 2018).
To further confirm our results and to elucidate a possible functional role for the nuclear Myo1f, we performed co-immunoprecipitations from chromatin fractions. Precipitated full length Myo1f-GFP was found to be associated with RNA polymerase II and the histone protein H3 ( Figure 2E). It should also be pointed out that the cleavage fragment of Myo1f is not detectable here as in this approach no bortezomib was added to prevent it from proteasomal degradation.  Of note, not only wild-type Myo1f-GFP, but also the variant with a mutated Tasp1-cleavage site was bound to chromatin and associated with the same nuclear proteins to a comparable extent ( Figure 2E). Moreover, association of Myo1f with active chromatin occurs regardless of whether WT Tasp1-HA is co-expressed or the proteolytically inactive Tasp1 mutant Tasp1-T234V-HA. This suggests that cleavage of Myo1f is not a prerequisite for its nuclear localization, which is in line with our microscopic data.

Myo1f induces filopodia formation
To understand the biological effects of Myo1f cleavage, we aimed to gain more insights into its general cellular functions. Various cell lines (293T, HeLa, SW480) were transfected with Myo1f-GFP and analyzed by confocal microscopy. Myo1f-GFP was predominantly localized at the plasma membrane and in the cytoplasm of transfected cells ( Figures 3A and 3D). Interestingly, expression of Myo1f-GFP ( Figures 3A and 3D, left panel) led to impressive formation of thin plasma membrane protrusions reminiscent of filopodia. Staining with phalloidin revealed that F-actin and Myo1f-GFP clearly co-localized at the rim of the cell and within the membrane protrusions, indicating that the Myo1f-provoked structures are indeed filopodia ( Figure 3D, left panel). Filopodia quantification with FiloQuant software (Jacquemet et al., 2017) showed a significant increase in relative filopodia density of Myo1f-GFP expressing HeLa cells compared to GFP expressing control cells ( Figure 3B), demonstrating that the observed effect can be attributed to Myo1f overexpression. Moreover, only Myo1f-GFP expressing cells display filopodia longer than 5 mm ( Figure 3C).
The extent of Myo1f-containing cell protrusions varies widely among the three different cell lines (Figure 3D): The non-cancerous 293T cell line exhibited the highest incidence of cell protrusions, whereas the cancerous cell lines HeLa and SW480 revealed a less and the least pronounced effect, respectively. To test whether differences in the Tasp1 expression level might account for this varying abundance and size of filopodia, we analyzed the different cell lines for endogenous protein amounts by immunoblot (Figure 3F). Indeed, Tasp1 was undetectable in 293T cells, whereas the 50 kDa proenzyme could be detected in lysates of HeLa cells and the highest expression level was found in SW480 cells.
As Myo1f is referred to be mostly expressed in innate leukocytes, we next compared THP-1 cells differentiated with PMA (phorbol 12-myristate 13-acetate) into macrophage-like cells (THP-1 macrophages) with their direct precursors -THP-1 monocytes. During the maturation of monocytes into macrophages, cells undergo distinct morphological changes. In particular, they enlarge, become adherent and develop membrane ruffles and filopodia. Treatment with PMA indeed induced such phenotypic alterations of THP-1 cells. Immunostaining of endogenous Myo1f revealed that it is localized in the filopodia of THP-1 macrophages. In THP-1 monocytes, lower amounts of endogenous Myo1f were detected and no Myo1fcontaining cell protrusions were observed ( Figure 3E). Likewise, immunoblot analyses demonstrated a reduced expression of full length Myo1f in THP-1 monocytes compared to macrophages in which Tasp1 concentration was decreased ( Figure 3G).
The small Rho GTPase Cdc42, has been shown to control the formation of actin based filopodia and dynamic cell adhesions called focal complexes to the extracellular matrix (Nobes and Hall, 1995). Thus, we next investigated a potential interplay of Myo1f and Cdc42 in filopodia formation, by using wild-type mCherry-Cdc42 and a constitutively active mutant thereof, mCherry-Cdc42-Q61L. In the latter one, a Q61 to L amino acid substitution results in the loss of the intrinsic GTPase activity thereby preventing iScience Article inactivation (Krengel et al., 1990;Nalbant et al., 2004). Both Cdc42 variants are at least partially co-localized with Myo1f-GFP at the plasma membrane and at the base of filopodia ( Figure S3A). These findings were further substantiated by co-immunoprecipitation of Myo1f and the constitutively active form of Cdc42 ( Figure S3B).

Cleavage of Myo1f by Tasp1 antagonizes filopodia formation
Next, we analyzed whether Tasp1's proteolytic cleavage activity is responsible for the observed inverse correlation of Tasp1 expression and the amount of Myo1f-provoked filopodia. Therefore, we tested whether Tasp1-mCherry co-expression affects Myo1f-GFP-induced filopodia formation. An uncleavable Myo1f mutant (Myo1f-245 AA 246 -GFP) served as control. Filopodia amount and length were quantified in fluorescent confocal images of a sufficient number of HeLa and 293T cells (n > 80) expressing the respective Myo1f variant with or without Tasp1-mCherry. Of note, two classes of filopodia were categorized according to length: < 5 mm and > 5 mm. The amount of filopodia smaller than 5 mm seemed to be stable within each Myo1f expressing cell line and independent of Tasp1 co-expression. Therefore, filopodia longer than 5 mm were quantified and assessed by pairwise comparison via t-test among the cells expressing the same Myo1f variant with or without Tasp1 co-expression. Indeed, in Myo1f-GFP-expressing cells the number of filopodia longer than 5 mm was significantly reduced by Tasp1-mCherrry co-expression in both Hela-and 293T cells ( Figure 4). Interestingly, we observed that expression of the uncleavable Myo1f mutant (Myo1f-245 AA 246 -GFP) also promotes filopodia formation ( Figure S4), underlining the relevance of unprocessed, full-length Myo1f for the promotion and/or maintenance of distinct cell protrusions. But in this case, the presence of Tasp1-mCherry did not result in a perturbed filopodia formation ( Figure 4). This indicates that the inverse correlation of the cellular Tasp1 level and Myo1f-induced filopodia is indeed causally linked to the specific proteolysis of Myo1f. The latter is supposed to result in the elimination of the ATP binding site located in the head domain ( Figure 1C). As such, it is conceivable that after the potential to produce mechanical energy is lost by Tasp1 cleavage, the ability of Myo1f to initiate or elongate cell protrusions decreases. This might in turn account for the observed inverse correlation of endogenous Tasp1 levels and the capability of Myo1f to efficiently induce, stabilize or elongate filopodia formation.
Migrating cells often rely on filopodia to sense the environment and to support cell adhesion and movement, thus enabling guided cell migration. Indeed, Myo1f-GFP containing filopodia of transfected HeLa cells are actively involved in probing the environment as sequential TIRF images demonstrate a dynamic alternation of extending and retracting Myo1f-GFP containing filopodia ( Figure 5A). Next, we examined whether the Myo1f-induced filopodia alter the cell adhesion capacity. Myo1f-GFP expressing 293T cells iScience Article adhered more efficiently to fibronectin-coated cell culture wells than non-transfected control cells. Co-expression of proteolytically active Tasp1-mCherry clearly abrogated this effect ( Figure 5B). Furthermore, cell motility of Myo1f-GFP expressing cells was determined with a cell migration assay in collagen-coated Boyden chambers with 8 mm pore size. Migratory cells moved towards the chemoattractant in the lower compartment and adhered to the bottom side of the membrane. Here again, significantly more Myo1f-GFP expressing cells migrated through the pores compared to non-transfected cells or cells that co-express Tasp1-mCherry ( Figure 5C). These data thus confirm our hypothesis that Myo1f-induced cell protrusions facilitate increased adhesion and expand the migratory capacity of 293T cells. Moreover, concurrent Tasp1 expression antagonizes this effect, supposedly by proteolytic removal of the ATP binding site, which is a part of the short N-terminal cleavage fragment. Even though the C-terminal cleavage fragment still contains the actin-binding site ( Figure 1C), the generation of force and motion essentially requires ATP binding.
A potential role of the myo1f/Tasp1-axis during immune cell development As Myo1f is found to be mostly expressed in innate leukocytes, we next compared different types of blood cells isolated from human peripheral blood mononuclear cells (PBMCs). In particular Lincells, representing an enriched stem cell fraction, were obtained by depletion of cells that express lineage markers. These enriched stem cells were compared to CD14 + CD16cells and macrophages. CD14 + CD16cells represent classical monocytes, whereas macrophages were generated from isolated human monocytes by human iScience Article serum-induced differentiation. Blood cell differentiation occurs by a stepwise developmental progression from hematopoietic stem cells towards lineage commitment. The monocyte-macrophage lineage represents one differentiation pathway in hematopoiesis, where HSPC-derived monocytes are precursors of macrophages and dendritic cells. Interestingly, differentiation from hematopoietic stem and progenitor cells (HSPCs) to monocytes and finally macrophages was indeed accompanied by decreased Tasp1 expression coinciding with increased concentrations of full-length Myo1f in the more specialized cell types ( Figures 6A and S5).
This again strongly supports that the inverse correlation of Tasp1 and Myo1f concentration and respectively Myo1f-induced filopodia is also detectable in the physiological context of immune cell differentiation. We conclude that Tasp1 counteracts the filopodia formation capability of Myo1f in hematopoietic stem cells and macrophage precursors and we thus suggest the following model ( Figure 6B): During hematopoiesis, full-length Myo1f levels increase while Tasp1 levels decline with a higher state of cellular differentiation. HSPCs reside in the bone marrow, where high Tasp1 expression enables suppression of Myo1f-induced filopodia formation. This allows to retain hematopoietic stem cells in the bone marrow until developmental signals induce their mobilization, and the emanating monocytes start circulating in the bloodstream. The intermediate developmental stage requires active cell protrusions, yet less than terminally differentiated macrophages which rely on filopodia for ECM-and tissue-invasion as well as phagocytosis. In sum, Tasp1-mediated proteolysis might represent a fine-tuning mechanism to modulate the biological function of Myo1f by increasing filopodia length and density during immune cell differentiation.

DISCUSSION
The unconventional Myo1f was suggested as a potential, previously unknown substrate in a genome-wide bioinformatic screen for the human Tasp1 degradome (Bier et al., 2011b). Our study experimentally verified Myo1f as a substrate of Tasp1. Myo1f cleavage by Tasp1 results in the removal of the anterior 245 amino acids of the N-terminal motor domain containing the ATP-binding site. The resulting main cleavage fragment (Myo1f 246-1098) is strongly destabilized and destined for degradation. iScience Article Myo1f is the first confirmed substrate of Tasp1 with a predominant subcellular localization outside the nucleus. The apparent contradiction that a nuclear/nucleolar protease cleaves a cytosolic and plasma membrane-associated protein was resolved by our finding that Myo1f is a nucleo-cytoplasmic shuttle protein equipped with a functional nuclear localization sequence and multiple potential nuclear export signals. Moreover, Myo1f can be added to the growing group of nuclear myosins which have been attributed to functions such as transcriptional activation and relocation of chromatin (Philimonenko et al., 2004;Kulashreshtha et al., 2016).
A contribution of Myo1f to such nuclear key processes is indeed conceivable as we demonstrated an association of Myo1f with RNA-polymerase II and histones. Inside the nucleus, Myo1f might facilitate transcription of genes that are complementary to its pro-migratory function at the cell membrane, although this has to be elucidated in future studies.
In addition, our study unravels that the expression of Myo1f is associated with a massive formation of F-actin rich, cellular membrane protrusions exhibiting characteristic features of filopodia. Successful coimmunoprecipitation of Myo1f and a constitutively active form of Cdc42 further indicates an involvement of Myo1f in filopodia formation. Interestingly, Myo1f-induced filopodia formation was accompanied by increased substrate adhesion and improved migration ability. This is in agreement with previous findings revealing that Myo1f mediates neutrophil migration (Kress et al., 2007;Salvermoser et al., 2018) and with data showing that mouse Myo1f enhances intercellular adhesion capability of macrophages, thereby triggering M1 polarization (Piedra-Quintero et al., 2018).
We thus hypothesize that the unconventional Myo1f executes a specific function to promote the formation of filopodia in the immune response, as Myo1f expression was reported to be particularly elevated in natural killer cells, macrophages, dendritic cells and neutrophils (Kim et al., 2006).
Further, we could demonstrate an inverse correlation between the abundance of filopodia and the amount of cellular Tasp1. The ability of Myo1f to induce, stabilize or elongate filopodia and the concomitant effects of increased cell adhesion and migration are significantly reduced upon Tasp1 co-expression. Indeed, this inverse correlation could be causally attributed to Tasp1-mediated cleavage of Myo1f. Proteolytic processing removes the N-terminal ATP-binding site, rendering the main cleavage fragment incapable of inducing or elongating filopodia. Although the actin-binding site is retained in the C-terminal fragment, ATP binding and hydrolysis are prerequisite to generate force and thus motion (Muretta et al., 2015). Hence, Tasp1mediated proteolysis is expected to pivotally impact Myo1f functionality and stability.
Different studies have demonstrated that the initiation of filopodia formation in cancer cells from solid tumors leads to an increased migration ability of the transformed cells and accelerated metastatic processes (Gupton and Gertler, 2007;Jacquemet et al., 2015). At first glance, our data seemingly conflict with prior findings, namely that the cleavage of Myo1f by the cancer-promoting Tasp1 prevents Myo1finduced filopodia formation, which would admittedly avert the acceleration of metastatic processes.
However, Myo1f is predominantly expressed in cells of the innate immune response, therefore the observed effects have to be evaluated in the context of leukocyte development and leukemia. Innate immune cells patrol in the blood and migrate to sites of infection or inflammation in response to certain stimuli. To efficiently migrate in the extracellular matrix and to invade tissues, they rely on filopodia as these structures sense the environment (Davenport et al., 1993) and mediate attachment to the ECM via filopodia tips and shafts which contain adhesion molecules such as integrins and cadherins (Letourneau and Shattuck, 1989;Zhang et al., 2004).
Neutrophils, macrophages and dendritic cells are moreover phagocytic cells which ensure efficient removal of pathogens and other potentially harmful cells. Here, the formation of filopodia is particularly important, since filopodia support the physical capturing of pathogens and mediate the contact to infected or cancerous cells (Kress et al., 2007).
An impaired Myo1f-induced filopodia formation, e.g., resulting from higher Tasp1 expression, might easily disturb the ability of white blood cells to adhere to ECM and to enter tissue and thus even enhance the cancerous potential of white blood cells. As Tasp1 is indeed overexpressed in many leukemic cells ll OPEN ACCESS iScience 25, 104355, June 17, 2022 iScience Article (Takeda et al., 2006), its newly discovered Myo1f cleavage ability may deliver a further explanation for the relevance of Tasp1 in cancerogenesis.
Moreover, TASP1 is a disease-related gene as human TASP1 gene defects unambiguously result in developmental disorders and congenital immunodeficiency (Suleiman et al., 2018(Suleiman et al., , 2019Balkin et al., 2019). The clinical phenotype of infants with homozygous deletion or homozygous loss-of function mutations of the TASP1 gene includes developmental delay, anomalies related to skeletal and organ systems, recurrent respiratory infections, anemia and a lack of early erythroid precursors combined with an impaired megakaryopoiesis. A comparable clinic-pathologic pattern was observed for Tasp1 deficient mice. Besides massive early postnatal lethality, Tasp1 À/À animals are smaller in size and display profound skeletal abnormalities coupled with an impaired cell proliferation and a decrease in thymocytes and hematopoietic stem cells (Takeda et al., 2006;Niizuma et al., 2015Niizuma et al., , 2021. Those observations are consistent with the results of our study, as we provide first evidence that Tasp1 activity might play a role in developmental processes of immune cells. We detected the most prominent Tasp1 levels in an enriched stem cell fraction, whereas the amounts of Tasp1 are decreased in monocytes and further reduced in terminally differentiated macrophages. Inversely, in mature macrophages we detected substantially higher amounts of full-length Myo1f protein, correlating with an augmented filopodia formation. It is thus tempting to speculate that Tasp1-mediated cleavage of Myo1f might enable the adjustment of filopodia quantity during the development of macrophages, and maybe also of neutrophils and dendritic cells. However, this correlation must be validated in comprehensive functional studies in different immune cell populations to experimentally prove a causative role of Tasp1.
In conclusion, Tasp1-mediated proteolytic cleavage might serve as a fine-tuning mechanism to regulate cellular Myo1f concentration and activity post-transcriptionally, thereby dynamically modulating filopodia density and length, e.g., essential during differentiation of innate immune cells.''

Limitations of the study
Some of our observations were primarily based on Myosin1f overexpression studies.
To overcome this limitation we tried to perform Myosin1f and Taspase1 loss-of-function studies in immune cells but we observed transfection efficiencies of monocytes and macrophages less than 10% reflected by a minimal reduction of respective protein expression levels in Myo1f-siRNA and Tasp1-siRNA transfected cells compared to the non-silencing control siRNA approach. A meaningful comparison of filopodia content was therefore not possible. However, we were able to demonstrate in an alternative experimental approach that endogenous Myosin1f indeed fulfills a physiological function in filopodia formation: untagged, endogenous Myosin1f was visualized by immunofluorescence microscopy with a Myosin1f antibody in THP-1 monocytes and PMA-differentiated THP-1 macrophages. Endogenous Myosin1f was primarily detected in the filopodia of macrophages, and thus resembled the localization pattern observed for overexpressed Myosin1f-GFP.
Although we were able to demonstrate that Myosin1f enhances cellular filopodia level, the underlying mechanism remains to be determined. Moreover, future studies are needed to elucidate the exact function of nuclear Myo1f.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

ACKNOWLEDGMENTS
We thank M. Gunzer who kindly provided THP-1 cells and M. Ehrmann who kindly provided SW480 cells. We acknowledge the use of the imaging equipment and the support in microscope usage and image analysis by the Imaging Center Campus Essen (ICCE), Center of Medical Biotechnology (ZMB), University of Duisburg-Essen, and especially thank J. Koch for his expertise. We thank A. Spierling for critically reading the manuscript. Moreover, we would like to thank the Institute of Transfusion Medicine of the University Hospital Essen for providing us with Buffy Coats. Parts of the data were generated during the doctoral dissertation of L. Kö nig (formerly L. Kunst) entitled ''Funktionelle Analyse der Protease Tasp1 und ihrem Zielprotein Myo1f''. This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) -project number 409946033.

METHOD DETAILS Cloning
Eukaryotic expression constructs encoding wild-type Tasp1 and a cleavage-deficient mutant C-terminally tagged with mCherry and HA were described previously (Bier et al., 2011a) as well as the vector pET-22b encoding Tasp1-His for bacterial expression . Plasmids pc3-Myo1f-GFP and pc3-Myo1f -245 AA 246 -GFP are based on a Myo1f coding sequence amplified from human head and neck tumor cDNA, and cloned into the expression vector pc3-GFP as described (Bier et al., 2011b).
The coding region of the N-terminal Tasp1 cleavage fragment, comprising amino acids 1 to 245, was amplified from plasmid pc3-Myo1f-GFP, as well as the C-terminal part of Myo1f-GFP downstream of the Tasp1 cleavage site, comprising amino acids 246 to 1098 (primer sequences are depicted in the key resources table). PCR products were cloned into the vector pcDNA3.1-GFP via KpnI/NheI. Both resulting constructs were additionally equipped with a N-terminal myc epitope tag by insertion of annealed oligonucleotides via KpnI as described (Bier et al., 2011b), resulting in the eukaryotic expression plasmids pc3-myc-Myo1f-Nterm-GFP and pc3-myc-Myo1f-C-term-GFP. All plasmids were finally confirmed by restriction enzyme digestion and sequencing. Plasmids pcDNA3-mCherry-Cdc42 wt and pcDNA3-mCherry-Cdc42 Q61L were derived from the corresponding GFP-tagged constructs described previously (Nalbant et al., 2004).

Cell fixation, microscopy and imaging
HEK293T-, HeLa-and SW480 cells were seeded in ibiTreat m-slide 8-well chamber slides (ibidiâ) and transfected with Lipofectamineâ 2000 or left untreated. After 24 h, cells were washed with PBS and fixed with Rotiâ-Histofix 4% (4% phosphate buffered formaldehyde solution; Roth) for 20 min at room temperature. Cells were washed with PBS and stained with Hoechst33342 (10 mg/mL in H 2 O, 1:1000) for 15 min in DPBS (1% BSA, 0.3% Triton) or with Phalloidin-Tetramethylrhodamine B (Sigma-Aldrich) for 45 min. Before incubation with the staining reagent, fixed cells were incubated with 0.1% Triton X-100 in PBS for 15 min to increase permeability.
Confocal laser scanning microscopy of fixed cells was performed on a TCS SP8X Falcon Confocal Microscope (Leica Microsystems) using the HC PL APO 633/1.2 W motCORR CS2 water-immersion objective. The laser lines of the white light laser used for excitation were 488 nm (EGFP) and 561 nm (mCherry and rhodamine). Fluorescence was detected using HyD detectors (490-540 nm) and (575-650 nm), respectively. Images were acquired using the ''LAS X'' software (Leica Microsystems) and assembled with ''Canvas 5'' (ACD Systems).
For live cell imaging HeLa cells were seeded on 35mm glass bottom dishes coated with fibronectin (10 mg/ mL, Corning) and transfected as mentioned above. Live cell TIRF and spinning disc confocal images were then acquired with an Eclipse Ti-E8Nikon) inverted microscope with TIRF illuminator Unit, an AOTC Laser Combiner, a laser dual spinning disc scan Head (CSU-X1: Yokogawa), and an iXon3 897 single-photondetection EMCCD camera. Laser lines 488 and 561 nm were used. Images were acquired using an Apo TIRF 603/1.49 NA oil immersion objective with EM gain of 50-100 and 1*1 binning. A CSU Quad Dichroic mirror set and TIRF Dual line Beamsplitter zt 488/561 rpc were used. Acquisition was controlled with Andor IQ Software and images were processed using ImageJ. iScience Article the populations of mean fluorescence intensity of both treated and untreated nuclei were fitted to a Gaussian frequency distribution in order to better visualize the difference between these populations. For every experiment n z 100 treated as well as untreated cells were captured and a total of four independent experiments were performed.
Tasp1 influence on Myo1f induced filopodia formation (Figure 4) was analyzed by confocal imaging of a sufficient number of HeLa cells as well as 293T cells (n > 80) expressing Myo1f-GFP or Myo1f-245 AA 246 -GFP together with or without Tasp1-mCherry. 3D z-stacks of whole cells were acquired and filopodia amount and length was analyzed by manual evaluation using Fiji tools. To ensure unbiased quantification, image file names were randomized using the ''Filename_randomizer.txt" macro for ImageJ.
Two classes of filopodia were defined by categorizing filopodia according to length: < 5 mm and > 5 mm. The amount of filopodia smaller than 5 mm seemed to be stable within each Myo1f-GFP expressing cell line and independent of Tasp1 co-expression. Therefore, filopodia longer than 5 mm were quantified and assessed by pairwise comparison via t-test among the cells expressing the same Myo1f variant with or without Tasp1 co-expression.

STATISTICAL ANALYSIS
Statistical analysis was performed with GraphPadPrism 9 software. Statistical relevance of the data was determined by applying unpaired Student's t-test (