Fluolab【Nat.Methods】破译细胞成像“卡脖子”难题！覆盖450-660nm全可见光谱的20+种新型抗原稳定荧光纳米抗体工具包
文章标题：Synthetic multicolor antigen-stabilizable nanobody platform for intersectional labeling and functional imaging
通讯作者：Axel Nimmerjahn & Vladislav V. Verkhusha
Fig. 1: A series of fluorescent nanobodies to GFP (VIS–FbGFP) spanning the visible spectrum.
a, Schematic of a nanobody with inserted red FP (PDB 1ZGO) bound to GFP (PDB 3OGO). CDRs are highlighted in violet. The position of mCherry insertion to the anti-GFP nanobody is indicated with a red arrow. Structural representations were generated using PyMOL. b, Fluorescence intensity of cells transfected with 13 different VIS–Fbs targeting GFP. The following FPs were coexpressed with VIS–Fbs as positive controls: mEGFP for mTagBFP2–FbGFP, mScarlet-I–FbGFP, mScarlet–FbGFP, mCherry–FbGFP, mNeptune2–FbGFP, mCardinal–FbGFP; mVenus for mTFP1–FbGFP; EBFP2 for mWasabi–FbGFP, mNeonGreen–FbGFP, mOrange–FbGFP, LSSmOrange–FbGFP, CyOFP1–FbGFP, LSSmScarlet–FbGFP. The following FPs were coexpressed with VIS–Fbs as negative controls: mCherry for mTagBFP2–FbGFP, mTFP1–FbGFP; mTagBFP2 for other VIS–Fbs. c, Fluorescence images of HeLa cells coexpressing VIS–Fbs from b with their cognate antigens. d, HeLa cells coexpressing NES–mTagBFP2–FbGFP with mEGFP, H2B–CyOFP1–FbGFP with mEGFP, Mito–mCardinal–FbGFP with mEGFP. e, Multicolor fluorescence images for HeLa cells coexpressing mEGFP antigen with three nanobodies in different compartments: NES–mTagBFP2–FbGFP in the cytoplasm, Mito–Cardinal–FbGFP in the mitochondria and H2B–CyOFP1–FbGFP in the nucleus. f, Change in fluorescence intensity of HeLa cells coexpressing PAmCherry-based VIS–FbGFP and mEGFP in response to 390 nm irradiation. g, Change in fluorescence intensity of representative cells coexpressing PAmCherry-based VIS–FbGFP and mEGFP in response to 390 nm irradiation. Change in fluorescence for three regions of interest (ROIs) from f is presented. h, Contrast of PAmCherry-based VIS–FbGFP coexpressed with either mTagBFP2 (n = 18) or mEGFP (n = 15) after 390 nm irradiation. i, Change in fluorescence intensity of HeLa cells coexpressing mEos4a–FbGFP (red and green forms) in response to 390 nm irradiation. j, Change in fluorescence intensity of representative cells coexpressing mEos4a–FbGFP and EBFP2 in response to 390 nm irradiation. Change of fluorescence for three ROIs from i is presented. k, Contrast of mEos4a–FbGFP coexpressed with either mTagBFP2 (n = 14) or EBFP2 (n = 20) after 390 nm irradiation. The maximal fluorescence of antigen-bound form for each VIS–FbGFP was assumed to be 100%. Data are presented as mean values ± s.d. for n = 3 transfection experiments. Scale bars, 40 μm.
文章概要
引言
多色荧光蛋白质与生物传感器的开发极大地推进了现代生物科学的发展，使科学家能够直接在活细胞及生物体内观察基因表达、蛋白质动力学和分子相互作用。然而传统的荧光蛋白直接融合技术常常会干扰靶蛋白的天然功能、细胞定位、表达水平以及正常代谢。虽然基于 camelid 抗体的重组纳米抗体提供了一种高特异性的替代方案，但将其与常规荧光蛋白融合形成的传统“显色体”在细胞内 constitutive 表达时，会产生大量未结合的游离荧光探针。这些游离探针会带来极高的背景荧光噪音，严重掩盖了真正目标抗原的荧光信号。为了彻底解决这一成像瓶颈，研究团队在此前开发的近红外抗原稳定纳米抗体的基础上，成功克服了绿色荧光蛋白类（GFP-like）蛋白质由于较大的 桶状结构及柔性两端对空间构象稳定性的巨大挑战，成功将这一抗原依赖性自动降解技术拓展至了整个可见光谱，开发出了全新的可见光抗原稳定荧光纳米抗体（VIS–Fbs）合成生物学平台。
Fig. 2: Multicolor imaging with three VIS–Fbs in HeLa cells.
a, Fluorescence intensity and images of cells expressing mCardinal–Fb59H10 or NIR–Fb59H10 for HIV p24 protein. Co-transfection with a plasmid encoding p24–sfGFP was used as a positive control (+), and co-transfection with pmEGFP-N1 was used as a negative control (−). b, Fluorescence intensity and images of cells expressing mTagBFP2–Fb2E7 or NIR–Fb2E7 for HIV gp41 protein. Co-transfection with a plasmid encoding gp41–sfGFP was used as a positive control (+), and co-transfection with pmEGFP-N1 was used as a negative control (−). c, Fluorescence intensity and images of cells expressing mTagBFP2–FbALFA, mStayGold–FbALFA, CyOFP1–FbALFA, TagRFP-T–FbALFA, mScarlet–FbALFA or NIR–FbALFA for ALFA-tag. Co-transfection with a plasmid encoding 24× ALFA-tag repeats was used as a positive control (+), and co-transfection with pmEGFP-N1 was used as a negative control (−). d, Multicolor images of HeLa cells coexpressing three different VIS–Fbs: mTagBFP2–Fb2E7, mScarlet–FbALFA/mStayGold–FbALFA and mCardinal–Fb59H10 with their cognate antigens fused to a NES, nuclear localization signal, clathrin localization signal or membrane-localization signal, respectively: NES–gp41–sfGFP-Tyr66Gly, H2B–ALFA-tag–sfGFP-Tyr66Gly/24xALFA-tag–clathrin-sfGFP-Tyr66Gly and p24–sfGFP-Tyr66Gly–CAAX. Scale bars, 40 μm. In a and b, the maximal fluorescence of antigen-bound form for each VIS–Fb was assumed to be 100%. In c, for VIS–FbALFAs, fluorescence of mTagBFP2–FbALFA was assumed to be 100%. In a–c, data are presented as mean values ± s.d. for n = 3 transfection experiments.
主要实验及结论
研究团队首先以绿色荧光蛋白纳米抗体为支架，通过系统性地删减和优化嵌入荧光蛋白的氨基酸 端与 端序列，成功构建出了首个抗原依赖型的 mCherry 红色荧光纳米抗体，并证实其荧光强度与抗原浓度呈现显著的剂量依赖性。在此成功经验的基础上，团队进一步实现了模块化设计的标准化。他们精准比对并调整了 12 种涵盖蓝光到深红光光谱的高亮度单体可见光荧光蛋白的末端构象，将其批量嵌入纳米抗体支架中，打造出了一系列对比度高达 10 到 110 倍的全新可见光荧光纳米抗体家族。为了进一步丰富该工具箱的功能多样性，研究人员还成功将光控激活荧光蛋白 PAmCherry 和光控转换荧光蛋白 mEos4a 融合进该平台，实现了极高时空分辨率的光控标记与荧光状态转换。
Fig. 3: Simultaneous detection of different metabolites with jRGECO1a–FbLAG30 in live GFP-expressing cells.
a, Change in fluorescence intensity of the cell coexpressing jRGECO1a–FbLAG30 (red) and mEGFP (green) in response to 5 μM ionomycin. b, Change in fluorescence intensity of the cell coexpressing jRGECO1a–FbLAG30 (red) and Green Pegassos pyruvate biosensor (green) in response to 1 mM pyruvate. c, Change in fluorescence intensity of the cell coexpressing jRGECO1a–FbLAG30 (red) and PyronicSF pyruvate biosensor (green) in response to 10 mM pyruvate. d, Change in fluorescence intensity of the cell coexpressing jRGECO1a–FbLAG30 (red) and iGlucoSnFr biosensor (green) in response to 20 mM pyruvate. Scale bars, 40 μm (a–d). e, Contrast of mEGFP coexpressed with jRGECO1a–FbLAG30 (n = 10) after addition of 5 μM ionomycin. f, Contrast of Green Pegassos only (n = 9) and Green Pegassos coexpressed with jRGECO1a–FbLAG30 (n = 9) after addition of 1 mM pyruvate. g, Contrast of PyronicSF only (n = 10) and PyronicSF coexpressed with jRGECO1a–FbLAG30 (n = 9) after the addition of 10 mM pyruvate. h, Contrast of iGlucoSnFr only (n = 11) and iGlucoSnFr coexpressed with jRGECO1a–FbLAG30 (n = 13) after the addition of 20 mM glucose. i, Contrast of jRGECO1a (n = 16) and jRGECO1a–FbLAG30 only (n = 10) or jRGECO1a–FbLAG30 coexpressed with Green Pegassos (n = 9), PyronicSF (n = 9) or iGlucoSnFr (n = 13). In a–i, data are presented as mean values ± s.d.
在应用层面上，这一多色平台展现出了极为惊人的多参数细胞成像与体内功能监测能力。通过给不同的 VIS–Fbs 融合特定器官定位信号，研究者实现了在同一细胞内对单一抗原在细胞质、细胞核以及线粒体等不同细胞器中的多色空间定位与分流操纵。此外，团队还成功跨越了结构体量的限制，将体积更大的钙离子生物传感器 jRGECO1a 以及多种代谢物生物传感器插入不同的纳米抗体位点，实现了在活细胞内对两种不同代谢产物在极微小空间内的实时同步功能成像，且完全不影响传感器本身的动力学响应特征。
Fig. 4: Performance of dTomato–FbLAG16 in HeLa cells and cultured primary neurons.
a, Red fluorescence intensity of cells transfected with mCherry–FbGFP, dTomato–FbLAG16 without linkers or dTomato–FbLAG16 with –GGS-linkers and coexpressed with mEGFP (right column (+)) or mTagBFP2 (left column (−)). b, Fluorescence images of HeLa cells coexpressing dTomato–FbLAG16 with mTagBFP2 (negative control) or mEGFP (positive control). c, Scheme of a VIS–Fb with a red FP (PDB 1ZGO) inserted into LAG16 anti-GFP nanobody (PDB 6LR7) bound to GFP-based biosensor GCaMP6m (PDB 3WLD). CDRs are highlighted in violet. The position of dTomato insertion to the anti-GFP nanobody is indicated with a red arrow. Structural representations were generated using PyMOL. d, Upper, representative image of HeLa cells coexpressing GCaMP6s and dTomato–FbLAG16. Three ROIs are indicated with white squares. Lower, changes in fluorescence intensity of the same cell coexpressing GCaMP6s (green) and dTomato–FbLAG16 (red) in response to 5 μM ionomycin. Fluorescence changes for three ROIs are shown. e, Upper, contrast of GCaMP6s only (n = 10) and GCaMP6s coexpressed with dTomato–FbLAG16 (n = 11) after addition of 5 μM ionomycin. Lower, contrast of dTomato–FbLAG16 (n = 11) for the data presented in the left graph. f, Coexpression of dTomato–FbLAG16 fused to RiboL1 tag and mEGFP in the soma of hippocampal neurons. The maximal fluorescence of antigen-bound form for dTomato(GGS)–FbLAG16 was assumed to be 100%. Data are presented as mean values ± s.d. for n = 3 transfection experiments. Scale bars, 40 μm (b and d) and 20 μm (f).
Fig. 5: Intersectional targeting of specific cell compartments and populations with dTomato–FbLAG16 in GCaMP6f reporter mice.
a, Schematic of the experimental approach. An AAV vector driving dTomato–FbLAG16 expression under the hSyn promoter and the soma-targeting peptide RiboL1 was stereotactically injected into the somatosensory cortex of Thy1-GCaMP6f mice with preferential calcium indicator expression in a subset of excitatory pyramidal neurons. b, Immunostaining validation. Left, confocal images showing GCaMP6f-expressing (gray) and dTomato-expressing (red) cells in a cortical tissue section. Center, zoom-in of the indicated region. Scale bars, 250 μm (left) and 50 μm (center). Right, population analysis (n = 12 tissue sections from four mice). Data are presented as mean values ± s.d. *P = 0.00008. c, In vivo validation. Left, representative dual-color two-photon time-lapse imaging showing GCaMP6f-expressing (gray) and dTomato-expressing (red) cells in the somatosensory cortex of a behaving Thy1-GCaMP6f mouse. Recording depth (z) from the pial surface and seven somatic ROIs are indicated. Center, fluorescence transients in the indicated ROIs are shown as Δ_R/R_ (blue) for the combined channels. The simultaneous mouse locomotor activity on a spherical treadmill is shown above. Scale bars, 50 μm (left), 50 mm s−1 and 200% (center). Right, zoom-ins of the two periods indicated in c (center). The in vivo time-lapse recording in c and the corresponding xy fluorescence image stack, including the recording site, are shown in Supplementary Videos 1 and 2. d, Schematic of the experimental approach. An AAV vector driving dTomato–FbLAG16 expression under the astrocyte enhancer 3xCore2(390m) was stereotactically injected into the somatosensory cortex of GFAP-GCaMP6f mice with preferential calcium indicator expression in astrocytes. e, Immunostaining validation. Left, confocal images showing GCaMP6f-expressing (gray) and dTomato-expressing (red) cells in a cortical tissue section. Center, zoom-in of the indicated region. Scale bars, 250 μm (left) and 50 μm (center). Right, population analysis (n = 6 tissue sections from two mice). Data are presented as mean values ± s.d. *P = 0.00001. f, In vivo validation. Left, representative dual-color two-photon time-lapse imaging showing GCaMP6f-expressing (gray) and dTomato-expressing (red) cells in the somatosensory cortex of a behaving GFAP-GCaMP6f mouse. Recording depth (z) from the pial surface and seven somatic ROIs is indicated. Center, fluorescence transients in the indicated ROIs are shown as Δ_R/R_ (blue) for the combined channels. The simultaneous mouse locomotor activity on a spherical treadmill is shown above. Scale bars, 50 μm (left), 50 mm s−1 and 200% (center). Right, zoom-ins of the two periods indicated in f (center). The in vivo time-lapse recording in f and corresponding xy fluorescence image stack, including the recording site, are shown in Supplementary Videos 3 and 4. g, Schematic of the experimental approach. An AAV vector driving dTomato–FbLAG16 expression under the DLX2.0 enhancer was stereotactically injected into the somatosensory cortex of Viaat-GCaMP6f mice with calcium indicator expression in inhibitory interneurons. h, Immunostaining validation. Left, confocal images showing GCaMP6f-expressing (gray) and dTomato-expressing (red) cells in a cortical tissue section. Center, zoom-in of the indicated region. Scale bars, 250 μm (left) and 50 μm (center). Right, population analysis (n = 6 tissue sections from two mice). Data are presented as mean values ± s.d. *P = 0.000082. i, In vivo validation. Left, representative dual-color two-photon time-lapse imaging showing GCaMP6f-expressing (gray) and dTomato-expressing (red) cells in the somatosensory cortex of a behaving Viaat-GCaMP6f mouse. Recording depth (z) from the pial surface and seven somatic ROIs is indicated. Center, fluorescence transients in the indicated ROIs are shown as Δ_R/R_ (blue) for the combined channels. The simultaneous mouse locomotor activity on a spherical treadmill is shown above. Scale bars, 50 μm (left), 50 mm s−1 and 200% (center). i, Zoom-ins of the two periods indicated in i (center). The in vivo time-lapse recording in i and the corresponding xy fluorescence image stack, including the recording site, are shown in Supplementary Videos 5 and 6. In b, e and h, statistical significance was assessed using a paired two-tailed t-test. Schematics in a, d and g created in BioRender; Barykina, N. https://biorender.com/5cdph4w (2026).
在活体动物实验中，研究团队利用腺相关病毒（AAV）载体将融合了细胞体靶向肽的红荧光纳米抗体导入表达钙指示剂的转基因小鼠大脑皮层中。借助于该探针无背景残留、抗原结合后高度稳定的特性，成功实现了高信噪比的比率计功能成像（Ratiometric Imaging）。这一技术不仅能精确区分小鼠在运动行为中大脑皮层内兴奋性锥体神经元、抑制性中间神经元以及星形胶质细胞的特定钙活动，还极大清除了神经纤维网背景信号的干扰。最后，研究团队将该技术推进到脊椎动物胚胎发育学研究中，开发出了针对内源性 -catenin 蛋白的绿色荧光纳米抗体并注射到斑马鱼受精卵中。在无需基因敲入的条件下，首次在活体斑马鱼胚胎早期发育及药物小分子干预过程中，以单细胞分辨率实时追踪到了内源性信号通路核心蛋白的动态时空分布与降解规律。
Fig. 6: In vivo tracking of endogenous β-catenin in zebrafish.
a, A scheme of experiment: the pTol2-sfGFP–FbBC2 plasmid encoding VIS–Fb for β-catenin was injected into wild-type zebrafish embryos at the one-cell stage. At 24 hpi, larvae were screened for green fluorescence and then divided into two groups. One group remained untreated (control), while the other one was incubated with 10 μM of β-catenin inhibitor IWR-1 for 6 h. At 30 hpi, larvae exhibiting mosaic expression of sfGFP–FbBC2 were mounted on a zWEDGI chamber and imaged using a spinning-disk confocal microscope for up to 17.5 h. Zebrafish pretreated with 10 μM IWR-1 were also imaged in continuous presence of 10 μM IWR-1. b, Upper part, three-dimensional (3D) view time-lapsed images from intravital spinning-disk fluorescent confocal microscopy of 30 hpi wild-type untreated zebrafish larvae previously injected at the one-cell stage with pTol2-sfGFP–FbBC2 for β-catenin. Time-lapse images acquired every 3.5 h are presented. Shown are zoom-ins for xy and xz projections for the sfGFP–FbBC2 channel, as well as xy projection of the merged brightfield and green fluorescence channels. Dashed white ROIs and orange arrows indicate cells displaying an increase in endogenous β-catenin signal over time. Blue arrows highlight a migrating cell. Lower part, 3D view time-lapsed images from intravital spinning-disk fluorescent confocal microscopy of 30 hpi wild-type zebrafish larvae pretreated with 10 μM of β-catenin inhibitor IWR-1, previously injected at the one-cell stage with pTol2-sfGFP–FbBC2 for β-catenin. Time-lapse images acquired every 3.5 h are presented. Shown are zoom-ins for xy and xz projections for the sfGFP–FbBC2 channel and xy projection of the merged brightfield and green fluorescence channels. Dashed white, orange and blue ROIs indicate cells showing a decrease in endogenous β-catenin signal over time in response to IWR-1 treatment. c, Fluorescence intensity of sfGFP–FbBC2 was quantified in control (untreated) zebrafish (n = 3) and IWR-1-treated zebrafish (n = 3) after 17.5 h of live imaging. Intensities at 17.5 h were normalized to the corresponding values at 0 h for each group. d, A scheme of experiment: the pTol2-sfGFP–FbBC2 plasmid encoding VIS–Fb for β-catenin was injected into wild-type zebrafish embryos at the one-cell stage. At 6 hpi, larvae were divided into three groups in a 24-well plate: control (untreated), treated with 10 μM of β-catenin inhibitor IWR-1 or 0.2 mM of β-catenin activator LiCl. After 24 h of incubation, zebrafish larvae were imaged using a fluorescence microscope. e, Representative fluorescent images of 24 hpi larvae with mosaic expression of sfGFP–FbBC2: untreated, or treated with 10 μM IWR-1 or 0.2 mM LiCl. White arrows indicate β-catenin-expressing cells. f, Proportion of larvae exhibiting fluorescence, no fluorescence and mortality in each treatment condition (control, 10 μM IWR-1 and 0.2 mM LiCl). g, Percentage of zebrafish larvae with developmental defects across treatment conditions (control, 10 μM IWR-1 and 0.2 mM LiCl). Scale bars, 100 μm, 20 μm, 40 μm (b) and 500 μm (e). The in vivo time-lapse recordings in b are shown in Supplementary Videos 7 and 8. Diagrams in a and d created inBioRender; Barykina, N. https://biorender.com/5cdph4w (2026).
总结及展望
该研究成功构建出了一个模块化、可量身定制且高度通用的可见光抗原稳定荧光纳米抗体合成生物学技术平台。该平台的核心设计逻辑在于利用精心工程化改造的连接肽，使荧光纳米抗体在游离状态下产生内部结构张力并暴露疏水区，从而触发细胞内泛素化机制使其迅速降解；而一旦与目标抗原特异性结合，这种结构张力即被释放，探针得以稳定并发出明亮荧光，从而彻底消除了背景噪音。这一突破性的技术完美避开了复杂的基因编辑流程，具备极高的安全性和极低的细胞毒性。未来，该设计策略不仅有望进一步推广至荧光素酶、转录激活因子和工程酶等更为广泛的功能蛋白质功能调控中，更将在细胞谱系追踪、超分辨显微成像、活体多靶点动态监测以及细胞治疗质控等诸多前沿医学与生物学工业化场景中发挥不可替代的核心工具价值。