Real-time precision opto-control of chemical processes in live cells

There is a critical need in biological science to control molecular activities at high spatial precision to resolve spatial distributions of targets, avoid off-target effects, and dissect subcellular functions. Conventional chemical treatments lack spatial selectivity; optical tweezers/trapping manipulate only a few pre-identified targets; laser ablation/manipulation rely on pre-imaging and manual targeting; and optogenetics requires pre-imaging and typically lacks subcellular precision. Because living systems are dynamic with moving organelles and migration, pre-imaging–based methods are inadequate. The goal of this work is to achieve simultaneous real-time detection and selective control of molecular targets with high chemical specificity and submicron spatial accuracy. The authors introduce real-time precision opto-control (RPOC), which detects chemical-specific optical signals during laser scanning and, based on preset criteria, triggers a control laser to interact only at those pixels, enabling precise, selective modulation of molecular activities and chemical reactions in live cells.

Existing approaches provide limited chemical selectivity and spatiotemporal precision for controlling intracellular targets. Chemical treatments are spatially nonspecific and may cause off-target effects. Optical tweezers and trapping can physically manipulate only a few pre-detected targets. Laser ablation/manipulation typically require pre-acquired images and manual operation, which is unsuitable for dynamic samples. Optogenetics controls neuronal functions via light-sensitive ion channels but generally requires pre-imaging and lacks subcellular precision. The field thus lacks a technique that can both detect and control molecular targets in real time with high chemical selectivity and submicron accuracy.

RPOC platform: During fast laser scanning, optical signals (e.g., fluorescence, SRS) from targets are detected at each pixel and compared to preset thresholds using high-speed comparator circuitry. When conditions are met, a TTL output drives an acousto-optic modulator (AOM) to gate a separate control laser to the same pixel in real time. Digital logic (AND/OR/NAND/NOR) enables multi-channel gating and intensity passband selection for active pixel (APX) determination.

• Optics: Dual-output femtosecond laser; 1045 nm output used as Stokes; tunable output used as pump for SRS; same sources provide TPEF excitation. Portions of both are frequency-doubled to visible for control beams (e.g., 522 nm, 400 nm). Control laser passes through an AOM; 1st order is coupled into the microscope with IR beams.
• Electronics and timing: Comparator box response ~15 ns; AOM ~7 ns; total RPOC response for fluorescence <50 ns (including path and cable delays). For SRS signals, the response is dominated by the lock-in time constant, affecting no more than ~1 pixel (~90 nm) in oversampling. Pixel dwell time during scanning is 10 µs.
• Spatial performance: SRS and TPEF spatial resolution ~373 nm; control laser spatial precision ~525 nm. Oversampling is used throughout (pixel size smaller than beam waist), so the laser interaction area is larger than the APX footprint.
• APX determination: APXs are pixels where the control laser is activated. Threshold tuning determines APX distribution; higher thresholds reduce APXs. Two comparator boxes can select an intensity passband (between lower V₁ and upper Vᵤ). Logic gating across detectors (e.g., SRS and TPEF) restricts APXs to specific structures (e.g., ER-associated compartments).
• Demonstrations of APX selection: Mixed fluorescent particles imaged in 570/60 nm channel under 800 and 1045 nm excitation; appropriate threshold V₁ aligns APXs with particles. Raman-selective APXs using hyperspectral SRS of PMMA and PS (e.g., 2955, 2990, 3060 cm⁻¹) by tuning laser frequencies to target specific vibrational modes. Live-cell APX tracking shows APX trajectories match lipid droplet (LD) motion using SRS at 2855 cm⁻¹.
• Chemical control (CMTE photochromic molecule): CMTE switches between open (1a) and closed (1b) forms; 1b has a strong Raman peak at 1510 cm⁻¹ (pump ~902 nm). SRS imaging can also convert CMTE to 1b; a 522 nm control laser (from 1045 nm frequency doubling) converts 1b to 1a within selected APXs. Cells (MIA PaCa-2) cultured with CMTE show CMTE accumulation in LDs. RPOC performed at ~10 µW control power; multi-frame or single-frame scans used. Quantification involves integrating SRS intensity changes at 1510 cm⁻¹ within APX-associated pixels.
• Digital logic control with two detectors: AND logic between SRS (CMTE, 1510 cm⁻¹) and TPEF (ER-Tracker, 540–600 nm) to select ER-associated aggregates for site-specific RPOC.
• Microtubule dynamics control (PST-1): Synthesis of photoswitchable photostatin PST-1 (trans inactive; cis active). UV-Vis validation under 405 and 532 nm matches literature. HeLa Kyoto cells stably expressing EGFP-EB3 used to visualize microtubule polymerization. TPEF imaging with femtosecond excitation (chirp rods bypassed). Cells treated with 4 µM trans-PST-1 for 15 min. RPOC uses 400 nm pulses (from 800 nm doubling) to locally convert PST-1 to cis within APXs. Imaging protocol: T1 (10 frames pre-RPOC), T2 (5 frames RPOC), T3 (10 frames post-RPOC), T4 (20 frames later); frames in each window averaged for analysis. Controls include whole-field blue scanning without PST-1.
• LD trafficking assay: Femtosecond SRS (CH₂ stretch) detects LDs; APXs determined to target LD-adjacent pixels for PST-1 activation. Time-lapse imaging before (0–220 s) and during RPOC (222–442 s). LD trajectories tracked; maximum displacements quantified to assess transport inhibition at APXs.

• System performance: Comparator response ~15 ns; AOM ~7 ns; overall RPOC latency <50 ns for fluorescence. For SRS, lock-in limits may affect ≤1 pixel (~90 nm) in oversampling. Spatial resolution for SRS/TPEF ~373 nm; control laser precision ~525 nm. Pixel dwell time 10 µs.
• APX control and chemical specificity: Threshold tuning aligns APXs with fluorescence signals; Raman-selective APXs demonstrated for PMMA vs PS using 2955 and 3060 cm⁻¹ peaks; non-resonant (e.g., 2990 cm⁻¹) yields no APXs. APX trajectories match LD motion in live cells, enabling real-time targeting of moving structures.
• Quantitative, site-specific control of CMTE: In CMTE-loaded MIA PaCa-2 cells, RPOC with ~10 µW at 522 nm selectively converts 1b→1a only at APXs, evidenced by reduced SRS at 1510 cm⁻¹ localized to APX pixels after 5 frames. With AOM constantly on for 20 frames, reduction occurs across the field (nonselective control). Integrated SRS intensity change per aggregate scales quadratically with the number of APXs; mean intensity change scales approximately linearly, reflecting oversampling geometry. By selecting passbands, RPOC targets either aggregate centers (V₁=0.15 V, Vᵤ=0.4 V) or edges (V₁=0.10 V, Vᵤ=0.12 V), achieving sub-aggregate spatial control.
• Organellar logic targeting: Using AND logic between SRS (CMTE 1510 cm⁻¹) and ER-Tracker TPEF, RPOC specifically converted CMTE on ER-associated aggregates while leaving non-ER aggregates unaffected.
• Microtubule polymerization control with PST-1: Blue light RPOC (400 nm) in cells treated with 4 µM trans-PST-1 disrupted EB3 dynamics (loss of high-intensity EGFP-EB3 streaks). In APX-rich centrosomal areas, T1 vs T3 showed strong disparities (change), whereas T3 vs T4 overlapped, indicating suppressed dynamics post-RPOC. In untreated controls (blue light scanning without PST-1), high differences persisted across time windows, indicating ongoing dynamics. The cis-PST-1 relaxes back to trans without blue light, consistent with recovery after RPOC.
• Lipid droplet trafficking: During RPOC, LDs associated with APXs exhibited markedly reduced long-distance transport compared to pre-RPOC and to non-APX LDs. Histograms of maximum displacement showed statistical reduction during RPOC for APX-associated LDs, demonstrating precise, local inhibition of microtubule-dependent transport.

RPOC enables, for the first time, real-time, chemically selective, submicron-precision control of molecular activities and reactions in living cells by coupling detection and actuation within each pixel during scanning. Adjustable comparator thresholds and digital logic across multiple detection channels provide versatile gating for complex biological contexts (e.g., restricting control to organelle-associated targets). Demonstrations with CMTE validate precise, quantifiable, and spatially confined photochemical switching, including sub-aggregate targeting. Biological application with PST-1 establishes site-specific inhibition of microtubule polymerization, evidenced by localized disruption of EB3 dynamics and reduced LD trafficking only at APXs. The approach is broadly applicable to photochromic vibrational probes, photoswitchable fluorophores, and light-sensitive reactions. Future system enhancements, such as improved control beam shaping, CW-laser implementations for cost and integration benefits, and acousto-optic tunable filters for flexible beam selection, could expand accessibility and functionality.

This work introduces RPOC, a platform that detects chemical-specific optical signals during scanning and instantaneously actuates a control laser at the same pixel, achieving selective, real-time manipulation of molecules and reactions in live cells with submicron precision. The system demonstrates: (i) precise and quantifiable control of a photochromic molecule (CMTE), including organelle-logic targeting; and (ii) site-specific inhibition of microtubule polymerization using a photoswitchable inhibitor (PST-1), leading to localized suppression of EB3 dynamics and LD transport. RPOC’s chemical specificity, speed, and spatial accuracy make it a powerful general tool for studying and modulating intracellular chemistry and dynamics. Future directions include optimizing control-beam precision, developing compact CW-based implementations, integrating with commercial microscopes, and adding programmable spectral selection (e.g., AOTFs) for multi-laser control.

• Oversampling geometry causes the laser interaction area to be larger than the APX footprint, so chemical changes may extend slightly beyond APX pixels; conversely, larger pixel sizes can make APXs equal to or exceed the interaction area.
• For SRS-based control, the lock-in time constant introduces a response lag up to approximately one pixel, which can slightly shift effective control locations.
• EGFP-EB3 fluorescence is weak and subject to photobleaching under femtosecond excitation, complicating direct quantification of dynamics; indirect LD-based assays are used to corroborate results.
• The active cis form of PST-1 thermally relaxes back to trans, limiting the duration of inhibition without continued activation.
• Proper threshold selection (V₁, Vᵤ) is required to align APXs with targets; suboptimal settings can under- or over-target regions.