Molecular insights into profilin 1-dependent regulation of cellular phosphatidylinositol (4,5)-bisphosphate

ABSTRACT

Phosphatidylinositol (4,5)-bisphosphate (PIP₂), the most abundant cellular poly-phosphoinositide (PPI) class of phospholipid, is a central plasma membrane (PM)-associated signaling hub that controls many cellular processes. In this study, we demonstrate that both deletion of the gene encoding actin-binding protein profilin 1 (Pfn1) and disruption of Pfn1–actin interaction leads to downregulation of PM PIP₂ content in cells. This is also phenocopied when F-actin is depolymerized, implying that Pfn1-dependent PIP₂ alteration is related to its actin-regulatory function. Phospholipase C (PLC) activity is crucial for Pfn1-deficient cells to exhibit the PIP₂-related phenotype. These findings, taken together with biochemical signatures of elevated PIP₂ hydrolysis (higher baseline PM diacylglycerol-to PIP₂ ratio and protein kinase C activity) exhibited by Pfn1-deficient cells, imply that PLC-mediated PIP₂ hydrolysis plays a role in Pfn1-dependent regulation of PM PIP₂. Furthermore, we unexpectedly found that Pfn1 loss leads to dramatic alterations in several other important forms of lipids, revealing a previously unrecognized role of Pfn1 as a broad regulator of cellular lipid environment that extends beyond PPI control. In conclusion, our study establishes Pfn1 as an important regulator of cellular lipid homeostasis.

Summary: Functional loss of profilin 1, a key regulator of the actin cytoskeleton, can trigger downregulation of plasma membrane content of PIP₂, an important class of phospholipid, in cells by promoting PLC-mediated PIP2 hydrolysis.

INTRODUCTION

Phosphatidylinositol (4,5)-bisphosphate [PI(4,5)P₂; hereafter referred to as PIP₂] is a poly-phosphoinositide (PPI) class phospholipid that plays a central role in the regulation of diverse cellular processes. Although it constitutes a minor component of the inner leaflet of the plasma membrane (PM), PIP₂ is the most abundant PPI and serves as a crucial signaling molecule and structural regulator in cells (reviewed in Czech, 2000; Dickson and Hille, 2019; Mandal, 2020). For example, its hydrolysis by phospholipases (PLCs) generates important second messengers, namely inositol trisphosphate (IP₃) and diacylglycerol (DAG), to promote the intracellular release of Ca²⁺ and activation of protein kinase C (PKC), respectively. Metabolic conversion of PIP₂ into D3-PPIs [e.g. PI(3,4,5)P₃ [PIP₃] and PI(3,4)P₂] and negative regulation of those pathways involving the actions of various site-specific kinases [e.g. phosphoinositide 3-kinase (PI3K)] and phosphatases [e.g. SH2-domain containing inositol phosphatase 2 (SHIP2), phosphatase and tension homolog deleted on chromosome 10 (PTEN) and inositol polyphosphate-4-phosphatse (INPP4)] also influences pathways of cell proliferation, survival, and metabolism. Additionally, PIP₂ directly interacts with a broad array of proteins involved in cytoskeletal remodeling (elaborated below), influencing a host of other processes including membrane trafficking (Thapa and Anderson, 2012), ion channel regulation (Harraz et al., 2020), and cell–cell and cell-extracellular matrix (ECM) adhesion (Akiyama et al., 2005; Chinthalapudi et al., 2014). Through these interactions, PIP₂ helps orchestrate dynamic cellular responses to environmental stimuli and maintains the spatial organization of signaling domains. Given its centrality in coordinating membrane-associated events, precise regulation of PIP₂ synthesis, localization and turnover is essential for normal cellular function and viability.

Many important actin-binding proteins (ABPs) are influenced by their direct interactions with PIP₂. For example, simultaneous binding of PIP₂ and the small GTPase Cdc42 to the Wiskott–Aldrich syndrome protein (WASP) and its homolog N-WASP induces conformational activation of these proteins (Higgs and Pollard, 2000). This activation promotes their interaction with the Arp2/3 complex, initiating actin polymerization. PIP₂ has also been shown to sequester cofilin proteins, ABPs responsible for actin filament disassembly, thereby inhibiting their depolymerizing activity and further enhancing actin filament stability (Zhao et al., 2010). Beyond direct modulation of actin filament dynamics, PIP₂ also influences cytoskeleton-associated signaling pathways. Notably, it releases the autoinhibited conformation of talin proteins, ABPs that links the actin cytoskeleton to the plasma membrane (PM), thereby facilitating integrin-mediated adhesion to the ECM (Ye et al., 2016). In a similar manner, PIP₂ can facilitate the interaction between the focal adhesion-associated protein vinculin and the Arp2/3 complex to initiate actin nucleation at the sites of nascent adhesions (Braun et al., 2021; James et al., 2025). These findings underscore a central role of PIP₂ in actin cytoskeletal regulation.

Profilins (Pfns) belong to a small family of G-actin-binding proteins. Pfns facilitate the nucleotide exchange from ADP- to-ATP-bound G-actin and deliver ATP-actin directly to the barbed ends of growing actin filaments. By virtue of their affinity for poly-L-proline (PLP) motifs, Pfns also interact with and delivers G-actin to several important PLP-domain-bearing actin assembly-inducing factors (example N-WASP, formins and Ena/VASP family proteins) to further enhance their ability to polymerize actin (Davey and Moens, 2020; Ding et al., 2012). These functions mean that Pfns are crucial regulators of actin cytoskeletal dynamics in cells. Pfns also interact with various PPIs (Lambrechts et al., 1997; Lassing and Lindberg, 1985) although the physiological significance of Pfn–PPI interaction has been largely understudied particularly in cellular settings. Lipid micelle- and/or model membrane-based studies have shown that profilin 1 (Pfn1; the most abundantly expressed cellular isoform of Pfn and the focus of the study) binds to PIP₂ as well as PPIs that are generated downstream of PIP₂ [such as PIP₃ and PI(3,4)P₂] (Lu et al., 1996; Sohn et al., 1995). As PIP₂ dissociates Pfn1 from G-actin, PIP₂ binding has been postulated as a potential mechanism of negatively regulating the sub-membranous Pfn1–actin interaction (Lassing and Lindberg, 1985); however, direct experimental evidence for this tenet in actual cells is still lacking in literature.

Complementing the widely accepted view of PIP₂ acting as a regulator of ABP functionality in cells, two previous Pfn1-related studies from our group have suggested that the reverse can be also true in a cellular setting. Specifically, we have demonstrated that acute Pfn1 knockdown leads to plasma membrane (PM) enrichment of PI(3,4)P₂ thereby influencing PI(3,4)P₂-dependent PM recruitment of lamellipodin–VASP complex and cell migration (Bae et al., 2010), providing evidence for Pfn1-dependent modulation of cell behavior via PM PPI control. In a recent study we further showed that EGF-induced production of PIP₃ [the PPI direct upstream of PI(3,4)P₂] at the PM as well as the PM residence time of SHIP2 [the enzyme responsible for the metabolic conversion of PIP₃ into PI(3,4)P₂] in cells are enhanced in acute Pfn1 knockdown condition (Ricci et al., 2024). Although PIP₂ serves as the immediate precursor PPI for PIP₃ generation, our overexpression and knockdown experiments reported in that study collectively supported a premise that Pfn1 positively influences the basal cellular PIP₂ content, and this is a generalizable feature across different cell types. The goal of the present study was to gain mechanistic insights into Pfn1-dependent regulation of cellular PIP₂ content. This study provides the first direct evidence for PLC-mediated PIP₂ hydrolysis as a mechanism for Pfn1-dependent regulation of PM PIP₂ content in a cellular setting. However, Pfn1-dependent changes in PIP₂ are driven by its role in actin dynamics rather than through direct PPI interaction. Our studies also unexpectedly identify Pfn1 as a broad regulator of cellular lipid environment that extends beyond PPI control.

RESULTS

Pfn1 depletion-induced PIP₂ reduction is not an acute cellular response

In our previous study (Ricci et al., 2024), we showed that transient knockdown of Pfn1 expression in several different types of human cells [e.g. breast cancer cell lines MDA-MB-231 (MDA-231) and BT-474; cervical cancer cell line HeLa] led to an appreciable (35–50% depending on the cell type) reduction in PIP₂ at the PM. Biochemical analyses of bulk lipids performed in MDA-231, BT-474 and MCF10A (a transformed mammary epithelial cell line) cells further indicated that total cellular PIP₂ content is reduced upon transient Pfn1 loss. Conversely, we reported that overexpression of Pfn1 (either transiently or in a stable manner) leads to PIP₂ enrichment. Given these findings, we first asked whether PM PIP₂ reduction is maintained or somehow adaptively restored under complete and/or prolonged loss of Pfn1. To address this question, we performed immunofluorescence-based assessment of PM PIP₂ in MDA-231 as well as B16F1 (a mouse melanoma cell line) cells that are engineered for CRISPR knockout (KO) of the Pfn1 gene versus their respective control counterparts. Note that Pfn1-knockout (KO) variant of B16F1 cells has permanent loss of Pfn1 expression serving us as a model system to study the impact of prolonged Pfn1 depletion on PIP₂. In contrast, Cas9 expression in MDA-231 cells is doxycycline (dox)-inducible, which allowed us to study the acute impact (5–6 days post-triggering of Cas9 induction) of complete loss of Pfn1 expression on PIP₂. Immunoblot data shown in Fig. 1A demonstrates the loss of Pfn1 expression in each of these two cell lines. As previously done in other studies (Hammond et al., 2006, 2009), for PM PIP₂ analysis, we specifically quantified the total rather than the average PM PIP₂ staining intensity for three reasons. First, PIP₂ is non-uniformly distributed across the PM, and therefore the average intensity calculation collapses a lot of biologically meaningful spatial information. Second, the average intensity calculation is impacted by significant cell shape and area differences that exist between cells within a group as well as between groups. Third, the integrated PM intensity is a better metric of how much total PIP₂ is available for metabolic turnover on a cell-by-cell basis. Our PIP₂ immunostaining experiments revealed that Pfn1 KO variants of both cell lines exhibited significantly lower total PM PIP₂ content versus their respective control counterparts (Fig. 1B,C). The average reductions in total PM PIP₂ staining intensity in MDA-231 and B16F1 cells resulting from Pfn1 loss were equal to 45% and 20%, respectively (we offer potential explanations for quantitative differences in Pfn1-dependent PIP₂ changes between the two cell lines in the Discussion section). These data demonstrate that PM PIP₂ reduction is maintained even under complete and/or prolonged loss of Pfn1 expression.

Fig. 1.

Pfn1 depletion-induced PIP₂ reduction is not an acute cellular response. (A) Representative Pfn1 immunoblots of cellular lysates from control and Pfn1 KO variants of B16F1 MDA 231 cells (GAPDH blot and Coomassie stained gels serve as loading controls). Blot representative of three repeats. (B) Representative confocal images of control and Pfn1 KO variants of B16F1 (60× objective) and MDA-231 (20× objective) cells immunostained for PIP₂ (magnified images of selected regions are shown alongside). Scale bars: 100 μm (upper row) and 40 μm (lower row). (C) Box-and-whisker plots summarizing the total PM PIP₂ staining of Pfn1 KO cells relative to their control counterparts for each cell line. Each box represents the data falling within the 25th to 75th percentile range, and a line denoting the median value of the data set. Whiskers represent 1.5× the interquartile range above and below the 25th and 75th percentiles. The data sets represent analyses of 200–300 and ∼400 cells pooled from three independent experiments for B16F1 and MDA-231 cell lines, respectively. *P<0.05; **P<0.001 (two-tailed unpaired t-test).

Disruption of actin-binding of Pfn1 phenocopies PIP₂ phenotype of Pfn1-deficient cells

Next, to determine which functionality of Pfn1 is linked to its regulation of cellular PIP₂, we performed functional rescue experiments where Pfn1 KO B16F1 cells were transfected with plasmids encoding Myc-tagged versions of either wild-type (WT) Pfn1 or various ligand binding-deficient point-mutants of Pfn1 before performing PIP₂ immunostaining. These mutants include: H119E-Pfn1 (∼98% reduction in actin-binding), H133S-Pfn1 (deficient in PLP-binding), and R88L-Pfn1 (deficient in PPI-binding). Note that owing to structural overlap between the actin- and PPI-binding regions of Pfn1, R88L-Pfn1 has a 50% reduction in actin-binding (Bae et al., 2010). Although PIP₂-binding of WT versus H119E-Pfn1 has never been directly quantified in biochemical assays, we have previously shown that the H119E substitution does not affect the membrane fraction of ectopically overexpressed Pfn1 in cells (Bae et al., 2010). Furthermore, in purified protein–lipid vesicle mixture settings, the analogous mutant H119D-Pfn1 inhibits PLCγ-mediated PIP₂ hydrolysis as efficiently as WT-Pfn1 (Sohn et al., 1995). These findings suggest that H119D/E-Pfn1 retains intact membrane PPI binding. All four rescue Pfn1 constructs were expressed in an internal ribosomal entry site (IRES)-GFP backbone vector that allowed identification of rescued cells by the presence of GFP fluorescence. As additional transfection controls, both control and Pfn1 KO cells were transfected with an IRES-GFP backbone vector. Surprisingly, PIP₂ immunostaining revealed that PM PIP₂ reduction in Pfn1-deficient cells is reversible by all Pfn1 rescue constructs except the actin-binding deficient mutant H119E-Pfn1 (Fig. 2A,B). Given that GFP and Pfn1 rescue constructs are linked by an IRES, we analyzed the GFP fluorescence intensity of cells selected for PIP₂ analyses as a surrogate measure for comparing the relative expressions of Pfn1 rescue constructs across the various groups. As per these analyses, the average GFP expression of cells chosen for PIP₂ analyses was found to be comparable between the various Pfn1 KO-rescue groups (Fig. 2C). Therefore, we argue that our observed phenotypic differences related to PIP₂ are not confounded by the expressions of various Pfn1 rescue constructs. Based on these results, we conclude that Pfn1-dependent changes in PIP₂ are primarily attributed to its actin-related function.

Fig. 2.

Disruption of actin-binding of Pfn1 phenocopies the PIP₂ phenotype of Pfn1-deficient cells. (A) Representative 20× imaging of PIP₂ immunostaining (red) of control and Pfn1 KO B16F1 cells with or without functional rescue by wild-type (WT) and the indicated mutants of Pfn1 GFP-positive cells indicate the transfected cells; empty vector GFP transfection serves as the control. Magnified PIP₂ staining images of selected GFP+ cells in each group are shown underneath. Scale bars: 100 μm. (B) Quantification of PM PIP₂ intensity in GFP-positive cells for various groups relative to control (control cells transfected with GFP). Each datapoint represents an independent experiment; a total of ∼300–450 cells were quantified pooled from three independent experiments. (C) Quantification of relative GFP expression for the various rescue groups (each data point is the mean readout of all GFP+ cells for a given experiment). ***P<0.001; NS, not significant (one-way ANOVA with Tukey HSD post-hoc test for multiple comparisons).

As Pfn1 loss generally leads to a reduced level of polymerized actin in cells including B16F1 cells (Fig. 3A,B), we next asked whether acute actin depolymerization could invoke a similar PIP₂ loss in cells. To test this possibility, we performed PIP₂ immunostaining of B16F1 cells subjected to acute treatment (5 min) of actin depolymerizing drug Latrunculin-B (LatB) versus DMSO (vehicle control). Cells were also counterstained with phalloidin to assess the changes in F-actin upon LatB treatment. Our experiments showed that global actin depolymerization by LatB treatment was also accompanied by a significant reduction in PM PIP₂ content of B16F1 cells (Fig. 3C,D). To further determine whether there is any temporal correlation between LatB-induced F-actin disruption and PIP₂ attenuation, we performed orthogonal live-cell imaging of HEK-293 cells co-transfected with PIP₂ biosensor GFP–PH-PLCδ along with a PM marker (iRFP–Lyn11) and FastActX (a probe for F-actin) before and after acute treatment of LatB (DMSO treatment served as control). Consistent with our observations in B16F1 cells, LatB treatment resulted in a rapid decrease in the PM-to-cytosolic fluorescence ratio (indicative of PM PIP₂ reduction) in HEK-293 cells that remarkably synchronized with actin depolymerization marked by the loss of F-actin fluorescence intensity (Fig. 3E–G). On a kinetic scale, LatB-induced PIP₂ decline occurred slightly slower and persisted longer than the F-actin response, with the mean half-lives of PIP₂ and F-actin changes equal to approximately 2.2 min and 1 min, respectively (Fig. 3H). In a separate experiment, we studied the impact of short-term (30 min) treatment of non-muscle myosin II inhibitor blebbistatin (versus DMSO as vehicle control) on PM PIP₂ content in MDA-231b cells (Fig. S1). The major effects of blebbistatin on actin cytoskeleton are disintegration of actin stress fibers, softening of cortical actin and transformation of lamellipodial actin into a loose network of accumulated amorphous actin structures that correspond to membrane ruffles (Shutova et al., 2012). These phenotypes were also recapitulated in our experimental settings (Fig. S1A). In general, blebbistatin-treated cells exhibited protrusive structures in random directions with PIP₂ enrichment in peripheral F-actin-rich regions (consistent with the LatB experimental data) and a trend toward higher (P=0.09) overall cell edge PIP₂ staining versus vehicle-treated cells, further underscoring a prominent impact of actin cytoskeletal perturbation on PM PIP₂ (Fig. S1B,C). Collectively, the findings from our Pfn1 KO-rescue and various actin-related pharmacological studies suggest that Pfn1-dependent PIP₂ alteration is likely a secondary consequence of peripheral actin cytoskeletal disturbances in cells.

Fig. 3.

Actin depolymerization triggers rapid loss of PM PIP₂. (A,B) Representative 60× fluorescence images (A) of control and or Pfn1 KO B16F1 cells stained with phalloidin. (B) The associated quantification with each dataset representing quantification of ∼45 fields pooled from three independent experiments. Scale bar: 20 μm. (C,D) Representative 20X fluorescence images (C) of B16F1 cells stained with phalloidin and anti-PIP₂ antibodies in DMSO (control) versus LatB-treated conditions (magnified images of PIP₂ staining of selected regions of interest are shown below). (D) Box-and-whisker plot in summarizing the LatB-induced changes in PM PIP₂ intensity (∼500 cells were quantified per treatment group pooled from three independent experiments). In B,D, the box represents the data falling within the 25th to 75th percentile range, and a line denoting the median value of the data set. Whiskers represent 1.5× the interquartile range above and below the 25th and 75th percentiles. *P<0.05 (two-tailed unpaired t-test). (E–H) Representative confocal fluorescence time-lapse images of HEK-293 cells expressing GFP–PH-PLCδ and treated with FastActX before and after LatB treatment (E). (F,G) Summary of temporal changes of F-actin (by FastActX fluorescence) and PM PIP₂ (by PM/cytoplasmic fluorescence of PH-PLCδ) in response to LatB versus DMSO as control [traces represent fluorescence mean±s.e.m. of all cells (DMSO, 50 cells; LatB, 45 cells) pooled from three individual experiments]. (H) The relative half-times of the F-actin and PIP₂ response curves (mean±95% c.i.).

PIP₂ attenuation upon loss of Pfn1 is reversible by PLC inhibition

Based on our observation that PM PIP₂ loss ensues rapidly (within minutes) in response to triggering actin depolymerization, we speculated that Pfn1-dependent changes are PIP₂ could be somehow related to perturbation of its turnover rather than its synthesis pathway. PIP₂ is turned over by two major pathways. One of these pathways involves PI3K-mediated conversion of PIP₂ into PIP₃ [followed by sequential conversion of PIP₃ into PI(3,4)P₂]. Our previous studies have provided evidence for enhanced EGF-induced production of D3-PPIs [i.e. PIP₃ and PI(3,4)P₂] in MDA-231 and HeLa cells upon Pfn1 depletion (Bae et al., 2010; Ricci et al., 2024). However, these D3-PPIs are transiently generated in response to acute activation of receptor tyrosine kinases (RTKs; EGFR, PDGFR and IGFR) and only a small fraction of the PM pool of PIP₂ is converted into PIP₃. Therefore, we reasoned that increased conversion of PIP₂ into PIP₃ is unlikely to explain the ∼40% decrease in the basal PIP₂ level observed in those cell lines upon Pfn1 depletion in steady-state serum-containing culture. A second pathway of PIP₂ turnover involves hydrolysis of PIP₂ into DAG and IP₃ by the action of PLC class of enzymes. Therefore, to test whether pharmacological PLC inhibition has any effect on Pfn1-dependent changes in the PM PIP₂ content, we performed PIP₂ staining of control and Pfn1 KO MDA-231 cells following overnight treatment with either U73122 (a broad-spectrum PLC inhibitor; Selleck Chem, #S8011) or its inactive analog U73343 (a negative control). As expected, in the inactive compound (U73343) treatment setting, Pfn1 KO cells exhibited lower PM PIP₂ content compared with their control counterparts. However, when treated with the active PLC inhibitor U73122, the total PM PIP₂ content was found to be conspicuously elevated in Pfn1 KO cells abolishing the PIP₂ differential between the two groups of cells (Fig. 4A,B). The two major classes of PLC enzymes that hydrolyze PIP₂ are PLCγ (activated downstream of RTKs) and PLCβ [activated downstream of G protein-coupled receptors (GPCRs)]. In model membrane studies utilizing purified proteins (Pfn1 and PLCγ), Goldschmidt-Clermont and colleagues have previously demonstrated that Pfn1 inhibits PLCγ-mediated hydrolysis of PIP₂ but this inhibitory action of Pfn1 on PIP₂ hydrolysis is abolished when PLCγ is tyrosine-phosphorylated (mimicking the activated state of PLCγ in response to RTK activation) (Goldschmidt-Clermont et al., 1991, 1990). In fact, in our previous study, we also failed to see any effect of Pfn1 depletion on the production of IP1 (an immediate metabolic byproduct of IP₃) in MDA-231 cells in an acute EGF stimulation setting (Ricci et al., 2024) suggesting that Pfn1 is unlikely to impact PIP₂ hydrolysis through RTK-activated PLCγ. Given that PLCβ activity is also stimulated by active ingredients in serum (e.g. various bioactive lipids, hormones and peptides), we next examined the effect of silencing the expression of PLCβ3 (the dominant isoform of PLCβ in MDA-231 cells) on PIP₂ content in control versus Pfn1-deficient MDA-231 cells. Consistent with our pharmacological inhibition data, these experiments also revealed that silencing PLCβ3 expression increased PIP₂ selectively in Pfn1-deficient cells, mitigating the PIP₂ differential between the two cell types (Fig. 4C–E). Collectively, these findings demonstrate that intact PLC activity is crucial for Pfn1-deficient cells to exhibit the PIP₂-related phenotype.

Fig. 4.

PIP₂ attenuation upon loss of Pfn1 is reversible by PLC inhibition. (A,B) Representative PIP₂ staining images (A; 40×) of control and Pfn1 KO MDA 231 cells, treated with either inactive (U73343) or active (U73122) pan-PLC inhibitor drug (magnified images of PIP₂ staining of selected regions of interest are shown alongside). (B) The associated PM PIP₂ quantification for the various groups (all data normalized to the average readout of control cells treated with the U73343 compound). A total of ∼450 cells/group were analyzed from three independent experiments. Scale bar: 40 μm. (C–E) Representative PLCβ3 and Pfn1 immunoblots of lysates of MDA-231 cells following transient transfection of the indicated siRNAs (non-targeting control siRNA transfection served as control); GAPDH blot serves as the loading control (C). Representative PIP₂ staining images fluorescence images (40×) of MDA-231 cells transfected with the indicated siRNAs (magnified images of PIP₂ staining of selected regions of interest of each group are shown underneath) (D). Scale bar: 80 μm. (E) Associated quantification to compare the PM PIP₂ intensity between the various groups of cells based on the results of three independent experiments. In B,E, the box represents the data falling within the 25th to 75th percentile range, and a line denoting the median value of the data set. Whiskers represent 1.5× the interquartile range above and below the 25th and 75th percentiles. *P<0.05; ***P<0.001 (one-way ANOVA with Tukey HSD post-hoc test for multiple comparisons).

Pfn1-deficient cells exhibit signatures of elevated PIP₂ hydrolysis

The results of our PLC inhibition studies can be interpreted in at least two different ways. One possible scenario is that the presence of Pfn1 somehow acts as a brake on PLC-mediated PIP₂ hydrolysis, and this inhibition is lost in Pfn1-deficient cells resulting in PM PIP₂ loss through accelerated PIP₂ hydrolysis. Alternatively (but not necessarily in a mutually exclusively manner), the signaling downstream of activated PLC could somehow be indirectly responsible for Pfn1-dependent modulation of PIP₂. We asked whether Pfn1-depleted cells bear any of the molecular signatures of increased PIP₂ hydrolysis. To address this, we first performed quantitative mass-spectrometry (MS)-based lipidomic analyses of control versus Pfn1 KO MDA-231 cells which revealed several interesting findings. First, the total lipid content (normalized to the total protein content) of Pfn1 KO cells was found to be ∼10% lower than control cells, and this difference was found to be significantly different (Fig. 5A). Second, analyses of differentially abundant lipids (out of a total of ∼900 lipids) revealed that 89 and 104 lipids were upregulated and downregulated, respectively, in Pfn1 KO cells relative to their control counterparts (Fig. 5B), suggesting that Pfn1 loss has a global impact on lipid profile that extends beyond PPI control. Differentially abundant lipids categorized in major classes are shown as a heat-plot in Fig. 5C. Major lipids that were altered upon loss of Pfn1 expression in a statistically significant manner include phosphatidylcholine (PC; reduced by ∼25%), cholesterol (reduced by ∼17%), free fatty acid (FFA; reduced by ∼26%), sphingomyelin (reduced by ∼23%) and triglycerides (TGs: increased by ∼12%) (Fig. 5D,E). Although the lipidomic panel utilized in our studies did not probe for PPIs, the PI content was found to be comparable between control and Pfn1 KO cells. Third, in four out of five experiments, we saw a general trend of elevated DAG (a direct hydrolysis product of PIP₂) content in Pfn1 KO samples; however, the large experiment-to-experiment variability in the absolute content as well as Pfn1-dependent changes in DAG precluded us from achieving statistical significance between the two groups. The large variability in the measured DAG content in our experiments is not totally surprising as cellular DAG level is known to fluctuate with growth and/or be impacted by unintended changes in the chemical parameters of culture conditions (Mondal et al., 2024; Van Veldhoven and Bell, 1988). Phosphatidic acid (PA) is a lipid byproduct of DAG that is generated by direct conversion of DAG by DAG kinase (DGK) to replenish the pool of cellular PIP₂ through a feedback signal. Encouragingly, we observed a near-statistically significant (P=0.06) increase in the absolute content of PA in Pfn1 KO cells relative to their control counterparts (Fig. 5D,E). Note that although PA can be also generated by phospholipase-D (PLD)-mediated conversion of PC (a lipid that is also decreased dramatically in Pfn1 KO cells), it is highly unlikely that PA increase in Pfn1-deficient cells is reflective of increased PLD-mediated conversion of PC for two reasons. First, the observed magnitude of change in the content of PA (∼3000 pmol/mg increase) was disproportionately lower than that of PC (>200,000 pmol/mg decrease) in response to Pfn1 KO. Second, monomeric actin directly binds to and inhibits the activity of PLD (Lee et al., 2001). Therefore, the increased G-to-F-actin ratio in Pfn1-deficient cells, if it had any effect at all, would be expected to result in diminished PLD activity and PLD-mediated conversion of PC into PA.

Fig. 5.

Pfn1 loss has a broad impact on cellular lipids. (A) Quantification of the total lipid content (in pmol/mg of total protein) based on mass-spectrometry based on lipidomic analyses of control and Pfn1 KO MDA-231 cells (five biological replicates/group with >10⁶ cells per sample were analyzed). (B) Volcano plot represents the number of differentially abundant lipid moieties measured in the lipidomic panel. (C) Heat-plot showing directional alterations in various lipid classes upon loss of Pfn1 in MDA-231 cells. (D) Quantitative comparison of the abundance (in pmlo/mg of total protein) of the indicated lipids between control and Pfn1 KO cells (PC, phosphatidylcholine; CHOL, cholesterol; FFA, free fatty acid, SM, sphingomyelin; TG, triglyceride; PI, phosphatidylinositol; PA, phosphatidic acid; DAG, diacylglycerol; lipids in red and blue exhibited trends in up- and down-regulation, respectively, upon Pfn1 KO). n=5. (E) Line plots showing percentage changes in the indicated lipids in Pfn1 KO relative to control cultures in all five biological replicates (note: near identical percentage changes in certain biological replicates led to overlapping line plots for some lipids).

Although the general trends of elevated DAG and PA in Pfn1 KO cells are conceptually consistent with a scenario of increased PIP₂ hydrolysis, a key limitation of whole-cell lipidomic analysis is that it fails to distinguish lipid changes in specific subcellular locations and compartments. This is particularly a relevant issue for DAG since: (1) the largest pool of DAG is in the endoplasmic reticulum (ER) and Golgi, (2) DAG is under constant metabolic flux, and (3) importantly, subtle changes in DAG resulting from PM PIP₂ hydrolysis might be difficult to capture by whole cell lipidomic measurements. To circumvent these issues, we performed a series of additional studies to gain further biological insights into the effect of Pfn1 on PIP₂ hydrolysis as described in the following sections.

First, a key downstream signature of PLC-mediated PIP₂ hydrolysis is DAG-mediated activation of serine-threonine kinase protein kinase C (PKC). Therefore, to assess Pfn1-dependent changes in PKC activation, we performed immunoblot analyses of total cell extracts prepared from control versus Pfn1-deficient MDA-231 cells with a phospho-PKC (pPKC) substrate antibody that recognizes all PKC-phosphorylated proteins. For these studies, cells were either serum-starved (representing the unstimulated ‘baseline’ condition) or subjected to acute stimulation with lysophosphatidic acid (LPA), a bioactive lipid and an active ingredient of serum that stimulates PLCβ activity via GPCR activation. The baseline (i.e. in an unstimulated condition) PKC activity signature (measured by the intensity of pPKC-substrate immunoreactive bands) was clearly more pronounced in Pfn1-depleted MDA-231 versus their control counterparts (Fig. 6A,B), a finding that was also reproducible in B16F1 cells (Fig. 6C,D). When we compared LPA-induced increase in the PKC activity readout relative to the baseline condition, the fold-change values (control cells, ∼3-fold increase from baseline; Pfn1-deficient cells, ∼1.3-fold increase from baseline) implied that Pfn1 loss did not necessarily augment LPA-induced PKC activation response. A similar trend was observed when we analyzed LPA-induced Ca²⁺ response (another event downstream of PLC-mediated PIP₂ hydrolysis triggered by IP₃) by live-cell fluorescence imaging of GCaMP (a Ca²⁺ biosensor) transfected in MDA-231 cells (data summarized in Fig. 6E–I). Both groups of cells exhibited a rapid LPA-induced Ca²⁺ spike (reached its peak at ∼30 s after stimulation) followed by a rapid decline reaching a post-stimulation resting level. To account for cell-to-cell variation in the actual expression of the biosensor, we baseline-corrected GCaMP fluorescence by normalizing each kinetic datapoint readout to the average pre-stimulation value on a cell-by-cell basis, for calculating the peak amplitude, integrated Ca²⁺ signal (area under the curve) and the post-stimulation resting value for each of the two groups. As per these analyses, we did not see any significant difference in either the peak amplitude or integrated Ca²⁺ signal between the control and Pfn1-knockdown groups, further underscoring the fact that Pfn1 loss does not necessarily confer cells an increased ability to respond to agonists (i.e. LPA-induced GPCR activation in this specific case). However, we noted that the post-stimulation resting Ca²⁺ signal was elevated in Pfn1-deficient cells relative to control cells (P<0.01), a feature that could result from increased basal PIP₂ hydrolysis and/or reduced re-uptake of cytosolic Ca²⁺ by ER and/or reduced efficiency of Ca²⁺ export.

Fig. 6.

Supportive evidence for Pfn1-dependent changes in PIP₂ hydrolysis signature. (A,B) Representative phospho-PKC (p-PKC) substrate and Pfn1 immunoblots of extracts prepared from control and Pfn1-depleted MDA-231 cells (A) following treatment with either 10 μM LPA or vehicle control (unstimulated) for 5 min; the stain-free gel image serves as the loading control (quantification of immunoblot results are shown in B; mean±s.d.; n=3 experiments). *P<0.05; **P<0.01; NS, not significant (one-way ANOVA with Tukey HSD). (C,D) Representative baseline p-PKC substrate immunoblot of control versus Pfn1 KO B16F1 cultures (C) and the associated quantification (D; mean±s.d.; n=3 experiments). *P<0.05 (one-sample t-test to compare relative to control). (E–I) Representative live fluorescence images (E) of MDA 231 cells with or without siRNA-mediated depletion of Pfn1, transfected with the Ca²⁺ ion probe GCaMP, treated with 10 μM of LPA over 8 min with 2 min of baseline imaging and 6 min of imaging post treatment (stimulation at t=120 s). Images depict t=0, 60, 120, 150, 180, 210, 300, and 450 s. Scale bar: 20 μm. (F) Quantification of average baseline-corrected GCaMP integrated fluorescence intensity over time normalized to the average of 2 min pre-stimulation of the treatment groups. Mean±s.d., n=18 cells. G–I show the comparisons of peak Ca²⁺ indicator intensity, Ca²⁺ spike area and the post-stimulation intensity (relative to the pre-stimulation value) between the two groups. The dataset represents analyses of 18 cells/group pooled from three independent experiments; mean±s.d. **P<0.01; ns, not significant (two-tailed unpaired t-test). (J–M) Lipid biosensor studies demonstrating elevated DAG:PIP₂ levels in HEK-293 cells in response to Pfn1 knockdown (KD). J shows representative confocal images of HEK-293 cells expressing Tubby_c-EGFP PIP₂ and mCherry-C1ab-PKD1 DAG biosensors, after treatment with non-targeting (ctrl) or Pfn1-directed siRNA pools. Scale bar: 10 μm. (K–M) Quantification of the PM to cytosolic intensity of the DAG and PI(4,5)P₂ biosensors, as well as the ratio of the two measurements. The box represents the data falling within the 25th to 75th percentile range, and a line denoting the median value of the data set. Whiskers represent 10th–90th percentiles. n=76 (ctrl) or 74 (Pfn1 KD) cells from three independent experiments (P-values show the results of Welch's t-test).

Second, we performed lipid biosensor studies in HEK-293 cells to assess the impact of Pfn1 knockdown on the relative DAG-to-PIP₂ content at the PM as a measure of intrinsic PIP₂ hydrolysis efficiency (data summarized in Fig. 6J–M). For these studies, HEK-293 cells were co-transfected with GFP–PH-Tubby_C (a C-terminal fragment of Tubby protein that binds to PM PIP₂ and, therefore, acts as a PIP₂ biosensor) and mCherry–C1ab (a construct with tandem DAG-binding C1A–C1B domains of PKD1 that serves as a DAG biosensor) along with a PM marker (iRFP–Lyn11). We first confirmed that Pfn1 depletion by siRNA transfection did not significantly change the PM PI4P (the direct precursor PPI for PIP₂ biosynthesis) content in HEK-293 cells (Fig. S2), mirroring our previous findings in HeLa cells (Ricci et al., 2024). However, it resulted in reduced PM PIP₂ content as indicated by a lower PM-to-cytoplasmic fluorescence ratio of GFP–Tubby_C in Pfn1 knockdown vs control cells (Fig. 6L). The absolute PM-to-cytosolic fluorescence ratio of the DAG reporter was slightly higher in Pfn1-knockdown cells than control cells, but this was not statistically significant (Fig. 6K). However, the difference was more pronounced and statistically significant when these DAG biosensor readouts were normalized to the PM-to-cytosolic ratio of PIP₂ biosensor readouts on a cell-by-cell basis before comparison (Fig. 6L). These data indicate that there is a higher DAG-to-PIP₂ ratio (a measure of intrinsic PIP₂ hydrolysis efficiency) at the PM in Pfn1-knockdown cells.

Third, we performed gene set enrichment analyses of bulk transcriptomic data (summarized in Fig. S3) of control versus Pfn1 KO MDA-231 cells. These analyses predicted enrichment of cytosolic IP₃ and/or IP₄ (downstream metabolites of PIP₂ hydrolysis) synthesis-related genes along with transcriptional upregulation of two PLC enzymes (PLCβ1 and PLCβ4) by ∼4-fold in Pfn1 KO cells. Collectively, these findings support the idea that Pfn1 has a negative impact on PIP₂ hydrolysis at the PM.

DISCUSSION

In our previous study, we first reported the phenomenon that perturbing Pfn1 expression has a profound consequence on the cellular PIP₂ content; however, the underlying mechanism remained unclear. The present study reports several new findings that shed novel biological insights into how the presence of Pfn1 might positively impact the PM PIP₂ content. First, we reveal that complete loss of Pfn1, either triggered acutely (as in MDA-231 cells) or in a prolonged manner (as in B16F1 cells) also leads to reduced PIP₂ level at the PM, recapitulating our previous observations in transient knockdown settings in various cell types. This suggests that Pfn1-dependent change in PIP₂ is not a transient cellular response that can be compensated for by other mechanisms in the long term. Second, based on findings from in vitro studies utilizing purified proteins (Pfn1 and PLC) and PPI-containing model membrane (Goldschmidt-Clermont et al., 1991, 1990), it has been previously postulated that Pfn1 can protect PIP₂ from PLC-mediated hydrolysis through its direct binding to PIP₂. However, this postulate was never directly tested in a cellular setting. Notably, our study unexpectedly reveals that Pfn1-dependent changes in PM PIP₂ are linked to its actin-regulatory function, highlighting that in vitro systems do not adequately recapitulate the complex cellular context in which Pfn1 operates. Third, we demonstrate that Pfn1-deficient cells exhibit biochemical signatures of elevated PIP₂ hydrolysis including higher baseline PM DAG-to PIP₂ ratio and protein kinase C activity, and that PLC activity is essential for Pfn1-dependent changes in PM PIP₂ content, supporting the idea that Pfn1 regulates PIP₂ via impacting its hydrolysis in a cellular context. Fourth, we demonstrate for the first time that Pfn1 loss can have a major influence on cellular lipids that extend beyond PPIs.

Between the two cell lines we studied herein, Pfn1-dependent changes in PM PIP₂ content were more prominent in MDA-231 cells (∼45%) relative to B16F1 (∼20%) cells. There could be at least two possible explanations for this observation. First, our MDA-231 studies were performed in an acute KO setting, and this contrasts the experimental setting of B16F1 cells with prolonged Pfn1 loss where one cannot rule out the possibility of compensatory mechanisms. Second, we have confirmed that B16F1 cells have appreciably higher expression of Pfn2 (a minor isoform of Pfn) than MDA-231 cells (data not shown). Therefore, it is possible that increased presence of Pfn2 might partly offset the phenotypic changes associated with loss of Pfn1 in B16F1 cells.

Our conclusion that the actin binding of Pfn1 rather than its PPI binding that is somehow responsible for Pfn1-dependent modulation of PIP₂ is derived from functional rescue studies, where re-expression of all Pfn1 constructs except the actin-binding deficient mutant (H119E-Pfn1) rescued the PIP₂ phenotype of Pfn1 KO cells. This was further supported by our LatB studies, which revealed that F-actin loss (a natural consequence of loss of Pfn1) leads to rapid PIP₂ loss at the PM in a temporally synchronized manner. Consistent with these results, we further showed that blebbistatin treatment resulted in accumulation of F-actin patches in cells and PIP₂ enrichment in peripheral F-actin-rich regions. Together these findings demonstrate that PM PIP₂ is significantly impacted by perturbation of the actin cytoskeleton. However, given prior demonstration of the ability of Pfn1 to protect PIP₂ from PLC-mediated hydrolysis in model membrane settings (Goldschmidt-Clermont et al., 1991, 1990), this begs a question as to whether we can completely disregard any contribution of the ability of Pfn1 to modulate PIP₂ by direct protein–lipid interaction in cells. First, it is noteworthy that in our previous study, we failed to see Pfn1 recruitment to the PM even when we artificially increased PIP₂ synthesis at the PM by a chemical genetics strategy (Ricci et al., 2024). Given that Pfn1 only displays transient, low-affinity interactions with PIP₂-rich model membranes (Senju et al., 2017), we could not completely rule out technical limitations in capturing low abundance or low affinity interactions between Pfn1 and PIP₂ at the PM in cells. Second, Pfn1 binds to actin and PPIs in a mutually exclusive manner. At least in a model membrane setting, Pfn1 can induce PIP₂ clustering causing destabilization and deformation of the membrane (Krishnan et al., 2009). Therefore, one can logically argue that in a KO-rescue setting, H119E-Pfn1 has more frequent interactions with membrane PIP₂ than either the WT or other mutant forms of Pfn1. As a result, local membrane deformation might render PIP₂ more hydrolysis-prone partly contributing to reduced PM PIP₂ content in the H119E-Pfn1-expressing cells. However, in our previous knockdown-rescue studies, we did not see any evidence for increased membrane content of H119E-Pfn1 mutant in MDA-231 cells (Bae et al., 2010). Therefore, the substantially diminished PM PIP₂ content (in the range of 25–50% depending on the cell type) of Pfn1-deficient cells is very unlikely to be explained by the loss of low abundance or low affinity Pfn1–PIP₂ interaction, prompting us to speculate that the contribution of the ability of Pfn1 to modulate PIP₂ by direct protein–lipid interaction in cells, if at all, might be insignificant.

Although our whole-cell lipidomic studies revealed trends of elevated DAG and PA in Pfn1-deficient cells, because of greatest abundance of DAG in ER/Golgi compartments and metabolic fluxing of DAG to other lipids (TG and PA), those findings failed to provide granularity on whether the absence of Pfn1 specifically alters PM PIP₂ hydrolysis. However, our proposed tenet of Pfn1-mediated regulation of PIP₂ via modulation of PIP₂ hydrolysis is supported by multiple orthogonal lines of evidence. First, our lipid biosensor studies revealed a greater PM DAG-to-PIP₂ ratio (a measure of the intrinsic PIP₂ hydrolysis efficiency) in Pfn1-deficient cells. Second, GSEA of transcriptomic data predicted enrichment of cytosolic IP₃ and IP₄ synthesis pathways in Pfn1-deficient cells. Third, cells lacking Pfn1 exhibited increased baseline PKC activity. Fourth, PLC inhibition resulted in reversal of Pfn1-dependent changes in PIP₂. Two previous studies have demonstrated that actin filament disruption increases PM mobility of PIP₂ (Andrade et al., 2015; Cho et al., 2005). There is also evidence for actin depolymerization-induced uncaging of PLC from the cortical actin network (Huang and Crain, 2009). Therefore, it is possible that Pfn1 deficiency leads to loosening of cortical actin network leading to increased PM diffusion of PIP₂ and/or uncaging of PLC, and this results in more frequent PLC–PIP₂ interaction and higher baseline PIP₂ hydrolysis, as schematically represented in Fig. 7.

Fig. 7.

Proposed schematic models of how Pfn1 loss could potentially accelerate PIP₂ hydrolysis in an actin-dependent manner. Proposed schematic models of how Pfn1 loss could potentially accelerate PIP₂ hydrolysis in an actin-dependent manner. We hypothesize that loosening of cortical actin cytoskeleton in Pfn1-depleted condition may promote PIP₂ diffusion and/or de-caging of PLC thereby allowing more frequent PLC-PIP₂ encounter (see main text for details). Created in BioRender by Orenberg, A., 2026. https://BioRender.com/wsg5ehx. This figure was sublicensed under CC-BY 4.0 terms.

Although Pfn1 depletion has no discernible impact on PM PI4P (the precursor PPI for PIP₂ synthesis) content (this is also consistent with of comparable PI content between control and Pfn1 KO cells), a limitation of the present study is that we have not explored the possibility of Pfn1-dependent alterations in localization and/or activity of cellular machinery (i.e. PIP5K) for PIP₂ biosynthesis. For example, a previous study reported that Pfn1 depletion in chondrocytes inhibits Rho GTPase activation (Bottcher et al., 2009). As the enzymatic activity of PIP5K is stimulated by activated Rho (Weernink et al., 2004), it is not outside of the realm of possibility that Pfn1 loss could still have a negative impact on the actual enzymatic activity of PIP5K via inhibition of Rho. Alternatively, loss of Pfn1 might influence the PM localization of PIP5K (crucial for PIP₂ synthesis) as we have previously shown for SHIP2 (Ricci et al., 2024). Even if some of these possibilities are true, this is still unlikely to be the dominant mechanism given that inhibition of PIP₂ hydrolysis pathway rescues the PIP₂ phenotype of Pfn1-deficient cells.

Finally, our proof-of-concept lipidomic profiling shows that Pfn1 modulates a broader range of lipid species that extend beyond PPIs in cells, revealing a previously unrecognized role of Pfn1 in lipid homeostasis. Of particular importance, we found evidence for a prominent decrease in PC, cholesterol and sphingomyelin, and increase in TG in cells when Pfn1 expression was suppressed. PC is the most abundant phospholipid making up 40–50% of the phospholipid content in cells and plays a major role in determining cell membrane structure (van der Veen et al., 2017). Cholesterol and sphingolipids are major molecular constituents of lipid rafts, the microdomains in the cell membrane that serve as platforms for protein localization and signaling hotspots (Lingwood and Simons, 2010). TG plays crucial functions as energy reservoirs and its hydrolysis via lipolysis into free fatty acids that are subsequently oxidized in mitochondria to generate ATP (Kwiterovich, 2000). Therefore, we think our lipidomic findings open completely new directions for future studies investigating whether and how Pfn1 regulates various biological processes via lipid control. Those findings could shed novel mechanistic insights linking Pfn1 dysregulation to disease progression.

MATERIALS AND METHODS

Cell culture, plasmids and siRNA transfection

HEK-293A cells (Thermo Fisher Scientific R705-07) were cultured in complete medium composed of low-glucose Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific 10567022), 10% heat-inactivated fetal bovine serum (Thermo Fisher Scientific 10438-034), 100 μg/ml penicillin, 100 μg/ml streptomycin (Thermo Fisher Scientific #15140122), and 0.1% chemically defined lipid supplement (Thermo Fisher Scientific #11905031). GFP/Luciferase expressing sublines of MDA-MB-231 (MDA-231; sourced from the ATCC) breast cancer cells [as described previously (Gau et al., 2022)] were cultured in in DMEM supplemented with 10% (v/v) fetal bovine serum, 100 μg/ml penicillin, 100 μg/ml streptomycin and geneticin (100 μg/ml, Thermo Fisher Scientific, #10131027). To generate the doxycycline-inducible Pfn1-knockout (KO) MDA-231 cell line, we utilized the Edit-R Inducible Lentiviral CRISPR/Cas9 system following manufacturer's protocol (Pfn1 sgRNA, VSGH10142-246635145; non-targeting sgRNA, VSGC10215; Cas9 nuclease, VCAS11227, Horizon). A multiplicity of infection (MOI) of 0.3 was utilized for sequential transduction of Cas9- and gRNA-encoding lentivirus for stable cell generation. Cas9-positive clones were selected and maintained with blasticidin (100 μg/ml). Pfn1 or non-targeting KO clones were selected and maintained in medium with puromycin (2.5 μg/ml; Thermo Fisher Scientific; #A1113803). Doxycycline was prepared at 1 mg/ml and administered to cells at a final concentration of 1 μg/ml to trigger Pfn1 KO, and experimental analyses were performed 6 days after initial doxycycline induction. Generation and culture of stable Pfn1 KO B16F1 melanoma cells have been described elsewhere (Tang et al., 2025 preprint). All cell lines were routinely screened for confirmation of mycoplasma-free status by PCR. Lipofectamine 3000 and RNAi Max were used for plasmid and siRNA transfection of cells, respectively, as per manufacturer's instructions. Various Myc-tagged Pfn1 constructs were cloned into an IRES-GFP backbone vector. For transient silencing studies, siRNAs targeting Pfn1 [as previously described (Joy et al., 2017)] and PLCβ3 (Thermo Fisher Scientific #AM16708) were transfected in cells and downstream analyses were performed at least 72 h after transfection. For intracellular Ca²⁺ measurement, live imaging of cells transfected with GCaMP, a genetically encoded Ca²⁺ sensor (Addgene #40753), was performed.

PPI immunostaining and quantification

For PIP₂ immunostaining, cells were seeded on glass substrates coated with either rat tail collagen or murine laminin 24 h prior to staining. Cells were rapidly fixed in 4% formaldehyde and 0.2% glutaraldehyde fixation buffer for 15 min at room temperature (20–24°C) before rinsing three times with PBS containing 50 mM NH₄Cl. Cells were then chilled on ice for at least 10 min before proceeding. All subsequent steps were performed on ice, with all solutions pre-chilled. Cells were blocked and permeabilized for 45 min with a solution of buffer A containing PBS with 5% (v/v) normal goat serum (NGS), 50 mM NH₄Cl and 0.5% saponin, before incubation with a 1:100 dilution of mouse monoclonal anti-PIP₂ antibody (Echelon Biosciences #Z-PO45) in buffer B containing PBS with 5% NGS and 0.1% saponin. Note that the selectivity of this PIP₂ antibody was previously established in immunostaining studies where the immunofluorescence signal was lost with antibody-competing PIP₂ or neomycin (which binds to PIP₂ with high affinity) and upon addition of pleckstrin-homology domain of PIP₂-binding protein PLCδ (Hammond et al., 2006, 2009). At the end of primary antibody incubation, cells were washed three washes in buffer B before incubation with a secondary antibody in buffer B for 45 min followed by four washes in buffer B. Cells were then post-fixed in 2% formaldehyde in PBS for 10 min on ice, before warming to room temperature for an additional 5 min. Formaldehyde was removed by three rinses in PBS containing 50 mM NH₄Cl, followed by one rinse in PBS. Fluorescence images were acquired on an IX83 Olympus (with Cicero confocal) microscope using either a 20× or 40× objective. For PI4P immunostaining, a mouse monoclonal anti-PI(4)P (Echelon Biosciences #Z-P004; 1:100 dilution) antibody was used with identical procedures.

To analyze PM fluorescence intensity, fluorescence images were saved as .tif files and processed using CellProfiler image analysis software (https://cellprofiler.org/). Background subtraction was performed by generating an illumination correction function, smoothing it with a Gaussian blur and subtracting it from the original image. Regions of interest (cell boundaries) were identified by an Otsu threshold algorithm for quantifying the integrated fluorescence intensities at the edge for each cell. For functional rescue experiments, this pipeline was supplemented with additional steps to identify vector expressing cells by saving multi-channel fluorescence images (GFP and RFP), manually segmenting GFP-positive cells, and then employing a distance-based propagation method to define ROIs corresponding to cells co-expressing both fluorophores and cell boundaries prior to integrated edge intensity measurement.

Lipid biosensor studies

HEK-293, co-transfected with Tubby_c–EGFP (Quinn et al., 2008, a biosensor for PIP₂), mCherry–C1ab-PKD1 (Zewe et al., 2020, a biosensor for DAG) and iRFP–Lyn11 (Hammond et al., 2014, a PM biomarker), were imaged with a confocal microscope for quantification of PM to cytosolic fluorescence intensities for various reporters.

F-actin staining and quantification

F-actin staining was performed by fixing cells in 3.7% formaldehyde for 15 min at 37°C. After three washes with Dulbecco's PBS (DPBS), samples were permeabilized with 0.5% Triton X-100 in PBS at room temperature for 5 min. Following another three washes with DPBS, samples were incubated with 1:400 FITC–phalloidin stain (Sigma-Aldrich P5282) in PBS for 1 h at room temperature in the dark, after which the staining solution was aspirated, and samples were washed 3x with PBS. The samples were imaged using either 20× or 60× objectives. For live imaging of F-actin, cells were labeled with SPY55FastActX (Cytoskeleton Inc.) as per the manufacturer's instruction.

Pharmacological studies

In experiments involving pan-PLC inhibitor U73122 and its inactive control U73343 (Medchem Express, #HY108630), the dry compounds were diluted in DMSO to a stock concentration of 10 mM and added to cells at a final concentration of 2 μM after 24 h of serum stimulation. In experiments using lysophosphatidic acid (LPA; Santa Cruz Biotechnology, CAS 325465-43-8), the agonist was diluted to a stock concentration of 10 mM and administered to cells after 24 h of serum starvation at a final concentration of 10 μM for 5 min prior to or during experimental analysis. For studies using latrunculin B (LatB; Cayman Chemical, #10010631), the compound was diluted to a stock concentration of 1 mM and administered to cells at a working concentration of 2 μM for 5 min prior to analysis. For studies involving inhibition of myosin, blebbistatin was diluted to stock concentrations of 10 mM and added to the cultures at a final concentration of 50 μM for 30 min prior to end-point analysis.

Lipidomic analysis

Quantitative lipidomic analysis was performed by Metaware Bio (Durham, NC, USA). Cells were grown to 80% confluency at which point they were trypsinized, washed twice with PBS, and aliquoted into pellets of >10⁶ cells per sample and immediately flash-frozen in liquid nitrogen and stored at −80°C until shipment. Upon receipt, Metaware Bio conducted comprehensive lipidomic profiling using liquid chromatography–mass spectrometry (LC-MS/MS). A detailed protocol of their quantitative lipidomic panel can be found at https://www.metwarebio.com/lipidomics/.

Immunoblotting

Total cell lysate was collected using a modified RIPA buffer [25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% (v/v) Nonidet P-40, 0.2% SDS, 5% (v/v) glycerol, 1 mM EDTA] supplemented with protease/phosphatase inhibitor cocktail (Thermo Fisher Scientific, #78440). Total cell lysate was run on an SDS-PAGE and immunoblotted using various antibodies: Pfn1 monoclonal antibody (Abcam, #EPR6304, 1:750), PLCβ3 polyclonal antibody (Invitrogen, #PA5-78109; 1:1000), GAPDH monoclonal antibody (Invitrogen, #MAS-15738; 1:2000) and Phospho(ser)PKC substrate polyclonal antibody (Cell Signaling Technology, #2261; 1:1000).

Bulk transcriptome sequencing by RNA-seq

For transcriptome sequencing, total RNA was extracted from MDA-231 cells (control versus Pfn1 KO) using a commercial kit (Qiagen). The mRNA was isolated and reverse-transcribed to cDNA and amplified using Illumina^® Stranded mRNA Prep from Illumina, Inc (San Diego, CA, USA). The library preparation processes such as adenylation, ligation and amplification were performed following the manual provided by the manufacturer. The quantity and quality of the libraries were assessed through Bioanalyzer 2100 and Qubit instruments. The procedure of 150 cycle paired-end sequencing in AVITI sequencer followed the manufacturer's manual (Love et al., 2014). Approximately 25×10⁶ sequencing reads were generated per sample, and these reads were trimmed using Trimmomatic to filter out low-quality reads and adapter sequences (Bolger et al., 2014). The surviving reads were then aligned to the reference genome, and the gene-level counts were quantified using STAR with the quantMode GeneCounts option. Differential expression analysis was performed on the gene-by-sample count matrix using the DESeq2 package in R/Bioconductor. Differentially expressed genes (DEGs) were defined by a false discovery rate (FDR) equal to 5% and an absolute fold change >1.5. These DEGs were subsequently used for gene set enrichment analyses (GSEA).

Statistics

All statistical tests were performed using GraphPad Prism 9 software (https://www.graphpad.com). One-way ANOVA with Tukey HSD post tests was used for more than two groups, and a two-tailed unpaired Student's t-test was used for two groups for comparing means between various experimental groups. A P-value less than 0.05 was statistically significant. All error bars in the figures represent standard deviations of data unless explicitly stated otherwise.

Peer review history

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.265025.reviewer-comments.pdf