【JACS】突破至1100纳米近红外区！从头设计SWIR荧光激活蛋白
文章标题： De Novo Design of Near-Infrared Fluorescence-Activating Proteins
通讯作者： Chunfu Xu, David Baker
文章概要
引言
基于蛋白质的荧光成像在生物医学领域应用广泛，但在穿透力更强的短波红外（SWIR，1000–2000 nm）窗口，依然极度缺乏高性能的基因编码蛋白质探针。为此，研究团队开创性地将计算蛋白质设计与有机化学合成相结合，成功从头设计出能够特异性结合合成梅罗cyanine染料（merocyanine dyes）的荧光激活蛋白。这种方法通过精确调控蛋白质内部环境，使染料与特定赖氨酸残基形成席夫碱共价键并使其质子化，从而引发明显的红移，突破了现有天然荧光蛋白的波长限制，为活体深层组织成像提供了前所未有的高对比度和灵敏度。
Figure 1. Computational design strategy. (a) General procedure for the design of fluorescence-activating proteins. Symmetric helical bundle oligomers with central pores suitable for ligand binding are generated using parametric backbone design and experimentally characterized. Those that are found experimentally to adopt the target structure are connected into single-chain proteins with short loops and the residues surrounding the binding pocket are optimized by RosettaDesign to bind merocyanine retinals. To enable simultaneous sampling of the ligand conformation and design of the amino acid identities of the surrounding pocket, we treat the merocyanine retinal-lysine conjugates as NCAAs in Rosetta and allow them to sample a wide range of rotameric states. (b) Design model of a far-red fluorescence-activating protein utilizing the MeroCy7 dye (chemical structure shown in the figure) based on a pentameric two ring helical bundle. (c) Design model of a fluorescence-activating protein predicted to emit in the SWIR range utilizing the bulkier MeroCy9 dye and based on a hexameric two ring helical bundle. Spectra are for illustrative purposes only.
主要实验及结论
研究人员首先聚焦于远红外荧光激活蛋白的构筑。他们以经结构验证、具有中心孔道的对称五聚体螺旋束作为支架骨架，通过短接环将其连接为单链以打破结构对称性。在此基础上，团队将Retinal-Lysine共价结合物作为非天然氨基酸（NCAA） 引入Rosetta计算框架中，实现了配体构象采样与周边口袋残基的同步优化。通过高通量筛选与点饱和突变技术，最终设计出远红外荧光激活蛋白MC7BP34。实验证明，该蛋白与MeroCy7染料结合后能发出685纳米的强荧光，不仅量子产率高达0.44且亮度超越FPbase数据库中同类光谱的所有荧光蛋白，更成功在U2OS细胞中实现了线粒体靶向的活细胞高对比度成像，且与HaloTag和SNAP-tag展现出完美的成像正交性。
Figure 2. Design and experimental validation of a far-red fluorescence-activating protein. (a) Crystal structure (color) of the designed pentameric two ring helical bundle matches well with the design model (gray) with a Cα-RMSD of 0.86 Å. (b) Crystal structure (teal) of the designed far-red fluorescence-activating protein MC7BP34 aligns with the design model (gray). (c) Ligand binding pocket in the design model. Modeling suggests the small molecule should fit closely with the surrounding protein residues, but it is not discernible in the crystal structure due to the low resolution. (d) Absorption spectra of MC7BP34 and the K57R mutant incubated with MeroCy7 dye compared to MeroCy7 alone. The absorption peak shifts from 505 nm for the unbound dye to 655 nm for the protein-dye complex; the knockout mutant K57R does not induce a shift in the absorption spectra, suggesting that the dye molecule binds specifically to the lysine residue as designed. The K57R mutant remains monomeric and folded (Figure S6). (e) Excitation and emission spectra of MC7BP34 and the K57R mutant incubated with MeroCy7 dye, or MeroCy7 alone. (f) U2OS cells expressing a MC7BP34-GFP fusion targeted to mitochondria incubated with 0.3 μM MeroCy7 for 15 min then washed and imaged live by spinning disk confocal microscopy (single plane). Scale bars: 10 μm.
随后，团队进一步挑战了波长更长、分子体积更大的近红外和短波红外光谱区。针对体积更大的MeroCy9染料，研究人员选用了内部通道更宽阔的六聚体两环螺旋束支架WSHC6作为设计基底。由于单链桥接易导致序列重复并引发蛋白质聚集，团队利用Rosetta HBNet专门构筑了特定的氢键网络打破重复模式，从而稳定了目标折叠状态。历经多轮核心区域重新设计与优化，成功培育出MC9BP72及优化版本MC9BP81。光谱分析显示，该新型蛋白与MeroCy9结合后的复合物光的吸收峰大幅红移近300纳米至904纳米，其荧光发射峰不仅红移至920纳米，其次级发射峰更是延伸至1100纳米以上的短波红外区域，其光谱红移幅度相较于临床常用染料ICG而言优势极其显著。
Figure 3. Design and experimental validation of a SWIR fluorescence-activating protein. (a) Superposition of the crystal structure (color) and the design model (gray) of the hexameric two ring helical scaffold used for the SWIR design. The larger internal channel of this scaffold is suitable to bind bulkier ligand compounds. (b) The AlphaFold predicted model (blue) of the MeroCy9 binding SWIR protein MC9BP72 aligns with the design model (gray) with a Cα-RMSD of 1.41 Å. (c) The ligand binding pocket in the design model is complementary in shape to the MeroCy9 dye. (d) Absorption spectra of free MeroCy9 dye (10 μM), MC9BP72 (10 μM), and ICG (1 μM). The protein-dye complex absorbs light at 904 nm, representing an almost 300 nm red shift from the free dye’s absorbance and a 100 nm red shift compared to ICG. (e) Fluorescent emission spectra for ICG (green) and MC9BP72-MeroCy9 complex (red) obtained at an excitation wavelength of 785 nm. The emission peak of the designed protein-dye complex shifts over 100 nm toward the longer wavelength compared to that of ICG.
为了验证该短波红外探针在实际生物成像中的应用潜力，研究团队将MC9BP81与表皮生长因子受体（EGFR）结合蛋白以及绿色荧光蛋白进行融合表达。活细胞实验结果表明，在小鼠白血病K562细胞表面，该融合表达体系展现出极高的特异性膜定位标记能力。在最后的动物活体实验中，研究人员将表达有对照探针iRFP720的细胞包裹于阿尔金酸钠微球中并植入小鼠腹腔内。在相同深度的活体组织穿透测试中，传统的iRFP720由于受到组织自体荧光干扰而无法清晰辨识微球边界，而使用892纳米激发的MC9BP81-MeroCy9体系则凭借极低的组织背底荧光和散射，成功实现了对微球的高对比度、超灵敏深层空间轮廓定位。
Figure 4. SWIR imaging using MC9BP81. (a,b) Purified MC9BP81-EGFP-EGFRc and MC9BP81-EGFP-EGFRn fusion proteins can be readily chromophorylated with MeroCy9 as observed using SWIR imaging. Both constructs exhibited bright fluorescence when a 900 nm long-pass filter was used. (c,d) After incubation with chromophorylated MC9BP81-EGFP-EGFRc and MC9BP81-EGFP-EGFRn proteins, the EGFR and iRFP720 expressing K562 cells displayed a clear fluorescence signal when measured using a 900 nm long-pass filter. Zoomed-in cell images showing membrane localization after labeling with chromophorylated MC9BP81-EGFP-EGFRc protein are shown in Figure S12e. (b,d) Measurements from different samples show good agreement (data are from a single measurement).
总结及展望
这项研究有力地证明了计算蛋白质设计与有机合成化学交叉融合的巨大威力。团队不仅成功研发出了兼具超高亮度和极佳正交性的远红外基因编码探针MC7BP34，更创制出首个发射光谱能够跨越至SWIR区域的从头设计荧光激活蛋白MC9BP81。尽管该体系在绝对亮度上仍有优化空间，且游离染料在水相中的稳定性仍需改善，但全α螺旋的独特结构为其转化为跨膜传感器奠定了坚实基础。这不仅为生物体内深层动态成像与神经科学电信号监测提供了崭新的原型工具，也为未来设计波长更长、性能更优的短波红外生物传感器彻底打开了技术大门。