Supplementary Figure 1 Supplemental correlative nanomanipulation-fluorescence traces probing nascent RNA and fluorescent Mfd during TCR initiation. Supplemental correlative nanomanipulation-fluorescence traces showing (a) loss of fluorescent DNA probe signal upon formation of the Mfd-RNAP repair intermediate and (b) increase in fluorescence signal of SNAP-Mfd over the lifetime of the intermediate. In one of the latter traces photobleaching is observed prior to resolution of the intermediate, but not before a marked increase in fluorescence of SNAP-Mfd can be recorded. In another of the traces in this series, the DY-549 fluorescence of labeled Mfd appears approximately 30s prior to formation of the intermediate, consistent with the ~20 s rate-limiting catalytic step identified for this system 13.
Supplementary Figure 2 Supplemental correlative nanomanipulation-fluorescence traces probing fluorescent RNAP during TCR initiation. Supplemental correlative nanomanipulation-fluorescence traces showing an increase or a decrease, respectively, in fluorescence of SNAP-RNAP during the lifetime of the intermediate when RNAP was originally directed to transcribe (a) towards or (b) away from the surface. Traces for which there is no break in the time-axis were obtained by monitoring the nanomanipulated DNA for fluorescence even absent RNAP, and here the moment of binding of RNAP is seen to be coincident with formation of the initially-transcribing complex. Traces for which there is a break in the time-axis were obtained via pulse-chase experiments where RNAP was first seen to load onto DNA in the mechanical channel, and then excess RNAP washed out (break) before fluorescence imaging was started.
Supplementary Figure 3 Supplemental correlative nanomanipulation-fluorescence traces probing fluorescent RNAP transcribing towards the surface. Nanomanipulation and fluorescence time-traces for transcription, directed towards the surface, of the 400 bp construct (TTS400, see Supplementary Note). Such a transcribed region represents 140 nm of DNA contour length, but only 90 nm of vertical rise through the TIR field at the low extending force (F=0.3 pn) used here. (a) Time-traces corresponding to ~2/3 of the optimal events (11/15) described in the Venn diagram (Supp. Fig. 4) [The remaining 4 events were nanomechanically validated and displayed fluorescence but presented no increase in fluorescence for the duration of the event and were not fitted are are not shown]. When a clear change in fluorescence signal is observed (SNR>2.5 at RNAP binding and followed by fluorescence increasing by at least a factor of two during elongation) it is possible to obtain an estimate of velocity which is in reasonable agreement with the nanomechanical signal. Thus in red are shown single-exponential fits to the fluorescence increase in the time interval between promoter escape and transcription termination (as determined from the nanomechanical channel, vertical grey guides). The time inscribed above the nanomechanical trace gives this time interval. The time inscribed in the fluorescence trace represents the time constant for a single-exponential rise in fluorsecence (red line) obtained via non-linear fitting of the background-corrected data with fitting error bars obtained from photon counting statistics. Because the TIR depth in this experiment was ~120 nm (see below) and the vertical rise in the TIR field during elongation is ~90 nm, this time has been corrected by 25% before being inscribed here; it therefore represents the time required for transcription of 400 bp as estimated from the exponential properties of the TIR field traversed by the fluorescent RNAP. (b) Calibration of the TIR field for these experiments gave a good fit to a single exponential with decay length of 118 +/- 7 nm (s.e.m). Error bars on the fluorescence signal represent the standard error on the mean for counting statistics; error bars on the extension signal are negligeable. We note that the decay length,, thus obtained is consistent with expectations based on the equation = in /4 [(n 1 sin ) 2 (n 2 ) 2 ] 1/2, where in is the input wavelength, n 1 the index of refraction of the glass, the angle of the incoming beam relative to the optical axis, and n 2 the index of refraction of the media in which the evanescent wave decays. For in = 532 nm, n 1 = 1.52, 68 o (calculated for the input beam displaced from the optical axis by 4.2 mm, using the fact that in this setup first TIR light is observed at the glass-water interface for a displacement of 3.1 mm from the optical axis) and n 2 = 1.33, we obtain l ~ 90 nm, in reasonable agreement with the value obtained experimentally. Deviation from ideal behavior may come from both extrinsic effects (eg an imperfectly focused input beam may cause light to enter the objective at slightly different effective NAs) and intrinsic effects (eg if the magnetic beads were to display nonlinear autofluorescence). (c) The distribution of RDe lifetimes from mechanical events (corresponding to the time taken to transcribe 400 bp and terminate, and for which the termination time is small compared to the elongation time 35 ). These events correspond to the uncorrupted nanomechanical traces of transcription obtained for this dataset (ie the 85 events depicted in the Venn diagram Supplementary Figure 4). A gaussian fit to the data gives a lifetime of 40 +/- 1 s. (standard deviation 9 bp/s), corresponding to a velocity of 10 bp/s, in agreement with rates expected for NTP concentrations (200 M) and temperature (T=28 o C) used here.
Supplementary Figure 4 Venn diagrams of events obtained and presented. Venn diagrams of events obtained and presented. The surface of the white box represents the total number (n) of nanomechanical events obtained for each condition. Numbers inscribed within the diagram specify the number of events represented by the delineated surface they are contained within. The surface of the yellow oval represents the number of nanomechanical events for which a flawless nanomechanical trace was obtained over the entire duration of the reaction i.e. for which 1) there are clearly-detectable transitions in the DNA extension corresponding to well-established transcription events, 2) there is no excessive mechanical drift or noise over ~20-30 minutes, and 3) the intermediate state of a first complex is not interrupted by stochastic loading of a second RNAP 13. This represents typically upwards of 2/3 of experiments. The surface of the pink oval represents the number of nanomechanical events for which singlemolecule fluorescence is concomittantly monitored flawlessly over the lifetime of the object of interest ie for which 1) the protein or probe turns out to indeed have a functionning fluorophore attached to it, 2) there is no premature photobleaching and 3) fluorescence is not altered by loading of a second fluorecent RNAP. The pink region of this oval and the number inscribed therein corresponds to flawless fluorecent events but corrupted nanomechanical events. The orange region of overlap between yellow and pink ovals corresponds to events presented in this paper, i.e. those events for which there is full monitoring of the process both in the nanomechanical channel and in the fluorescence channel. At the same time all of the data extracted from the partial or interrupted events not part of this orange overlap supports the observations based on data with full coverage in both mechanical and fluorescence channels.
Supplementary Note: Experimental setup, calibration, and DNA constructs The combined magnetic trap and TIRF microscope used for these experiments is described in part (a) of the figure included in this Note. The output of a 25 mw 532 nm laser (Crystalaser, CA-532-025-L) is controlled by passing the laser through a ½ wave plate (WP) followed by a polarizing beamsplitter (PBS); the beam is then steered using mirrors (M1, M2) into a telescope consisting of two lenses (T1, f = -25 mm, d = 0.5 in.; T2, f = 100, d = 1 in.) to expand it to a diameter of about 8 mm. The expanded beam is then steered using mirrors (M3, M4) into an achromatic doublet (L1, f1 = 400 mm, d = 2 in.) which focuses the beam into the back focal plane (BFP) of a 60x 1.49 NA oil immersion objective (Olympus APON 60x, 1.49 NA). The laser is guided to the BFP by two mirrors (M5, M6) and a long-pass dichroic (D1, Semrock DI-01- R532); M6 is mounted on a translation stage which is used to switch between epifluorescence and TIRF as a function of whether the beam enters the objective from the center or the edge, respectively. The achromatic doublet L1 is also mounted on a translation stage which serves to optimize collimation of the laser as it exits the front lens of the microscope objective. To image the magnetic beads for simultaneous particle tracking and fluorescence imaging, the sample is illuminated from the top using a near-ir resonant-cavity LED (650 nm, Roithner Laser, RC-LED-650-02). Both near-infrared light carrying bead images, and 565 nm light emitted by the DY-549 fluorophore used in these experiments, are collected by the objective, transmitted through the long-pass dichroic (D1), and reflected into the camera system by a mirror (M7). Near-IR and yellow light are separated using a second long-pass dichroic (D2, Semrock, FF-605). Fluorescence light is steered using the dichroic (D2) and mirror (M8) through an emission filter (F2, Chroma, 575/50) and then a tube lens (L2, f = 200 mm, d = 1 in.) which focuses light onto the detector of an EMCCD camera (ixon+ 897, Andor). Near-IR light is steered using mirrors (M8, M9) through a filter (F3, Semrock LP-02-635-RS) and a tube lens (L3, f=200 mm, d = 1 in.) which focuses light onto a CCD camera (CM140GE, JAI). L2 makes the EMCCD chip conjugate with the imaging plane of the TIRF objective, while L3 makes the CCD chip conjugate with a plane located a few microns deeper into the sample, allowing for multi-plane imaging of a defocused bead image (for particle tracking purposes) and of the in-focus glass surface (for TIRF formation and fluorescence collection). The CCD camera is mounted on a 50 mm translation stage (Polytec PI) to facilitate control of the imaging plane made conjugate to the CCD. Further optimization of the system focuses on increasing the surface area of the image plane adequately illuminated by the TIR field, currently a disc on the order of 30 m diameter. The z-position and autofluorescence for a DNA-tethered magnetic bead are both collected to calibrate the TIRF field penetration depth; raw data for TIR field calibration in the main text are presented in parts (b-g) of the figure which accompanies this Note. Parts (b-d) show, respectively, an image of the bead as seen using the CCD camera and near-ir illumination; a time-trace of bead z-position and fluctuations as a function of DNA supercoiling; and a time-averaged trace of bead z-position as a function of supercoiling. Parts (e-h) show the same measurements carried out by recording bead autofluorescence (using the EMCCD) rather than bead z-position. Raw data (green) for bead Z-position is acquired for 35 seconds for each rotation; standard deviation of bead Z-position fluctuations Is on the order of 55 nm. Raw data averaged over 4 seconds is also shown (red). From the raw data the bead s cutoff frequency is determined to be ~4 Hz, thus in 35 seconds there are approximately 140 independent measurements of bead position. For mean bead position vs. supercoiling, error bars representing standard error on the mean z-position of
the bead are +/- 4 nm, too small to plot here. For bead autofluorescence vs. supercoiling, 16 images (200 ms each) were acquired for each rotation point, and error bars represent standard error on the mean. Scale bars: 5 micrometers. 3 kbp DNA constructs studied using this system are schematically depicted in part (h) of the figure accompanying this Note; distances are to scale. The SbfI end of the molecule is tethered to the glass surface, the XbaI end of the molecule to the magnetic bead. The green rectangle represents the T5 N25 promoter, the arrow the transcription start site, the magenta line the stalling position, and the red rectangle the tr2 terminator.