Imaging the rapid yet transient accumulation of regulatory lipids , lipid kinases , and 1 protein kinases during membrane fusion , at sites of exocytosis of MMP-9 in cancer cells 2 3

20 21 Background: The control of exocytosis is physiologically essential. In vitro SNARE proteins are 22 sufficient to drive membrane fusion, but in cells there are additional proteins and lipids that work 23 together to drive efficient, fast, and timely release of secretory vesicle cargo. Growing evidence 24 suggests that regulatory lipids act as important lipid signals and regulate various biological 25 processes including exocytosis. Though functional roles of many of these regulatory lipids has 26 been linked to exocytosis, the dynamic behavior of these lipids during membrane fusion at sites of 27 exocytosis in cell culture remains unknown. 28 29 Methods: We used total internal reflection fluorescence microscopy (TIRF) to observe the spatial 30 organization and temporal dynamics of several lipids, and accessory proteins, like lipid kinases 31 and protein kinases, in the form of protein kinase C (PRKC) relative to single sites of exocytosis 32 of MMP-9 in living MCF-7 cancer cells. 33 34 Results: After stimulating exocytosis with PMA, we observed a transient accumulation of the 35 regulatory lipids (e.g. PIP, PIP2, and DAG), lipid kinases (e.g. PI4K2B, PI4K3A, and PIP5KA), 36 and protein kinases (e.g. PRKCA and PRKCE) at exocytic sites centered on the time of membrane 37 fusion, before rapidly diffusing away from the fusion sites. Additionally, the synthesis of these 38 regulatory lipids, degradation of these lipids, and the downstream effectors activated by these 39 lipids, are also achieved by the recruitment and accumulation of key enzymes at exocytic sites 40 (during the moment of cargo release), including lipid kinases, protein kinases, and phospholipases 41 that facilitate membrane fusion and exocytosis of MMP-9. 42 43 Conclusions: This work suggests that these regulatory lipids and associated effector proteins are 44 locally synthesized and/or recruited to exocytic sites, during the time of membrane fusion and 45 cargo release, and their enrichment at fusion sites serves as an important spatial and temporal 46 organizing “element” defining individual exocytic sites. 47 48 Background 49 50 Exocytosis is a fundamental behavior, ubiquitous across eukaryotes and a variety of cell 51 types. During exocytosis, vesicles fuse with the plasma membrane and result in secretion of 52 biomolecules (vesicle cargo) to the outside of cells. This important mode of cellular 53 communication can affect a variety of physiological processes, including polarized growth and 54 motility [1, 2], cancer progression [3-5], and diabetes [6]. As it relates to cancer, defects in 55 exocytosis have been implicated in many cancer types and are generally attributed to dysregulation 56 in proteins involved in the discrete steps of exocytosis. Cancer progression relies, in part, on 57 exocytosis to secrete a variety of protein factors, including growth factors, cytokines, proteases, 58 and exosomes for establishment of tumor growth. During cancer progression, up-regulation of 59 trafficking and secretion of several proteolytic enzymes, known as matrix metalloproteinases, 60 MMPs, are responsible for degrading the extracellular matrix (ECM). This degradation by these 61 proteases is a necessary step in tumor progression and metastasis. 62 Metastasis is a complex multicellular process that involves the dissemination of cancer 63 cells through a series of sequential steps [7, 8]. During metastasis, the tumor migrates from the 64 primary site to colonize distant (secondary) sites and represents the most common cause of cancer 65 death. During metastasis, the ECM is degraded by the matrix-degrading proteins, MMPs, which 66 are secreted by cells as a result of precisely organized intracellular cell signaling events. These 67 MMPs rely on cells that utilize exocytosis to deliver secretory vesicles containing these proteases 68 to the plasma membrane for subsequent release of the vesicle-associated cargo proteins to the 69 outside of the cell. Therefore, MMPs play an important role in cancer progression by altering cell 70 invasion, migration, metastasis, and tumorigenesis [7, 9-11]. Because of their important role in 71 ECM degradation, MMPs have often been used as major biological markers of metastasis and 72 acquisition of metastatic traits in cancer cells. Several studies indicate that increased expression of 73 MMPs correlates with aggressive forms of several cancers, including colorectal, breast, ovarian, 74 and melanoma [12-15]. Furthermore, this increased expression of several MMPs, including MMP75 9, MMP-2, and MT1-MMP (i.e. MMP-14), are secreted by a variety of metastatic cancer cells to 76 aid in ECM degradation [16, 17]. Therefore, it is clear that the release of MMPs depends on 77 exocytosis associated with a variety of cancer cells. 78 During exocytosis, cells use more than twenty-five different proteins and an unknown 79 number of lipids [18], participating in a cascade of protein-protein and lipid-protein interactions, 80 and lipid signaling leading to the externalization of the secretory vesicle cargo molecules [19]. The 81 minimal machinery required for fusion is a complex of three proteins, syntaxins, synaptosomal82 associated proteins (SNAPs), and vesicle-associated proteins (VAMPs), collectively called the 83 soluble N-ethylmaleimide–sensitive factor attachment protein receptors (SNAREs)[18, 20, 21]. 84 This complex of proteins coil together to pull the plasma membrane and vesicle membrane into 85 close apposition to drive fusion. For example, VAMP2, SNAP25, and syntaxin-1 were identified 86 in brain tissue, where they mediate synaptic vesicle fusion and neurotransmitter release [20]. While 87 VAMP3, SNAP23, and synatxin-13 were found to be involved in exocytosis of MMP-9 and MMP88 2, degradation of the ECM, and subsequent cell invasion in fibrosarcoma cells [22]. 89 Although the three SNARE proteins represent the minimal fusion machinery, fusion 90 mediated by SNAREs only is relatively slow and uncoordinated. Therefore, though the SNAREs 91 are sufficient to induce membrane fusion in vitro, there are dozens of other accessory proteins and 92 lipids that assemble together with the SNAREs to accelerate and regulate exocytosis in cells [18, 93 23]. The additional biomolecules essentially help to spatially and temporally coordinate the distinct 94 steps associated with secretory vesicle fusion and exocytosis. Moreover, the SNAREs and the 95 additional protein and lipid factors help to provide spatial and temporal cues or spatiotemporal 96 organization associated with sites of secretory vesicle fusion (i.e. exocytic sites). The role of spatial 97 and temporal organization at exocytic sites has been generally shown to occur through membrane 98 specialization or organized spatial regions that have elevated levels of these factors at exocytic 99 sites [24]. Specifically, the SNAREs, accessory proteins, and lipids form distinct organizing 100 “elements” near the cell membrane, which interact with components of secretory vesicles and help 101 facilitate exocytosis in cells. Fundamentally, these organizing “elements” can be grouped into 102 three distinct classes and include: (1) specialized scaffolding proteins; (2) specialized lipids; and 103 (3) the actin cytoskeleton network proteins [24]. 104 The role of specialized lipids has become of increasing importance and overwhelming 105 evidence suggests that specialized lipids, including regulatory lipids and other bioactive lipids, act 106 as lipid-signaling mediators and affect a variety of cellular functions (e.g. signaling and regulation) 107 [25], like the sequential stages underlying exocytosis (e.g. secretory vesicle trafficking, docking, 108 priming, vesicle fusion, and recapture) [19]. The most notable of specialized lipids is the regulatory 109 lipid, phosphatidylinositol-4,5-biphosphate (PI4,5P or PIP2) [26, 27, 28F148, 29, 30], which is a 110 prerequisite for Ca-dependent exocytosis [31], coordinates trafficking of secretory vesicles to 111 their docking sites on the plasma membrane [32], and primes secretory vesicles for exocytosis, by 112 recruiting accessory proteins and interacting with key components of the exocytic machinery (e.g. 113 SNARE proteins) [33]. Additionally, the major upstream lipid precursor of PIP2 synthesis, 114 phosphatidylinositol 4-phosphate (PI4P or PIP) is another distinct phosphoinositide (PI) lipid 115 species, which regulates exocytosis by promoting docking of specific subpopulations of exocytic 116 vesicles with the plasma membrane [34]. While the downstream degradation product of PIP2 117 hydrolysis, diacylglycerol (DAG) promotes exocytosis by recruiting vesicles to the immediate 118 releasable pool through regulation of the vesicle priming protein, Munc13-1, in neuroendocrine 119 cells [35]. 120 The regulatory lipid, DAG has two possible fates, either phosphorylation to become the 121 bioactive lipid, phosphatic acid (PA) or activation of protein kinase C (PRKC) via allosteric 122 modification. Both downstream cellular fates have been functionally linked to exocytosis. PA can 123 recruit proteins, like the SNAREs, thereby facilitating priming and vesicle fusion [36-38]. PRKCs 124 are critical regulators of exocytosis through phosphorylation of distinct protein targets effecting or 125 regulating the exocytotic machinery, including SNAP-25 and Munc18 [39-44]. The 126 interconversion of these various distinct specialized lipids, as well as, downstream activation by 127 DAG is achieved by specific lipid kinases, phospholipases, phosphatases, and protein kinases [45, 128 46]. Furthermore, these effector proteins have also been found to promote distinct steps involved 129 in exocytosis [47-49] and suggests that local synthesis and degradation of these specific lipid 130 species is achieved through the specific accumulation of these effector proteins at sites of 131 membrane fusion and exocytosis [24, 27, 45, 50-52]. Overall, this implies that PIP2-mediated 132 signaling: (1) is an important signal transduction pathway, involving a cascade of specialized 133 lipids, and a variety of effector proteins, that are functionally linked to exocytosis; and (2) involves 134 specialized lipids that possibly accumulate in defined microdomain regions, with some spatial and 135 temporal patterning, in order to recruit effector proteins and promote exocytosis. 136 More and more, total internal reflection fluorescence (TIRF) has been utilized to image the 137 spatiotemporal organization and dynamics associated with exocytosis in a variety of cellular 138 contexts, including: (1) exocytosis involved in neurite elongation [53]; (2) exocytosis associated 139 with cytoskeleton rearrangements and the formation of membrane fusion “hotspots” [54]; (3) viral 140 exocytosis [55]; (4) microvesicles exocytosis [56]; (5) dense core vesicle (DCV) exocytosis [57] 141 and; (6) secretory vesicle exocytosis [58]. We previously showed using two-color TIRF that it was 142 possible to study the spatiotemporal patterning and dynamic behavior of several red fluorescently143 labeled Rab GTPases, Rab effector proteins, and SNARE proteins (organizing “element #1; 144 specialized scaffolding proteins) during regulated exocytosis of MMP-9, at individual membrane 145 fusion sites or exocytic sites in MCF-7 adenocarcinoma cancer cells [58]. We found that many of 146 these proteins stably assembled on docked secretory vesicles before exocytosis, however, at the 147 moment of fusion, all of these components quickly diffused away and were lost from the exocytic 148

More and more, total internal reflection fluorescence (TIRF) has been utilized to image the 137 spatiotemporal organization and dynamics associated with exocytosis in a variety of cellular 138 contexts, including: (1) exocytosis involved in neurite elongation [53]; (2) exocytosis associated 139 with cytoskeleton rearrangements and the formation of membrane fusion "hotspots" [54]; (3)  and; (6) secretory vesicle exocytosis [58]. We previously showed using two-color TIRF that it was 142 possible to study the spatiotemporal patterning and dynamic behavior of several red fluorescently-143 labeled Rab GTPases, Rab effector proteins, and SNARE proteins (organizing "element #1; 144 specialized scaffolding proteins) during regulated exocytosis of MMP-9, at individual membrane 145 fusion sites or exocytic sites in MCF-7 adenocarcinoma cancer cells [58]. We found that many of 146 these proteins stably assembled on docked secretory vesicles before exocytosis, however, at the 147 moment of fusion, all of these components quickly diffused away and were lost from the exocytic 148 site. 149 Here, to visualize exocytic vesicle fusion, we again exploited two-color TIRF microscopy 150 to image the spatiotemporal organization and dynamics associated with specialized lipids 151 (organizing "element" #2) and effector proteins on exocytosis of MMP-9 from MCF-7 breast 152 cancer cells. To this end, we relied on the use of red fluorescently-tagged lipid-binding sensor 153 proteins, with known specificity for single distinct lipids [59][60][61][62][63], to monitor the dynamics of  were superimposed in post-processing by acquiring an image of 100 nm yellow-green beads 210 (Invitrogen) that were visible in both channels, mapping the position of six beads in both channels, 211 and then superimposing the channels using projective image transform. This protocol was 212 performed each day and the image transform was used to superimpose cell images. Images were 213 acquired with IQ2 software (Andor) successively in the green then red channels with an exposure 214 time of 500 ms and a 500 ms pause between pairs of images. Pixel size was 160 nm.

216
Structured Illumination Microscopy. TIRF-SIM imaging was performed on a DeltaVision 217 OMX SR Structured Illumination. An inverted microscope is used with a spatial light modulator 218 (SLM) that diffracts the beam using multiple grating patterns. In this system a 100X/1.49NA oil 219 immersion objective lens is used with 488 nm (GFP) and 561 nm (mCherry) excitation and passed 220 through a multi-band dichroic mirror (DM4). Red and green images were superimposed during 221 processing using OMX SR task builder.

223
Image Analysis. All analysis was performed using custom Matlab scripts and ImageJ. Correlation 224 analysis was performed as previously described [67]. All of the local maxima above a user-defined 225 threshold and >2 μm away from the edge of the cell boundary were identified in each whole-cell 226 green image (MMP9-GFP). Each local maximum was centered in a 4 μm × 4 μm cropped region.

227
These cropped regions from several cells were each normalized to the brightest pixel and averaged 228 together. The regions corresponding to the green local maxima were also excised from the red 229 images, normalized to the brightest pixel, and averaged together. In each average red image, 12 230 radial line scans were taken and averaged together. Randomly selected regions equal in number to 231 the excised regions were processed in the same way. The average random line scan was then 232 subtracted from the average red image line scan to account for background intensity statistics. Peak 233 height for each construct was determined by measuring the maximum intensity difference between 234 the first point (center of the image) and all other points in the averaged scan. Smaller 1.0 μm × 1.0 235 μm cropped, vesicle-centered red/green regions were used to calculate Pearson's c for each green 236 peak: where Gi and Ri represent pixel intensities in the two channels and G and R indicates average pixel 239 intensity. The region size was chosen based on the spacing of secretory vesicles in an attempt to 240 capture only one structure per correlation region on average. Correlations were then averaged 241 together, yielding an average C for each construct. Correlation values were calculated between 242 peak-centered green excised regions and randomly selected red regions as well.

244
Exocytic events were analyzed by hand-selecting event coordinates and stack position in the green 245 channel in ImageJ. Matlab scripts extracted a raw intensity mean (Fcenter) from a 480 nm square 246 box centered on the coordinates and a mean background value from a 1440 nm square surround 247 area (Fsurround); this was performed for both channels. The traces were then temporally aligned to 248 the frame before the maximum intensity decrease of the MMP9-GFP channel (t=0 in our 249 trajectories). We aligned to the maximum intensity decrease because this was the most robust   All bands from western blot were quantified using ImageJ software (NIH, Bethesda, MD). For 281 quantification, all band intensities were measure and averaged from four separate experiments.

285
Systematic spatial mapping of specialized lipids associated with secretory vesicles containing 286 MMP-9 at exocytic sites 287 We used MCF-7 breast cancer cells, which can be used as a model system for regulated   Figure 1 shows two separate lipid sensors tested, including the PIP2-sensor ( Figure 1A) and   To test for possible overexpression effects on our colocalization analysis, we confirmed by using 358 all of our lipid sensor and effector proteins that there was no relationship between expression and 359 correlation values (data not shown). Similar effects were also seen using a neuronal secretory  Since DAG is generated directly from PIP2 hydrolysis by PLC, it suggests that the DAG 404 might be locally synthesized at exocytic sites following recruitment and/or synthesis of PIP2 at 405 these sites. Because the PIP2-sensor we used contains the PH domains from PLC [63], we 406 predicted that exocytosis of MMP-9 would be altered in the presence of an inhibitor of PLC. We 407 observed that inhibition of PLC stops exocytosis of MMP-9, following the addition of 10 μM of 408 U-73122, a PLC inhibitor. There were no observable fusion events and no vesicle trafficking 409 between docked sites, in the presence of the PLC inhibitor. This suggested that PLC inhibition 410 blocked local hydrolysis of PIP2 and DAG synthesis, and subsequent loss of MMP-9 release. It is 411 worth noting that we cannot discount a non-specific effect that might be caused by the inhibition.

412
Overall, our TIRF data suggest that regulatory lipids PIP2 and DAG, have dynamic temporal  However, Rabphilin 3A (Ser234) antibody was available and used to detect phosphorylation of a 520 Rab effector protein known to have an association to exocytic machinery and is expressed in a 521 variety of secretory vesicle cell types, including both neuronal and non-neuronal cells lines. We 522 purified lysates that were either PMA-uninduced or PMA-induced using a phosphoprotein 523 purification kit, followed by conducting western blots using anti-phospho Rabphilin 3A (Ser234) 524 antibody, a probe specific for phosphorylated Rabphilin. We were not able to see enrichment of 525 phosphorylated Rabphilin following purification using the phosphoprotein purification kit. This is 526 likely because the purification results in large loss of proteins and subsequent lack of detection 527 using the kit ( Figure 7A). However, the western blots did show phosphorylated Rabphilin

546
This is consistent with our dynamic imaging data suggesting that a PRKC specific isoform (e.g. 547 PRKCE) may accumulate at exocytic sites and could potentially phosphorylate any number of 548 proteins associated with secretory vesicles containing MMP-9. We also examined the spatial 549 positioning of PRKC isoforms with Rab27 GTPase isoforms (organizing "element #1; specialized 550 scaffolding proteins), which revealed punctate localization and a close proximity of PRKCA and 551 Rab27b ( Figure 7D), following PMA induction. There was no punctate localization or close 552 proximity associated with PRKCA and Rab27a ( Figure 7C) with any isoform and we observed no 553 direct localization with Rab27b and PRKCG and PRKCE (data not shown). This result implies 554 that a PRKC specific isoform (e.g. PRKCA) is highly localized with the Rab27 GTPase specific 555 isoform, Rab27b and suggests that the targets of phosphorylation might be associated with 556 Rab27b-specific secretory vesicles.

Mutations effecting prenylation (a lipid modifications) of Rab27 isoforms alter MMP-9
583 release dynamics at exocytic sites 584 We also examined the role of prenylation of Rab27 isoforms, Rab27a and Rab27b, at sites 585 of exocytosis of MMP-9. Protein prenylation is a specific type of lipid modification made to some  (Figure 8A), while Rab27b prenylation mutants slowed by approximately six-fold, 598 from τ = 1.1 ± 0.0 s (WT) to τ = 6.2 ± 0.3 s (prenylation) ( Figure 8C). We also examined the 599 protein dynamics of the Rab27 isoform-specific prenylation mutants (red channel) at these sites 600 of exocytosis of MMP-9. When Rab27a-and Rab27b-specific prenylation mutants were 601 overexpressed in PMA-induced MCF-7 cells, we found localization to sites before fusion that 602 diffuse away following exocytosis of MMP-9, like their WT counterparts ( Figure 8B and 8D).   The process of regulated exocytosis has been a topic of research for several decades and 710 has led to the discovery of a host of different proteins and lipids involved throughout this process.

811
We hypothesize that a similar scenario exists in MCF-7 cells, where PRKC is locally recruited to 812 exocytic sites, during the moment of membrane fusion. This would allow the PRKCs to 813 phosphorylate a variety of proteins associated with fusogenic secretory vesicles containing MMP-814 9. We are currently trying to identify potential targets of phosphorylation, in response to the 815 recruitment of PRKC that we observed at exocytic sites. Identification of targets of 816 phosphorylation by PRKC, in non-neuronal cells, like MCF-7 cells, has proven to be rather 817 difficult due to the unavailability of phospho-specific antibodies used to detect phosphorylated 818 target proteins, as compared to those available for the same family of proteins in neuronal cells, 819 like PC12 cells. Although, we are not able to report multiple targets of phosphorylation by PRKC, 820 our blots suggest that Rabphilin, a Rab effector protein, is a potential target (Figure 7). Moreover, 821 we showed using super-resolution TIRF-SIM that after stimulation with PMA, there is close 822 proximity (i.e. colocalization) between PRKCE and secretory vesicles containing MMP-9 ( Figure   823 7). Overall, our data showed isoform-specific PRKC locally accumulated at exocytic sites of 824 MMP-9 and suggests a possible spatial and temporal role of PRKC at exocytic sites in MCF-7 825 cells.

826
Our TIRF-based imaging data show that the PIP2-mediated signaling pathway is involved 827 in the spatiotemporal organization during membrane fusion, at exocytic sites of MMP-9 in MCF-828 7 cells. More importantly this cascade of lipids in the pathway may serve as organizing "elements", 829 spatially and temporally coordinating regulated exocytosis in these breast cancer cells. We 830 systematically mapped out the dynamic behavior of the lipid signaling cascade from PI to PRKC.

831
The lipids, which include PIP, PIP2, and DAG, are regulated at the level of synthesis and 832 degradation, through the interconversion by specific kinases and phospholipases, that locally  Additionally, our results reiterate the utility and broad applicability of this two-color TIRF 852 imaging-based approach for systematically mapping the molecular composition, spatial 853 organization, and dynamic temporal behavior of discrete cellular processes like that surrounding 854 regulated exocytosis [56][57][58]67]. It is important to note that all of the spatiotemporal information 855 we acquired using this TIRF method implies only correlative associations. Specifically, the 856 recruitment of lipid sensors and tagged kinase enzymes (i.e. lipid kinases and protein kinases) 857 involved in PIP2-mediated signaling pathway to exocytic sites is correlated with the exocytosis of 858 MMP-9 from secretory vesicles in MCF-7 cells and does not prove a functional or mechanistic 859 connection with exocytosis. The goal of this imaging-based TIRF method was to establish the 860 spatial organization and temporal dynamics of specialized lipids (organizing element #2) 861 associated with membrane fusion, at exocytic sites of MMP-9 in MCF-7 cells. This work is meant 862 to complement, not replace, traditional functional studies used to probe the spatiotemporal 863 organization associated with these exocytic sites. Future biochemistry studies will be essential in