|
Short Communication |
Use
of Three-dimensional Fluorescence Deconvolution Microscopy for Study of Spatial
Distribution of Secretory Vesicles in Living Cells
LI Dong-Dong， GUO Xue-Bin， QU An-Lian*， XU Tao
(
Institute of Biophysics and Biochemistry, School of Life Science and
Technology, Huazhong University of Science and Technology, Wuhan 430074, China
)
Abstract A
three-dimensional image of a living cell is helpful for cell secretion study.
In this report , the three-dimensional fluorescence deconvolution microscopy
for observing living cells was studied, because this technique can obtain a
quick three-dimensional imaging with minimal fluorescence quenching and
cytotoxicity for living cell observation. The property of three-dimensional
point spread function (PSF) of imaging system was analyzed. The relationship
between experimental and theoretical PSF was illustrated, and the theoretical
PSF was proved that it could reflect the principle of imaging system with NA
1.65 objective in use. Three-dimensional deconvolution algorithm in this report
was proved effective by well-defined three-dimensional specimens. Furthermore,
the rat pancreatic β cell secretory vesicles labeled by acridine orange was
observed by using this technique. Results showed that the blurring induced by
out-of-focus light was removed by the deconvolution algorithm effectively,
under current experiment conditions (with NA 1.65 objective) the experimental
PSF approximated the theoretical PSF very well, and deconvolved living cell
images exhibited the spatial distribution of the secretory vesicles clearly.
Key words ecretory vesicle;
three-dimensional fluorescence deconvolution microscopy; point spread function
_________________________________________
Received: January 24, 2003 Accepted: May 12, 2003
This work was supported by grants of a
General Program (No.60071030) and a Key Program (No.30130230) from the National
Natural Science Foundation of China, and a grant from National Science Fund for
Distinguished Young Scholars (No. 30025023)
*Corresponding author: Tel,
86-27-87543104-86; Fax, 86-27-87542499；e-mail, [email protected]
以三维荧光反卷积显微技术研究活体细胞中分泌囊泡的空间分布
李栋栋 郭学彬 瞿安连* 徐涛
(
华中科技大学生命科学与技术学院生物物理与生物化学研究所, 武汉430074 )
摘要
获得活体细胞三维图像以观察细胞内分泌囊泡的空间分布有助于细胞分泌机制的研究。 三维荧光反卷积显微技术可以为活体细胞观察提供低荧光漂白, 低毒副作用的快速三维成像。 研究了显微成像系统实验测定和理论计算点扩展函数之间的关系，
并且实验验证了NA 1.65物镜条件下， 理论计算点扩展函数可以较好地反映显微成像系统的特性。
然后使用已知物理结构的三维样本对反卷积算法的有效性进行了研究。 进而对使用吖啶橙 (acridine orange) 标记的大鼠胰腺β细胞分泌囊泡进行观察。 结果显示， 反卷积算法可以有效地去除原始图像中因为焦外光影响产生的模糊， 处理后图像清晰地显示了细胞内分泌囊泡的空间分布。
关键词 分泌囊泡; 三维荧光反卷积显微技术; 点扩展函数
内分泌细胞和神经元通过分泌囊泡的胞吐释放激素和神经递质。
使用生物化学和电生理方法对分泌过程的研究已取得显著进展［1~3］， 但是对于分泌囊泡胞吐前阶段的研究只能通过动力学模型推测。
对活体细胞分泌囊泡进行直观观察将有助于深入了解分泌囊泡在细胞内的整个动态过程。 利用全内反射显微技术对细胞膜附近分泌囊泡动态过程的定量研究已经发现了很多有价值的信息［4~6］， 然而这种技术无法观察发生在细胞内部较深处的分泌囊泡运动过程。
共聚焦显微技术可以获得细胞的三维图像, 但其高强度激发光使荧光染料被快速漂白并对细胞造成毒副作用。共聚焦系统的逐点扫描成像方式可能导致图像不能正确反映在扫描时间内发生在细胞内的动态变化过程［7］。
三维荧光反卷积显微技术以光学切片方法获得样本的三维原始图像, 然后利用反卷积图像恢复技术去除焦平面外荧光造成的模糊。 这种技术可以为活体细胞观察提供低荧光漂白, 低毒副作用的快速三维成像。 点扩展函数（point spread function, PSF）表征了成像系统的特性，
在本文中我们使用实验方法分别测定了NA 1.35物镜和NA 1.65物镜条件下成像系统的PSF， 同时研究了PSF的数学模型， 证实在本文成像条件下理论计算PSF的正确性。 使用已知物理构造的三维样本讨论了一种反卷积算法的有效性。
对大鼠胰腺β细胞内分泌囊泡的观察结果显示反卷积处理后的细胞三维图像清楚地展现了分泌囊泡在β细胞内的分布情况。
1 材料和方法(Materials and
Methods)
1.1 数据采集系统
宽场荧光显微镜(IX70; Olympus, 日本), 物镜(APO
100×/1.65 NA Oil HR; UPIanApo 100×/1.35 NA Oil Iris; Olympus), 单色光源及控制器(Polychrome Ⅳ; Till
Photonics GmBH, 德国), 电荷偶合器件CCD (SensiCam, SVGA; PCO; 德国), 计算机(P4 1.7 G, 1 G RAM), 压电z轴扫描控制器(E-662.
LR; Physik Instruments, 德国)和图像采集软件(Till
vision 4.0; Till Photonics GmBH)组成。 将光轴定义为z轴。 使用压电z轴扫描控制器以0.2 μm步进间隔控制物镜沿光轴移动， 采集得到的一系列二维切片图像被用来重建三维图像［8］， 此时成像系统焦平面沿光轴步进的间隔是物镜步进间隔的75%［9］。
1.2
PSF的实验测定
取通体发光小荧光珠, λ表示发射荧光波长, NA表示物镜数值孔径, 当荧光珠直径小于显微镜分辨率即D≤0.61 λ/ NA时［10］, 它可用来充当点源, 其三维图像就是成像系统的PSF。 本实验使用了两种荧光珠， 直径0.05 μm荧光珠(FluoresbriteTM Carboxylate
Microspheres, 直径=0.0481 μm, 标准差=0.0052 μm, Polyscience, Inc)和直径为0.175 μm荧光珠(PS -SpeckTM, Green; Molecular
Probes)。实验中将荧光珠悬浮液滴在和物镜相匹配的玻片上(NA 1.65: 折射率=1.78, 厚度=0.15 mm, Olympus ,Inc； NA 1.35: 0#)静置晾干， 然后采集其三维图像。
1.3 PSF数学模型及与实验测定PSF的比较
三维显微成像系统的PSF可以由下式描述［11］：
式中C为幅度常数， J0(*)为零阶贝塞尔函数, 波数k=2π/λ, λ为发射光的波长, OPD为光程差。 将理论计算PSF和上述实验测定PSF进行下列比较： x-y切面和x-z切面图像；
三维光学传递函数［optical transfer function,
OTF ( u, v,η)］在不同η的径向频谱； 实验测定和理论计算PSF荧光分布和频谱。
1.4
反卷积算法
成像系统记录的原始图像是样本函数和PSF的卷积。 采集得到的三维原始图像因受焦外光的影响而变得模糊。
根据得到的原始图像和PSF求得样本函数的过程称为反卷积运算。 本文中使用的反卷积算法是基于最大似然的最大期望迭代算法(ML-EM)［12,13］。 我们使用直径15 μm荧光珠(FocalCheckTM
Fluorescent Microsphere, Mounted on slides; Molecular Probes, 美国) 验证反卷积算法，
激发光照射下约0.7 μm厚度的球壳发出荧光, 内部不发荧光。 因此理想三维图像的x-y切面和x-z切面应该是一定宽度的亮环。 亮环图像局部近似为二维高斯分布［4,14］， 亮环处最亮点荧光强度为Im, 荧光强度大于0.8 Im的像素为亮环宽度。 0.175 μm荧光珠的焦平面图像直径约为0.134 μm， 因此在图像中检测到亮环宽度接近0.53 μm时认为图像反映了样本结构。
1.5 细胞准备和囊泡标记
大鼠胰腺β细胞的原代培养方法本研究所已有报道［15］。 细胞贴壁后, 使用3 μmol/L吖啶橙溶液在21 ℃ 条件下温育15 min［16,17］, 然后用β细胞标准外液将细胞冲洗干净。 图像中的荧光亮点指示出分泌囊泡的空间位置。
2 结果 (Results)
2.1 成像系统PSF
0.05 μm荧光珠更接近点源， 但是它的荧光强度过弱， 图像信噪比较差。 用0.175 μm荧光珠可以在适当曝光时间下得到具有较满意信噪比的三维PSF。 结果显示实验测定和理论计算PSF对应x-y切面的衍射光环直径基本相等， 两者的x-z切面沿z轴的张角基本相等［图1 (A)］。 对PSF频谱［图1 (B)］和荧光强度分布［图1 (D)］进行威尔科克秩和检验(显著性水平α=0.05)证明两者没有显著性差异。 在三维OTF (u, v, η)不同轴向频率η处的二维径向频谱显示当η升高时信号径向频谱在低频段的损失快速上升［图1 (C)］。 本文对NA1.35物镜的实验结果和其他研究者的结论一致［11］， 不再列出。
在后面的研究中我们使用理论计算PSF进行反卷积处理。
Fig.1 Experimental
and theoretical PSF
(A) Comparison between experimental and theoretical PSF x-y and z
section image with equal defocus distance. (B) OTF of 2D section with equal
defocus distance (0.4 μm) from experimental and theoretical PSF. Radial cutoff
frequency is 6.2 μm－1. Curve 1, theoretical PSF section; curve 2,
the average of 5 experimental PSFs (±sx ). (C) Planar section at several axial
frequencies of the 3D optical-transfer function (100×1.65 NA): (Curve 1)η=0.0
μm－1; (Curve 2) η=0.1 μm－1; (Curve 3) η=0.15
μm－1; (Curve 4) η=0.27 μm－1. Radial cutoff
frequencies is 6.2 μm－1, axial cutoff frequencies is 1.53 μm－1. (D) Fluorescence
intensity distribution of 2D sections with equal defocus distance (0.3 μm) from
experimental and theoretical PSF. Curve 1, theoretical PSF; curve 2, the
average of 5 experimental PSFs (±sx). Objective Olympus APO
100×/1.65 NA Oil HR was used in this experiment.
2.2 反卷积算法
原始图像中亮环宽度大于荧光珠实际发光层宽度, z方向切面图像的下部顶点几乎无法分辨, 上部顶点过亮并有形变。 反卷积运算使图像得到显著改善［图2 (A)］。 亮环处荧光强度分布［图2 (B)和(C)］和处理前后图像的灰度直方图［图2 (D)］指示出反卷积运算去除模糊的过程。 较少的反卷积运算次数可以使x-y切面的亮环宽度接近真实值， 而z方向两个顶点的处理需要较多的反卷积运算次数 (图3)。
Fig.2 3D fluorescence deconvolution
(A) x-y and x-z section image of 15 μm fluorescent bead. Olympus
UPIanApo 100×/1.35 NA Oil Iris was used, calibration bar is 2 μm. Sampling
distance was 0.067 μm×0.067 μm×0.2 μm (x, y, z). Deconvolution iteration number
is 1600. (B) Profile of medial section, starting point is the center of
section. Curve 1, original image; curve 2, deconvolved result of iteration 800;
curve 3, deconvolved result of iteration 1600. (C) Fluorescence distribution of
two poles of x-z section. Three curves have the same meaning as in (B). (D)
Histogram of one 2D section of the 15 μm fluorescent bead. Curve 1, original
image; curve 2, deconvolved result of iteration 1600.
Fig.3
Deconvolution process
(A) Ring width in x-y section. Its shape approached the real structure
of 15 μm bead with the increase of deconvolution number. (B) Both of the poles
along z direction. Their shape approached real structure though their
distortion in original image were serious. Curve 1, the low pole of x-z
section; curve 2, the upper pole of x-z section. Both of which were displayed
in Fig.2 (A). All sections image were medial sections of 3D image. Every curves
was the average of 6 beads(±sx ). The definition about ring width was
demonstrated in our paper.
2.3
观察大鼠胰腺β细胞内分泌囊泡
反卷积运算使原始图像中的模糊被逐渐去除, 处理后图像显著地将发光亮点分离出来［图4 (A)和(B)］。 原始图像中指示分泌囊泡的荧光亮点因为受到焦外荧光的影响而不能反映分泌囊泡的实际大小，
反卷积运算使囊泡大小接近实际值约300 nm［8,9,18］［图4 (C)］。
Fig.4 3D
distribution of secretory granules in single rat pancreatic β cells
(A) 3D image of secretory granules labeled by acridine orange.
Deconvolved results displayed secretory granules distribution in 3D space of
cells. Calibration bar is 2 μm. Olympus APO 100×/1.65 NA Oil HR was used.
Sampling distance in 3D direction was 0.067 μm×0.067 μm×0.3 μm(x, y, z). (B)
Single granule fluorescence distribution. (C) Single granule image approaching
its real diameter with deconvolution curve was the average of 15 granules from
3 different cells (±sx).
3 讨论 (Discussion)
PSF对反卷积处理有重要影响［19,20］。 根据每一次活体细胞实验的条件实验测定PSF会给图像分析带来不便， 实际上也很难保证实验条件的完全一致。 根据具体实验的各个参数使用数学模型得到理论计算PSF有利于图像的反卷积处理， 前提是需要验证数学模型可以较好地反映实验测定值。 本文实验结果为在NA 1.65物镜条件下使用理论计算PSF提供了实验依据。 使用较高数值孔径物镜可以在相同曝光时间内获得更多样本信息, 减少曝光时间, 并使成像速度得到提高。 同时因为高数值孔径物镜提高了原始图像的分辨率, 所以相比使用较低数值孔径物镜的情况［12,13,21,22］， 反卷积运算能够以较少的运算次数得到较为理想的处理后图像。
本文使用已知物理结构的三维样本验证了反卷积算法。
结果显示荧光珠原始图像的x-z切面有一个顶点过亮, 这与以前研究［22］有所不同, 可能是因为本文选用了与以前研究不同的样本和成像系统。 由于原始图像z方向两个顶点处的轴向频率(η)较高， 因此顶点处图像位于OTF的低频截止区［图1 (C)］, 低频成分的丢失使原始图像z方向的两个顶点发生模糊和形变。但是相比以前研究, 反卷积算法对图像的改善基本一致［图3 (B)］。 相比共聚焦系统反卷积显微技术对活体细胞观察时可在一些方面取得相对较优的观察效果, 图像的空间分辨率可以高于共聚焦系统［7,23］。
反卷积结果显示大鼠胰腺β细胞分泌囊泡的大小约为0.3 μm， 这和肾上腺嗜铬细胞分泌囊泡的大小近似［8,9］。 切片1和切片2［图4 (A)］相距约1.7 μm, 两切片的囊泡分布基本没有联系， 细胞核在切片2中显现得较为充分， 切片中囊泡荧光强度的不同指示出囊泡的主体可能位于其他切片；切片1位于玻片和细胞膜界面附近， 切片2较远离玻片， 分泌囊泡在切片1中的密度大于切片2中的密度， 说明大鼠胰腺β细胞中分泌囊泡主要位于细胞膜附近， 这和嗜铬细胞［24］和PC-12细胞［9,25］中分泌囊泡的分布近似。 细胞需要通过待触发库(readily releasable pool, RRP)释放使分泌囊泡分泌［26,27］， 三维荧光反卷积显微技术可以研究分泌囊泡在细胞内三维空间的整个动态过程，
从而进一步阐明其中涉及的具体分子机制。
References
1 Zhu
H, Hille B, Xu T. Sensitization of regulated exocytosis by protein kinase C.
Proc Natl Acad Sci USA, 2002, 99(26): 17055－17059
2 Moser
T, Neher E. Rapid exocytosis in single chromaffin cells recorded from mouse
adrenal slices. J Neurosci, 1997, 17(7): 2314－2323
3 Xu
T, Bajjalieh SM. SV2 modulates the size of the readily releasable pool of
secretory vesicles. Nat Cell Biol, 2001, 3(8): 691－698
4 Oheim
M, Stuhmer W. Tracking chromaffin granules on their way through the actin
cortex. Eur Biophys J, 2000, 29(2): 67－89
5 Ohara-Imaizumi
M, Nakamichi Y, Tanaka T, Ishida H, Nagamatsu S. Imaging exocytosis of single
insulin seretory granules with evanescent wave microscopy: Distinct behavior of
granule motion in biphasic insulin release. J Biol Chem, 2002, 277(6): 3805－3808
6 Steyer
JA, Almers W. A real-time view of life within 100 nm of the plasma membrane.
Nat Rev Mol Cell Biol, 2001, 2(4): 268－275
7 Rizzuto
R, Carrington W, Tuft RA. Digital imaging microscopy of living cells. Trends
Cell Biol, 1998, 8(7): 288－292
8 Rehberg
M, Graf R. Dictyostelium EB1 is a genuine centrosomal component required for
proper spindle formation. Mol Biol Cell, 2002, 13(7): 2301－2310
9 Lang
T, Wacker I, Wunderlich I, Rohrbach A, Giese G, Soldati T, Almers W. Role of
actin cortex in the subplasmalemmal transport of secretory granules in PC-12
cells. Biophys J, 2000, 78(6): 2863－2877
10 Castleman
KR. Digital Image Processing. Beijing: Tsinghua University Press, 1998, 368－371
11 Gibson
SF, Lanni F. Experiment test of an analytical model of aberration in an
oil-immersion objective lens used in three-dimensional light microscopy. J Opt
Soc Am A, 1991, 9(1): 154－166
12 Markham
J, Conchello JA. Tradeoffs in regulated maximum-likelihood image restoration.
In: Cogswel C J, Conchello JA, Wilson T eds. 3D Microscopy: Image Acquisition
and Processing IV. Proceedings of the 1997 SPIE’s Biomedical Optics Symposium,
BiOS97, Bellingham, Washington, U.S.A.: SPIE――The Inter-national Society for
Optical Engineering, 1997, SPIE code No 298：4－18
13 Conchello
JA, McNally JG. Fast regularization technique for expectation maximization
algorithm for computational optical sectioning microscopy. In: Cogswel C J,
Kino G, Wilson T, eds. Three-Dimensional Microscopy: Image Acquisition and
Processing. Proceedings of the 1996 IS&T/SPIE Symposium on Electronic
Imaging: Science and Technology, Bellingham, Washington, U.S.A.: SPIE――The
International Society for Optical Engineering, 1996, SPIE code No 2655: 199－208
14 Cheezum
MK, Walker WF, Guilford WH. Quantitative comparison of algorithms for tracking
single fluorescent particles. Biophys J, 2001, 81(4): 2378－2388
15 Zeng
X, Xu J, Zhou Z, Qu A. A practical method of primary culture of rat pancreatic
β-cell. J Huazhong Univ of Sci Tech, 1998, 26(5): 24－26
16 Oheim M, Loerke D, Stuhmer W, Chow RH.
Multipile stimulation-dependent processes regulate the size of the releasable
pool of vesicles. Eur Biophys J, 1999, 28(2): 91－101
17 Steyer
JA, Almers W. Tracking single secretory granules in live chromaffin cells by
evanescent-field fluorescence microscopy. Biophys J, 1999, 76(4): 2262－2271
18 Lang
T, Wacker I, Steyer J, Kaether C, Wunderlich I, Soldati T, Gerdes HH, Almers W.
Ca2+-triggered peptide secretion in single cells imaged with green fluorescent
protein and evanescent-wave microscopy. Neuron, 1997, 18(6): 857－863
19 Carrington
WA, Lynch RM, Moore ED, Isenberg G, Fogarty KE, Fay FS. Superresolution
three-dimensional images of fluorescence in cells with minimal light exposure.
Science, 1995, 268(5216): 1483－1487
20 Boutet de Monvel, Le Calvez S, Ulfendahl
M. Image restoration for confocal microscopy: Improving the limits of
deconvolution, with application to the visualization of the mammalian hearing
organ. Biophys J, 2001, 80(5): 2455－2470
21 McNally
JG, Preza C, Conchello JA, Thomas Jr LJ. Artifacts in computational
optical-sectioning microscopy. J Opt Soc Am A, 1994, 11(3): 1056－1067
22 McNally
JG, Markham J, Conchello JA. Comparing methods for 3D microscopy. In: Cogswell
CJ, Conchello JA, Lerner JM, Lu T, Wilson T eds. Three-Dimensional and
Multidimensional Microscopy: Image Aquisition and Processing V. Proceedings of
SPIE, Bellingham, Washington, U.S.A.: SPIE――The International Society for
Optical Engineering, 1998, 3261: 108－116
23 Littlefield
R, Fowler VM. Measurement of thin filament lengths by distributed deconvolution
analysis of fluorescence images. Biophys J, 2002, 82(5): 2548－2564
24 Steyer
JA, Horstmann H, Almers W. Transport, docking and exocytosis of single
secretory granules in live chromaffin cells. Nature, 1997, 388(6641): 474－478
25 Pouli
AE, Emmanouilidou E, Zhao C, Wasmeier C, Hutton JC, Rutter GA.
Secretory-granule dynamics visualized in vivo with a phorgrin-green fluorescent
protein chimaera. Biochem J, 1998, 333(Pt 1): 193－199
26 Olofsson
CS, Gopel SO, Barg S, Galvanovskis J, Ma X, Salehi A, Rorsman P et al. Fast
insulin secretion reflects exocytosis of docked granules in mouse pancreatic
β-cells. Pflugers Arch, 2002, 444(1-2): 43－51
27 Xu
T, Binz T, Niemann H, Neher E. Multiple kinetic components of exocytosis
distinguished by neurotoxin sensitivity. Nat Neurosci, 1998, 1(3): 192－200