Brought to you by Portland Press Ltd.
Published on behalf of the International Federation for Cell Biology
Cancer Cell death Cell cycle Cytoskeleton Exo/endocytosis Differentiation Division Organelles Signalling Stem cells Trafficking
Cell Biology International (2006) 30, 885–894 (Printed in Great Britain)
Kinectin participates in microtubule-dependent hormone secretion in pancreatic islet β-cells
Ji‑Zhong Bai, Yu Mon and Geoffrey W. Krissansen*
Department of Molecular Medicine & Pathology, Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand


Abstract

Kinectin (KNT) is a candidate membrane receptor for kinesin in the movement of intracellular organelles along microtubules. Isoforms of KNT exist containing different combinations of six small (residues 23–33) variable domains (vd) vd1–6 within the C-terminus. Here we investigate a role for KNT and its isoform KNTvd4 in the transport of amylin and insulin-containing secretory vesicles in the pancreatic islet β-cell line RINm5F. KNTvd4 lacks vd4 that forms the kinesin-binding domain, and hence its role in the cell is an enigma. We report that amylin-containing vesicles also contained insulin, and exhibited microtubule, and small G-protein-dependent secretion. Knockdown of KNT by small interference RNA (siRNA) inhibited amylin expression and secretion. In contrast, recombinant KNTvd4 overexpressed in RINm5F cells associated with amylin-containing vesicles and inhibited amylin secretion, but had no discernible affect on amylin expression. The data suggests that both KNT and KNTvd4 participate in microtubule-dependent secretion of amylin in islet β-cells.


Keywords: Kinectin, Intracellular transport, Amylin, Hormone secretion, Islet β-cells.

*Corresponding author. Tel.: +64 9 373 7599x86280; fax: +64 9 373 7492.


1 Introduction

Kinectin (KNT) is a 160kDa intracellular membrane receptor essential for kinesin-dependent movement of intracellular organelles (reviewed by Kumar et al., 1995; Burkhardt, 1996; Ong et al., 2000; Toyoshima et al., 1992). It was cloned from chick embryonic brain (Yu et al., 1995), human lymphocytes (Futterer et al., 1995; Print et al., 1994), mouse testis and spleen (Leung et al., 1996), and fox testis (Xu et al., 2002), and predicted to possesses a short N-terminal hydrophilic domain, a transmembrane (TM) domain, and a large C-terminal coiled-coil domain (Print et al., 1994; Yu et al., 1995). KNT has been localized to the endoplasmic reticulum (ER) of chick fibroblasts, astroglia, Schwann cells, and dorsal root ganglion cells (Futterer et al., 1995; Toyoshima et al., 1992), lysosomes of REF-52 and COS-7 cells (Vignal et al., 2001), the purified phagosomes of chicken monocytes (Blocker et al., 1997), and melanosomes of human melanocytes (Vancoillie et al., 2000).

An antibody against the cytoplasmic domain of KNT blocked kinesin-binding to microsomes, and inhibited kinesin-driven plus-end, and dynein-powered minus-end directed vesicle motility (Kumar et al., 1995). KNT progressively accumulates on maturing phagosomes that display high motility, and whose motility is bi-directionally inhibited by KNT fragments derived from the central stalk domain, but not the C-terminus of chicken KNT (Blocker et al., 1997). It interacts with the cargo-binding tail of kinesin at its head-binding and myosin-Va-binding domains via its C-terminus (residues 1188–1288) (Ong et al., 2000). The kinesin-binding domain of KNT enhances microtubule-stimulated kinesin ATPase activity, and over-expression of this domain disrupts kinesin-dependent lysosome transport (Ong et al., 2000). KNT mediates microtubule-dependent RhoG activity by a direct interaction of RhoG with its central domain (Vignal et al., 2001), whereas the C-terminus of KNT (residues 630–935) interacts with small G-proteins including RhoA, Rad and Cdc42 (Hotta et al., 1996).

KNT exists as two protein species of 160 and 120-kDa on SDS-PAGE (Kumar et al., 1998; Vignal et al., 2001). The 120kDa species lacks the N-terminal 232 residues of the 160kDa full-length form, and has been proposed to form a heterodimer with full-length KNT (Kumar et al., 1998). The 160kDa species is mainly concentrated in the ER, anchored there via its transmembrane domain and is essential for ER membrane extension (Santama et al., 2004). In contrast the 120kDa species is specifically associated with mitochondria, and its interaction with kinesin influences mitochondrial dynamics. Multiple differentially spliced isoforms of KNT appear to exist containing different combinations of at least six small (residues 23–33) variable domains (vd) vd1–6 within the C-terminus of KNT (Print et al., 1994; Leung et al., 1996; Xu et al., 2002; Santama et al., 2004). The minimum kinesin-interacting domain of KNT lies between residues 1188 and 1288, shared by inserts V3 and V4 (Ong et al., 2000). An isoform that contains vd1, vd2, vd3, vd5, and vd6, but lacks vd4, fails to interact with uKHC, and alternatively functions to anchor translation elongation factor-1δ to the ER (Ong et al., 2003). KNT isoforms may display alternative cellular and tissue distributions, and possess distinct functions (Santama et al., 2004). One isoform containing vd2 was found overexpressed in cancerous tissues (Wang et al., 2004). The role of KNT as a universal kinesin receptor for vesicle motility has been challenged by the finding that the trafficking of lysosomes, phagosomes, and mitochondria is unimpaired in mice deficient in KNT (Plitz and Pfeffer, 2001).

Here we address the role of KNT, and a kinesin-binding domain deficient (vd4-deficient) KNT isoform, in the trafficking of amylin-containing secretory vesicles in a pancreatic islet β-cell line. While kinesin has been shown to play an essential role in insulin secretion (Meng et al., 1997; Varadi et al., 2002a,b), we report here for the first time that KNT is associated with both amylin and insulin-containing vesicles, and plays an essential role in mediating hormone secretion.

2 Materials and methods

2.1 Antibodies and plasmids

Living colour pER1S2/EGFP cDNAs encoding enhanced green fluorescent protein (GFP), and its derivative cyan fluorescent protein (CFP) were obtained from ClonTech Laboratories (Palo Alto, CA). Expression vectors, pcDNA3.1/Neo(+) and pcDNA3.1/Bsd(+), and a rabbit polyclonal anti-GFP antibody (Living Colour A. v. Peptide) were purchased from ClonTech Laboratories. Peroxidase-conjugated donkey anti-rabbit IgG, and a mouse monoclonal anti-α-tubulin antibody were obtained from Sigma (Poole, UK). A guinea pig polyclonal anti-insulin antibody was purchased from CALTAG Laboratories (Burlingame, CA). TRITC-conjugated sheep anti-guinea pig IgG, and Alexa Fluor 594 donkey anti-mouse IgG were from Molecular Probes (Eugene, USA). A rabbit polyclonal antibody raised against a λ lysogen expressing nucleotides 1378–3304 of human KNT clone L25616 has been described previously (Print et al., 1994).

2.2 Construction of expression plasmids

PCR primers used in the construction of plasmids are detailed in Table 1. To construct an expression vector encoding KNTvd4CFP, a 3610bp BamHI-EcoRI fragment encompassing nt 825–4416 of human KNT (GenBank Accession Number L25616) including the stop codon (Print et al., 1994) was subcloned from pUC18 into the mammalian expression vector pcDNA6/V5-hisC (Invitrogen, Carlsbad, CA) to generate the vector pcDNA6/Bsd-KNT©. A KNT shuttle cassette was generated in order to fuse ECFP coding sequences in-frame between nt 631 and 632 of human KNT cDNA by overlap PCR. The CFP tag would then reside between amino acid residues 197 and 198 that separate the TM domain from the large coiled-coil domain. The large tandem repeats of the related p180 molecule are naturally inserted by splicing into the same position in p180 (Langley et al., 1998), and hence this arrangement was considered less likely to disrupt the function of the molecule. The shuttle cassette was generated using pUC18-CG1 (Print et al., 1994) as a template and the primer sets #1–#4. The shuttle cassette generated by overlap PCR with the #1 and #4 primers was directionally inserted into the pcDNA6/Bsd-KNT© vector at XbaI (5′) and BamHI (3′) sites to replace the corresponding wild-type sequence. A NheI-XbaI CFP fragment was amplified from the pECFP-C1 plasmid (Clontech) using primers #5 and #6. It was inserted into the KNT cassette that had been digested with NheI to generate full-length KNTvd4CFP, which was authenticated by DNA sequencing.


Table 1.

PCR primers used in the construction of expression plasmids

Primer no.Primer sequence
#1(S)5′-CGCGTCTAGA(XbaI)TCTACCATGGAGTTTTATGAGTC-3′
#2(AS)5′-CTCTTGGCTAGC(NheI)GTGGAGGGGCAATGCTTCT-3′
#3(S)5′-GCCCCTCCACGCTAGC(NheI)CAAGAGACTAAACAAGAAAG-3′
#4(AS)5′-GCATTTTCTTTGTCAGTTTCGG-3
#5(S)5′-GGTCTAGAGCTAGC(NheI)GTGAGCAAGGGCGAGGAGCTG-3′
#6(AS)5′-GGATGCTCTAGA(XbaI)CTTGTACAGCTCGTCCATGCCG-3′
#7(S)5′-CTAGAGAACCCACTGCTTACTGGC-3′
#8(AS)5′CCCTCGAG(XhoI)ATCTGCAGAATTC(EcoRI)ATATGTATTGGATCCCACATTGGTTGGTGG-3′
#9(S)5′-TATGAATTC(EcoRI) TGCAGATCTCGAG(XhoI)GAAGAGGAATGTGGCAGAGGATC-3′
#10(AS)5′-AAGGGAAGAAAGCGAAAGGAGCGGG-3′
#11(S)5′-GCGAGATCTGAATTC(EcoRI)GTGAGCAAGGGCGAGGAGCTGTTC-3′
#12(AS)5′-CTTCCCTCGAG(XhoI)ATCTCTTGT ACAGCTCGTCCATGCCGAG-3′


To generate the amylinGFP pcDNA3.1/Neo(+) expression vector, a shuttle cassette was generated to fuse the 3′-end of amylin in-frame to the coding sequence of GFP. The cassette was generated by overlap PCR using pcDNA3.1/Neo(+)-rat amylin (GenBank Accession Number J04544, kindly provided by Drs Garth Cooper and Shaoping Zhang from School of Biological Sciences, the University of Auckland, Auckland) as a template and the following primer sets: #7 plus #8, and #9 plus #10. The overlap PCR product was inserted into HindIII and XbaI restriction sites in the pcDNA3.1/Neo(+)-amylin vector replacing the original amylin cDNA. The modified pcDNA3.1/Neo(+)-amylin vector was digested with EcoRI and XhoI to insert an EcoRI-XhoI GFP fragment that had been amplified from the pER1S2/EGFP vector (Clontech) using primers #11 and #12. The coding sequences of each vector were verified by automated DNA sequencing (Department of Molecular Medicine & Pathology DNA Sequencing Facility, the University of Auckland, Auckland).

2.3 Generation of stable amylinGFP and KNTvd4CFP-expressing cell lines

Insulin-secreting RINm5F cells (passage number 30–45) were cultured at 37°C in air/CO2 (19:1) in RPMI 1640 medium supplemented with 10% (v/v) fetal calf serum, 290μg/ml l-glutamine, 100units/ml penicillin, and 100μg/ml streptomycin, as previously described (Bai et al., 1999). Cells cultured overnight in 24-well plates were transfected at 80% confluence with the amylinGFP plasmid or control vectors using Lipofectamine™ 2000 (Invitrogen Life Technologies Inc., CA). Clonal cell lines were isolated by serial dilution in geneticin selection medium (0.4mg/ml) added 2 days after transfection. Transfectants were analysed by confocal microscopy for autofluorescence and expression of GFP. Those expressing amylinGFP were further transfected with the KNTvd4CFP plasmid, or control vectors encoding free CFP. Stable cell clones were selected with a combination of geneticin (Invitrogen Life Technologies Inc., CA), and blasticidin (InvivoGen, San Diego, CA) at 400, and 16μg/ml, respectively.

2.4 Disruption of the cytoskeleton

Transfectants were incubated in 10 or 20μM nocodazole for 30min at 37°C to depolymerise microtubules. They were preincubated with 20 or 40μM cytochalasin D for 30min at 37°C to disassemble actin filaments. Mock-treated control cells were treated under the same conditions with 20 or 40μM DMSO.

2.5 RT-PCR analysis of expression of KNT isoforms

Total cellular RNA was isolated from RINm5F islet β-cell cultures using the RNeasy Mini kit from Qiagen (Valencia, CA). RT-PCR was performed using the SuperScript™ III one-step RT-PCR system (Invitrogen Life Technologies Inc., CA, USA). The oligonucleotide primers vd2f 5′ ctctgca atgagttaga gtctttgaag 3′ and vd4r 5′ gccttg tgcaaatcat cagctacctt ct 3′ were designed to bracket vd2 and vd4 of rat KNT. The RT-PCR products were analysed on a 2.5% agarose gel, and sequenced.

2.6 Western blot analysis

Transfectants were lysed for 30min at 4°C in lysis buffer containing 50mM Tris–HCl (pH 7.4), 150mM NaCl, 1% (v/v) Triton X-100, 1% (w/v) Na deoxycholate, 1% (w/v) SDS, and protease inhibitors (20μg/ml pepstatin A, 20μg/ml leupeptin, 20μg/ml aprotinin and 1mM PMSF). Protein samples were resolved by SDS-PAGE, and transferred to nitrocellulose membrane. Membranes were blocked overnight at 4°C with 10% (w/v) non-fat milk powder in T-TBS buffer (1% Tween in 20mM Tris–HCl, pH 7.6, 137mM NaCl). They were incubated for 1h at room temperature with either rabbit polyclonal anti-GFP antibody (1:100) which cross-reacts with CFP, rabbit polyclonal anti-KNT antibody (1:50), or mouse anti-tubulin mAb (1:500), and immunoreactivity visualized by incubation for 1h with peroxidase-conjugated goat anti-rabbit IgG (1:10,000) or anti-mouse IgG (1:50,000), respectively, and development by enhanced chemiluminescence (Pierce, Rockford, IL).

To detect the effect that cytoskeletal disruptors have on secretion of amylinGFP, stable transfectants were grown for 20h in 24-well plates at 4×105cells/ml. They were incubated with cytochalasin D (40, 20μM) or nocodazole (20, 10μM) for 30min, with lovastatin (50, 25μM) for 24h at 37°C, or with DMSO (40, 20μM), as indicated. The media was replaced with fresh media without the cytoskeletal modulators, and the cells incubated for a further 3h in 300μl KRB buffer [120mM NaCl, 5mM KCl, 3mM CaCl2, 1.2mM MgSO4, 1mM KH2PO4, 5mM NaHCO3, 10mM HEPES, 0.1% (w/v) BSA (fraction V), and 2.8mM glucose (pH 7.4)]. Cells were removed by centrifugation, and 20μl aliquots of supernatants loaded onto 8–12% (w/v) polyacrylamide SDS gels. Gels were subjected to Western blot analysis as above.

2.7 Immunocytochemistry

Transfectants (2×105) were grown on coverslips in 24-well tissue culture plates, and fixed with 4% paraformaldehyde for 20min at room temperature. The cells on coverslips were mounted on glass slides with Citifluro (Life Technologies, Auckland) and examined using a Leica TC SP2 confocal microscope as described below.

For immunofluorescence assays, cells seeded overnight on coverglasses in RPMI supplemented with 10% FCS, l-glutamine, penicillin, and streptomycin were fixed as above and permeabilized with 0.02% Triton X-100 in PBS for 20min at room temperature. The cells were blocked with 5% BSA in PBS for an hour prior to an overnight incubation with either the guinea pig polyclonal anti-insulin antibody, or mouse monoclonal anti-α-tubulin antibodies. They were incubated for 1h with the secondary antibodies, TRITC-conjugated sheep anti-guinea pig IgG, and Alexa Fluor 594 goat anti-mouse IgG. The coverglasses were mounted in Citifluor supplemented with 1% DAPI for nuclear staining, and examined using a Leica TCS SP2 confocal microscope at 488nm for GFP and 543nm for TRITC or Alexa Fluor 594 for dual-colour determination, and subsequent UV imaging of DAPI-stained nuclei. Images were assembled from confocal data files using Adobe Photoshop software.

2.8 Confocal microscopy

Expression profiles of transfectants were analysed using a Leica TCS SP2 confocal microscope (Leica, Heidelberg, Germany) equipped with high numerical aperture lenses (63×NA 1.32 OIL HC×PL APO CS; 20×NA 0.7 IMM/CORR HC×PL APO CS). GFP-tagged proteins were excited with the 488-nm line of an Ar/ArKr laser at 12% intensity. CFP fusions were excited with the 458-nm line of the same laser at 100% intensity. All the emission lines were collected through the prism spectra-photometer with emission band passes of 500–525nm for GFP, 460–480nm for CFP, and 590–650nm for TRITC or Alexa Fluor 594. The GFP and TRITC or Alexa Fluor 594 emissions were collected simultaneously, and CFP emission was obtained subsequentially to eliminate residual cross-talk between GFP and CFP dual-channel-imaging. Images were acquired in a series of continuous spatial frames using the 63×NA 1.32 OIL HC×PL APO CS oil-immersion objective at a scanning speed of 400Hz with line average 16, and a resolution of 512×512 pixels (8-bits per sample) in XYZ mode. Processing was done with in-house software to align the individual channels (GFP, CFP, TRITC, and Alexa Fluor 594) in multicolour. Finally, the separate channels were pseudo-colour encoded and combined to an RGB sequence with software Photoshop 7.0. All images were acquired at room temperature. The acquisition software used was Leica Confocal Software (LCS, Leica). Adjustment of brightness, contrast, and colour balance were made to whole images without any change to gamma settings. Image analysis to measure colocalization was performed using ImageJ image analysis software (http://rsb.info.nih.gov/ij/).

2.9 siRNA knockdown

The siRNA oligonucleotides siKNT (5′-AATGGAAGCAGAGATAGCTCATT-3′), and siGFP (5′-GGCUACGUCCAGGAGCGCACC-3′) were purchased from Dharmacon (Lafayette, CO). Transfectants singly expressing amylinGFP were seeded into 24-well plates at 1×105cells/ml in culture media. Cells were transfected with Lipofectamine™ 2000 containing 100nM of each siRNA. After a 7h incubation at 37°C, the cell media was removed and transfectants were cultured in normal growth media not containing the siRNA for a further 65h, and then analysed for intracellular levels of endogenous KNT by Western blot analysis. Adobe photoshop gel images were taken to document the results (unnecessary background was removed from the image recording intracellular KNT). Alternatively, the media was again removed and replaced and cells incubated for a further 3h, and then the culture media harvested to measure levels of secreted amylinGFP. Otherwise, the cells were fixed in 4% paraformaldehyde following removal of the media, and examined by confocal microscopy to determine the intracellular distribution of amylinGFP-containing vesicles.

3 Results

3.1 Visualization of the amylin secretion pathway

Rat amylin cDNA was C-terminally fused to an enhanced green fluorescent protein (GFP) tag (Fig. 1A), and stably transfected into rat pancreatic RINm5F islet β-cells. AmylinGFP was expressed throughout the cytosol on vesicles that were mostly uniformly distributed throughout the cytoplasm, and reached the cell membrane (Fig. 2A). Staining of the cells with an anti-insulin antibody revealed that most amylinGFP-containing vesicles expressed insulin, and vice versa many insulin-containing vesicles expressed amylin (Fig. 2A). This result is in accord with a previous demonstration that amylin and insulin are co-secreted from islet β-cells (Moore and Cooper, 1991). Staining of cells with an anti-tubulin antibody revealed that amylinGFP-containing vesicles were distributed within the microtubule network, from the slow growing minus end in the vicinity of nucleus to the fast growing end in the cell periphery (Fig. 2B). Treatment of cells with the microtubule-depolarizing agent nocodazole at 10μM for 30min led to the disintegration of the microtubular network, such that microtubules disappeared and were replaced by a diffuse tubulin staining. AmylinGFP-containing vesicles, which had been evenly distributed, collapsed into large clumps indicating that the intracellular distribution of amylinGFP-containing vesicles in the RINm5F β-cell line is microtubule-dependent (Fig. 2C).


Fig. 1

Schematic representation of rat amylinGFP and KNTvd4CFP expression constructs. (A) Mature rat amylin was C-terminally tagged with GFP. SP, signal peptide; GFP, green fluorescent protein. (B) The KNT isomer KNTvd4 lacks vd2 and vd4, where the latter variable domain forms part of the kinesin-binding site. CFP was fused in-frame between amino acid residues 197 and 198 of KNT, which correlates with the site in the related P180 molecule into which a large tandem repeat domain is inserted. The coiled-coil domain is expanded to illustrate the variable domains in the human KNTvd4 cDNA clone L25616 compared to the human KNT cDNA clone Z22551 that contains all six known variable domains. Structural domains, and amino acid positions are indicated. CFP, cyan fluorescent protein.


Fig. 2

AmylinGFP colocalizes with insulin in vesicles whose distribution is microtubule-dependent. (A) RIN5mF cells expressing amylinGFP (green) were immunostained for insulin (red) with simultaneous DAPI staining of the nucleus (blue), as indicated. The images were merged to colocalize amylin and insulin, where colocalization generates a yellow colour. RIN5mF cells expressing amylinGFP (green) were immunostained for tubulin (red) (B), or were additionally treated with 10μM nocodazole for 30min, which led to loss of the microtubular network (C). The images shown represent single confocal layers across a group of intact cells.



Nocodazole, cytochalasin D, which disassembles filamentous actin, and lovastatin, which inhibits small G-proteins, were examined for their effects on the secretion of amylin. As determined by Western blot analysis, treatment of cells with 10 and 20μM nocodazole inhibited amylinGFP release into the culture media over a 3h period by 63 and 85%, respectively, compared to control cells treated with DMSO (Fig. 3), in accord with the earlier result that nocodazole disrupted the subcellular distribution of amylinGFP-containing vesicles (Fig. 2C). Cytochalasin D inhibited amylinGFP release by 58 and 65% at 20 and 40μM, respectively, whereas lovastatin inhibited release by 31 and 59% at 25 and 50μM, respectively. Thus cytoskeletal disruptors and small G protein inhibitors significantly impair amylin release.


Fig. 3

Cytoskeletal modulators inhibit amylinGFP secretion. AmylinGFP-expressing islet β-cells were pre-treated with cytochalasin D (40, 20μM) or nocodazole (20, 10μM) for 30min, with lovastatin (50, 25μM) for 24h at 37°C, or with DMSO (40, 20μM), as indicated. The media was replaced with fresh media without the cytoskeletal modulators, and the relative levels of amylinGFP released into KRB buffer during a 3h incubation were determined by quantitative immunoblotting analysis of 20-μl aliquots of culture supernatants using a polyclonal antibody against GFP (Upper panel). The blots were scanned to record band intensity, and the data expressed as percentage of band intensity relative to DMSO controls. The data are representative of at least three independent experiments.


3.2 RNA interference of endogenous KNT expression inhibits amylin expression and secretion

Knockdown by RNA interference (RNAi) was used to silence KNT expression in RINm5F cells in order to determine whether KNT is critical for amylin secretion. A synthetic 21-nucleotide siRNA duplex (siKNT) with symmetrical 2-nt 3′ overhangs was prepared against the region encompassing nt 1477–1500 of rat KNT (GenBank accession # XM-341305). siKNT binds to a region just upstream of vd1, and hence would be expected to knockdown all KNT isoforms. A control siRNA (siGFP) against nt 174–194 of GFP was used to target the GFP tag in amylin. The siRNA duplexes were transfected into islet β-cells singly transfected to express amylinGFP . Western blot analysis of cell lysates prepared from cells 72h after treatment with the siRNAs revealed that siKNT at 100nM completely abolished expression of endogenous KNT (Fig. 4A), and reduced (by 70%) the cellular level of amylinGFP within cells (Fig. 4A), and also inhibited (by 92%) amylinGFP secretion into the cell media during a further 3h incubation (Fig. 4B). In contrast, silencing of KNT had negligible effect on the levels of tubulin, suggesting that general protein synthesis was not inhibited. This is in accord with the finding that the cells remained healthy and viable. The siRNA against GFP as expected selectively inhibited the expression of cellular and secreted amylinGFP, but had no detectable effect on native KNT levels, or on the expression of the tubulin control protein (Fig. 4A, B). The successful targeting of endogenous KNT and exogenous amylinGFP by the cognate siRNA duplexes demonstrates that RNAi is functioning in islet β-cells. These results establish that KNT influences the expression of amylin and the release of amylin-containing secretory granules from islet β-cells.


Fig. 4

Knockdown of KNT expression by siRNA duplexes inhibits amylinGFP expression and secretion. RIN5mF cells (1×105cells/ml) seeded in 24-well plates were incubated in media containing either 100nM of the KNT-specific siRNA siKNT, a GFP-specific siRNA (siGFP) to target amylinGFP, or vehicle control (Mock), as indicated. Cells were incubated in the latter media for 7h, and then the cell media was removed and transfectants cultured in normal growth media not containing the RNAi for a further 65h. Cells were harvested to determine the levels of intracellular KNT and amylinGFP by SDS-PAGE and Western blot analysis of cell lysates (40μg protein per lane). Blots were stained with an antibody against KNT (upper panel), or a combination of antibodies against the GFP tag and the control tubulin (lower panel) (A). Alternatively, after the 65h incubation the media was again removed and replaced and cells incubated for a further 3h. The cell culture media was harvested and a 20μl aliquot was resolved by SDS-PAGE and Western blotted with an antibody against GFP to detect the levels of amylinGFP secreted into the culture media (B). The blots were scanned to record band intensity, and the data expressed as percentage of band intensity relative to mock-transfected controls. The data are representative of at least three independent experiments.


3.3 RINm5F cells express KNT isoforms lacking the kinesin-binding domain

To establish that kinesin binding-domain-deficient forms of KNT play a role in amylin secretion in RINm5F cells, it was essential to first demonstrate whether such forms were naturally present in RINm5F cells. PCR primers flanking vd2 and vd4 were used to RT-PCR KNT transcripts from total cellular RNA in order to determine whether isoforms lacking vd3 and/or vd4, and were expressed in RINm5F β-cells. Gel analysis (Fig. 5A) followed by sequencing of the PCR products revealed the presence of two vd4-deficient isomers, and two vd4-replete isomers (Fig. 5B). Both the former isoforms contained vd3, but one lacked vd2 in addition to vd4. Thus, kinesin binding-domain-deficient (vd4-deficient) forms of KNT are present in RINm5F cells.


Fig. 5

A KNT isoform lacking vd2 and vd4 is expressed in RINm5F cells. A region encoding the C-terminal variable domains 2 and 4 of KNT was amplified by RT-PCR from total RNA of RINm5F cells to identify KNT isomers expressed in RINm5F cells. (A) Agarose gel analysis of RT-PCR products obtained using primers that flank vd2 and 4 (lane 2). Lane 1, 100bp DNA ladder. (B) Schematic representation of KNT isoforms identified by sequencing the PCR products in A.


3.4 A vd4-deficient KNT isoform associates with amylinGFP-containing granules and inhibits amylin secretion

The above experiments demonstrated that KNT influences the expression and secretion of amylin, but the role(s) of kinesin binding-domain-deficient (vd4-deficient) forms of KNT in this secretory process are an enigma. In order to investigate a role for the latter forms, the human KNT isoform (clone L25616) lacking vd2 and vd4 (Print et al., 1994) was fused to a cyan fluorescent protein (CFP) tag generating the protein designated KNTvd4CFP (Fig. 1B). AmylinGFP-expressing RINm5F β-cells were transfected with KNTvd4CFP. KNTvd4CFP was stably expressed as a &007E;180kDa protein (Fig. 6A). Transfectants stably overexpressing both KNTvd4CFP and amylinGFP were analysed to determine whether the KNTvd4 isoform influences amylin secretion. Two independent KNTvd4CFP transfectants were identified that stably expressed KNTvd4CFP, albeit at different levels. Both the latter transfectants had similar levels of cellular amylinGFP expression (Fig. 4A). They both displayed a marked 80% reduction of amylinGFP release into the cell culture media compared to the parental cells, as determined by Western blot analysis (Fig. 4B). Thus, over-expression of KNTvd4CFP markedly inhibits amylin secretion.


Fig. 6

The KNTvd4CFP isoform colocalizes with a population of amylinGFP-containing granules and inhibits amylinGFP secretion. Two double transfectants (lanes 1 and 2) were engineered to stably coexpress KNTvd4CFP and amylinGFP. Cell lysates (50μg protein per lane) and culture supernatants were resolved by SDS-PAGE and Western blotted with an antibody against the GFP tag (which cross-reacts with CFP) to detect intracellular KNTvd4CFP and amylinGFP (A), and amylinGFP secreted into the culture media (B). Both transfectants expressed KNTvd4CFP of the correct size, and displayed similar levels of intracellular amylinGFP (A), but produced significantly reduced levels of secreted amylinGFP as compared to a transfectant (lane labelled amylin) singly expressing amylinGFP (B). (C) Double transfectants stably expressing amylinGFP and KNTvd4CFP were analysed by confocal microscopy. Images were merged to colocalize amylinGFP with KNTCFP , as indicated. The blue pseudo-colour of KNT was changed to red in order to more clearly discriminate colocalization of amylinGFP with KNTCFP, which is depicted by yellow/brown coloration. Scale bar: 8μm. Boxes mark a higher magnification image of part of the cell shown in the inset.


Overexpressed recombinant KNTvd4CFP fluoresced blue and was pseudo-coloured to red in order to more clearly discriminate its colocalization with green amylinGFP . The combination of green with red gave a yellow coloration as a marker that KNTvd4CFP was closely associated with amylin. Confocal microscopy revealed that 46% of amylinGFP colocalized with KNTvd4CFP marked cytosolic vesicles (Fig. 6C). In contrast, free CFP expressed in RINm5F β-cells as a control was diffusely distributed throughout the cytosol and nucleus (data not shown). The data suggest that overexpressed KNTvd4 is closely associated with a population of amylin secretory vesicles, and selectively inhibits amylin secretion, but not the expression of the hormone.

4 Discussion

Here we confirmed that both KNT and kinesin binding-deficient isoforms of KNT were endogenously expressed in RINm5F cells. Silencing of KNT expression inhibited amylinGFP expression, and abolished amylinGFP secretion into culture media. The data suggest that KNT is essential for the translocation of amylin-containing vesicles to the periphery for secretion. A two-step mechanism of vesicle transport involving kinesin-dependent long-distance and actin-myosin-mediated short-distance transport (Huang et al., 1999) has been suggested for Ca++-regulated exocytosis in wounded sea urchins eggs (Bi et al., 1997) and for insulin release in islet β-cells (Donelan et al., 2002; Varadi et al., 2002a). This notion is in accord with the finding that both kinesin-microtubule and myosin-actin filament disruptors (nocodazole and cytochalasin D, respectively) blocked amylin secretion in RINm5F cells. The results obtained with lovastatin further suggest that small G-proteins participate in regulating amylin secretion. As a correlate, KNT interacts with the small G-proteins Rac1 and RhoA (Alberts et al., 1998; Hotta et al., 1996) and acts as a key effector of microtubule-dependent RhoG activity (Vignal et al., 2001). Further, lovastatin has been observed to inhibit nutrient-induced insulin secretion from islet β-cells (Li et al., 1993).

While the above considerations propose a role for KNT in the transport and secretion of amylin-containing vesicles in islet β-cells, we cannot rule out the possibility that KNT is also involved in the synthesis of amylin. Indeed, silencing of KNT led to significantly decreased levels of amylinGFP expression, which ultimately has effects on amylin secretion. The explanation for selective inhibition of amylin synthesis is not clear, particularly since the expression of amylinGFP is driven from a plasmid. It was proposed that KNT can support protein synthesis by anchoring the EF-1 complex to the ER via EF-1δ (Ong et al., 2003). EF-1δ plays an important role in the regulation of protein synthesis by participating in the elongation step during the translation of mRNA. However, our results suggest that silencing of KNT expression did not have a general effect on protein synthesis. Recent studies suggest that the translation of proteins transported by secretory vesicles is co-ordinately regulated with the biogenesis of the secretory granule through mRNA stabilization (Knoch et al., 2004). Thus, inhibition of vesicle movement by silencing KNT expression may feedback to reduce vesicle biogenesis, which in turn leads to post-translational inhibition of amylin synthesis.

Santama et al. (2004) reported that 12 of 16 isoforms identified in the mouse nervous system lacked the whole or part of the kinesin-binding domain, and the only ubiquitous KNT isoform was that lacking all the C-terminal variable inserts. The relative abundance and wide-spread distribution of vd4-deficient forms of KNT is interesting given that such isoforms would not be expected to bind to kinesin. To date there has been no attempt to understand the functions of specific KNT isoforms. We have shown that transcripts encoding vd4-deficient forms of KNT are expressed in RINm5F cells. Overexpressed recombinant KNTvd4 was localized to amylin/insulin secretory vesicles, but was not detectable on the ER where KNT has often previously been localized (Futterer et al., 1995; Toyoshima et al., 1992). Overexpressed KNTvd4 inhibited sustained insulin secretion, but unlike the effect obtained with silencing of KNT expression, it did not influence the cellular levels of amylinGFP. Further, it did not appear to influence the distribution of amylin-containing vesicles. The mechanism by which the KNTvd4 isoform inhibits amylin secretion is a mystery. Overexpressed KNTvd4 had no observable effect on either the microtubular or ER network (data not shown). In contrast overexpression of KNT isoforms containing the kinesin-binding domain induced a pronounced collapse of the ER network, in accord with suggestion that KNT is involved in the maintenance of the ER (Santama et al., 2004).

KNT has been predicted to form a dimer and there is a twofold molar excess of KNT over kinesin in unextracted vesicles (Kumar et al., 1998). It is thus plausible that two molecules of vd4-replete KNT are required to form a functional bond with kinesin. In this case, vd4-deficient isoforms of KNT might dimerize with vd4-replete isoforms and thereby antagonize the kinesin-binding function of KNT. The pairing of different isoforms with one another would considerably expand the potential number of different isoforms from the predicted 32 theoretically possible zero- to five-insert permutations (Santama et al., 2004). Each isoform pair may exhibit different cellular distributions and functions, and pairing provides a mechanism by which one isoform could antagonize or enhance the effects of another. Since overexpressed KNTvd4 and silencing of KNT expression had different outcomes in terms of amylin expression and vesicle distribution, it would seem that KNTvd4 does not act simply as a dominant-negative inhibitor of KNT.

The above explanations for KNTvd4-mediated inhibition of amylin secretion are predicated on the demonstration that kinesin mediates the secretion of amylin. Should this not necessarily be the case, then KNTvd4 itself may mediate amylin secretion by binding other motor protein partners through binding domains residing outside the conventional kinesin-binding motif. Overexpressed KNTvd4 could overwhelm the secretory system, and thereby hinder amylin secretion.

In conclusion, our findings demonstrate that KNT is essential for amylin secretion in pancreatic islet β-cells. The KNTvd4 isoform selectively associates with secretory vesicles, and regulates amylin secretion via a mechanism that appears to differ from KNT-mediated secretion of amylin. These results provide additional insights into the fascinating pathways by which cargoes destined for secretion are transported inside the living cell.

Acknowledgements

We thank the staff of the Biomedical Imaging Unit, Faculty of Medical and Health Sciences, University of Auckland, for their technical assistance. We are grateful to Dr Peter Barling for his critical review of the manuscript. This work was supported in part by grants from the Marsden Fund, the Health Research Council of New Zealand, the Royal Society of New Zealand, and the Wellcome Trust, UK. We are grateful to the Lottery Grants Board, and the Freemasons (Lodge Discovery) for the supply of equipment.

References

Alberts AS, Bouquin, N, Johnston, LH, Treisman, R. Analysis of RhoA-binding proteins reveals an interaction domain conserved in heterotrimeric G protein beta subunits and the yeast response regulator protein Skn7. J Biol Chem 1998:273:8616-22
Crossref   Medline   1st Citation  

Bai JZ, Saafi, EL, Zhang, S, Cooper, GJ. Role of Ca2+ in apoptosis evoked by human amylin in pancreatic islet beta-cells. Biochem J 1999:343:53-61
Crossref   Medline   1st Citation  

Bi G-Q, Morris, RL, Liao, G, Alderton, JM, Scholey, JM, Steinhardt, RA. Kinesin- and myosin-driven steps of vesicle recruitment for Ca2+-regulated exocytosis. J Cell Biol 1997:138:999-1008
Crossref   Medline   1st Citation  

Blocker A, Severin, FF, Burkhardt, JK, Bingham, JB, Yu, H, Olivo, JC. Molecular requirements for bi-directional movement of phagosomes along microtubules. J Cell Biol 1997:137:113-29
Crossref   Medline   1st Citation   2nd  

Burkhardt JK. In search of membrane receptors for microtubule-based motors-is kinectin a kinesin receptor? Trends Cell Biol 1996:6:127-31
Crossref   Medline   1st Citation  

Donelan MJ, Morfini, G, Julyan, R, Sommers, S, Hays, L, Kajio, H. Ca2+-dependent dephosphorylation of kinesin heavy chain on beta-granules in pancreatic beta-cells. Implications for regulated beta-granule transport and insulin exocytosis. J Biol Chem 2002:277:24232-42
Crossref   Medline   1st Citation  

Futterer A, Kruppa, G, Kramer, B, Lemk, H, Kronke, M. Molecular cloning and characterization of human kinectin. Mol Biol Cell 1995:6:161-70
Medline   1st Citation   2nd   3rd  

Hotta K, Tanaka, K, Mino, A, Kohno, H, Takai, Y. Interaction of the Rho family small G proteins with kinectin, an anchoring protein of kinesin motor. Biochem Biophys Res Commun 1996:225:69-74
Crossref   Medline   1st Citation   2nd  

Huang JD, Brady, ST, Richards, BW, Stenolen, D, Resau, JH, Copeland, NG. Direct interaction of microtubule- and actin-based transport motors. Nature 1999:397:267-70
Crossref   Medline   1st Citation  

Knoch K-P, Bergert, H, Borgonovo, B, Saeger, H-D, Altkrüger, A, Verkade, P. Polypyrimidine tract-binding protein promotes insulin secretory granule biogenesis. Nat Cell Biol 2004:6:207-14
Medline   1st Citation  

Kumar J, Yu, H, Sheetz, MP. Kinectin, an essential anchor for kinesin-driven vesicle motility. Science 1995:267:1834-7
Crossref   Medline   1st Citation   2nd  

Kumar J, Erickson, HP, Sheetz, MP. Ultrastructural and biochemical properties of the 120-kDa form of chick kinectin. J Biol Chem 1998:273:31738-43
Crossref   Medline   1st Citation   2nd   3rd  

Langley R, Leung, E, Morris, C, Berg, R, McDonald, M, Weaver, A. Identification of multiple forms of 180-kDa ribosome receptor in human cells. DNA Cell Biol 1998:17:449-60
Crossref   Medline   1st Citation  

Leung E, Print, CG, Parry, DA, Closey, DN, Lockhart, PJ, Skinner, SJ. Cloning of novel kinectin splice variants with alternative C-termini: structure, distribution and evolution of mouse kinectin. Immunol Cell Biol 1996:74:421-33
Crossref   Medline   1st Citation   2nd  

Li G, Regazzi, R, Roche, E, Wollheim, CB. Blockade of mevalonate production by lovastatin attenuates bombesin and vasopressin potentiation of nutrient-induced insulin secretion in HIT-T15 cells. Probable involvement of small GTP-binding proteins. Biochem J 1993:289:379-85
Medline   1st Citation  

Meng YX, Wilson, GW, Avery, MC, Varden, CH, Balczon, R. Suppression of the expression of a pancreatic beta-cell form of the kinesin heavy chain by antisense oligonucleotides inhibits insulin secretion from primary cultures of mouse beta-cells. Endocrinology 1997:138:1979-87
Crossref   Medline   1st Citation  

Moore CX, Cooper, GJ. Co-secretion of amylin and insulin from cultured islet beta-cells: modulation by nutrient secretagogues, islet hormones and hypoglycemic agents. Biochem Biophys Res Commun 1991:179:1-9
Crossref   Medline   1st Citation  

Ong LL, Lim, AP, Er, CP, Kuznetsov, SA, Yu, H. Kinectin-kinesin binding domains and their effects on organelle motility. J Biol Chem 2000:275:32854-60
Crossref   Medline   1st Citation   2nd   3rd   4th  

Ong LL, Er, CPN, Ho, A, Aung, MT, Yu, H. Kinectin anchors the translation elongation factor-1delta to endoplasmic reticulum. J Biol Chem 2003:278:32115-23
Crossref   Medline   1st Citation   2nd  

Plitz T, Pfeffer, K. Intact lysosome transport and phagosome function despite kinectin deficiency. Mol Cell Biol 2001:21:6044-55
Crossref   Medline   1st Citation  

Print CG, Leung, E, Harrison, JEB, Watson, JD, Krissansen, GW. Cloning of a gene encoding a human leukocyte protein characterised by extensive heptad repeats. Gene 1994:144:221-8
Crossref   Medline   1st Citation   2nd   3rd   4th   5th   6th   7th  

Santama N, Er, CP, Ong, LL, Yu, H. Distribution and functions of kinectin isoforms. J Cell Sci 2004:117:4537-49
Crossref   Medline   1st Citation   2nd   3rd   4th   5th   6th  

Toyoshima I, Yu, H, Steuer, ER, Sheetz, MP. Kinectin, a major kinesin-binding protein on ER. J Cell Biol 1992:118:1121-31
Crossref   Medline   1st Citation   2nd   3rd  

Vancoillie G, Lambert, J, Mulder, A, Koerten, HK, Mommaas, AM, Van Oostveldt, P. Kinesin and kinectin can associate with the melanosomal surface and form a link with microtubules in normal human melanocytes. J Invest Dermatol 2000:114:421-9
Crossref   Medline   1st Citation  

Varadi A, Ainscow, EK, Allan, VJ, Rutter, GA. Molecular mechanisms involved in secretory vesicle recruitment to the plasma membrane in beta-cells. Biochem Soc Trans 2002:30:328-32
Crossref   Medline   1st Citation   2nd  

Varadi A, Ainscow, EK, Allan, VJ, Rutter, GA. Involvement of conventional kinesin in glucose-stimulated secretory granule movements and exocytosis in clonal pancreatic beta-cells. J Cell Sci 2002:115:4177-89
Crossref   Medline   1st Citation  

Vignal E, Blangy, A, Martin, M, Gauthier-Rouviere, C, Fort, P. Kinectin is a key effector of RhoG microtubule-dependent cellular activity. Mol Cell Biol 2001:21:8022-34
Crossref   Medline   1st Citation   2nd   3rd   4th  

Wang HC, Su, YR, Han, KJ, Pang, XW, Peng, JR, Liang, B. Multiple variants and a differential splicing pattern of kinectin in human hepatocellular carcinoma. Biochem Cell Biol 2004:82:321-7
Crossref   Medline   1st Citation  

Xu J, Bird, PH, Bradley, MP, Janssens, PA, Hardy, CM. Molecular cloning and characterization of fox testis kinectin. Mol Reprod Dev 2002:62:37-46
Crossref   Medline   1st Citation   2nd  

Yu H, Nicchitta, CV, Kumar, J, Becker, M, Toyoshima, I, Sheetz, MP. Characterization of kinectin, a kinesin-binding protein: primary sequence and N-terminal topogenic signal analysis. Mol Biol Cell 1995:6:171-83
Medline   1st Citation   2nd  


Received 20 September 2005/22 April 2006; accepted 2 June 2006

doi:10.1016/j.cellbi.2006.06.008


ISSN Print: 1065-6995
ISSN Electronic: 1095-8355
Published by Portland Press Limited on behalf of the International Federation for Cell Biology (IFCB)