【预约试拍】随机光学重建显微镜STORM显示的内皮细胞表面糖萼（ESG）成分和超微结构

智能文字提取功能测试中

为了在血管功能中作为血流的机械传感器和作为经血管交换的屏障发挥不同的作用，内皮表面糖萼（ESG）应该具有有组织的结构。由于光学显微镜和电子显微镜的局限性，直到最近发展的超分辨光学显微镜STORM才揭示了ESG的超微结构。实验结果表明：HA是一种长链分子，编织成网状覆盖在内皮细胞表面.相比之下，HS是垂直于细胞表面的较短分子。HA和HS在内皮管腔表面彼此部分重叠。研究者还定量了内皮表面顶部、中部和底部区域的HS的长度、直径、方向和密度。一 导言排列在我们血管系统中的内皮细胞（EC）的腔表面包被有由蛋白聚糖和糖胺聚糖（GAG）组成的膜结合大分子的糖萼。由于其独特的位置，在循环血液和血管壁的界面处，已经发现ESG在维持正常血管功能方面起重要作用。因此，自20世纪60年代以来，通过各种方法对其组成、结构和机械性能进行了广泛的研究。然而，由于空间分辨率的限制或样品制备过程中的人为因素，静电陀螺仪的空间化学组织仍然不清楚。最近开发的随机光学重建显微镜（STORM）是三种类型的单分子定位显微镜之一，采用有机染料和荧光蛋白作为光可切换发射器，以将时间分辨率转换为超空间分辨率，这比传统的共聚焦显微镜高一个数量级。STORM和其他类型的超分辨率光学显微镜能够在纳米尺度上发现和研究细胞结构，从单个蛋白质到天然细胞环境中的整个细胞器。ESG的作用基于其分子组成和组织。ESG的组成部分已在中进行了深入研究。ESG主要由带有短酸性寡糖和末端唾液酸（SA）的糖蛋白和蛋白聚糖（PG）如硫酸乙酰肝素PG（HSPG）（包括具有长糖胺聚糖（GAG）侧链的多配体聚糖和磷脂酰肌醇聚糖核心蛋白）组成。带负电荷的GAG结合蛋白质、生长因子、阳离子和其他血浆组分。血管EC中的主要GAG是硫酸乙酰肝素（HS）、硫酸软骨素（CS）和透明质酸（HA）。在这三者中，最丰富的是HS，占总GAG的50-90%，其余由CS、HA和SA组成。HS和CS与PG共价结合，而HA不与PG核心蛋白结合。HA是一种非硫酸化GAG，其与其表面受体CD 44和HA介导的运动性受体（RHAMM）结合。除了其生化组成外，ESG的厚度和超微结构决定了其功能。通过电子显微镜（EM）对ESG的第一次可视化使用阳离子染料钌红，其结合酸性粘多糖并在四氧化锇存在下产生电子密度。随后的研究使用金胶体和免疫过氧化物酶标记。Adamson和Clough然后使用大的带电标记蛋白（无法穿透ESG），阳离子化铁蛋白（分子量≈ 450 kDa）证明，在不存在血浆蛋白的情况下，ESG会崩溃，可能是由于消除了与血浆蛋白的分子内相互作用，并且其未受干扰的厚度比钌红观察到的20 nm大几倍。所有这些方法都可能遭受与水性固定剂相关的脱水假象，水性固定剂可能溶解除蛋白聚糖的蛋白质核心之外的所有蛋白质，并破坏固有的水合结构。研究人员开发了一种使用碳氟化合物作为四氧化锇的非水载体来保存水溶性结构的方法，该方法被应用于微血管，以消除其中的一些限制。Rostgaard和Qvortrup对碳氟化合物-戊二醛固定方法的进一步阐述揭示了毛细血管壁上的丝状刷状表面涂层，但层厚度小于50 nm，表明更表面的基质结构裂解。所有前述EM研究揭示具有小于IOOnm的厚度的ESG。最近研究者发现大鼠左心室心肌毛细血管上的ESG厚度为0.2-0.5 μm。活体光学显微镜的差的空间分辨率限制了ESG厚度的准确测量。因此，通过采用激光扫描共聚焦显微镜和多光子显微镜，以及针对HS或HA结合蛋白或麦胚凝集素的荧光标记抗体来标记ESG的主要组分，已经开发了新的成像方法。这些新方法的应用揭示了大血管中厚得多的ESG：在小鼠颈总动脉中为4.3-4.5 μ π ι ，在小鼠颈内动脉中为2.2 μ π ι ，在颈外动脉中为2.5 μ π ι 。Ebong等人首次展示了体外静电陀螺仪的冷冻电镜图像，该图像避免了常规电镜的脱水伪影，并观察到厚度大于5 μm（最大可达11 μm）的结构。最近，使用高灵敏度和分辨率共聚焦显微镜和原位/体内单微血管和离体主动脉免疫染色，Yen等人揭示了大鼠肠系膜和小鼠提睾肌毛细血管和毛细血管后小静脉上ESG的厚度为1-1.5 μm。大鼠和小鼠主动脉的ESG厚度为2-2.5 μ m。Betteridge等人通过比较在体内单个微血管中质膜标记和麦胚凝集素（WGA）标记ESG（SA残基）之间的距离，发现与Yen等人相同类型的微血管中ESG厚度为0.17-3.02 μm，这取决于标记和分析方法。ESG的超微结构组织及其与细胞骨架组分的关系（使用电子断层扫描进行3D重建得到证实。通过他们的EM方法观察到的ESG的厚度为〜 100nm，类似于先前在青蛙肠系膜微血管上发现的厚度。克服脱水工件在新兴市场和传统荧光显微镜的空间分辨率的限制,在目前的研究中,通过使用STORM,研究者发现第一次超限分辨图像ofHS和HA组件的环境、社会和治理在支架表面培养bEnd3老鼠大脑微血管内皮细胞单层,并量化ultra-structural参数ofHS和直径HA在EC支架表面的不同区域。二 实验方法2.1样品制备细胞培养：小鼠脑微血管内皮细胞（bEnd 3），来自ATCC（Manassas，VA）在Dulbecco改良Eagle培养基/营养混合物F-12 Ham培养基中培养。（DMEM/F-12）、2 mM L-谷氨酰胺和100 U/mL青霉素和1 mg/mL链霉素（均来自Sigma-Aldrich，St. Louis，MO），补充有10%胎牛血清，并在37 °C下在具有5%C02的潮湿气氛中孵育。免疫荧光染色：将bEnd 3细胞胰蛋白酶化并以600个细胞/mm 2的密度接种在50 yg/mL纤连蛋白包被的1.5号玻璃底培养皿上并培养3-4天直至汇合。然后用1%BSA/PBS轻轻洗涤细胞，用2%多聚甲醛和0.1%戊二醛固定20分钟，用0.1%NaBH4还原7分钟，用1%BSA/PBS洗涤3次，并用2%正常山羊血清（NGS）封闭30分钟。对于HS染色，将固定的细胞与小鼠抗硫酸乙酰肝素10 e4在4 ℃孵育过夜，然后与ATTO 488缀合的山羊抗小鼠IgG孵育;对于HA染色，将固定的细胞与生物素化的透明质酸结合蛋白在4 °C孵育过夜，然后与Alexa Fluor 647缀合的抗生素孵育。最后，将样品用2%多聚甲醛和0.1%戊二醛后固定10分钟，然后保持在1%BSA/PBS中。2.2STORM图像采集使用STORM对样品进行成像。使用由405 nm波长激光以交替照明激活的ATTO 488和Alexa Fluor 647的多个报告物来获取HS和HA的3D图像。基于每个报告子以19 ms/帧的捕获速度捕获的40，000个ofEM-CCD图像，获得细胞顶端（腔）表面上的三个256 × 256（40 × 40 μm2）的HS和HA视野，位于顶部、中部和底部。顶部区域聚焦于细胞核上方的细胞表面，底部区域聚焦于两个相邻细胞的谷，中间区域聚焦于顶部和底部区域中间的平面。闪烁点的原始图像电影由伴随STORM系统的分析软件处理，并在图1中展示。2.3数据分析应用单发射器质心算法来估计数据电影中激活的荧光团的3D位置，并产生3D STORM图像，其空间分辨率分别为侧向平面中的20nm和轴向方向上的50nm。通过使用混合高斯模型从STORM图像估计细胞表面上的组分中的地面实况荧光团位置。然后，在估计的荧光团位置的基础上估计组件的超微结构参数。对于HS元件，我们假设它是一个圆柱体，其等效直径具有与真实的形状相同的横截面积（椭圆状）。其他超微结构参数、长度、HS的取向和相邻HS元件之间的距离（中心到中心）在图2的插图中定义。由于难以识别在细胞腔表面处编织成网络的单个HA元件，我们从图3中所示的HA区段的2D图像估计HA直径。三.研究结果3.1HS和HA的组织图1A-C展示了STORM揭示的HS和HA的组织。图1A是在顶部区域观察到的，图1B是在中间区域观察到的，图1C是在细胞腔（顶端）表面的底部区域观察到的。对于所有图，第二列是HS图像（绿色），第三列是HA图像（红色），第一列是第二列和第三列的叠加。第一行是2D顶视图，第二行是34 μm × 34 μm视场的3D视图。第三和第四行是第一行中具有绿色虚线的区域的放大的2D和3D视图。第五和第六行是来自第三行中具有蓝色虚线的区域的进一步放大的2D和3D视图。第二列和第三列中三维视图左侧的颜色条表示长度比例。从这些图中可以看出，HA（红色）是编织成2D网络片的长分子，其覆盖细胞腔表面并且与细胞腔表面在同一平面中。相比之下，HS是垂直于细胞腔表面的较短分子。HS和HA在细胞腔表面部分重叠。3.2HS超微结构参数和HA直径图1 STORM揭示的ESG成分和超微结构HS和HA在单元(A),在细胞细胞底部中间(B)和(C)从细胞腔的表面。所有的插图,第二列是HS图像(绿色),第三列是HA图像(红色)和第一列是第二和第三列的叠加。第一行是2 d顶视图,第二行是3 d视图的字段∼34μm×34μm。第三和第四行放大2 d和3 d视图从该地区绿色虚线的第一行。第五和第六行进一步扩大2 d和3 d视图从该地区的蓝色虚线第三行。左边的颜色条的3 d视图HS(第二列)和HA(第三列)代表深度(长度)。表1 bEnd3细胞腔表面HS超微结构的定量表2 在bEnd3细胞腔表面的HA直径表2总结了HA直径估计从2 d图像图3所示。直径是185.3±44.7 nm, 155.5±57.2 nm,和156.9±56.1 nm,分别在顶部、中部和底部区域的细胞腔的表面。这在HA直径没有显著差异这三个地区(p > 0.1)。比较HA直径与HS直径没有区别哈哈直径和HS直径顶部和底部区域(p > 0.2),但在中部地区有显著差异(p < 0.01)。四.讨论使用STORM，研究者揭示了在纳米尺度的bEnd 3单层的ESG中HS和HA的空间化学组织。直接和间接证据表明，ESG在维持血管功能方面至少发挥三种重要作用。1)它可以作为EC的机械传感器和换能器来感知血流；2）它通过形成分子筛维持正常的微血管通透性，允许水和小分子通过但限制血浆蛋白等大分子;3）它在循环红细胞（RBC）和形成血管壁的EC之间产生润滑缓冲液以避免RBC损伤，并产生屏障以防止循环白细胞、血小板和肿瘤细胞粘附到血管壁。它还屏蔽EC表面受体，防止其超活化。STORM揭示的ESG中HS和HA的组织和分布为ESG扮演这三个角色提供了直接的结构可行性。HS元件垂直于细胞腔表面，能够感测血流并用作EC的机械传感器和换能器。在细胞腔表面的底部处的更密集的HS分布（其可以覆盖细胞间连接）沿着由HA形成的编织网络有利于ESG成为用于经血管交换的分子筛、循环细胞和形成血管壁的内皮细胞粘附的屏障以及EC表面受体的屏障。总之，ESG在bEnd3细胞表面的空间化学组织被揭示，从STORM采集的图像中还获得了HS的超微结构参数和HA的直径。在本研究中，研究者主要借助STORM超分辨率显微镜来研究巨噬细胞足体簇的纳米结构和调控，目前在国内，随机光学重建显微镜STORM已成功实现商用，有需要STORM成像技术进行实验研究的专家老师们，请文末填写问卷，即可预约获得 iSTORM 超高分辨率显微成像系统试拍服务~力显智能现已发布的超高分辨率显微成像系统 iSTORM，成功实现了光学显微镜对衍射极限的突破，使得在20纳米的分辨率尺度上从事生物大分子的单分子定位与计数、亚细胞及大分子复合物结构解析、生物大分子生物动力学等的研究成为现实，从而给生命科学、医学等领域带来重大突破。关于我们     宁波力显智能科技有限公司（INVIEW）是专业从事超高分辨率显微技术和产品研发的科技企业，依托复旦大学的自动控制、新一代信息技术及香港科技大学的生物、光学、图像处理等的技术，拥有光学、生物、自控、机械、信息技术等多领域交叉学科技术团队，将2014年诺贝尔化学奖技术产业化，推出了超高分辨率显微成像系统iSTORM、细胞智能监控助手赛乐微等一系列产品，帮助人们以前所未有的视角观察微观世界，突破极限，见所未见。ResearchGateSee discussions， stats， and author profiles for this publication at： https：//www.researchgate.net/publication/332768318 Biorheology 56 (2019)77-8877DOI 10.3233/BIR-180204IOS Press Endothelial surface glycocalyx(ESG) components and ultra-structure revealed by stochastic optical reconstruction microscopy (STORM) Article in B i or h eo l ogy· Apri l 2019 DOI：10.3233/B I R-180204 CI T AT IO NS 21 READS 191 5 authors， i n c l u din g： J ie F an U niv er s i ty of M ic higan -Dearborn 42 PU B L I C A TIO N S 718 C ITAT IO N S S E E PROFIL E Y if an X i a City Col l ege of New York 15 PU B L ICATIONS 194 CIT A TIONS S E E P ROFI L E Some of the authors of thi s publ i cation are also working on these re l ated projects： Pr o ject Tra nspo r t ac r o ss th e Bl o od-b ra in barr i er ， thro u g h t he CSF ， a nd in the br a in ti s s u e View pr o j e c t Pr o je c t C ell c h i r a li ty Vie w p r oje c t Endothelial surface glycocalyx (ESG)公一 components and ultra-structure revealed by stochastic optical reconstruction microscopy (STORM) Jie Fan"， Yi Sun ， Yifan Xia ， John M. Tarbell and Bingmei M. Fua .* aDepartment of Biomedical Engineering， The City College ofthe City University of New York，New York， NY，USA Department ofElectrical Engineering， The City College of the City University of New York， New York，NY， USA Received 19 November 2018 Accepted 10 April 2019 Abstract. BACKGROUND： In order to play different roles i n vascular functions as a mechanosensor to blood flows and as a barrier to transvascular exchange， the endothelial surface glycocalyx (ESG) should have an organized structure. Due to the l i mitations of optica l and electron microscopy， t he ul t ra-structure of ESG has not been revealed unt il the recent development of super-resolution optica l microscopy， STORM. OBJECTIVES： To investigate the ESG components and their organization on bEnd3 (mouse brain microvascular endothelial cells) monolayer. METHODS： ESG was immunolabeled with anti-heparan sulfate (HS)， followed by an ATTO488 conjugated goat anti-mouse IgG， and with biotinylated hyaluronic acid (HA) binding protein， followed by an AF647 conjugated anti-biot i n. The ESG was then i maged by the STORM. RESULTS： HA is a long molecule weaving into a network which covers t he endothelial luminal surface. In contrast， HS is a shorter molecule ， perpendicular to the cell surface. HA and HS are partially overlapped with each other at the endothelial luminal surface. We also quanti f ied the length ， diameter， orientation， and density of HS at the top， middle and bottom regions of the endothelial surface . CONCLUSIONS： Our results suggest that HS plays a major role in mechanosensing and HA plays a major role in the molecular sieve . Keywords： Mouse brain microvascular endothelial cells (bEnd3)， heparan sulfate， hyaluronic acid， i mmunolabeling 0006-355X/19/$35.00 O2019-IOS Press and the authors. All rights reserved 1. Introduction The luminal surface of endothelial cells (ECs) that line our vasculature is coated with a glycocalyx of membrane-bound macromolecules comprised of proteoglycans and glycosaminoglycans (GAGs) [1].Due to i ts unique location， at the interface of the circulat i ng blood and the vessel wall ， the ESG has been found to play important roles in maintaining normal vascular functions [2]. I ts composition， structure， and mechanical property have thus been widely studied since the 1960's by a variety of methods. However，due to the limitations of spatial resolutions or artifacts in sample preparations in the previous techniques，the spatio-chemica l organization of the ESG remains unclear. The recently developed Stochastic Optical Reconstruction Microscopy (STORM)， one of three types of single molecule localization microscopy [3]， employs organic dyes and fluorescent proteins as photo-switchable emit t ers to trade temporal resolution for a super spatial resolution (~20 nm lateral and~50 nm axial at present )， which is an order of magnitude higher than conventional confocal microscopy. STORM and other types of super-resolution optical microscopy enable the discovery and investigation of cellular structures at the nanometer scale， from individual proteins to entire organelles in the native cellular environment [4-6]. The roles of ESG are based on its molecular composition and organization [2，7，8]. The components of the ESG have been investigated in depth in [9，10]. The ESG is mainly composed of glycoproteins bearing short acidic oligosaccharides and terminal sialic acids (SA)， and proteoglycans (PG) l ike heparan sulfate PGs (HSPG) including syndecan and glypican core proteins with long glycosaminoglycan (GAG) side chains. The negatively charged GAGs bind proteins， growth factors， cations， and other plasma components The predominant GAGs in vascular ECs are heparan sulfate (HS)， chondroitin sulfate (CS) and hyaluronic acid (HA). Of t he t hree， the most abundant is HS， accounting for 50-90% of the total GAGs， the rest being comprised of CS， HA and SA [11]. HS and CS are covalent l y bound to PGs， whereas HA does not bind to a PG core protein. HA is a non-sulfated GAG， which binds with i ts sur f ace receptors CD44 and receptor for HA-mediated motility (RHAMM) [8]. In addition to i ts biochemical composition， the thickness and ultra-structure of the ESG determine its functions. The f irst visualization of the ESG by electron microscopy (EM) used the cationic dye ruthenium red that binds to acidic mucopolysaccharides and generates electron density in the presence of osmium tetroxide [12]. Subsequent studies [13，14] used gold colloids and i mmunoperoxidase labeling. Adamson and Clough [15] then demonstrated using a large charged marker protein (unable to penetrate the ESG)，cationized ferritin (molecular weight ~450 kDa)， that i n the absence of plasma proteins， the ESG would collapse， presumably due to elimination of intramolecular interactions with plasma proteins， and that its undisturbed thickness was several t imes greater t han the 20 nm observed with ruthenium red [12].All of these methods may suffer from dehydration artifacts associated with aqueous fixatives that l ikely dissolve all but the protein cores of proteoglycans and collapse the inherently hydrated structures. A method developed to preserve water-soluble structures using fluorocarbons as non-aqueous carriers of osmium tetroxide was applied to microvessels to obviate some of these limitations by Sims and Horne [14].Further elaborations of the fluorocarbon-glutaraldehyde fixation methods by Rostgaard and Qvortrup [16]revealed a f i l amentous brush-like surface coating on capillary walls， but a layer thickness of less than 50 nm， suggesting a cleavage of more superficial matrix structures. Al l of the foregoing EM studies revealed an ESG with a thickness less than 100 nm. Recently， van den Berg et al. [17] used a new approach to stabilize the anionic carbohydrate structures on the ESG by Alcian blue 8GX. They found that the ESG thickness was 0.2-0.5 um on rat left ventricular myocardial capillaries. To overcome the artifact by EM， the amount of ESG was estimated based on the biophysical principles from in vivo observations [18]. A direct in vivo measurement of the ESG thickness wi t h the dye-exclusion technique was developed by Vink and Duling [19]. Using a 70 kD FITC-dextran plasma tracer， which they showed was sterically excluded by the ESG， they were able to provide the first estimate of the in vivo thickness of the ESG in capi l laries of hamster cremaster-muscle to be ~0.4-0.5 um. Later， the ESG thickness was also estimated as ~0.5 um in rat mesenteric post-capillary venules by FITC-dextran labeling with intravital microscopy [20]. The poor spatial resolution of an intravital optical microscope limits the accurate measurement of the ESG thickness [21]. New imaging methods have thus been developed by employing laser scanning confocal microscopy and multi-photon microscopy， and f luorescently tagged antibodies to HS or HA binding protein， or wheat germ agglutinin to label major components of the ESG. Application of these new methods has revealed a much thicker ESG in large blood vessels： 4.3-4.5 um in the mouse common carotid artery [22]， 2.2 um in the mouse internal carotid artery [23]， and 2.5 um in the external carotid artery [24]Ebong et al. [25] presented the first cryo-EM images of in vitro ESG that avoided the dehydration artifacts of conventional EM and observed structures greater t han 5 um in thickness (up to ~11 pm). Most recently using high sensitivity and resolution confocal microscopy and i n situ/in vivo single microvessel and ex vivo aorta i mmunostaining， Yen et al. [26] revealed that t he thickness of the ESG on rat mesenteric and mouse cremaster capillaries and post-capillary venules is 1-1.5 pm. The ESG thickness is 2-2.5 um on rat and mouse aorta. By comparing the distance between the plasma membrane l abeling and the label i ng of ESG (SA residues) by wheat germ agglutinin (WGA) in a single microvessel in vivo， Betteridge et al. [27]found that the ESG thickness is 0.17-3.02 um in the same type of microvessels as in Yen et al. [26]，depending on the labeling and analyzing methods. The ultrastructural organization of the ESG and its relation to the cytoskeleton components (e.g.F-actin) of ECs was first investigated by Squire et al. [28] Using computed autocorrelation functions and Fourier transforms of EM images of frog mesenteric microvessels， they estimated a quasi-periodic substructure in the ESG， which is a 3D fibrous meshwork with characteristic spacing ~20 nm. The recent study by Arkill et al. [29] observed similar ESG structures in mammalian microvessels of choroid，renal tubules， glomerulus， and psoas muscle， and was confirmed by a 3D reconstruction using electron tomography by employing a tannic acid staining method and a novel staining technique of Lanthanum Dysprosium Glycosamino Glycan adhesion (the LaDyGAGa method) [30]. The thickness of the ESG observed by their EM method is ~100 nm， similar to what was previously found on frog mesenteric microvessels [15]. 2. Experimental methods 2.1.Working principle of STORM The di f fraction of a single f luorescence molecule can be described as the point spread function (PSF).When the light of wavelength A excites the fluorophore (emitter)， the intensity profile of the spot is defined as the PSF with the width ~0.6A/NA， NA is the numerical aperture of t he objective. As a result， two identical emitters separated by a distance less than the width of the PSF would appear as a single object， not distinguished from each other (top plot at the right panel in Fig. 1A). The diffraction-limited image resolution， for a high numerical aperture objective lens， is ~200 nm in the lateral direction and ~500 nm in the axial direction， for a conventional f luorescence microscope [32].The key idea of the single-molecule localization microscopy is to light t he molecule， i n turn， to achieve the nanometer-l evel accuracy of their position and reconstruction i nto a super-resolution image， such as Stochastic Optical Reconstruction Microscopy (STORM). Figure 1A demonstrates the working principle of STORM by using photo-switching mechanisms to stochastically activate individual molecules (photo-switchable or photoactivatable fluorophores) within the diffraction-limited region at different times. Then images wi t h sub-diffraction l imit resolution are reconstructed from the measured positions of i ndividual fluorophores [3]. To localize a molecule， the emitter at that molecule should not have any overlap with its neighbors， so that it can be isolated and f itted into the PSF (middle and bottom plots in Fig. 1A). One way to keep the fluorescence molecule emitting in turn i s to make the photo-switchable fluorophore blink To achieve t he photo-blinking of the f luorophore is to switch between l ight and dark states， which are usually called “on”and“off’.“On’ is the state in which the fluorophore is able to be excited and i ts emission can be detected by the camera. The “off"state is in which the f luorophore cannot be excited by the laser， i ncluding the temporarily blinking or permanently bleaching. The most important factor for STORM to accompl i sh sub-di f fraction resolution is that the photo-switchable fluorophores are used to maintain neighboring molecules in different states， “on"or “off， enabling to be distinguished f rom each other [33]. Figure 1B shows a representative optical set up of STORM for the f luorophore ATTO 488From one optical path， the 405 nm wavelength laser activates the photo-switchable fluorophore ATTO 488， enabling the excitation. Then through another optical path， the 488 nm l aser exc i tes the activated “on”fluorophore and its emission is recorded by a camera. The number of emitters and the i mage collecting time are the determinants of the efficiency and accuracy of any microscopy. The higher population of the emitter would have a higher image collection efficiency， however， the distance between adjacent emitters must be greater than their PSF to distinguish each other for conventional microscopy. To trade the super spatial resolution (accuracy)， STORM sacrifices i ts temporal resolution (efficiency) by switching the state and sequentially exciting the emitters at high density. Rust et al. [3] employed organic dyes and fluorescent proteins as photo-switchable emitters to t rade temporal resolution for a super spatial resolution (~20 nm lateral and ~50 nm axial at present， can go down to a couple of nanometers if using smaller peptides or antibody fragments instead of currently used whole anti-bodies)， which is an order of magnitude higher t han conventional confocal microscopy.STORM can perform 2D [3]， 3D [32]， and multicolor [34] imaging. In each frame of a STORM movie，a set of emitters is randomly ac t ivated so that locations of single emitters isolated within their PSFs can be spatially resolved. Fig. 1. Schematics of the working principle and the optical se t -up of the stochastic optica l reconstruction microscopy (STORM).(A) The top row shows that the two f luorophores (located along the two dashed l ines) activated and exc i ted simultaneously by conventional microscopy. Due to the diffraction l imi t ， t he two f luorophores cannot be separated， resulting in a blurring image.The middle and bottom rows show the principle of STORM for the localization of a single molecule to a nanometer accuracy. The f l uorophores are activated and exc i ted not simultaneously， but sequentially to be localized and separated with each other. This technique can overcome the diffraction barrier for the conventional fluorescence microscopy. (B) A representative optical set-up of STORM f or the f l uorophore ATTO 488. (a) The optica l path i n ac t ivation process . The 405 nm wavelength laser act i vates the photo-switchable f luorophore ATTO 488， enabling the excitation. (b) The optica l path in excitation and imaging processes. The 488 nm laser excites the activated“on”fluorophore ATTO 488 and its emission is recorded by a camera. 2.2. Sample preparation Cell culture： Mouse brain microvascular endothelial cells (bEnd3) from ATCC (Manassas， VA) were cultured in Dulbecco’s Modified Eagle’s Medium/Nutrient Mixture F-12 Ham (DMEM/F-12)，2mM L-glutamine， and 100 U/mL penicillin and 1 mg/mL streptomycin (al l from Sigma-Aldrich， St. Louis，MO)， supplemented with 10% f etal bovine serum (FBS， Atlanta Biologicals， Flowery Branch， GA)， and incubated in the humidified atmosphere with 5% Co， at 37℃[35]. Immunofluorescence staining： The bEnd3 cells were trypsinized and seeded on the 50 pg/mL fibronectin coated No.1.5 glass-bottom dish (MetTek， Ashland， MA) at a density of 600 cells/mm’and cultured for 3-4 days until confluent. Then the cells were gently washed with 1% BSA/PBS， fixed with 2% paraformaldehyde and 0.1% glutaraldehyde for 20 min， reduced with 0.1% NaBH4 for 7 min， washed 3x with 1% BSA/PBS and blocked with 2% normal goat serum (NGS) for 30 min. For HS staining， the fixed cells were incubated with mouse anti-heparan sulfate 10e4 (1：100， Amsbio， Cambridge， MA) at 4℃ overnight， followed by an ATTO 488 conjugated goat anti-mouse IgG (1：100， Sigma-Aldrich， St Louis， MO)； for HA staining， the fixed cells incubated with Biotinylated hyaluronic acid binding protein (50 pg/mL， Amsbio， Cambridge， MA) at 4C overnight， followed by an Alexa Fluor 647 conjugated anti-biotin (1：200， Jackson ImmunoResearch， West Grove， PA). Finally， the samples were post-f ixed with 2%paraformaldehyde with 0.1% glutaraldehyde for 10 min and then kept in 1% BSA/PBS. 2.33..Image acquisition by STORM The samples were i maged using t he N-STORM system (Nikon Instruments INC.， Melville， NY)， with a 100×/1.49 oil i mmersion lens. Mul t iple Reporters of ATTO 488 and Alexa Fluor 647 ac t ivated by 405 nm wavelength laser with alternating illumination were used to acquire the 3D images of HS and HA. Three fields of 256 ×256 (40×40 um²) of HS and HA on the cell apical (luminal) surface， located at the top，middle and bottom (Fig.2) were obtained based on 40，000 of EM-CCD captured i mages for each reporter at a capturing speed of 19 ms/frame. The top region is focused at t he cell surface above the cell nucleus，the bottom region is focused at the valley of two adjacent cells， and the middle region is focused at a plane mid-way i n between top and bottom regions. The x-y plane shown i n Fig. 2 is the focal plane (horizontal plane) and the z-direction is perpendicular to the x-y plane (vertical direction). The raw image movies of blinking dots were processed by the analyzing software accompanied with the N-STORM system and demonstrated in Fig. 3. 2.4.Data analysis The single-emitter centroid algorithm [3] is applied to estimate the 3D locations of activated fluo-rophores i n the data movie and produces a 3D STORM image with the spatial resolutions of ~20 nm i n the lateral plane and ~50 nm in the axial direction， respectively. The ground-truth f luorophore locations in a component on a cell surface are estimated from the STORM image by using the mixed Gaussian model. The ultrastructure parameters of the component are then estimated on the basis of the estimated fluorophore locations. For the HS element， we assumed that it is a circular cylinder with the equivalent diameter having the same cross-sectional area as the real shape (elliptic-like). Other ultra-structural parameters， length， the orientation of an HS and the distance between adjacent HS elements (center to center) are defined in the i nset of Fig. 2. Due to a difficulty in recognizing individual HA elements， which are weaved into a network at the cel l luminal surface， we estimated the HA diameter from the 2D images of HA segments shown in Fig. 3. Fig. 2. Schemat i cs of image acquisition locations and defini t ions for t he ultra-struc t ura l parameters of HS elements. The focal plane x-y is the horizontal plane. The z -direction is perpendicular to the x-y plane wi t h the origin at the local cel l luminal sur f ace. The raw image movies (blinking dots) were collected from the cell luminal surface (origin) up to ~1.2 um from the origin. STORM images for 3 f ields of 40 um x 40 um were collected at the cell top (r ight above t he cel l nucleus)， at the cel l bottom (thinnest part of a cell ) and at the cell middle (mid-way between the top and bottom) from t he cell l uminal surface. The enlarged i nset shows t he def i nition for the diameter， length， and orientation of an HS element ， and the distance between adjacent HS elements. 3. Results 3.1. Organization of HS and HA Figures 3A-C demonstrate the organization of HS and HA revealed by STORM. Figure 3A is for that observed at the top region， Fig. 3B for that at the middle region and Fig. 3C for t hat at the bottom region of the cell luminal (apical) surface (see Fig. 2 for the definition of each region). For all the figures， the second column i s the HS image (green)， the third column i s the HA image (red) and the first column is the overlay of the second and the third columns. The f irst row is the 2D top view and the second row is the 3D view of the field of ~34 um ×34 um. The third and fourth rows are the enlarged 2D and 3D views from the region with the green dashed line in the f irst row. The fifth and sixth rows are the further enlarged 2D and 3D views from the region with the blue dashed line i n the third row. The color bar at the left of 3D views in the second and third columns represents the length scale. One can see from these figures that HA (red) is a long molecule weaving into a 2D network sheet which covers the cell luminal surface and is in the same plane as the cell luminal surface. In contrast， HS is a shorter molecule which is perpendicular to the cel l luminal surface. HS and HA are partially overlapped at the cell luminal surface. 3.22.. Ultrastructural parameters ofHS and diameter ofHA We employed the single-emitter centroid algorithm [3] and a mixed Gaussian model to estimate the ultra-structure of HS. Its diameter， length， and the orientation， as well as the density， distance between the Fig. 3. Illustrations of STORM images for HS and HA at the cell top (A)， at the cel l middle (B) and at the cell bottom (C) from the cell luminal surface. In all the illustrations， the second column is t he HS image (green)， the third column is the HA image (red) and the first column is the overlay o f the second and the t hird columns. The first row i s the 2D top view and the second row is t he 3D view of the f ield o f ~34 um×34 um. The t hird and fourth rows are the enlarged 2D and 3D views from t he region with the green dashed l i ne in the f irst row. The fifth and sixth rows are the further enlarged 2D and 3D views from the region with the blue dashed l ine i n the third row. The color bar at the lef t of the 3D view for HS (second column) and HA (third column)represent s the depth (length) scale. Table 1Quantification of HS ultra-structure at the bEnd3 cell luminal surface p value HS Top (T) Middle (M) Bottom (B) Tvs M T vs B M vs B Density (elements/um-) 0.3977 0.345 0.665 N/A Distance (nm) 1494.40±628.13 1844.8±758.5 1221.90±450.71 >0.1 <0.05 <0.05 Diameter (nm) 190.99±45.96 284.31±71.06 184.16±59.62 <0.001 >0.6 <0.001 Length (nm) 620.96±75.71 650.97±117.96 575.21±10.86 >0.1 >0.1 >0.1 α() 92.92±1.88 88.65±8.20 96.23±10.86 <0.001 >0.05 <0.001 B() 91.68±4.80 88.33±6.57 88.85±7.93 <0.001 >0.05 <0.001 Y(°) 4.70±2.33 9.54±5.71 11.81±7.85 <0.001 >0.05 <0.001 Data shown are mean ± STD. adjacent HS elements (see Fig. 2 inset for the definition of these parameters) are summarized in Table 1It should be noted that t he estimated parameters of t he HS or HA are for the conjugates used to visualize these structures， not the parameters for a single molecule. For example， the spatial diameter of a single HA is less than 1 nm [36]. The diameter of the observed HS is 191.0±46.0 nm， 284.3±71.1 nm， and 184.2±59.6 nm， respectively， at the top， middle and bottom regions of the cel l lumina l surface. There are no significant differences between the HS diameters at the top and bottom regions (p>0.6)， but there is a signif i cant difference between the estimated HS diameters at the top/bottom and middle regions (P <0.001). This difference i s most l i kely due to the til t ing between the cell lumina l surface and the focal plane (x-y plane) at the middle region. The length of the HS is 621.0±75.7 nm， 651.0±118.0 nm， and 575.2±105.6 nm， respectively， at the top， middle and bottom regions. No significant difference among the HS lengths in these 3 regions (p >0.1). For the HS orientation， its angle with the x-axis， α， is 92.9+1.9，88.7±8.2， and 96.2 ± 10.9 degree， respectively， at the top， middle and bottom regions； its angle with the y-axis， β， is 91.7±4.8，88.3±6.6， and 88.9±7.9 degree， respec t ively， at the top， middle and bottom regions； its angle with t he z-axis， y， is 4.7±2.3，9.5±5.7， and 11.8±7.9 degree， respectively， at the top，middle and bottom regions. There is no significant difference in the HS orientation (for all a， B，y angles)between those at t he top and bottom regions (p>0.05)； however， there is a significant difference between those at the top/bottom and middle regions (p<0.001). For t he HS distribution， the average density i s 0.398 elements/um²， 0.345 elements/um’ and 0.665elements/um²， respectively， at the top， middle and bottom regions of the cel l luminal surface. The distance between the adjacent HS is 1694.4±628.1 nm， 1844.8±758.5 nm， and 1221.9±450.7 nm， respectively，at the top， middle and bottom regions. No significant di f ference between the HS distributions at the top and middle regions (p >0.1) but there is a significant difference between the top/middle and bottom regions in the HS distribution. Table 2 summarizes the HA diameter which was estimated from the 2D images shown in Fig. 3. The diameter of the HA segments is 185.3 ± 44.7 nm， 155.5±57.2 nm， and 156.9±56.1 nm， respectively，at the top， middle and bottom regions of the cell luminal sur f ace. This are no significant di f ferences i n the HA diameter among these 3 regions (p>0.1). Comparison of the HA diameter with the HS diameter shows no di f ference between the HA diameter and the HS diameter at the top and bottom regions (p>0.2)but there i s a significant difference at the middle region (p<0.01). Table 2Diameter of HA at the bEnd3 cel l luminal surface p values HA Top (T) Middle (M) Bottom (B) Tvs M T vs B M vs B Diameter (nm) 185.27±44.72 155.51±57.17 156.86±56.08 >0.05 >0.05 >0.05 Data shown are mean ± STD. 4. Discussion Using the recently obtained N-STORM system， we revealed the spatio-chemical organization of HS and HA in the ESG of bEnd3 monolayer at the nanometer scale. Our results show that HS elements are perpendicular to the cell luminal surface， which is the x-y plane at the top and bottom regions， and is the plane a l ittle t i lt from the x-y plane at the middle region (see Fig. 2). The angle of the HS element with the cell luminal surface is in the range from 78 to 85 degree (90 degree is perfectly perpendicular). The length of the HS element is ~600 nm. This length is comparable to the estimated thickness of the ESG previously observed in vivo and i n vitro by conventional fluorescence microscopy [19，26，27，35]. HS is the glycosaminoglycan side chain of HSPG (heparan sulfate proteoglycan)， 40-300 sugar residues，~20-150 nm long [11]. The length of the HS element revealed i n this study most l ikely represents the length of a HSPG. In contrast， t he HA elements are long molecules which l ie on t he cel l luminal surface and form a hexagonal-like network. This observation is consistent with the f inding that HA is a non-sulfated GAG，which binds with its surface receptors CD44 and receptor for HA-mediated motility (RHAMM) [8]. Since HA l ies in the same plane as the cell surface， its coverage seems to be higher t han HS， which is shorter and perpendicular to the cell surface (Fig. 3). The higher coverage of HA favors it to be the molecular sieve. Direct and indirect evidence indicates that the ESG should play at least three important roles in maintaining vascular functions. 1) It can serve as a mechanosensor and transducer for the ECs to sense the blood flow； 2) It maintains normal microvessel permeability by forming a molecular sieve， which allows the passage of water and smal l molecules but restricts large molecules such as plasma proteins； and 3) It generates a lubrication buffer between circulating red blood cel l s (RBCs) and ECs forming the vessel wall to avoid RBCs damage， and a barrier to prevent adhesion of circulating leukocytes， platelets and tumor cells to the vessel wall [2]. It also shields EC surface receptors， preventing their hyperactivation [37].The revealed organization and distribution of HS and HA in the ESG by STORM provides the direct structural feasibility for the ESG to play these three roles. HS elements which are perpendicular to the cell luminal surface， are able to sense the blood flow and serve as a mechanosensor and transducer of the ECs The denser HS distribution at the bottom of the cell luminal surface， which may cover the intercellular junctions， along with the weaving network formed by HA， favor the ESG to be a molecular sieve for transvascular exchange， a barrier to the adhesion of circulating cells and the endothelial cells forming the vascular wall ， and a shield for t he EC surface receptors. To reveal the structure and organization of HS and HA in the ESG of bEnd3 cel l surface， we used the primary antibody or specific binding protein to recognize the HS and HA， and fluorescence-conjugated secondary antibodies for the optical imaging. Since the size of an antibody is ~10 nm， the observed diameters of the HS and HA elements (~150-200 nm diameter) overest i mate the real size by at least ~10%. The predominant GAGs in vascular ECs are HS， HA and chondroitin sulfate (CS). Of the three，the most abundant is HS， accounting for 50-90% of the total GAGs， the rest being comprised of HA and CS， and SA (sialic acids) [11]. Due to unavailability of proper probes for fluorescence labeling CS and SA at t he present for STORM i maging， we were not able to uncover the organization and structure of these components of the ESG. The study will be continued in t he future when the probes are available. As a molecular sieve to restrict proteins of 6-7 nm (serum albumin)， the gap size of t he HA network，~0.5 um，currently observed by STORM， is too large. The other components， such as CS and SA， may f i ll the gap and form the proper molecular sieve to albumin. In addition， other more sensitive antibodies against HS (e.g.JM403， Amsbio， Cambridge， MA) wi l l be used to reveal the HS structure. Specific enzymes wil l be used to degrade/modify either HS or HA to see how the resultant organization would be when only one of these elements is present . In summary， the spatio-chemical organization of the ESG at the bEnd3 cell surface was revealed， for the f irst time， by employing the super-resolution STORM. The ultra-structure parameters of HS and the diameter of HA were also obtained from the collected images. Acknowledgements This work is supported in part by NIH SC1CA153325-01，RO1NS101362-01，1UG3TR002151-01and NSF MRI CBET 1337746 References [1] Tarbell JM， Ebong EE. The endothelial glycocalyx： A mechano-sensor and -transducer. Science Signaling. 2008；1(40)：pt8 [2] Fu BM， Tarbel l JM. Mechano-sensing and transduction by endothel i al surface glycocalyx： Compos i tion ， structure， and function. Wiley Interdisciplinary Reviews Systems Biology and Medicine.2013；5(3)：381-90. [3] Rust MJ， Bates M， Zhuang X. Sub-diffraction-l i mit imaging by stochast i c optical reconstruction microscopy (STORM).Nature Methods . 2006；3(10)：793-5. [4] Patterson G， Davidson M， Manley S， Lippincott-Schwartz J. Superresolution i maging using single-molecule localization.Annual Review of Physical Chemistry. 2010；61：345-67. [5] Jones SA， Shim SH， He J， Zhuang X. Fast， three-dimensional super-resolution imaging of live cells. Nature Methods.2011；8(6)：499-508. [6]SSydor AM， Czymmek KJ， Puchner EM， Mennella V. Super-resolution microscopy： From single molecules to supramolec-ular assemblies. Trends in Cell Biology. 2015；25(12)：730-48. [7]1TTarbel l JM， Pahaki s MY. Mechanotransduction and the glycocalyx. Journal of Internal Medic i ne. 2006；259(4)：339-50. [8] Zeng Y. Endothelial glycocalyx as a critical signal l ing plat f orm integrat i ng the extracellular haemodynamic forces and chemical signalling. Journal of Cellular and Molecular Medic i ne. 2017；21(8)：1457-62. [9] Zeng Y， Ebong EE， Fu BM， Tarbell JM. The structural stabil it y of the endothel i al glycocalyx after enzymatic removal of glycosaminoglycans. PLoS One.2012；7(8)：e43168. [10]]TTarbel l JM， Cancel LM. The glycocalyx and its signi f icance in human medicine. Journal of Internal Medic i ne.2016；280(1)：97-113. [11] S Sarrazin S ， Lamanna WC， Esko JD. Heparan sulfate proteoglycans. Cold Spring Harbor Perspectives in Biology. 2011；3(7). [12] Luft JH. Fine structures of capillary and endocapillary layer as revealed by ruthenium red. Fed Proc . 1966；25(6)：1773-83. [13] B E aldwin AL， Winlove CP. Effects of perfusate composition on binding of ruthenium red and gold colloid to glycocalyx of rabbit aortic endothelium. The Journal of Histochemistry and Cytochemistry： Official Journal of the Histochemistry Society . 1984；32(3)：259-66. [14] S S ims DE， Horne MM. Non-aqueous fixative preserves macromolecules on the endothelial cel l surface： An in situ study.European Journal of Morphology.1994；32(1)：59-64. [15] Adamson RH， Clough G. Plasma proteins modify the endothel i al cel l glycocalyx of frog mesenteric microvessels . J Physiol .1992：445：473-86. [16] Rostgaard J， Qvortrup K. Electron microscopic demonstrations of f ilamentous molecular sieve plugs in capillary fenestrae.Microvasc Res. 1997；53(1)：1-13. [17] van den Berg BM， Vink H， Spaan JA. The endothel i al glycocalyx protects against myocardial edema. Circ Res.2003；92(6)：592-4. [18] Klitzman B， Duling BR. Microvascular hematocrit and red cell flow in resting and contracting striated muscle. Am J Physiol.1979；237(4)：H481-90. [19]Vink H， Duling BR. Identification of dist i nct luminal domains f or macromolecules， e rythrocytes， and l eukocytes within mammal i an capil l aries. Circ Res . 1996；79(3)：581-9. [20] Long DS， Smith ML， Pries AR， Ley K， Damiano ER. Microviscometry reveal s reduced blood viscosity and al t ered shear rate and shear stress profiles i n microvessels after hemodilution. Proc Natl Acad Sc i USA. 2004；101(27)：10060-5. [21] Pries AR， Secomb TW， Gaehtgens P. The endothelial surface layer . Pflugers Arch. 2000；440(5)：653-66. [22] van den Berg BM， Spaan JA， Vink H. Impaired glycocalyx barrier propert i es contribute to enhanced i nt i mal low-density lipoprotein accumulation at t he carotid artery bifurcat i on i n mice. Pflugers Arch. 2009；457(6)：1199-206. [23] Reitsma S， oude Egbrink MG， Vink H， van den Berg BM， Passos VL， Engel s W et al . Endothelial glycocalyx structure in the intac t carotid artery： A two-photon laser scanning microscopy study. Journal of Vascular Research. 2011；48(4)：297-306. [24] Megens RT ， Reitsma S， Schiffers PH， Hilgers RH， De Mey JG， Slaaf DW et al . Two-photon microscopy of vital murine elastic and muscular arteries. Combined structural and functional i maging with subcellular resolution. Journal of Vascular Research. 2007；44(2)：87-98. [25] Ebong EE， Macaluso FP， Spray DC， Tarbell JM. Imaging the endothe l ial glycocalyx i n vitro by rapid freezing/freeze substitut i on transmi s sion electron microscopy. Arterioscler Thromb Vasc Biol. 2011；31(8)：1908-15. [26]Yen WY， Cai B， Zeng M， Tarbell JM， Fu BM. Quantification of t he endothelia l surface glycocalyx on rat and mouse blood vessels . Microvasc Res. 2012；83(3)：337-46. [27] Betteridge KB， Arkill KP， Nea l CR， Harper SJ， Foster RR， Satchell SC et al . Sialic acids regulate microvesse l permeabil i ty， revealed by novel in vivo studies of endothelial glycocalyx structure and function. J Physiol. 2017；595(15)：5015-35.[28] Squire S . A chance to grow. Professional Nurse. 2001；16(8 Suppl)：S2. [29] Arkill KP， Knupp C， Michel CC， Neal CR， Qvortrup K， Rostgaard J et al. Similar endothel i al glycocalyx struc-tures in microvessels from a range of mammalian tissues： Evidence for a common fi l tering mechanism?. Biophys J.2011；101(5)：1046-56. [30]A Arkil l KP， Neal CR， Mantel l JM， Michel CC， Qvor t rup K， Rostgaard J et al. 3D reconstruction of the glycocalyx structure in mammalian capi l laries using electron tomography. Microc i rculation. 2012；19(4)：343-51. [31] Xia Y， Fu BM. Investigation of endothelial surface glycocalyx components and ul t rastructure by single molecule local-ization microscopy： Stochastic optica l reconstruct i on microscopy (STORM). The Yale Journal of Biology and Medic i ne.2018；91(3)：257-66. [32] Huang B， Wang W， Bates M， Zhuang X. Three-dimensional super-resolution imaging by stochastic optical reconstruction microscopy. Science. 2008；319(5864)：810-3. [33] Hel l SW. Far-field optical nanoscopy. Science. 2007；316(5828)：1153-8. [34] Bates M， Huang B， Dempsey GT ， Zhuang X. Multicolor super-resolution imaging with photo-switchable fluorescent probes.Science.2007；317(5845)：1749-53. [35] Fan J， Fu BM. Quanti f ication of malignant breast cancer cell MDA-MB-231 transmigration across brain and lung microvascular endothelium. Ann Biomed Eng. 2016；44(7)：2189-201. [36] Atkins E， Meader D， Scott J. Model for hyaluronic acid incorporat i ng four intramolecular hydrogen bonds. Int J Biol Macromol. 1980：2：318-9. [37] Zullo JA， Fan J ， Aza r TT ， Yen W， Zeng M， Chen J et al . Exocytosis of endothel i al lysosome-related organelles hair-tr i ggers a patchy loss of glycocalyx at the onset of sepsis. Am J Pathol. 2016；186(2)：248-58.