Gβγ signaling regulates microtubule-dependent control of Golgi integrity

Abstract

Gβγ subunits regulate several non-canonical functions at distinct intracellular organelles. Previous studies have shown that Gβγ signaling at the Golgi is necessary to mediate vesicular protein transport function and to regulate mitotic Golgi fragmentation. Disruption of Golgi structure also occurs in response to microtubule depolymerizing agents, such as nocodazole. In this study, we use siRNA against Gβ1/2 or specific Gγ subunits to deplete their expression, and show that their knockdown causes a significant reduction in nocodazole-induced Golgi fragmentation. We establish that knockdown of Gβγ or inhibition of Gβγ with gallein resulted in decreased activation of protein kinase D (PKD) in response to nocodazole treatment. We demonstrate that restricting the amount of free Gβγ available for signaling by either inhibiting Gα_i activation using pertussis toxin or by knockdown of the non-GPCR GEF, Girdin/GIV protein, results in a substantial decrease in nocodazole-induced Golgi fragmentation and PKD phosphorylation. Our results also indicate that depletion of Gβγ or inhibition with gallein or pertussis toxin significantly reduces the microtubule disruption-dependent Golgi fragmentation phenotype observed in cells transfected with mutant SOD1, a major causative protein in familial amyotrophic lateral sclerosis (ALS). These results provide compelling evidence that Gβγ signaling is critical for the regulation of Golgi integrity.

1. Introduction

Heterotrimeric G proteins (Gαβγ) transduce signals from G protein-coupled receptors (GPCRs) by binding to various downstream effectors to regulate a wide range of cellular functions. Activation of the Gα subunit, which involves its conversion from an inactive GDP bound state to an active GTP bound form, results in the dissociation of free Gβγ from the complex. Classically, this free Gβγ dimer was thought to function at the plasma membrane (PM) by binding to effectors, such as ion channels and phospholipase C-β (PLC-β) to carry out downstream signaling. Studies in recent years have shed light on several non-canonical signaling functions mediated by Gβγ heterodimers at distinct intracellular locations, such as the nucleus, mitochondria, endoplasmic reticulum (ER), and the Golgi complex [1–3].

The Golgi is an essential cellular organelle that is involved in the secretion and transport of proteins synthesized at the ER membranes to destined locations, such as PM and various organelles. Gβγ signaling at the Golgi, through generating a pool of diacylglycerol (DAG) and facilitating subsequent phosphorylation of protein kinase D (PKD), was found to be critical for the recruitment of several key enzymes and proteins that participate in this vesicular transport pathway [4–19].

The Golgi is a dynamic organelle that can undergo reversible fragmentation. Structurally, the Golgi is arranged as several flattened stacks called cisternae, which are linked to each other by tubular connections [20, 21]. Although overexpression of protein transport pathway components, such as, Gβγ, PKD, or PKCη can cause fragmentation of Golgi [6, 10, 18, 22, 23], the Golgi complex undergoes physiological fragmentation. A prominent example of this occurs during the late G2 stage of mitosis, in which fragmentation serves as a key cell cycle checkpoint, and is followed by reassembly in telophase to ensure its equal distribution to the daughter cells [24–28]. PKD signaling was shown to activate the Raf/MEK1 pathway, resulting in Golgi fragmentation during mitosis [29]. In addition, we recently established that Gβγ is required for mitotic Golgi fragmentation by showing that knockdown or inhibition of Gβγ results in substantial decrease in late G2 phosphorylation of PKD and causes delayed mitotic progression [30].

The maintenance of Golgi structure is dependent on intact microtubule organization [31–33], and Golgi fragmentation is observed in response to certain microtubule depolymerizing drugs, such as nocodazole [34]. Interestingly, a study demonstrated that PKD is activated in response to nocodazole and depletion of PKD blocks nocodazole-promoted Golgi fragmentation [35]. A study in our lab further established that Gβγ mediates Golgi fragmentation in response to microtubule disruption with nocodazole [12]. Thus, there are surprising similarities in the requirement for signaling proteins, such as Gβγ and PKD, that regulate the generation at the Golgi of secretory transport vesicles and that regulate Golgi fragmentation occurring during mitosis and in response to nocodazole.

Golgi fragmentation also occurs as a cellular response to certain pathophysiological conditions, such as cancer and neurodegenerative diseases [36–40]. Cellular expression of mutants of superoxide dismutase 1 (SOD1), the major causative protein in familial amyotrophic lateral sclerosis (ALS), results in increased expression of stathmin proteins which destabilize microtubules and trigger Golgi fragmentation [41]. Therefore, it is interesting to delineate whether Gβγ is involved in the mutant SOD1-mediated Golgi fragmentation.

Here, we sought to extend the analysis of a role for Gβγ in microtubule disruption-dependent Golgi fragmentation. Although it was previously shown that Gβγ plays a role in nocodazole-promoted Golgi fragmentation, we now show that Gβγ is also required for nocodazole-promoted activation of PKD. In addition, we begin to define a role for specific Gγ subunits and examine upstream regulators of Gβγ signaling in Golgi fragmentation caused by microtubule depolymerization. We find that knockdown or inhibition of Gβγ causes decreased activation of PKD in response to nocodazole treatment. In addition, we show that activation of Gα_i is necessary for the nocodazole induced phosphorylation of PKD and subsequent fragmentation of Golgi. Our results also suggest a role for non-GPCR guanine nucleotide exchange factor (GEF) GIV/Girdin in regulating nocodazole-induced Golgi fragmentation and modulating PKD activity. Lastly, our data shows that mutant SOD1-mediated Golgi fragmentation is also significantly assuaged by inhibiting Gβγ signaling. This study delineates the signaling function of Gβγ heterodimers in Golgi fragmentation occurring due to microtubule disruption and provides insights on the potential upstream regulators of the pathway.

2. Materials and Methods

2.1. Reagents, antibodies, siRNA and plasmids

Gallein (Cat # 3090) and Pertussis toxin (Cat # 3097) were purchased from Tocris biosciences. Nocodazole was purchased from Sigma Aldrich. YM-254890 (Cat # 257-00631) was obtained from Wako Chemicals. DMSO (Cat# BP231-1) was from Fisher. The anti-Gβ1 (Cat # 137635) and anti-Gβ2 (Cat # 108504) antibodies were obtained from Abcam. The antibodies for PKD/PKCμ (D4J1N) (Cat # 90039), PKD2 (Cat # 8188S), phospho-PKD/PKCμ (Ser 916) (Cat # 2051S) and GIV (Cat # 14200S) were purchased from Cell Signaling. The Hsp90 antibody (Cat # sc-7947) was obtained from Santa Cruz. The anti-GM130 mouse antibody (Cat # 610822) was from BD Biosciences. The anti-GFP rabbit (Cat # 50430-2-AP), anti- GAPDH (Cat # 60004-1) anti-GM130 rabbit antibodies (Cat # 11308-1-AP) were from Proteintech. YFP-Gγ3 and YFP-Gγ9 constructs were obtained from Addgene. The GFP-SOD1(G93A) plasmid construct was provided by Dr. Piera Pasinelli (Thomas Jefferson University). The siRNAs against Gβ1/2 were as described previously [10]. The Control siRNA (ON-TARGET plus non targeting siRNA), GIV-siRNA, Gγ3, Gγ9 and Gγ12 siRNA pools were obtained from Horizon discovery.

2.2. Cell culture and Transfection

HeLa or HEK293 cells were grown in DMEM supplemented with 10% FBS. Cells were grown and maintained in 100 mm dishes. The siRNA was transfected in cells grown on 6-well plates using Lipofectamine RNAi Max reagent (Invitrogen) as per the manufacturer’s instructions and typically the cells were harvested 48 h post transfection. The YFP-Gγ plasmid constructs were transfected 24 h post Gγ-siRNA transfection using Lipofectamine 2000 (Invitrogen) as per the manufacturer’s instructions and were harvested the next day (after a total of 48 h since siRNA transfection). For analyzing mutant SOD1-mediated Golgi fragmentation, the GFP-SOD1(G93A) plasmid construct was transfected using Lipofectamine 2000 (Invitrogen) as per the manufacturer’s instructions. 24 h post transfection, cells were replated on to coverslips and then processed for staining with GM130 antibodies.

2.3. Western Blotting

Cells were lysed in 1X SDS gel-loading buffer containing 1% BME. Lysates were typically resolved in 10% SDS PAGE gels followed by transfer to nitrocellulose membrane. The blotted membranes were blocked in 5% non-fat milk for 30 min and were incubated overnight with indicated primary antibodies. On the following day, secondary antibodies conjugated to HRP or IRDye were added and the images were captured either by AI 680 Imager (GE Healthcare) or by Li-Cor Odyssey Imaging system, respectively. For assessing the PKD phosphorylation, membranes were blocked with 5% BSA followed by incubation with phospho-PKD/PKCμ (Ser 916) antibody (Cell Signaling) in blocking buffer overnight, followed by incubation with secondary IRDye 680 RD anti-rabbit antibody (Fisher-Scientific) and visualized by Li-Cor Odyssey Imaging system.

For analyzing the lysates for GIV knockdown, the samples were run on 8% SDS-PAGE gels followed by transfer to PVDF membrane at 300 Volts for 2.5 h. The blotted membrane was processed as described above. To analyze the lysates for Gγ12 knockdown, the samples were resolved on 15% SDS-PAGE gels followed by transfer to PVDF membrane at 100 Volts for 45 min. The membrane was then processed by a modified protocol [42]. Briefly, the blotted membrane was washed in TBST (Tris-buffered Saline with 0.1% Tween 20) and then was treated with blocking buffer (0.1% BSA and 1% milk and in TBST) for 5 min. Later, the membrane was washed with PBST (Phosphate-buffered Saline with 0.1% Tween 20) for 3min, and was fixed with 0.2% Glutaraldehyde solution in PBST for 15min. The fixed membrane was washed 3 times with PBST and then retrieval was performed by soaking the membrane in citrate buffer (10mM citric acid at pH 6.0, 1mM EDTA, 0.05% Tween 20) and microwaving for 10 min. After cooling membrane to room temperature, the membrane was immersed in quenching buffer (200mM glycine in PBST) for 10min. The membrane was then blocked in 0.1% BSA and 1% milk and in TBST for 30min and probed with anti-Gγ12 antibody overnight at 4 °C. Following day, membrane was washed with TBST and incubated with secondary anti- rabbit antibody conjugated to HRP for 1hr. After 4 X TBST washes for 6 min each, chemiluminescence reagent (Thermo Scientific) was added to the membrane and the images were captured by AI 680 Imager (GE Healthcare).

2.4. Immunofluoresence Microscopy

Cells plated on coverslips were collected at indicated time points, and were fixed with 3.7% formaldehyde in PBS for 15min. Cells were blocked in buffer containing 2.5% milk in TBST for 30min, and were then probed with the anti-rabbit GM130 antibody in blocking buffer for 1 h. The cells were washed thrice with TBST and incubated with 1:200 dilution of either the goat anti- rabbit Alexa 488 or goat anti-rabbit Alexa 594 secondary antibodies (Invitrogen) for 1 h. Subsequently, cells were washed thrice with TBST and stained with DAPI solution for 5 min. The coverslips were then washed with PBS, rinsed in distilled water, and mounted on glass slides with Prolong Anti-fade reagent (Invitrogen). Images were acquired using an Olympus IX83 microscope with a 60x oil immersion objection and an ORCA Fusion sCMOS camera (Hamamatsu) controlled by Olympus cellSens software.

2.5. Assessing Golgi Fragmentation

Golgi status was assessed by immunofluorescence microscopy by staining with the GM130 antibody after treatment with 5 μg/ml nocodazole or DMSO control for indicated time points or after expression of mutant or WT SOD1. Cells were viewed and individually scored for intact Golgi due to a compact GM130 staining pattern or for a fragmented Golgi phenotype as defined by the loss of compact Golgi structure and a diffuse pattern of cytoplasmic puncta. In cells transfected with GFP-tagged mutant (G93A) or wild-type (WT) SOD1 protein, only cells with a GFP signal were scored for intact or fragmented Golgi. Cells displaying a fragmented Golgi phenotype are shown as a percentage of total cells.

2.6. Statistical Analysis

All data was analyzed using the GraphPad Prism software. The data across the groups were compared using either Unpaired t test, One-way or Two-way ANOVA as mentioned in the figure legends.

3. Results

3.1. Gβγ signaling is essential for nocodazole-induced activation of PKD

Activation of PKD was shown to be required for nocodazole-induced Golgi dispersal [35], and in a previous study, we had shown that knockdown of Gβ subunits result in reduced nocodazole-induced Golgi fragmentation [12]. Thus, we wanted to test whether Gβγ regulates PKD activity in response to nocodazole treatment. Towards this, we used siRNA to deplete the most abundantly expressed Gβ isoforms, Gβ1 and Gβ2, in HEK293 cells. In accordance with data reported by Fuchs et al. [35], in control cells treated with a non-targeting siRNA, we found a rapid increase in PKD phosphorylation as early as 15 min of nocodazole treatment. The pPKD levels continued to rise until 60 min post treatment (Fig. 1A and 1B). When we used siRNA against Gβ1/2 subunits, we found that their depletion resulted in substantial reduction of pPKD activity at all indicated time points (Fig. 1A and 1B). Next, we pretreated the cells with a pharmacological inhibitor of Gβγ, gallein, to impede its activity, and as a control, cells were pretreated with DMSO. Addition of nocodazole in control DMSO treated cells, resulted in a rapid rise of PKD phosphorylation in 15 min and the pPKD levels continued to increase until 60 min post nocodazole treatment (Fig. 1C and 1D). However, in cells pretreated with gallein, we found that pPKD levels were significantly lower at all the time points when compared to the control (Fig. 1C and 1D). These results indicate that PKD phosphorylation occurring in response to microtubule disruption caused by nocodazole, is regulated upstream by Gβγ.

Figure 1: Knockdown/Inhibition of Gβγ blocks nocodazole-induced activation of protein kinase D (PKD)-.

(A) HEK293 cells were transfected with either control or Gβ1/2 siRNA, and 48h later cells were treated with 5 μg/ml nocodazole for the indicated times. Lysates were immunoblotted with anti-pPKD and anti-PKD antibodies and the efficiency of knockdown was checked by Gβ1/2 antibodies. (B) Intensities of pPKD signals were quantified using ImageJ software and normalized by dividing by total PKD signals. Graph shows mean. +/− s.d. for three independent experiments. Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, p<0.0005 (***), p<0.005 (**) (C) Cells were treated with 5 μg/ml nocodazole for the indicated times, −/+ pretreatment for 2h with 10 μM gallein, as indicated. Lysates were immunoblotted with anti-pS910-PKD and anti-PKD antibodies. Representative immunoblots are shown. (D) Graph shows mean. +/− s.d. for three independent experiments. Intensities of pPKD signals were normalized by dividing by PKD signals. Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, p<0.0005 (***), p<0.005 (**), p<0.05 (*).

3.2. Specific Gγ subunits regulate Golgi fragmentation and PKD activation in response to nocodazole

Having established that Gβγ plays a role in linking microtubule depolymerization to PKD activation and Golgi disruption, we were interested in examining whether specific Gγ subunits that have been shown to function in Gβγ signaling at the Golgi also regulate nocodazole-promoted PKD activation and Golgi disruption. Recent work used expression of YFP-tagged Gγ subunits in HEK293 cells to show that among all 12 members of the Gγ family, only Gγ₉, within a Gβγ dimer, displayed strong translocation from the PM to the Golgi upon stimulation of cells with the CXCR4 agonist SDF1α, and depletion of Gγ₉ but not Gγ₃ prevented Golgi-localized activation of ERK1/2, PI3Kγ and ARF1 [43, 44]. Hence, we used siRNAs against Gγ₉ or Gγ₃ (as a control) to deplete their expression in HEK293 cells and analyzed for PKD activity. Our results indicate that knockdown of Gγ₉ but not Gγ₃ brought about a significant reduction in PKD phosphorylation upon addition of nocodazole when compared to control cells (Fig. 2A and 2B). Due to the unavailability of effective antibodies to detect the endogenous Gγ₃ and Gγ_9, proteins, the efficiency and specificity of Gγ₃ and Gγ₉ knockdown by the respective siRNAs was tested by expressing YFP-Gγ₃ or YFP-Gγ₉ constructs (Fig 2C). Compared to HEK293 cells, HeLa cells have a compact Golgi morphology and are thus a preferable model cell for visualizing changes in Golgi structure; however, Gγ₉ is expressed at very low levels in HeLa cells [45]. In contrast, Gγ₁₂ is one of the most highly expressed Gγ subunits in HeLa cells [45], and, importantly, Gγ₁₂ was identified as the critical Gγ subunit in a Gβγ dimer that couples GPRC5A to PKD activation at the Golgi [46]. We thus used siRNA against Gγ₁₂ to deplete its expression and investigated the integrity of Golgi complex after nocodazole treatment. Non-targeting siRNA or Gβ1/2 siRNA were also transfected in cells to visualize Golgi morphology. In accordance with the results observed by our group previously [12], addition of nocodazole caused Golgi fragmentation in ~80% of HeLa cells treated with non-targeting siRNA which was significantly inhibited by Gβ1/2 knockdown (Fig 2D and 2E). Interestingly, in cells where Gγ₁₂ was depleted, we observed a substantial decrease in the number of cells with fragmented Golgi (Fig 2D and 2E). We then assayed for PKD activity in HeLa cells and found that PKD was robustly phosphorylated within 10 min of addition of nocodazole and remained activated until 20 min (Fig 2F). When compared to control siRNA transfected cells, nocodazole-induced PKD phosphorylation was significantly lower in Gγ₁₂ depleted cells (Fig 2F and 2G). We also noted that the reduction in Golgi fragmentation and pPKD activity in Gγ₁₂ depleted cells was comparable to the decrease observed in Gβ1/2 depleted cells, suggesting that Gγ₁₂ is crucial for signaling at the Golgi in HeLa cells (Fig 2G). Taken together, these results emphasize that definitive Gγ subunits are essential to mediate the Gβγ signaling function at the Golgi. The specificity of Gγ subunits expressed by different cell types that translocate to Golgi add another level of complexity in the regulation of Gβγ function.

Figure 2: Knockdown of Gγ9 or Gγ12 reduces nocodazole-induced Golgi fragmentation and PKD activation-.

(A) HEK293 cells were transfected with either control (Csi), Gβ1/β2, Gγ3, or Gγ9 siRNA and, after 48 hrs, nocodazole was added to cells at indicated time points. Lysates were immunoblotted with the indicated antibodies. Representative immunoblots are shown. (B) Intensities of pPKD signals were normalized by dividing by total PKD signals. Graphs represent mean +/− s.d. for three independent experiments. Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, p<0.0005 (***), p<0.0001(****). (C) Cells were transfected with 100ng of YFP-Gγ3, or YFP-Gγ9, along with the Gγ3, or Gγ9 siRNA and, 48 hrs later, cells were lysed and the lysates were probed with anti-GFP antibody to check the efficacy of knockdown. The anti-HSP90 antibody was used as a loading control. (D) HeLa cells were transfected with control (Csi), Gβ1/β2, or Gγ12 siRNA and, after 48 hrs, nocodazole was added to cells for 25 min. Cells were fixed and Golgi integrity was observed with GM130 antibody. Representative cells depicting intact Golgi and fragmented Golgi are shown. (E) Cells, as described above (D), were scored for fragmented Golgi. Graphs represent mean +/− s.d. for three independent experiments. More than 100 cells were counted for each experiment. Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, p<0.005 (**), p<0.05 (*). (F) HeLa cells were transfected with either control (Csi), Gβ1/β2 (βsi), or Gγ12 siRNA and, after 48 hrs, nocodazole was added to cells at indicated time points. Lysates were immunoblotted with the indicated antibodies. Representative western blots are shown. Graph shows mean. +/− s.d. for three independent experiments. (G) Intensities of pPKD signals were quantified using ImageJ software and normalized by dividing by total PKD signals. Graph shows mean. +/− s.d. for three independent experiments. Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, p<0.0001(****).

3.3. Gα_i activation is necessary for nocodazole-dependent Golgi fragmentation and PKD phosphorylation

Gβγ at the Golgi can be activated by translocation from PM to Golgi or by GPCRs that are localized at the Golgi [46–48], and a previous report showed that Gα_i activation at the Golgi is required for regulation of Golgi transport function [49]. We sought to address whether the activation of a Gα_iβγ heterotrimeric complex is necessary for the regulation of nocodazole-induced fragmentation. To understand this, we used pertussis toxin (PTX), which is known to ADP-ribosylate a cysteine residue in the C-terminus of Gα_i and thereby prevent Gα_i activation by impeding its binding to GPCRs [50]. HeLa cells were pretreated with either gallein or pertussis toxin and the Golgi morphology was visualized by GM130 staining upon addition of nocodazole. We observed that nocodazole-induced dispersal of Golgi stacks was inhibited by pretreatment with PTX (Fig 3A and 3B). Next, we assessed for pPKD in HEK293 cells in response to Gα_i inhibition by PTX, and found that nocodazole-induced PKD phosphorylation was substantially reduced in response to pretreatment with either gallein or PTX (Fig 3C and 3D). To investigate whether inhibition of Gα_q or Gα₁₁ would have an effect on nocodazole-induced pPKD, we used a known inhibitor YM-254890 [51, 52]. Treatment of cells with YM-254890 brought about a slight reduction in phosphorylation of PKD upon addition of nocodazole, but this change was not found to be statistically significant (Supp. Fig 1). These findings therefore indicate a dependence on the activation of Gα_i which would in turn release free Gβγ for downstream signaling.

Figure 3: Pertussis toxin inhibits nocodazole-induced Golgi fragmentation and activation of PKD –

(A) HeLa cells were pretreated with 10 μM Gallein for 2h, or with 0.1 μg/ml Pertussis toxin (PTX) for 4h. Later, 5 μg/ml nocodazole was added to cells for 25 min. Cells were fixed and Golgi integrity was observed with GM130 antibody. Representative cells depicting intact Golgi and fragmented Golgi are shown. (B) Cells, as described above (A), were scored for the fragmented Golgi. Graphs represent mean +/− s.d. for three independent experiments, over 100 cells were counted for each experiment. Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, p<0.0001(****). (C) Cells were treated with 5 μg/ml nocodazole for the indicated times, −/+ pretreatment with 10 μM gallein (for 2 h), or with 0.1 μg/ml PTX (for 4h or overnight (O/N) as indicated. Lysates were immunoblotted with anti-pPKD and anti-PKD antibodies. Representative immunoblots are shown. (D) Graph shows mean. +/− s.d. for three independent experiments. Intensities of pPKD signals were normalized by dividing by PKD signals Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, p<0.005 (**), p<0.05 (*).

3.4. Knockdown of the non-GPCR GEF GIV/Girdin inhibits Golgi fragmentation and PKD activation in response to microtubule disruption

A group of proteins containing the Gα-binding-and-activating (GBA) domain was shown to act as non-GPCR guanine-nucleotide exchange factors (GEFs) that activate Gα subunit and mediate the dissociation of Gβγ from the heterotrimer [53–55]. One such GBA motif-containing protein, GIV (Gα-interacting vesicle-associated protein) also known as Girdin [53, 54], can transduce signals to activate heterotrimeric G proteins and regulate functions such as intracellular trafficking and maturation of autophagosomes [56]. Specifically, it was shown that GIV localizes predominantly to COPI vesicles on the Golgi [57] and activates Gα_i to regulate Gβγ-dependent Arf1 signaling and ER-Golgi protein transport [49]. We therefore wanted to test whether GIV might also function upstream of Gβγ to regulate Golgi fragmentation. Towards this, we used GIV siRNA to deplete endogenous GIV in HeLa or HEK293 cells to probe for changes in Golgi structure or modulation of PKD activity. We found that upon addition of nocodazole, in cells transfected with non-targeting siRNA, ~80% of the cells had fragmented Golgi, which was significantly inhibited by knockdown of GIV/Girdin (Fig 4A and 4B). We then assessed for pPKD activity in cells transfected with GIV siRNA, and observed that depletion of GIV resulted in a significant reduction in pPKD levels in response to nocodazole, which is comparable to the decrease seen in Gβ1/2 depleted cells (Fig. 4C and 4D). These findings indicate a role for GIV in regulating Golgi fragmentation occurring in response to microtubule disruption caused by nocodazole.

Figure 4: Knockdown of GIV/Girdin inhibits nocodazole-induced Golgi fragmentation and activation of PKD –

(A) HeLa cells were transfected with control (Csi), Gβ1/β2, or GIV siRNA and, after 48 hrs, nocodazole was added to cells for 25 min. Cells were fixed and Golgi integrity was observed with GM130 antibody. Representative cells depicting intact Golgi and fragmented Golgi are shown. (B) Cells, as described above (A), were scored for fragmented Golgi. Graphs represent mean +/− s.d. for three independent experiments, over 100 cells were counted for each experiment. Statistical analysis was done by one-way ANOVA, p<0.0001(****). (C) Cells were transfected with control (Csi), Gβ1/β2, or GIV siRNA and, after 48 hrs, nocodazole was added to cells at indicated time points. Lysates were immunoblotted with the indicated antibodies. Representative western blots are shown. Graph shows mean. +/− s.d. for three independent experiments. (D) Intensities of pPKD signals were normalized by dividing by total PKD signals. Graphs represent mean +/− s.d. for three independent experiments. Statistical analysis was done by one-way ANOVA followed by Tukey’s multiple comparison test, p<0.0005 (***).

3.5. Gβγ signaling plays a role in mutant SOD1 mediated Golgi fragmentation

Fragmentation of Golgi occurs in certain cancers and neurodegenerative diseases and is thought to contribute to the disease phenotypes [38, 40, 58]. Mutated SOD1, the major causative protein in familial amyotrophic lateral sclerosis (ALS) was shown to cause 3-fold increased expression of stathmin proteins in motor neurons which result in destabilization of microtubules and cause Golgi fragmentation [41]. Consistent with previous reports [41], we observed that HeLa cells transfected with GFP-SOD1-WT had intact Golgi structure whereas cells expressing GFP-SOD1-G93A exhibited Golgi fragmentation (Fig. 5A). Interestingly, when we depleted the expression of Gβ1 and Gβ2 subunits, we saw a significant reduction (~50%) in mutant SOD1-mediated Golgi fragmentation (Fig. 5A and 5B). We then transfected the cells with GFP-SOD1-G93A and treated the cells with gallein 5 hours post transfection to inhibit Gβγ.

Figure 5: Depletion of Gβ1/2 and inhibition of Gβγ attenuates mutant SOD1-mediated Golgi fragmentation -.

(A) HeLa cells were transfected with either Control or Gβ1/2 siRNA. After 24hrs, cells were transfected with either GFP-SOD1-WT or GFP-SOD1(G93A) constructs. 24h post SOD1 transfection, cells were replated onto coverslips and the following day, the cells were stained with GM130 antibody. Representative SOD1 transfected cells depicting intact Golgi and fragmented Golgi are shown. (B) Western blot analysis of cell lysates from (A) showing the expression of GFP-SOD1-WT or GFP-SOD1 (G93A) and the efficiency of knockdown with Gβ1/2 siRNA. (C) Cells were transfected with either control or Gβ1/2 siRNA and with GFP-SOD1(G93A) construct were stained with GM130 antibody and scored for the fragmented Golgi phenotype. Graph shows the mean. +/− s.d. for three independent experiments, over 50 GFP positive cells were counted for each experiment. Statistical analysis was done by one-way ANOVA, p<0.0001(****). (D) Cells were transfected with GFP-SOD1(G93A) construct and 4h post transfection, cells were treated with 10 μM gallein, or with 0.1 μg/ml Pertussis toxin (PTX). 24h post transfection, the cells were fixed and Golgi integrity was observed with GM130 antibody. Representative cells depicting intact Golgi and fragmented Golgi are shown. (E) Cells, as described above (D), were scored for the fragmented Golgi. Graphs represent mean +/− s.d. for three independent experiments, over 50 GFP positive cells were counted for each experiment. Statistical analysis was done by one-way ANOVA, p<0.0001(****).

We observed a substantial decrease in the number of mutant SOD1 expressing cells that contained fragmented Golgi (Fig. 5C and 5D). Also, as we had observed an effect of Gα_i inhibition by PTX on nocodazole-induced Golgi fragmentation (Fig. 4A and B), we wanted to test whether PTX could also inhibit mutant SOD1-mediated Golgi fragmentation. Towards this, PTX was added to the cells 5 hours post transfection with GFP-SOD1-G93A, and Golgi fragmentation was observed in these cells. Similar to the results observed with gallein, inhibition of Gα_i by PTX also significantly reduced the mutant SOD1-mediated Golgi fragmentation (Fig. 5C and 5D). These results provide initial evidence that Gβγ signaling couples disruption of microtubules to fragmentation of the Golgi in response to the familial ALS mutant SOD1.

4. Discussion

It is well established that microtubule depolymerization causes a rapid disassembly of the Golgi complex [27, 31, 59]. However, the signaling pathways that link microtubule disruption to Golgi fragmentation remain elusive. Previous reports established that PKD is required for nocodazole-induced Golgi fragmentation [12, 35]. In our study, we have discovered that Gβγ is required for nocodazole-dependent activation of PKD. We indicate that knockdown of individual Gβ or Gγ subunits or pretreatment with a chemical inhibitor of Gβγ, gallein, significantly reduced the pPKD levels in response to nocodazole treatment. Our results show that inhibition of Gα_i activation by pertussis toxin inhibited nocodazole-mediated Golgi fragmentation and PKD activation. In addition, we find that knockdown of Girdin, a non-GPCR GEF, resulted in decreased PKD phosphorylation. Lastly, our results demonstrate that Gβγ signaling regulates Golgi fragmentation which is triggered by expression of mutant SOD1 protein, a causative factor of familial ALS.

There is increasing evidence that the versatile Gγ subunits regulate multiple functions at intracellular locations. The sequence variation and tissue specific expression of the Gγ isoforms render them to bind differentially to various intracellular membrane locations, thereby allowing differential Gβγ interaction with downstream effectors and asserting functional specificity to Gβγ signaling [3, 18, 60, 61]. Few studies have previously addressed the role of specific Gγ isoforms function at the Golgi. A recent report showed that Gγ₉ but not Gγ₃ translocated to the Golgi and that the Golgi-localized Gβγ activates mitogen activated protein kinase (MAPK) pathway [43]. Our results indicate a function for Gγ₉ or Gγ₁₂ subunits in HEK293 cells and HeLa cells, respectively, in regulating nocodazole-induced Golgi fragmentation and activation of PKD. These data collectively indicate that the diversity of the Gγ isoform expressed in different cell types determines the functional regulation/translocation of Gβγ heterodimers to cellular compartments.

In our study, we found that inhibition of Gα_i activation by PTX significantly minimized nocodazole-induced Golgi fragmentation and PKD phosphorylation. Our results also showed that the Gαq/11 inhibitor, YM-254890 slightly reduced the nocadazole-induced PKD phosphorylation (Sup. Fig. 1). Although the effect of YM-254890 did not reach statistical significance, this result raises the possibility that Gq/11 heterotrimer activation could contribute to Golgi fragmentation. Overall, our results shed light on how Gβγ is regulated at the Golgi, as well as suggesting that Gi heterotrimer activation frees the Gβγ heterodimers which are responsible for activating the PKD pathway. Two independent studies have correlated the ability of a PM-localized GPCR to promote translocation of Gβγ from the PM to Golgi with enhanced regulation of secretory protein transport [18, 62]. Earlier reports have noted the presence of GPCRs on intracellular membranes [63]. However, whether a PM-localized GPCR promotes Gβγ translocation to the Golgi or an intracellular GPCR at the Golgi membrane is involved in initiation of this signaling pathway is still unclear. Further studies are necessary to determine whether any of the GPCRs which were previously reported to partially localize or get activated at the Golgi such as dopamine D1 receptor [64], β₁ adrenergic receptor [48], the orphan receptor GPRC5A (GPCR class C group 5 member A) [46] or the δ-opioid receptor [65] are involved in the microtubule dependent regulation of Golgi structure. Also, the cues or ligands needed to activate these specific GPCRs to regulate Golgi fragmentation need to be further elucidated. Alternatively, in our study we found that the non-GPCR activator of heterotrimeric G protein GIV/Girdin can play a role in nocodazole-promoted PKD activation and Golgi fragmentation. However, it has been reported that PTX has no effect on GIV-mediated activation of the Gα subunit [66]. Other proteins, such as Daple and AGS family proteins, have also been described as intracellular non-GPCR activators of G proteins [67–69]. Similar experiments as described for GIV, will have to be carried out for such candidates to delineate their role in Golgi fragmentation. It will be interesting to identify the pathways that link microtubule disruption to the choice of activation by either GPCRs or intracellular GEFs that efficiently release free Gβγ for downstream signaling at the Golgi. Further experiments are necessary to identify other key signaling proteins downstream of Gβγ and PKD that link microtubule depolymerization to alterations of Golgi structure. It is imperative to identify the key components and the underlying signaling mechanisms in order to better comprehend how microtubule destabilizing drugs affect cell function.

Golgi fragmentation also occurs in cancer and neurodegenerative diseases [38, 40, 58]. Although the direct contribution of Golgi fragmentation to a particular disease state remains to be fully elucidated, a study in a mouse model demonstrated that loss of the Golgi tether protein, GM130, caused Golgi fragmentation in neurons and ataxia in the mouse [70], supporting the idea that Golgi dysregulation and fragmentation can contribute to neurodegenerative diseases. However, there is little insight into pathways regulating Golgi fragmentation in response to dysregulated proteins in disease. An important step forward in this regard was taken in a study demonstrating that mutants of the SOD1 protein, that cause familial ALS, promote Golgi fragmentation via the microtubule destabilizing proteins stathmin 1/2 [41]. Since we have determined a role for Gβγ in linking the nocodazole-promoted disruption of microtubules to Golgi fragmentation, we asked whether Gβγ is also involved in mutant SOD1/stathmin-regulated Golgi fragmentation, a pathway which also depends on microtubule disruption. Indeed, we find that Gβγ regulates mutant SOD1-mediated Golgi fragmentation (Fig. 5). Particularly, we observe that depleting Gβ or treating the cells with either gallein or PTX minimized the Golgi dispersal caused by expression of mutant SOD1 protein. This novel finding suggests common pathways are used by diverse drivers of Golgi fragmentation. We observe that even when the methods used in this study (treatment with nocodazole or expression of mutant SOD1) to cause microtubule depolymerization are different, Gβγ signaling emerges as a common regulator involved in the Golgi fragmentation pathway. Further understanding of the signaling pathways that regulate microtubule-dependent Golgi dispersal may reveal potential therapeutic targets for neurodegenerative diseases.

5. Conclusions

In our study we delineated a role for Gβγ in microtubule disruption-dependent Golgi fragmentation induced either by nocodazole or by the expression of mutant SOD1 (Figure 6). We show that Gβγ signaling regulates nocodazole-induced activation of PKD. We attempted to define the specific Gγ subunits that regulate Golgi fragmentation caused by microtubule destabilization. In addition, we show that inhibition of Gα_i signaling or knockdown of non-GPCR guanine nucleotide exchange factor (GEF) GIV/Girdin significantly impedes nocodazole-induced Golgi fragmentation and PKD phosphorylation. This study outlines the signaling function of Gβγ heterodimers in regulating Golgi fragmentation occurring due to microtubule depolymerization and identifies the potential upstream regulators of the pathway.

Figure 6: Model depicting the microtubule-dependent regulation of signaling by free Gβγ at the Golgi:

In this model, disruption of microtubules by either nocodazole or by expression of proteins such as mutant SOD1 in disease states results in activation of GPCRs or non-GPCR GEFs such as GIV by yet unknown mechanisms. This activation would promote the release of free Gβγ from Gα_iβγ heterotrimeric complexes. Then, Gβγ signaling at the Golgi initiates a downstream signaling cascade which further activates protein kinase D (PKD) and ultimately results in fragmentation of the Golgi. This illustration was created with BioRender.com.

Supplementary Material

1

Supplementary Figure 1: (A) HeLa cells were treated with 5 μg/ml nocodazole for the indicated times, −/+ pretreatment with 10 μM gallein (for 2 h), or with 1μM YM-254890 (YM) (for 2h or overnight (O/N) as indicated). Lysates were immunoblotted with anti-pPKD and anti-PKD antibodies. Representative immunoblots are shown. (B) Graph shows mean. +/− s.d. for three independent experiments. Intensities of pPKD signals were normalized by dividing by PKD signals Statistical analysis was done by 2-way ANOVA followed by Tukey’s multiple comparison test, ns (nonsignificant, p > 0.05), p<0.005 (**), p<0.05 (*).

Highlights.

• Gβγ signaling regulates nocodazole induced Golgi fragmentation and PKD activation.

• Knockdown of GIV/Girdin diminishes nocodazole induced PKD phosphorylation.

• Inhibition of Gα_i impedes Golgi fragmentation caused by disruption of microtubules.

• Gβγ signaling modulates mutant SOD1 mediated Golgi fragmentation.

Abbreviations:

GPCR

G protein-coupled receptor

GTP

guanosine triphosphate

GDP

guanosine diphosphate

PKD

protein kinase D

MAPK

mitogen activated protein kinase

ERK

extracellular signal-regulated kinase

PI3K

phosphatidylinositol-3 kinase

Arf1

ADP (Adenosine Diphosphate) Ribosylation Factor-1

SOD1

superoxide dismutase 1

ALS

amyotrophic lateral sclerosis

DAG

diacylglycerol

PTX

pertussis toxin

GEF

guanine Nucleotide Exchange Factor

GBA

Gα-binding-and-activating

GIV

Gα-interacting vesicle-associated protein

YFP

yellow fluorescent protein

GFP

green fluorescent protein

DMSO

dimethylsulfoxide

DMEM

Dulbecco’s modified Eagle medium

HEK

human embryonic kidney

BSA

bovine Serum Albumin

BME

beta-mercaptoethanol

SDS-PAGE

sodium dodecyl sulfate polyacrylamide gel electrophoresis

Data availability

Data will be made available on request.