Secretory proteins are generally known to be transported through the endoplasmic reticulum (ER) and the Golgi apparatus. After this process, they are directed to the plasma membrane or released from the cell. Recently, it has been discovered that numerous substrates also reach these destinations through unconventional pathways. Unconventional protein secretion (UcPS) is a complex process that includes cargoes without a signal peptide or transmembrane domain, allowing them to translocate across the plasma membrane. Additionally, these cargoes reach the plasma membrane by bypassing the Golgi apparatus, even after entering the ER [1]. The UcPS is unique in that it involves the extracellular release of proteins bypassing the conventional ER-Golgi pathway, while conventional protein secretion follows this classical route within cells [2].

Several clinically important transmembrane proteins, such as the △F508 mutant (deletion of phenylalanine at position 508) of the cystic fibrosis transmembrane conductance regulator (CFTR), the H723R mutant of pendrin, and the Spike (S) protein of coronaviruses, have been demonstrated to reach the cell surface via UcPS under specific cellular conditions [3,4]. CFTR functions as an anion channel, and mutations in it cause cystic fibrosis, with △F508-CFTR being the most common worldwide [5]. The H723R-pendrin protein is generated by a mutation in the SLC26A4 gene, which encodes pendrin, an iodide/chloride/bicarbonate transporter expressed in the inner ear, and is the most common cause of congenital hearing loss in East Asia [6]. The S protein of the coronavirus is an envelope protein that plays a critical role in the entry into host cells. It has been shown that the assembly and cellular exit of coronaviruses share the unconventional secretion pathway with other budding viruses and lipoproteins [4].

In conventional Golgi-mediated protein secretion, N-glycosylated proteins such as CFTR and pendrin undergo core-glycosylation at the ER (known as band B) and complex-glycosylation at the Golgi apparatus (known as band C). Since the misfolded △F508- CFTR and H723R-pendrin are recognized by the ER quality control system and eliminated via the ER-associated degradation pathway, the core-glycosylated △F508-CFTR and H723R-pendrin hardly reach the cell surface under normal conditions [3,4,7]. Notably, upregulation of UcPS pathways resulting from GRASP55 overexpression or IRE1α kinase activation has been shown to directly induce cell surface expression of the ER form of core-glycosylated CFTR and pendrin, bypassing the Golgi. This suggests that UcPS activation could be a strategy to treat diseases caused by trafficking-defective CFTR and pendrin [3,6].

The molecular mechanisms of the UcPS pathway, particularly the molecular machineries involving cytosolic motor proteins, remain largely elusive. Kinesin and cytoplasmic dynein are microtubule-based motor proteins that transport materials within the cell [8]. Kinesin facilitates cargo transport by moving along microtubules toward the plus ends, whereas dynein transports cargo toward the minus ends of microtubules. The kinesin superfamily (KIF) consists of fourteen large families [9], whereas the dynein family has two major branches; cytoplasmic dynein and axonemal dynein [10]. In general, cytoplasmic dynein transports cargo vesicles to localize the Golgi apparatus to the center of the cell [11]. Dynactin, an activator of dynein, docks cargo to the motor protein and enhances dynein processivity [12]. These motor proteins may interact with the UcPS substrate proteins to facilitate UcPS. In this study, we identified that specific kinesins, such as KIF1A and KIF5A, along with cytoplasmic dynein, play a critical role in the UcPS using CFTR and the SARS-CoV-2 S protein as substrate cargo proteins.

Cell culture, plasmids, siRNAs, and transfection

HeLa and HEK293 cells were maintained in Dulbecco’s modified Eagle medium (Gibco, #11995-065) containing 10% fetal bovine serum (Gibco, #26140-079) and 1% penicillin/streptomycin (Gibco, #15140-122). The cells were grown in a 5% CO₂ incubator at 37°C.

pCMV-CFTR and pcDNA3.1-Spike have been described previously [4,6]. The coding region of ARF1-Q71L-HA was synthesized and inserted into pcDNA3. eGFP-KIF1A was a gift from Juan Bonifacino (Addgene plasmid # 172206; http://n2t.net/addgene:172206; RRID:Addgene_172206) pCMV-KIF5A-FLAG was commercially purchased (Sino biological, HG19049-CF).

Control (scrambled) siRNA and siRNAs targeting each gene were purchased from Bioneer (AccuTarget Genome-wide Predesigned siRNAs; KIF1A gene ID:547, KIF13A gene ID:63971, KIF13B gene ID:23303, DHC1 gene ID:1778, DIC1 gene ID:1780, DLIC1 gene ID:51143) and Dharmacon (ON-TARGET plus Human KIF5A siRNA gene ID:3798).

Plasmids were transfected into HeLa and HEK293 cells by using Lipofectamine LTX Reagent (Invitrogen, #15338-100), and siRNAs were transfected into HEK293 cells by using RNAiMAX transfection reagent (Invitrogen, #13778-150), according to the manufacturer’s protocol.

Antibodies

Anti-CFTR (Alomone Labs, #ACL006), anti-aldolase A (Santa Cruz, #sc-390733), anti-HA (Cell Signaling Technology, #2367), anti-SARS-CoV-2 S2 (GeneTex, #GTX632604), anti-KIF1A (Novus, #NBP1-80033), anti-KIF5A (Santa Cruz, #sc-376452), anti-KIF15 (Santa Cruz, #sc-100948), anti-DHC1 (Santa Cruz, #sc-514579), anti-DYKDDDDK (Cell Signaling Technology, #8146) anti-α-tubulin antibody (Invitrogen, #MA1-80017) were purchased commercially.

RNA extraction and quantitative polymerase chain reaction (qPCR)

RNA was extracted from the cells using an RNA extraction kit (Bioneer, #K-3140), according to the manufacturer’s protocol. cDNA synthesis from the extracted RNA was performed using cDNA EcoDry Premix (Takara, #639549), according to the manufacturer’s protocol.

The primer sequences used for qPCR were as follows: GAPDH, forward primer 5′- GTC TCC TCT GAC TTC AAC AGC G-3′, reverse primer 5′- ACC ACC CTG TTG CTG TAG CCA A -3′, KIF1A, forward primer 5′- GCC GAC ATC TTC TGC CAG TTC A-3′, reverse primer 5′-GAA GGA CTT GGT CAC CTC CAC T-3', KIF5A, forward primer 5′- AGG AGA AGA GCC AGC AGA ACC A-3′, reverse primer 5′-GT TTT CGC TGG TGT CCA CTG A-3', KIF13A, forward primer 5′- ATC CAT CGG CTG TGT AAC TGC C-3′, reverse primer 5′- GCA TCT GAC CAC CTC TCC CTT A-3′, KIF13B, forward primer 5′- CGG TGA TTG TCG CTC CTG AAG T-3′, reverse primer 5′- CCA ATG CTC CTT CGG AAA GAT GC-3′, DHC1, forward primer 5′- GCC ATC AGC AAA GAC CAC CTC T-3′, reverse primer 5′- AAG ACG ATC CAC TGG CGC TTC T-3′, DIC1, forward primer 5′- TCA GCC AAG TCT GGC AAG CAC T -3′, reverse primer 5′- GAG TCC AAG ACA CAA TCC TGC C-3′, DLIC1, forward primer 5′- GCA AAG CAA CCA CCA ACT GCA G-3′, reverse primer 5′- ATG GGT GAC ACG CTG GCA ACA T -3′, ARL8b, forward primer 5′- GAT ACC CAC AGT GGG CTT CAA C-3′, reverse primer 5′- TGA CTC CTC TGC AAT ACC GCT C -3′, SKIP, forward primer 5′- CCG AGA AAA CGA AGA GCA GCT G-3′, reverse primer 5′- TCC ACC AGG TAT GGC TTC TCT G -3′, FYCO1, forward primer 5′- CCA GAA GAT GCT TGC TGA CCT G-3′, reverse primer 5′- AGC CAT CTC TCC TCA GCA CTG A -3′.

Surface biotinylation and immunoblotting

To perform the biotinylation assay, HEK293 cells were first washed twice with cold phosphate-buffered saline (PBS) to remove any unbound proteins. The transmembrane proteins of the cells were then biotinylated with a reagent called Sulfo-NHS-SS-Biotin (Thermo Pierce, #21331) in cold PBS (0.3 mg/ml biotin in PBS) for 30 min on ice in the dark. After the biotinylation reaction, the cells were incubated with a quenching buffer containing 1% bovine serum albumin (BSA) in cold PBS for 10 min in the dark. The cells were then washed three times with cold PBS. The cells were then harvested and lysed with a buffer that ontaining 150 mM NaCl, 20 mM Tris (pH 7.4), 1 mM EDTA, 1% (v:v) NP40, 0.5% (v:v) sodium deoxycholate, and a protease inhibitor (Roche, #04693159001). The harvested cells were homogenized with a sonicator for 20 sec and centrifuged at 13,200 rpm for 20 min at 4°C. The supernatant of the lysate was collected, and 300 µg of the lysate was incubated at 4°C overnight with 300 µl of 10% streptavidin agarose resin (Thermo Pierce, #20349). The biotinylated protein-bound resin was centrifuged and washed four times with lysis buffer. Biotinylated proteins were eluted with 2× sodium dodecyl sulfate (SDS) sample buffer containing dithiothreitol (0.02 g/ml) at 38°C for 40 min and then separated by SDS-polyacrylamide gel electrophoresis (PAGE). The separated proteins were transferred to a nitrocellulose membrane, and the membrane was blotted with the appropriate primary antibodies and HRP-conjugated secondary antibodies in 5% skim milk.

Immunocytochemistry

HeLa cells were grown on 18 mm round coverslips and fixed using 4% paraformaldehyde (diluted in PBS) for 7 min at room temperature. Following fixation, the cells underwent incubation with a PBS solution containing a 2% BSA solution and 0.1% Triton X-100 at room temperature for 20 min. After the blocking and permeabilization processes, primary antibodies were applied and allowed to incubate for 1 h, followed by three washes with PBS solution. Subsequently, the cells were treated with secondary antibodies conjugated to fluorophores, with another three rounds of PBS solution washing. DAPI or Hoechst 33342 solution was then applied to the cells for 5 min at room temperature, followed by two washes with PBS solution. Finally, the cells on coverslips were mounted onto glass slides, and images were captured using a laser scanning confocal microscope (LSM 780; Carl Zeiss). The following antibodies were used for staining: anti-CFTR clone M3A7 (Millipore, #05-583), anti-HA (Cell Signaling Technology, #2367), Alexa Fluor 488 goat anti-mouse (Thermo Fisher Scientific, # A-11001), Alexa Fluor 568 goat anti-rabbit (Thermo Fisher Scientific, # A-11036).

Statistical analysis

For Western blot quantification, densitometric analyses were performed using Multi Gauge V3.0 software. The results of multiple experiments are presented as the mean ± standard error of the mean (SEM). Statistical analysis was performed using two-sided Student’s t-tests or one-way analysis of variance followed by Tukey’s multiple comparison test, using GraphPad Prism 8 (GraphPad Software, Inc.). p < 0.05 was considered statistically significant.

KIF1A and KIF5A participate in the UcPS, but not in the conventional protein secretion, of CFTR

In a previous study, we have shown that recycling endosomes and KIF5A play a critical role in the UcPS of ∆F508-CFTR induced by ARF1-Q71L [7]. Overexpression of dominant-inhibitory ARF1 mutant ARF1-Q71L blocks the ER-to-Golgi pathway and induces unconventional trafficking of ∆F508-CFTR [3]. The KIF5B (kinesin-1) and KIF1A (kinesin-3) proteins, members of the kinesin motor superfamily that transport cargo vesicles toward the plus end of microtubules, play crucial roles in orchestrating the directed movement of late endosomes and lysosomes toward the cellular periphery [13]. These two motor proteins employ a multi-subunit complex named BORC as a master regulator to achieve this transport, highlighting a conserved mechanism for regulating organelle positioning. From this background, we first assessed the knockdown effect of KIF5A, one of three homologous kinesin-1 motor proteins (KIF5A, KIF5B and KIF5C) to reaffirm the previous result and kinesin-3 motor KIF1A on the unconventional secretion of ∆F508-CFTR.

To investigate how silencing KIF1A or KIF5A influences the unconventional secretion of ∆F508-CFTR induced by ARF1-Q71L, we employed a surface biotinylation assay. The results show that knockdown of KIF1A and KIF5A genes suppressed the unconventional secretion of core-glycosylated ∆F508-CFTR (band B, Fig. 1A, B). Next, a surface biotinylation assay was performed to assess the knockdown effect of KIF5A and KIF1A on conventional secretion of wild-type (WT) CFTR (Fig. 1C, D). The silencing of the KIF1A and KIF5A did not affect the conventional secretion of the complex glycosylated form (band C) of WT-CFTR. The qPCR analysis of KIF1A and KIF5A mRNAs confirmed the gene silencing effects of siRNAs against these genes (Fig. 1E).

Figure 1. KIF1A and KIF5A are involved in the unconventional trafficking of ∆F508-CFTR but not in the conventional trafficking of WT-CFTR. Effects of KIF1A and KIF5A gene silencing on the unconventional trafficking of ∆F508-CFTR (A) and conventional trafficking of WT-CFTR (C). Cell surface biotinylation assay was conducted using HEK293 cells transfected with control (scrambled) siRNA or target siRNAs (50 nM each, 48 h) together with plasmids encoding ∆F508/WT-CFTR (24 h). ER-to-Golgi transport was inhibited by ARF1Q71L-HA overexpression in some cells. Surface-proteins versus cell lysates were quantified respectively (B, D). (E) Efficacy of each target gene knockdown. Bar graph data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by Tukey’s multiple comparison test. b, band B (core-glycosylated) CFTR; c, band C (complex-glycosylated) CFTR; CFTR, cystic fibrosis transmembrane conductance regulator; WT, wild-type; ER, endoplasmic reticulum; ns, not significant. **p < 0.01.

In the following step, our attention was directed towards kinesin-3 motor KIF13, a protein known for its involvement in intracellular transport and transportation of endosomes. KIF13A has been observed to localize to various tubular endosomes, facilitating the recycling of cargo and transport to developing melanosomes [14]. Furthermore, KIF13A is known to form either homodimers or heterodimers with KIF13B on early/sorting endosomes, leading to the formation of recycling endosomes [15]. In neurons, KIF13B has also been documented to interact with PIP₃ through centaurin-α1, transporting PIP₃-containing vesicles to the tips of axons, covering a long-distance journey [16].

In the same manner as KIF1A and KIF5A, we investigated the knockdown effect of KIF13A and KIF13B using surface biotinylation assay. The results show that knockdown of KIF13 has no effect on either unconventional (Fig. 2A, B) and conventional (Fig. 2C, D) secretion of CFTR. The qPCR analyses of KIF13A and KIF13B mRNAs confirmed the gene silencing effects of siRNAs against these genes (Fig. 2E).

Figure 2. KIF13A and KIF13B are not involved in both unconventional and conventional trafficking of ∆F508-CFTR. Effects of KIF13A and KIF13B gene silencing on the unconventional trafficking of ∆F508-CFTR (A) and the conventional trafficking of WT-CFTR (C). Cell surface biotinylation assay was conducted using HEK293 cells transfected with control (scrambled) siRNA or target siRNAs (50 nM each, 48 h) together with plasmids encoding ∆F508/WT-CFTR (24 h). ER-to-Golgi transport was inhibited by ARF1Q71L-HA overexpression in some cells. Surface-proteins versus cell lysates were quantified respectively (B, D). (E) Efficacy of each target gene knockdown. Bar graph data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by Tukey’s multiple comparison test. b, band B (core-glycosylated) CFTR; c, band C (complex-glycosylated) CFTR; CFTR, cystic fibrosis transmembrane conductance regulator; WT, wild-type; ER, endoplasmic reticulum; ns, not significant. **p < 0.01.

Motor proteins interact directly with adaptor proteins which act as molecular bridges between cargo and the cytoskeleton [17]. Some adaptor proteins are known to interact with KIF5A or KIF1A to transport axonal endo-lysosomes, which include SKIP, Arl8, and FYCO1 [18]. SKIP acts as an adaptor that connects lysosomes to kinesin-1 [19]. The BORC–Arl8–SKIP–kinesin-1 complex helps to couple lysosomes to kinesin-1, which allows lysosomes to be transported along microtubules [20]. The small GTPase Arl8, a member of Arf-like (Arl) family of proteins, has two paralogs, Arl8a and ARL8b in vertebrates. ARL8b recruits kinesin-1 to lysosomes, thereby facilitating the movement of lysosomes toward the cell periphery [21]. FYCO1 is a Rab7 effector that promotes plus-end directed lysosome transport [22]. It helps recruit kinesin-1, a motor protein, to endo-lysosomes, which allows endo-lysosomes to be transported along microtubules [23]. Based on this information, we examined the effect of knockdown of these adaptor proteins on unconventional transport of ∆F508-CFTR (Fig. 3A). When ARL8b was silenced, the levels of cytosolic ∆F508-CFTR protein were significantly reduced, hindering the accurate determination of CFTR cell surface trafficking (Fig. 3B). Therefore, it was excluded from the analysis of its effect on UcPS. Notably, knockdowns of FYCO1 and SKIP significantly reduced the UcPS of ∆F508-CFTR (Fig. 3C), implying that these adaptor proteins may participate in the KIF5A- or KIF1A-mediated trafficking of CFTR-containing UcPS vesicles. The qPCR analysis of FYCO1, SKIP, and ARL8b mRNAs confirmed the gene silencing effects of siRNAs against these genes (Fig. 3D).

Figure 3. FYCO1 and SKIP, acting as adaptor proteins, are involved in the unconventional trafficking of ∆F508-CFTR. Effects of SKIP, ARL8b, and FYCO1 gene silencing on the unconventional trafficking of ∆F508-CFTR. (A) Cell surface biotinylation assay was conducted using HEK293 cells transfected with control (scrambled) siRNA or target siRNAs (100 nM each, 48 h) together with plasmids encoding ∆F508-CFTR (24 h). ER-to-Golgi transport was inhibited by ARF1Q71L-HA overexpression in some cells. (B) ∆F508-CFTR in lysate were quantified respectively. (C) Surface-proteins versus cell lysates were quantified respectively. (D) Efficacy of each target genes knockdown. Bar graph data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by Tukey’s multiple comparison test. CFTR, cystic fibrosis transmembrane conductance regulator; ER, endoplasmic reticulum. *p < 0.05, **p < 0.01.

Subsequently, the study was extended to assess the effects of suppressing these adaptor proteins on conventional transport, aiming to determine whether they exclusively regulate only unconventional transport or both unconventional and conventional transport (Supplementary Fig. 1A). Decreased levels of cytosolic WT-CFTR were observed upon the suppression of ARL8b, consistent with the findings in ∆F508-CFTR (Supplementary Fig. 1B). Notably, the knockdown of FYCO1 and SKIP did not affect the conventional secretion of WT-CFTR (Supplementary Fig. 1C). This suggests that FYCO1 and SKIP exclusively play crucial roles as adaptor proteins in UcPS.

Cytoplasmic dynein is involved in the UcPS of CFTR, and dynein heavy chain (DHC) is the key subunit among the dynein motor subunits

We then investigated the role of dynein, which transports cargo vesicles toward the minus end of the microtubules, in the UcPS of ∆F508-CFTR. Among the two cytoplasmic dynein molecules, dynein-1 and -2 [10], dynein-1 is mainly involved in intracellular trafficking events, such as mitosis, cell migration, and intracellular transport, whereas dynein-2 mediates cilia biogenesis and signaling [24]. Dynein-1 consists of a homodimer of DHC, dynein intermediate chain (DIC), dynein light intermediate chain (DLIC), and three dynein light chain (DLC) families [25].

We evaluated the knockdown effect of each subunit of dynein-1: DHC1, DIC1, and DLIC1. The silencing of DHC1 and DLIC1 reduced the unconventional secretion of ∆F508-CFTR, with DHC1 exhibiting the most significant inhibitory effect among them (Fig. 4A, B). However, the silencing of any dynein subunit did not affect the conventional secretion of complex glycosylated form (band C) of WT-CFTR (Fig. 4C, D). The qPCR analysis of DHC1, DIC1, and DLIC1 mRNAs confirmed the gene silencing effects of siRNAs against these genes (Fig. 4E).

Figure 4. Dynein is involved in the unconventional trafficking of ∆F508-CFTR but not in the conventional trafficking of WT-CFTR. Effects of each dynein subunit; DHC1, DIC1 and DLIC1 gene silencing on the unconventional trafficking of ∆F508-CFTR (A) and conventional trafficking of WT-CFTR (C). Cell surface biotinylation assay was conducted using HEK293 cells transfected with control (scrambled) siRNA or target siRNAs (100 nM each, 48 h) together with plasmids encoding ∆F508/WT-CFTR (24 h). ER-to-Golgi transport was inhibited by ARF1Q71L-HA overexpression in some cells. Surface-proteins versus cell lysates were quantified respectively (B, D). (E) Efficacy of each target gene knockdown. Bar graph data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by Tukey’s multiple comparison test. b, band B (core-glycosylated) CFTR; c, band C (complex-glycosylated) CFTR; CFTR, cystic fibrosis transmembrane conductance regulator; WT, wild-type; ER, endoplasmic reticulum; ns, not significant. **p < 0.01.

Morphological features reveal the knockdown effects of key motor proteins in UcPS

We also investigated the impact of motor protein silencing on the cell surface presentation of ΔF508-CFTR through immunofluorescence analysis, focusing on the distribution of surface and intracellular CFTR. For the morphological assay, HeLa cells were utilized instead of HEK293 cells due to their stronger attachment to coverslips and greater resistance to cell loss and disruption during immunostaining procedures [26,27]. In control cells, ΔF508-CFTR proteins were predominantly confined to the perinuclear region (Fig. 5A). Co-expression with ARF1-Q71L-HA induced the cell surface expression of CFTR proteins in cells expressing ΔF508-CFTR (Fig. 5B). Notably, the administration of siRNA (100 nM) targeting KIF1A, KIF5A, and DHC1 reduced the UcPS of ΔF508-CFTR, with ΔF508-CFTR proteins concentrated in the perinuclear ER area (Fig. 5C–E).

Figure 5. Cell surface expression of ∆F508-CFTR is suppressed by knockdown of specific motor proteins. Immunofluorescent analysis was conducted on ΔF508-CFTR in cells subjected to siRNA treatment targeting KIF1A, KIF5A, and DHC1. Hela cells underwent sequential transfections with the indicated siRNAs (48 h) and plasmids (24 h), followed by immunostaining using anti-CFTR (green, Alexa Fluor 488) and anti-HA (red, Alexa Fluor 568) antibodies. The anti-HA immunofluorescence signal indicates the expression level of ARF1-Q71L. (A) ΔF508-CFTR was predominantly observed in perinuclear regions. (B) Co-transfection with ARF1Q71L in Hela cells resulted in an overall intracellular distribution and cell-surface expression of ΔF508-CFTR. Arrows represent the cell-surface rescue of ΔF508-CFTR by ARF1-Q71L. (C) Treatment with siRNAs against KIF1A (siKIF1A), (D) treatment with siRNAs against KIF5A (siKIF5A), and (E) treatment with siRNAs against DHC1 (siDHC1) abolished the ARF1Q71L-induced cell surface expression of the mutant CFTR. Three independent experiments yielded consistent results. Scale bar: 10 µm. CFTR, cystic fibrosis transmembrane conductance regulator.

Specific kinesins and dyneins are involved in the UcPS of the SARS-CoV-2 S protein

We conducted additional experiments on SARS-CoV-2 S, which is known to be transported via a Golgi-independent pathway [4], to examine whether specific motor proteins that have been identified in the UcPS pathway of ∆F508-CFTR are also involved in the UcPS pathway of other cargoes. The S protein of SARS-CoV-2 consists of two subunits, S1 which is responsible for receptor binding and S2, which is for membrane fusion [28]. The S protein is cleaved into S1 and S2 fragments by furin which is enriched in the Golgi apparatus [29].

To assess the effects of specific motor proteins on the UcPS of the S protein, we conducted a surface biotinylation assay with silencing of kinesins and subunits of dynein involved in the UcPS of ∆F508-CFTR. Interestingly, in line with previous results, the silencing of KIF1A and KIF5A (Fig. 6) and that of DHC1, DIC1, and DLIC1 (Fig. 7) suppressed the Golgi-bypassed unconventional secretion of uncleaved S protein but not the Golgi-mediated conventional secretion of cleaved S protein.

Figure 6. KIF1A and KIF5A are involved in the unconventional trafficking of S protein but not in the conventional trafficking of S protein. Effects of KIF1A and KIF5A gene silencing on the unconventional trafficking of S protein (A) and conventional trafficking of S protein (C). Cell surface biotinylation assay was conducted using HEK293 cells transfected with control (scrambled) siRNA or target siRNAs (50 nM each, 48 h) together with plasmids encoding S protein (24 h). ER-to-Golgi transport was inhibited by ARF1Q71L-HA overexpression in some cells. Surface-proteins versus cell lysates were quantified respectively (B, D). Bar graph data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by Tukey’s multiple comparison test. UC, Golgi-bypassed uncleaved form of S protein; C, Golgi-dependent cleaved form of S protein; ER, endoplasmic reticulum; ns, not significant. **p < 0.01.

Figure 7. Dynein is involved in the unconventional trafficking of S protein but not in the conventional trafficking of S protein. Effects of each dynein subunit; DHC1, DIC1, and DLIC1 gene silencing on the unconventional trafficking of S protein (A) and conventional trafficking of S protein (C). Cell surface biotinylation assay was conducted using HEK293 cells transfected with control (scrambled) siRNA or target siRNAs (100 nM each, 48 h) together with plasmids encoding S protein (24 h). ER-to-Golgi transport was inhibited by ARF1Q71L-HA overexpression in some cells. Surface-proteins versus cell lysates were quantified respectively (B, D). Bar graph data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by Tukey’s multiple comparison test. UC, Golgi-bypassed uncleaved form of S protein; C, Golgi-dependent cleaved form of S protein; ER, endoplasmic reticulum; ns, not significant. **p < 0.01.

Intracellular protein trafficking within the cell is essential for maintaining cellular function and homeostasis. Microtubules, cytoskeletal elements acting as intracellular highways, play a crucial role in this process by facilitating the long-distance transport of various cargoes. Specialized motor proteins, such as kinesins and dyneins, navigate along these microtubule tracks and deliver vesicles and organelles containing essential cargo to their designated destinations. While significant progress has been made in identifying motor proteins responsible for specific cargo transport in the Golgi-mediated conventional secretion pathway, the transport of unconventionally secreted proteins, which bypasses the Golgi apparatus, remains poorly understood. KIF1A has been reported for its role in the transport of dense-core vesicles and a subset of synaptic proteins [30], and KIF5A is known for its relevance of intracellular transport such as TI-VAMP vesicles [31], collagen [32], and GABA-A receptor [33]. In our investigation, we identified that these two kinesins play a role in the unconventional secretion of ∆F508-CFTR and S protein.

Since knockdown of KIF1A and KIF5A reduced the UcPS of CFTR and S protein, these proteins appear to be involved in the trafficking of CFTR and S protein to the cell periphery as motor proteins that mediate the plus-end trafficking of microtubules. Supporting this possibility, our results also revealed that the knockdown of FYCO1 and SKIP, kinesin-associated adaptor proteins participating in vesicular trafficking of late endosomes and lysosomes, led to a reduction in the unconventional secretion of ∆F508-CFTR and S protein. On the other hand, KIF13A and KIF13B, which were experimentally targeted as being involved in the formation of tubular endosomes, had no effect on the unconventional secretion of ∆F508-CFTR and S protein when subjected to siRNA knockdown experiments. Notably, previous reports indicated that SARS-CoV-2 utilizes lysosomes for egress instead of the conventional secretion pathway [34]. The vesicle type responsible for carrying ∆F508-CFTR and S protein during the UcPS pathway remains unclear. In this scenario, we wonder whether a common vesicle might transport both ∆F508-CFTR and S protein from the ER to the cell surface, sharing previously identified motor and adaptor proteins during the UcPS pathway. The precise molecular mechanisms underlying the transport of UcPS cargo from the ER to the cell surface by kinesin and dynein motors, bypassing the Golgi apparatus, as well as the selective cargo recognition process, remain the subject of future studies. Furthermore, additional gain-of-function experiments, such as exogenous supplementation of KIF1A and KIF5A, may complement the present loss-of-function experiments.

The findings, demonstrating that the silencing of cytoplasmic dynein hinders the unconventional transport of ∆F508-CFTR, suggest that the process of cargo movement from the ER to the cell surface is not solely characterized by centripetal (outward) movement alone. Instead, it involves dynein, which facilitates movement toward the minus end. In essence, as the cargo traverses the UcPS pathway, it has the potential for it to follow specific segments along the minus end axis. Alternatively, one could infer that the coordinated interplay between kinesin and dynein plays a pivotal role in facilitating cargo movement, a hypothesis supported by previous studies elucidating the complexity of intracellular transport mechanisms [35].

Dynein, a motor protein facilitating the movement of organelles and cargo toward the cell center by walking along microtubules in the minus-end direction, is crucial for various cellular processes [25], ranging from simple cargo transport [36] to axonal transport [37], cell division, and MTOC positioning [38]. Dynein, together with its accessory proteins, assists in the movement of COPII vesicles from the ER to the Golgi complex. This occurs via a minus-end-directed mechanism from dispersed ER exit sites to the Golgi complex [39]. However, since UcPS involves a process that bypasses the Golgi complex, vesicular trafficking from the ER to a certain platform might utilize dynein-mediated mechanisms for the UcPS pathway, similar to that of COPII vesicles to the Golgi complex. It has been proposed that cargo trafficking from multivesicular bodies to recycling endosomes might contribute to the unconventional secretion of ∆F508-CFTR, as evidenced by the significant reduction in ARF1Q71L-induced UcPS of ∆F508-CFTR upon knockdown of RAB8A, a protein involved in recycling endosomes [7]. Considering the discovery in this study that specific motor proteins (KIF1A, KIF5A) and adaptor proteins (FYCO1, SKIP) associated with late endosomes are involved in UcPS, we cautiously speculate that the "certain platform" mentioned could potentially be a form of endosomes. The compartment for unconventional protein secretion (CUPS), implicated in UcPS during cellular starvation in yeast [40] may have a counterpart in mammalian cells, possibly also serving as the intermediate platform between the ER and the cell surface either.

In addition to direct vesicular trafficking, these motor proteins may also be involved in ER dynamics, with ∆F508-CFTR being transported along with the changing shape of the ER as it translocates to the cell periphery. Kinesin-1 and cytoplasmic dynein are known to be involved in the extension of ER tubules [41]. It is noteworthy that a recent study has demonstrated that anchoring lysosomes and the growth tips of the ER can induce elongation and connection of ER tubules [42]. Moreover, KIF5A and KIF1A, both involved in the UcPS of ∆F508-CFTR and S protein in this study, also play a role in lysosome positioning [13,43,44]. Therefore, it is also possible that KIF5A and KIF1A facilitate the movement of lysosomes to the cell periphery, triggering the redistribution of the ER and the extension of ER structures to the cell periphery.

Our results provide insight into the previously unexplored functions of distinct kinesin and dynein motor proteins within the UcPS pathway of CFTR and SARS-CoV-2 S protein. We have pinpointed specific kinesin isoforms that mediate the transport of UcPS cargo proteins directly from the ER to the plasma membrane, bypassing the Golgi apparatus. Furthermore, our data also demonstrate the involvement of dynein in the UcPS pathway, suggesting a potential cooperative role between kinesin and dynein. Further research is warranted to elucidate the precise mechanisms underlying the selective cargo recognition and coordinated action of both kinesin and dynein in this unconventional secretion pathway. Once delivered to the cell surface, core-glycosylated CFTR and pendrin are known to retain their ability to transport anions [3,45]. Therefore, the selective activation of UcPS without inducing hazardous cellular stress represents a potential therapeutic strategy for the treatment of diseases arising from defects in the cell surface presentation of the misfolded proteins. It is anticipated that further studies on KIF1A and KIF5A, as discovered in this research, could lead to the development of new treatments that activate UcPS, thereby providing another therapeutic option for diseases such as cystic fibrosis and Pendred syndrome. Conversely, the identification of methods to selectively inhibit UcPS could provide new treatment options for conditions such as SARS-CoV-2 infection and inflammatory diseases associated with abnormal IL-1β secretion [34,46].

Supplementary data including one figure can be found with this article online at https://doi.org/10.4196/kjpp.2024.28.5.435

The authors express their gratitude to individuals who contributed to this work.

This research was supported by a grant of the MD-Phd/Medical Scientist Training Program through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea. This research was supported by grants (2021R1I1A1A01049255, 2022R1A2C3002917, and 2022R1A4A1031336) from the National Research Foundation (NRF) of Korea funded by the Ministry of Education and the Ministry of Science and ICT, Republic of Korea.

The authors declare no conflicts of interest.