Does 405 nm light set off phosphorescence in the eye?

Does 405 nm light set off phosphorescence in the eye?

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

While playing with a blue/violet laser (5 mW, 405 nm) tonight, I noticed that on dark, non-fluorescing surfaces, the beam reflection was trailed by what looked like a dimmer reflection of the beam. The effect was most easily observed when I moved the laser in a circular motion. The apparent 2nd reflection was not produced by a side lobe of the laser diode, as it always trailed the main reflection, regardless of clockwise or counterclockwise rotation. Also when I stop moving the beam, the second 'reflection' quickly merged with the main beam reflection.

I doubt it is caused by phosphorescence of the material (cloth) I was lasering, as it also showed up when I pointed the laser at the floor, other sorts of cloth and painted walls.

So, do rods, cones, visual purple or other eye components phosphoresce when hit with deep blue light?

Instead of phosphorescence, a more parsimonious explanation is the perception of afterimages. Afterimages regularly appear after viewing a bright stimulus. They occur in the same location in the visual field as the original stimulus, but lack clarity. Afterimages depend on the intensity and contrast of the original stimulus (i.e., they are more pronounced with bright stimuli in a dark environment), the time of fixation (i.e., longer fixation generate more persistent afterimages), and retinal adaptation (i.e., dark environments enhance afterimages).

After images are almost always the complementary color (negative afterimage) to the original stimulus, but can very briefly be the same color (positive afterimage) when viewing an exceptionally bright stimulus. A stimulus consistently produces the same afterimage, which varies in size based on the distance between the person and the background. After images are often revived by blinking.

Afterimages are thought to derive mainly from photobleaching of the retina, although cerebral processes may contribute too.

Phosphorescence in the retina has to be induced by injecting phosphorescent markers (Wanek et al., 2013). As far as I know, the retina is not autophosphorescent.

- Gersztenkorn & Lee, Survey Ophthalmol, (2015): 1-35
- Wanek et al., Curr Eye Res (2012); 37(2): 132-137

Fluorescence and Phosphorescence

Fluorescence and phosphorescence are types of molecular luminescence methods. A molecule of analyte absorbs a photon and excites a species. The emission spectrum can provide qualitative and quantitative analysis. The term fluorescence and phosphorescence are usually referred as photoluminescence because both are alike in excitation brought by absorption of a photon. Fluorescence differs from phosphorescence in that the electronic energy transition that is responsible for fluorescence does not change in electron spin, which results in short-live electrons (<10 -5 s) in the excited state of fluorescence. In phosphorescence, there is a change in electron spin, which results in a longer lifetime of the excited state (second to minutes). Fluorescence and phosphorescence occurs at longer wavelength than the excitation radiation.

Imaging and Spectroscopic Analysis of Living Cells

Mikko Koskinen , . Pirta Hotulainen , in Methods in Enzymology , 2012

2.3.3 Example experiment

In this photoactivation experiment, rat hippocampal neurons from E17 embryos were cotransfected with PAGFP-actin and mCherry at DIV 10, and the photoactivation analysis was performed at DIV 14 according to the microscope setup and settings described in Section 2.3.1 . In the photoactivation assay, the optimal transfection time was longer than what we normally use (normally one day). The longer transfection time increased the activated PAGFP intensity but did not affect neuron survival. The representative cell and spine were selected based on mCherry fluorescence ( Fig. 3.3 A ). The ROI for photoactivation was set around the selected spine ( Fig. 3.3 A), and the photoactivation analysis was recorded using the FRAP wizard. The 488 nm excited time frames before (− 1.6 s) and after (from + 1.6 to + 1000 s) photoactivation are shown in Fig. 3.3 A as heat maps (white-highest intensity, blue-lowest intensity). The photoactivation protocol was recorded from four spines from three different cells. Only the spines containing a stable pool of F-actin fluorescence observable 300 s after activation were taken into further analysis. The averaged fluorescence decay curve shows (see the dotted line in Fig. 3.3 B) that approximately 10% of total F-actin belongs to a stable pool of F-actin in these DIV 14 spines.

Figure 3.3 . (A) Hippocampal neurons from E17 rats were cotransfected with mCherry and PAGFP-actin at DIV 10 and imaged at DIV 14. Cells were visualized by mCherry excitation, and ROI for activation (red circle) was set based on mCherry fluorescence. Scale bar, 1 μm. (B) The averaged fluorescence decay of PAGFP-actin from four spines that contained the stable pool visible after 5 min postactivation reveals that the size of the stable pool is approximately 10% of the total actin. Error bars represent SEMs. (C) A heat map of the time-lapse images taken before (− 1.6 s) and after (from 1.6 to 1000 s) the photoactivation shows the increase (1.6 s) and then decay of the PAGFP-actin fluorescence. The spine outlining is based on simultaneously acquired mCherry fluorescence.

What does an absorption spectrum look like

The diagram below shows a simple UV-visible absorption spectrum for buta-1,3-diene - a molecule we will talk more about later. Absorbance (on the vertical axis) is just a measure of the amount of light absorbed. The higher the value, the more of a particular wavelength is being absorbed.

You will see that absorption peaks at a value of 217 nm. This is in the ultra-violet and so there would be no visible sign of any light being absorbed - buta-1,3-diene is colorless. You read the symbol on the graph as "lambda-max". In buta-1,3-diene, CH2=CH-CH=CH2, there are no non-bonding electrons. That means that the only electron jumps taking place (within the range that the spectrometer can measure) are from pi bonding to pi anti-bonding orbitals.

Does 405 nm light set off phosphorescence in the eye? - Biology

At either end of the rainbow there is light we can't see. Below the red end is near infrared light, shorter in wavelength than the infrared we feel as heat. Above the violet end of the rainbow is near ultraviolet, longer in wavelength than the ultraviolet light that causes sunburn.

Because these invisible lights are so close to the visible light we can see, they act in many ways just like normal light. They can be focused with lenses, and cameras can see them.

Fluorescent light

Seeing in the infrared

A source of near infrared light is as close as your television remote control.

Infrared remote control

Look directly into the front of the remote control and push some buttons. You can't see anything change. But now point the remote at an inexpensive digital camera or video camera and push the buttons. Now, in the camera's screen, you see a brightly flashing beacon of light. You can use it as a flashlight to read by, as long as you look in the camera screen.

Infrared remote control

Digital camera sensors can easily see in the infrared. In fact, they are more sensitive to infrared than they are to normal light. Camera manufacturers try to fix this by adding infrared blocking filters in front of the sensor. In cheap cameras, these are not very effective. In more expensive cameras, such as the digital single-lens reflex camera used to take the shot below, the infrared light is much dimmer than in the pocket camera shots above.

Infrared remote control

Seeing in the ultra-violet

Cheap ultraviolet light sources use a mercury vapor lamp and some dark purple glass that blocks most of the visible light from shining through. Unfortunately, rather a lot of blue and violet light gets through.

A better source is an ultraviolet light emitting diode. While some of the cheap LEDs called "ultraviolet" are actually really just purple LEDs, you can obtain good LEDs whose wavelength is 395 nanometers or below. At the time of this writing, 395 nm LEDs can be had for a few dollars, while a 375 nm LED penlight goes for $25, and a 255 nm LED will set you back $600.00.

It is not advisable to look into an ultraviolet LED, any more than it is to look at the sun. But a quick glance shows that a 395 nm LED looks dimly blue-violet to the human eye. But when we look at it through the camera, we can see that the camera sensor is again rather good at seeing this otherwise mostly invisible light. The light looks blue-white to the camera because the camera and its software interpret ultraviolet as blue, and this light is very bright in the ultraviolet.

Ultraviolet light emitting diodes

Bees can see in the ultraviolet. Flowers take advantage of this, and have ultraviolet pigments that bees can see but humans cannot. But when we take such a flower, in this case a California Poppy, and illuminate it with our 395 nanometer ultraviolet LED, and take a picture, we can see some of these markings.

In the photo below, I took the picture on the left in normal light, and the picture on the right using the UV LED. You can see that the edges of the flower are lighter in color, and the center is darker, except for the stamens, which are very bright.

Poppy in visible and ultraviolet

Invisible ink

Invisible ink has many uses beyond secret communications between spies or criminals. One such use is as an anti-counterfeiting measure in money, as seen by the photo below, where a U.S. twenty dollar bill is lit from behind by one of our ultraviolet LEDs.

Twenty dollar bill in ultraviolet light

In the bright green band of fluorescent ink, you can read the words "USA TWENTY".

How to see invisible ink

  1. Invisible ink (either in a vial or in an invisible ink pen).
  2. A source of ultraviolet light, such as a UV light emitting diode.
  3. (Optional) Blue-blocking sunglasses (yellow goggles).

The invisible ink pen is sold as a security marker, so you can invisibly write your name or driver's license on your belongings.

The fluorescent dyes can be used as ink in a fountain pen or a quill pen, or you can just dip a cotton swab into the dye and write with that.

They can also be poured into ink pads and used as hand stamps at dances or events, to permit re-entry without paying again.

To see the ink, put a small button cell lithium battery between the two leads of the ultraviolet LED. If it doesn't light up, turn the battery over (LEDs are diodes, and only light up if the current is flowing in the right direction). You can use two batteries to get a brighter light. Just stack the batteries one on top of the other, and place the LED wires on the top and bottom as if it were one battery.

Ultraviolet LED with two CR2032 batteries

Aim the light at the invisible writing. You may want to do this in a dark place for better readability.

The blue-blocking sunglasses are useful for enhancing the contrast. They block most of the blue and violet light that leaks through the purple glass filters of mercury vapor lamp ultraviolet lights. They are still useful with the ultraviolet LEDs, even though those emit much less visible light.

How to make your own fluorescent dye

You can make your own fluorescent dye easily, after a trip to the kitchen for some powdered turmeric spice, and the bathroom for some rubbing alcohol.

Put a teaspoon of turmeric into a glass, and add a few tablespoons of alcohol. Fold a paper towel into quarters to make a funnel, and filter the liquid into another glass. The photo below shows the liquid in a tiny glass vial, lit by our UV LEDs.

You can use a cotton swab to write things with your new ink. It will appear yellow on the paper, and bright yellow-green when lit with ultraviolet light.

Tiny vial of turmeric dye

How does it do that?

A typical fluorescent dye (several are shown below) has many conjugated bonds, where double bonds and single bonds alternate in the molecule.

Curcumin molecule

What is actually going on in the molecule is more complicated than the simple diagrams suggest, because the electrons that form those bonds aren't pinned down in one place, but tend to wander or smear out along the whole molecule. The electrons have many possible energy states. The lowest energy state (the ground state) is one where the electrons are closest to the nucleus of the atom, and are paired up, with each electron in the pair spinning in the opposite direction. We say that one electron is spin up, and the other is spin down.

Ethidium bromide molecule

At room temperature, the electrons have enough heat energy to bounce around and bump one another between energy levels at random. But most of the electrons will be in the lower energy levels at any given time.

Fluorescein molecule

The electrons in the molecule have energy levels of different types. The basic energy levels are the electronic states (the ground state, the excited ground states, and the exited triplet state). But layered on top of these are vibrational and rotational energy states.

Propidium iodide molecule

These extra vibrational states allow the electron to absorb photons of different energies that raise the energy of the electron to different levels. Since the energy of a photon is related to its wavelength (the color of the light), many slightly different colors of light can be absorbed by the electron to push it into higher energy states.

Sulforhodamine B molecule

Once an electron has absorbed a photon and has jumped into a higher energy state, it can lose some of the energy easily by releasing the vibrational energy as heat. The electron then settles into the lowest vibrational energy level at that excited electonic state.

The three electronic states are diagrammed below. In the ground state, electrons are paired, one with spin up, and one with spin down. In the excited state, one of the electrons has a higher energy, but still has the opposite spin. In the third (triplet) state, one of the electrons has flipped its spin, so both electrons have the same spin.

Electron states

All of these energy levels and states come together in what is called the Jablonski diagram, which is used to show what is happening during fluorescence.

Jablonski diagram for fluorescence

The diagram shows an electron in the ground state absorbing a photon and rising to the second excited singlet state (S2). This transition is shown in purple. Because the photon had more energy than was necessary to achieve the second electronic excited state, the electron is sitting at one of the vibrational energy levels in that electronic state. As time goes by, the electron loses energy as heat and settles into the bottom vibrational level of the second excited singlet state. It can lose more energy as heat and move down to the bottom of the excited singlet state 1, either in one step or in many small steps, stopping at various vibrational energy levels. These energy drops are shown in red in the diagram.

It is possible for the electron to lose energy further, either by emitting a photon of light, or by what are called non-radiative means, where the energy ends up as heat. When a photon is emitted, we say the molecule has fluoresced. This is shown in yellow in the diagram.

Jablonski diagram for phosphorescence

Sometimes the electron will lose energy by flipping its spin, and jump to the excited triplet state. It can then jump back and fluoresce like before, in which case it is called delayed fluorescence. Or it can flip again, and jump to one of the vibrational energy levels just above the singlet ground state. We call this phosphorescence. Flipping electron spins is statistically unfavorable, and this causes a delay in releasing the energy as a photon. Phosphorescent materials glow for a while after the exiting light is removed (glow in the dark paints are phosphorescent).

Adding heat energy makes it more likely that electrons will shake loose from their current states and fall down into lower states. Heating a glow in the dark toy will make it appear brighter, but the glow will fade way faster.


A very long lasting phosphorescent material has recently been discovered. It glows bright green for over 12 hours (all night long) after charging in the afternoon sun for an hour or two. It lasts so long, and glows so brightly, it has been named Kryptonite, after the glowing green mineral in the Superman comics and movies.

Kryptonite in the light

You can charge this stuff up by setting it next to a lamp for a while, then turn off the lamp and have a free night-light, or read Superman comics with it under the covers.

Kryptonite in the dark

Imagine sprinkling sand made from Kryptonite onto sidewalks before the concrete hardens. Now you won't need street lights. How many other uses can you think of? Post your ideas on our message board.


In epi-illumination, a dichroic beamsplitter is used to deflect the incident light towards the specimen. The spectral characteristics of the dichroic mirrors have been designed in such a way that only the desired excitation wavelengths are deflected downwards through the objective onto the specimen, while the unwanted wavelengths are transmitted by the dichroic mirror and collected in a light trap [15] behind the dichroic mirror. The elimination of this unwanted excitation light results in a significant decrease of stray light and thus improves the image contrast. The dichromatic mirror deflects the desired (short wavelength) excitation light through the objective onto the specimen, but is transparent to the longer fluorescence wavelengths. The suppression filter (barrier filter) absorbs (or reflects) the excitation light reflected from the specimen and the lens surfaces of the objective, but is highly transparent to the fluorescence, which can thus reach the eyepieces. The efficiency of epi-illumination is related to the fourth power of the numerical aperture (NA) of an objective, serving in epi-illumination first as a condenser and then for observation as a light-collecting lens. At the time of marketing the first multi-wavelengths epi-illuminators only high-power objectives (x70, x100) were available with a high NA (0.95, 1.30). Following suggestions by Ploem [21, 22], Leitz was the first manufacturer to produce moderate power objectives like the oil-immersion x40 objective with a NA of 1.30 (Figure 9). This new type of objective, especially designed for epi-illumination fluorescence microscopy, resulted in very bright images permitting short exposure times in routine fluorescence microphotography.

Right: Fig. 9: Leitz (Leica) early (1967) prototype ("Versuchs") oil immersion objective x40 with a NA 1.30 developed for trial experiments in fluorescence epi-illumination technology.


Protein engineering

The design of K-GECO is based on well-established GECI designs reported previously [18, 33, 45,46,47]. The initial construction of mKate2 and FusionRed-based Ca 2+ indicators was done by overlapping the assembly of the four DNA parts encoding the following protein fragments: the N-terminal (1–145) and C-terminal (146–223) parts of mKate2 or FusionRed, the RS20 peptide, and the CaM of R-GECO1. The fragments were amplified by PCR from mKate2, FusionRed (a kind gift from Michael Davidson), and R-GECO1 DNA. The overlap region and restriction sites were encoded in the primers. DNA encoding ckkap was synthesized by Integrated DNA Technologies (IDT). Purified PCR products were pooled and assembled in an overlapping PCR reaction. The resulting assembled PCR product was purified, digested with XhoI and HindIII (Thermo Fisher Scientific), and then ligated into a similarly digested pBAD/His B vector (Thermo Fisher Scientific). The ligation product was transformed into electrocompetent E. coli strain DH10B cells. Plasmids were purified with the GeneJET miniprep kit (Thermo Fisher Scientific) and then sequenced using the BigDye Terminator Cycle Sequencing kit (Thermo Fisher Scientific).

EP-PCR amplifications were performed to construct random mutagenesis libraries. The EP-PCR products were digested with XhoI and HindIII, and then ligated into a similarly digested pBAD/His B vector (Thermo Fisher Scientific). To construct site-directed mutagenesis and saturation mutagenesis libraries, QuikChange site-directed mutagenesis Lightning Single or Multi kit (Agilent Technologies) was used according to the manufacturer's instructions. The resulting variant libraries were transformed into electrocompetent E. coli strain DH10B cells and incubated overnight at 37 °C on 10-cm petri dishes with lysogeny broth (LB) agar supplemented with 400 g/mL ampicillin (Sigma) and 0.02% (wt/vol) L-arabinose (Alfa Aesar).

A custom imaging system was used for screening K-GECOs on plate with E. coli colonies expressing the variants [48]. When screening, fluorescence images of E. coli colonies were taken for each petri dish with an excitation filter of 542/27 nm and an emission filter of 609/57 nm. The colonies with the highest fluorescence intensity in each image were then picked and cultured in 4 mL liquid LB medium with 100 μg/ml ampicillin and 0.02% L-arabinose at 37 °C overnight. Proteins were then extracted using B-PER reagents (Thermo Fisher Scientific) from the liquid culture. The protein extraction was used for a secondary screen of the Ca 2+ -induced response test using Ca 2+ -free buffer (30 mM 3-(N-morpholino)propanesulfonic acid (MOPS), 100 mM KCl, and 10 mM EGTA at pH 7.2) and Ca 2+ -buffer (30 mM MOPS, 100 mM KCl, and 10 mM Ca-EGTA at pH 7.2) in a Safire2 fluorescence microplate reader (Tecan).

In vitro characterization

To purify K-GECO variants for in vitro characterization, the pBAD/His B plasmid encoding the variant of interest was used to transform electrocompetent E. coli DH10B cells and then plated on LB-agar plate with ampicillin (400 μg/mL). Single colonies were picked and inoculated into 5 mL LB medium supplemented with 100 g/mL ampicillin. Bacterial subcultures were incubated overnight at 37 °C. Then, 5 mL of bacterial subculture was added into 500 mL of LB medium with 100 μg/mL of ampicillin. The cultures were incubated at 37 °C to an OD of 0.6. Following induction with L-arabinose to a final concentration of 0.02% (wt/vol), the cultures were then incubated at 20 °C overnight. Bacteria were harvested by centrifugation at 4000 g for 10 min, resuspended in 30 mM Tris-HCl buffer (pH 7.4), lysed using a French press, and then clarified by centrifugation at 13,000 g for 30 mins. Proteins were purified from the cell-free extract by Ni-NTA affinity chromatography (MCLAB). The buffer of purified proteins was exchanged into 10 mM MOPS, 100 mM KCl, pH 7.2. Absorption spectra were recorded on a DU-800 UV-visible spectrophotometer (Beckman) and fluorescence spectra were recorded on a Safire2 fluorescence plate reader (Tecan).

To determine the quantum yield, the fluorescent protein mCherry was used as a standard. The detailed protocol has been described previously [18]. Briefly, the fluorescence emission spectra of each dilution of the protein solution of mCherry and K-GECO variants were recorded. The total fluorescence intensities were obtained by integration. The integrated fluorescence intensity versus absorbance was plotted for both mCherry and K-GECOs. The quantum yield was determined from the slopes of mCherry and K-GECOs. The extinction coefficient was determined by first measuring the absorption spectrum of K-GECO variants in a Ca 2+ -free buffer and a Ca 2+ -buffer. The absorption was measured following alkaline denaturation. The protein concentration was determined with the assumption that the denatured chromophore has an extinction coefficient of 44,000 M -1 cm -1 at 446 nm. The extinction coefficient of K-GECO variants was calculated by dividing the peak absorbance maximum by the concentration of protein.

For the Ca 2+ Kd determination, the purified protein solution was diluted into a series of buffers, which were prepared by mixing Ca 2+ -buffer and Ca 2+ -free buffer with free Ca 2+ concentration in a range from 0 to 3900 nM. The fluorescence intensity of K-GECO variants in each solution was measured and subsequently plotted as a function of Ca 2+ concentration. The data were fitted to the Hill equation to obtain Kd and the apparent Hill coefficient.

Two-photon excitation spectra and cross sections were measured as previously reported [49], with the following adjustments. For the two-photon excited spectra (2PE), the fluorescence was collected through a 694/SP filter for K-GECO1 (Semrock). To correct for wavelength-to-wavelength variations in the laser parameters, a correction function using rhodamine B in MeOH and its known 2PE spectrum was applied [50]. Two-photon cross sections were measured at 1100 nm for K-GECO1, with rhodamine B in MeOH as a reference standard. The fluorescence for cross sections were collected through a narrow bandpass filter, 589/15 (Semrock), and differential quantum efficiencies were obtained at 582 nm with a PC1 ISS spectrofluorimeter (this wavelength corresponded to the bandpass center of the above filter when used in the MOM Sutter Instruments microscope due to its tilted position). Since the filter (694/SP) used for the 2PE spectra measurements covers the fluorescence of both the neutral and anionic forms of the chromophore, the spectrum of a particular Ca 2+ state of a protein represents a combination of the unique 2PE spectra of the neutral and anionic forms, weighted to their relative concentrations (ρ, the concentration of one form divided by the total chromophore concentration) and quantum yields. The y-axis of the total 2PE spectrum is defined by F2(λ) = σ2,N(λ) φN ρN + σ2,A(λ) φA ρA, where σ2(λ) is the wavelength-dependent two-photon cross section and φ is the fluorescence quantum yield of the corresponding form (N for neutral or A for anionic in the subscript). At the wavelengths used to measure the cross sections (1060 and 1100 nm), σ2,N is assumed to be zero, and φA and ρA were independently measured to give a value for F2 (Goeppert-Mayer, GM). The relative concentrations of the neutral and anionic forms were found by measuring the absolute extinction coefficients of each respective form in the Ca 2+ -free and the Ca 2+ -bound states. These differ from the effective extinction coefficients reported in Additional file 2: Table S1, which are weighted by the relative concentrations of both forms of the chromophore.

For the fluorescence correlation spectroscopy measurement of two-photon molecular brightness, dilute protein solutions (50–200 nM) in Ca 2+ buffer (30 mM MOPS, 100 mM KCl, 10 mM CaEGTA, pH 7.2) were excited at 1060 nm at laser powers from 1 to 25 mW for 200 s. At each laser power, the fluorescence was recorded by an avalanche photodiode and fed to an Flex03LQ autocorrelator ( The measured autocorrelation curve was fitted to a simple diffusion model with a custom Matlab program [36] to determine the average number of excited molecules 〈N〉 in the excitation volume. The two-photon molecular brightness (ε) at each laser power was calculated as the average rate of fluorescence 〈F〉 per emitting molecule 〈N〉, defined as ε = 〈F〉/〈N〉 in kilocounts per second per molecule. As a function of laser power, the molecular brightness initially increases as the square of the laser power, then levels off and decreases due to photobleaching or saturation of the protein chromophore in the excitation volume. The maximum or peak brightness achieved, 〈emax〉, represents a proxy for the photostability of a fluorophore.

To measure the photoswitching of K-GECO1, R-GECO1, and RCaMP1h in vitro, the purified protein in Ca 2+ buffer (30 mM MOPS, 100 mM KCl, 10 mM CaEGTA, pH 7.2) or EGTA buffer (30 mM MOPS, 100 mM KCl, 10 mM EGTA, pH 7.2) were made into aqueous droplets with octanol in a 1:9 ratio and mounted on a presilanized coverslip. A single droplet was focused under the AxioImager microscope (Zeiss) with a 20× 0.8 NA objective and photoswitched by different laser excitations of 561, 405, and 488 nm. Fluorescence emission was detected using a SPCM-AQRH14 fiber coupled avalanche photodiode (Pacer).

Protein crystallography

K-GECO1 DNA was cloned into pRSET-A with a short N-terminal hexahistidine purification tag (MHHHHHHGSVKLIP…, tag underlined). K-GECO1 was expressed in T7 Express E. coli cells (New England Biolabs) for 36 h in autoinduction medium [51] supplemented with 100 mg/L ampicillin. E. coli pellets were lysed in B-PER (Thermo Fisher Scientific) supplemented with 1 mg/mL lysozyme followed by sonication. Insoluble cell debris was removed from the lysate by centrifugation for 20 min at 25,000 g, and soluble K-GECO1 protein was purified by immobilized metal affinity chromatography with nickel-charged Profinity resin (Bio-Rad), washed with 10 mM imidazole and eluted with 100 mM imidazole in Tris-buffered saline. K-GECO1 was further purified by size exclusion chromatography using a Superdex 200 column (GE Healthcare Life Sciences) with 10 mM Tris, 100 mM NaCl, pH 8.0, as the mobile phase. Purified K-GECO was concentrated to 10 mg/mL for crystallization using centrifugal concentrators (Sartorius Vivaspin, 10,000 molecular weight cut-off (MWCO)). Purified K-GECO1 protein at 10 mg/mL in 10 mM Tris, 100 mM NaCl, pH 8.0, was mixed with an equal volume of a precipitant solution containing 100 mM BIS-TRIS, 20% w/v polyethylene glycol monomethyl ether 5000, pH 6.5, at room temperature in a sitting-drop vapor diffusion crystallization tray (Hampton Research). Crystals were cryoprotected in the precipitant solution supplemented with 25% ethylene glycol. X-ray diffraction data were collected at 100 K on beamline 8.2.1 of the Advanced Light Source. Diffraction data were processed using the HKL software package [52]. The structure was solved by molecular replacement using Phaser [53], searching first for two copies of the fluorescent protein domain fragment using a single molecule of mKate (PDB ID 3BXB) as the search model, followed by two copies each of the separated N- and C-terminal lobes of the Ca 2+ -bound calmodulin domain using fragments of PDB ID 3SG3. Iterative model building in Coot [54] and refinement in Refmac [55] produced the K-GECO1 model, with two copies of K-GECO1 in the asymmetric unit. The K-GECO1 model was deposited at the PDB with the accession code 5UKG.

Cell culture and imaging

To characterize the K-GECO variants in HeLa cells, the cells were maintained in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific), penicillin-streptomycin (Thermo Fisher Scientific), GlutaMAX (Thermo Fisher Scientific) at 37 °C with 5% CO2. To construct the mammalian expression plasmid, pcDNA3.1(+) and the K-GECO variant were both digested with XhoI and HindIII, and the digested plasmid backbone and insert were purified by gel electrophoresis, followed by ligation and sequencing confirmation. Transient transfections of pcDNA3.1(+)-K-GECO plasmids were performed using Lipofectamine 2000 (Thermo Fisher Scientific). HeLa cells (60–70% confluency) on 35 mm glass bottom dishes (In vitro Scientific) were transfected with 1 μg of plasmid DNA, using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. The cells were imaged 24 h after the transfection. Immediately prior to imaging, cells were washed twice with Hanks balanced salt solution (HBSS) and then 1 mL of 20 mM HEPES buffered HBSS (HHBSS) was added. Cell imaging was performed with an inverted Eclipse Ti (Nikon). The AquaCosmos software package (Hamamatsu) was used for automated microscope and camera control. Cells were imaged with a 20× objective lens. To image the histamine-induced Ca 2+ dynamics, cells were imaged with a 200 ms exposure acquired every 5 s for a duration of 30 min. Approximately 60 s after the start of the experiment, histamine (10 μL) was added to a final concentration of 5 mM. The oscillation was imaged for 20 min, EGTA/ionomycin (40 μL) in HHBSS was added to a final concentration of 2 mM EGTA and 5 μM of ionomycin. After 5 min, Ca 2+ /ionomycin (40 μL) in Ca 2+ and Mg 2+ -free HHBSS was added to a final concentration of 5 mM Ca 2+ and 5 μM of ionomycin.

To characterize K-GECO variants in cultured dissociated neurons, the procedure was done as previously reported [29]. Dissociated E18 Sprague–Dawley hippocampal cells were purchased from BrainBits LLC. The cells were grown on a 35-mm glass-bottomed dish (In Vitro Scientific) containing NbActiv4 medium (BrainBits LLC) supplemented with 2% FBS, penicillin-G potassium salt (50 units/ml), and streptomycin sulfate (50 mg/ml). Half of the culture media was replaced every 4 or 5 days. Cells were transfected on day 8 using Lipofectamine 2000 (Thermo Fisher Scientific) following the manufacturer’s instructions with the following modifications. Briefly, 1–2 μg of plasmid DNA and 4 μl of Lipofectamine 2000 (Thermo Fisher Scientific) were added to 100 μl of NbActive4 medium to make the transfection medium and incubated at room temperature for 10–15 min. Half of the culture medium (1 ml) from each neuron dish was taken out and combined with an equal volume of fresh NbActiv4 medium (supplemented with 2% FBS, penicillin-G potassium salt, and streptomycin sulfate) to make a 1:1 mixture and incubated at 37 °C and 5% CO2. Then, 1 ml of fresh conditioned (at 37 °C and 5% CO2) NbActiv4 medium was added to each neuron dish. After the addition of transfection medium, the neuron dishes were incubated for 2–3 h at 37 °C in a CO2 incubator. The medium was then replaced using the conditioned 1:1 mixture medium prepared previously. The cells were then incubated for 48–72 h at 37 °C in a CO2 incubator before imaging. Fluorescence imaging was performed in HHBSS on an inverted Nikon Eclipse Ti-E microscope equipped with a 200 W metal halide lamp (PRIOR Lumen), 60× oil objectives (numerical aperture, NA = 1.4 Nikon), a 16-bit QuantEM 512SC electron-multiplying CCD camera (Photometrics), and a TRITC/Cy3 filter set (545/30 nm excitation, 620/60 nm emission, and a 570LP dichroic mirror, Chroma). For time-lapse imaging, neurons were imaged at an imaging frequency of 100 Hz with 4 × 4 binning. For photoactivation comparison, cells expressing K-GECO1 and R-GECO1 were stimulated with pulses of blue laser light (405 nm, 5 mW/mm 2 ).

To compare the K-GECO1 and red GECIs in stimulated cultured neuron cells, the procedure was done as previously reported [24]. Briefly, red GECIs were expressed after electroporation into rat primary hippocampal neurons (P0) using the Nucleofector system (Lonza). For stimulation, action potentials were evoked by field stimulation. The TxRed filter set (540–580 nm excitation, 593–668 nm emission, and 585-nm-long pass dichroic mirror) was used for illumination. Responses were quantified for each cell as the change in fluorescence divided by the baseline fluorescence before stimulation. The signal-to-noise ratio was quantified as the peak fluorescence signal over the baseline, divided by the standard deviation of the fluorescence signal before the stimulation.

iPSC-CMs were purchased from Axol Bioscience. Cells were plated in two wells of a six-well plate and cultured for 4 days in Cardiomyocyte Maintenance Medium (Axol Bioscience) to 60–80% confluency. Cells then were then transferred to fibronectin-coated (1%) coverslips and imaged in Tyrode’s buffer. Cells were transfected using transfection reagent Lipofectamine 2000 (Invitrogen). An inverted microscope (Zeiss) equipped with a NA 1.4, 63× objective lens (Zeiss) and a pE-4000 multi-wavelength LED light source (CoolLED) was used. Blue (470 nm) and green (550 nm) excitation were used to illuminate ChR2-EYFP and red GECIs, respectively. The green fluorescent protein filter set (excitation 480/10 nm, 495 nm long pass dichroic mirror, emission 525/50 nm) and the RFP filter set (excitation 545/30, 565 nm long pass dichroic mirror, emission 620/60 nm) were used to visualize ChR2-EYFP and K-GECO or R-GECO, respectively. Optical stimulation was achieved with the 470-nm LED light at a power density of 0.19 W/cm 2 and a pulse duration of 150 ms. Fluorescence signals were recorded using an ORCA-Flash4.0LT sCMOS camera (Hamamatsu) controlled by ImageJ [56].

Organotypic hypothalamic rat brain slice imaging

To prepare organotypic brain slices, experiments were done on neonatal rat coronal brain slices containing the VMN of the hypothalamus. In brief, postnatal 0–1-day-old Sprague–Dawley rats were anesthetized with 2–3% isoflurane until the paw reflex disappeared. Following decerebration, the brain was isolated in ice-cold divalent cation-free HBSS (Thermo Fisher Scientific) with 1 mM CaCl2 and 1.3 mM MgSO4. The brain was glued caudal side down to a metal plate and serial sections of 400 μm thickness were made using a vibratome (Leica Microsystems). Sectioning was stopped when the third ventricle became visible and two VMN-containing slices of 250 μm thickness were cut. Individual slices were placed on a sterile 0.4-μm-pore-membrane cell culture insert (Millipore). The insert and slice were then transferred to a 35-mm-diameter culture dish (Corning) containing 1.5 ml of NbActiv4 medium (BrainBits) supplemented with 5% FBS, penicillin-G potassium salt (50 units/ml), and streptomycin sulfate (50 μg/ml). Slices were cultured at 37 °C in an incubator (Thermo Fisher Scientific) under gassing with 5% CO2.

For transfection of organotypic slices, after 8–10 days of organotypic slice culturing, the VMN areas were transfected with an electroporation technique as previously described [47]. Specifically, the insert with the slice was placed on a platinum plate petri dish electrode (Bex Co Ltd) and electroporation buffer (HBSS with 1.5 mM MgCl2 and 10 mM D-glucose) was filled between the electrode and the membrane. Plasmids of pcDNA3.1-K-GECO1 were dissolved in the electroporation buffer at a concentration of 1 μg/ml and 10 μl of this solution was added to just cover the slice. Then, a square platinum electrode (Bex Co Ltd) was placed directly above the slice. Five 25-V pulses (5 ms duration and interval 1 s) were applied twice (the second time with reversed polarity) using a pulse stimulator (Sequim) and an amplifier (Agilent). The electroporation buffer was replaced with supplemented NbActiv4 medium and slices were returned to the incubator.

To image the cytosolic Ca 2+ dynamics using K-GECO1, an upright FV1000 confocal microscope equipped with FluoView software and a 20× XLUMPlanF1 water immersion objective (NA 1.0) was used (Olympus). The Millipore insert containing a transfected brain slice was placed in a custom-made chamber and mechanically fixed with a platinum harp. The slices were then perfused at 31 °C with artificial cerebrospinal fluid containing (in mM) 120 NaCl, 3 KCl, 1 CaCl2, 1.3 MgSO4, 26 NaHCO3, 1.25 NaH2PO4, and 10 D-glucose (the pH was adjusted to 7.4 by gassing with 95% O2 plus 5% CO2), at a flow rate of 5 ml/min using a peristaltic pump (Watson-Marlow). For single-color confocal Cai imaging, K-GECO-transfected VMN neurons were exposed to excitation with 543-nm laser light and emissions were collected from 560 to 660 nm using a variable barrier filter. Images were acquired at × 1–3 digital zoom at a frame resolution of 512 × 512 and with a 2 μs/pixel scanning rate resulting in image acquisition at 1.12 frames/s. To monitor the drug-evoked cytosolic Ca 2+ rises approximately 60 s after the start of image acquisition, 100 μM ATP (Sigma-Aldrich) was added to the artificial cerebrospinal fluid for 90 s. To compare the K-GECO1 signal with that of a chemical Ca 2+ fluorescent dye, transfected slices were stained with the membrane-permeant (AM) variant of green Fluo-4 by focal application. In brief, 0.5 mM of Fluo-4-AM was filled into a broken patch pipette with an outer diameter of

10 μm and subsequently pressure-injected (25–50 mmHg) for 10 min [57, 58] at 30–50 μm depth into the slice in the vicinity of the K-GECO1-transfected VMN neurons. This led to the uniform staining of cells in a radius of 150–200 μm from the injection site. For dual-color imaging of K-GECO1- and Fluo-4-based Ca 2+ responses, double-labeled neurons were excited with a 488-nm laser and emissions were simultaneously collected in two channels from 500 to 520 nm for Fluo-4 and 570 to 670 nm for K-GECO1 using variable barrier filters.

Imaging of zebrafish spinal sensory neurons

Mitfa w2/w2 roy a9/a9 (Casper) zebrafish were maintained under standard conditions at 28 °C and a 14:10 hr light:dark cycle. Embryos (cell stage 1–2) of Tg (elavl3:GAL4-VP16) [59] were injected with 25 ng/μl DNA plasmids encoding the K-GECO variants under the control of the 10xUAS promoter, and 25 ng/μL Tol2 transposase mRNA diluted in E3 medium. Three-day post-fertilization embryos showing expression in spinal sensory neurons (RB cells) were paralyzed by a 5-min bath application of 1 mg/ml a-bungarotoxin (Sigma, 203980). Larvae were mounted on their side in a field stimulation chamber (Warner, RC-27NE2) with 1.5% low-melting-point agarose and imaged using a custom-built two-photon microscope equipped with a resonant scanner. The light source was an Insight DS Dual femtosecond laser (Spectra-Physics) running at 1140 nm. The objective was a 25× 0.95 NA water immersion lens (Leica). Functional images (512 × 256 pixels) were acquired using ScanImage 5 ( at 7.5 Hz. The approximate laser power at the sample was measured using a slide power meter (Thorlabs) and 3 and 20 mW were used for functional imaging. Trains of 1, 2, 5, 10, and 20 field stimuli (1 ms pulse width at 50 Hz) were applied with a stimulator (NPI ISO-STIM). The stimulation voltage was calibrated to elicit an identifiable response to a single pulse in RB cells without stimulating muscle cells. Regions of interest (ROIs) were selected manually, and data were analyzed using MATLAB (MathWorks).

Mouse V1 imaging

For in vivo mouse V1 imaging, the procedure was done as previously reported [24]. Briefly, AAV injection was used for expression of K-GECO1 in mouse V1 neurons. After injection of the virus, a cranial window was implanted. The animal was then placed under a microscope at 37 °C and anesthetized during imaging. A custom-built two-photon microscope was used for imaging with a 1100-nm pulse laser as light source and a 16× 0.8 NA water immersion lens as objective. The laser power was 100–150 mW at the front aperture of the objective lens. The moving grating stimulus trial consisted of a blank period followed by a drifting sinusoidal grating with eight drifting directions with 45° separation. The gratings were presented with an LCD screen placed in front of the center of the right eye of the mouse. For the fixed tissue analysis, the mice were anesthetized and transcardially perfused. The brains were then removed and post-fixed. Sections of the brains were coverslipped and imaged using confocal microscopy (LSM 710, Zeiss).

Statistical analysis

All data are expressed as means ± standard deviation. Sample sizes (n) are listed for each experiment. For V1 functional imaging, the ANOVA test (p = 0.01) was used to identify responsive cells for each of the grating stimuli.


Patient recruitment and protocol

Institutional Review Board (IRB) approval was gained at the Beth Israel Deaconess Medical Center (BIDMC) under protocol number 2016P000352. Female patients over the age of 18 undergoing skin-sparing mastectomy and deep inferior epigastric artery perforator (DIEP) flap reconstruction were recruited. A number of common comorbidities negatively affect vasculature reperfusion and can lead to total or partial flap failure—such as smoking, diabetes, obesity, peripheral artery disease, history of venous thromboembolism, anemia or hypotension, coronary artery disease/myocardial infarction/stroke, and hypertension. A DIEP flap–free tissue transfer is risky for patients with such vascular comorbidities, and therefore, only patients without any of these risk factors were recruited for this study. Written consent for participation was acquired from all subjects, and patients were given the option to donate their discarded abdominal tissue (fig. S5) for the purpose of calibration, although only one patient in this study did so. Five women were enrolled over a period from March to September 2017. An a priori sample size power (α = 0.05, β = 0.95) calculation using in vivo animal data from a study published in Plastic and Reconstructive Surgery (PRS) (43) comparing the bandage to a Clark electrode and ViOptix monitor side by side determined a minimum need of four bandages, where 15 individual measurements are made per patient over 48 hours. The SD values for this calculation from the preliminary porcine study indicated that there exists an intrinsic SD of 10% or less across 15 identical sites. A fifth patient was added for recruitment to account for the possibility of equipment or surgical complications that could prevent completion of the full protocol. A total of seven bandages were used in the final analysis, as two of the five patients’ cases were bilateral, exceeding the sample size in the initial power calculation for a total of n = 101 image sets (table S1). Note that n is the number of phosphorescence images, not the number of subjects, meaning a total of n = 101 images was the sample size for this study as the purpose was to compare the readings from the two oximeters, not to detect flap failure.

A schematic of the DIEP flap reconstruction surgery and postoperative assessment for this study is shown in Fig. 2. A volume of tissue (skin and fat) is dissected from the lower abdomen, in the same area that an abdominoplasty or “tummy tuck” would typically be performed. It consists of skin and subcutaneous fat but not the underlying rectus abdominis musculature. Usually, two or three perforating arteries, branches of the deep inferior epigastric artery, which perfuse the flap, are identified, dissected, and included in the dissection, as are perforating veins. Simultaneously, the breast is excised down to the pectoralis muscles, and the internal mammary artery/vein is dissected. The flap is then moved from the abdomen to the breast, and the perforators are anastomosed to the internal mammary vessels. The donor and recipient sites are repaired with sutures, and perfusion of the flap was assessed perioperatively with a NIRS ViOptix device placed directly on the skin paddle of the flap, where an approximate drop in oxygenation of 30% or more triggers an alarm. Each patient underwent routine postoperative care at the BIDMC for postoperative monitoring, with no alterations to the clinical workflow other than painting, an approximately 1 cm by 1 cm area of skin on the patient’s flap(s) with the oxygen-sensing liquid bandage upon arrival to the postoperative recovery area, before dressing the flap with Tegaderm as is standard. After allowing the liquid bandage to dry into a thin film for 1 to 2 min, a clear dressing (Tegaderm, 3M) was applied as a barrier film to prevent interference from room air. Tegaderm is commonly used in postoperative recovery to help adhere flaps and grafts to the body during uptake. Like many transparent wound dressings, Tegaderm is water resistant and semi-oxygen permeable, meaning it acts as a barrier to prevent room air from interfering with the measurement of tissue pO2. When compared to other transparent medical films, Zimmermann et al. (54) found that Tegaderm (a polyethylene-based film with adhesive backing) has relatively low oxygen permeability. The kinetics of the oxygen flux during tissue and bandage equilibration under a Tegaderm barrier has been previously described by our group (51).

DIEP flaps taken from the abdomen were used to reconstruct the breast, and postoperative monitoring was performed for 48 hours with two oximeters placed onto each flap as shown: the ViOptix (NIRS, %stO2) and the paintable bandage (phosphorescence, pO2). Photo credit: Juan Pedro Cascales, MGH.

Synthesis of oxygen sensing Clickaphor metalloporphyrins and liquid bandage preparation

Reagents and solvents were purchased from Sigma-Aldrich and Thermo Fisher Scientific, with the exception of l -azidoglutamic acid mono-tert-butyl ester CHA salt [N3-Glu(OtBu)-OH.CHA], which was obtained from ChemPep Inc. All compounds were used without further purification. The synthesis of the palladium-porphyrin core was performed as previously described by Roussakis et al. (42). Deviations from the published protocol for the Williamson-type alkylation step, such as substantial dilution of the reaction mixture and a large increase in the excess of the reagents, led to a large increase in the yield for the synthesis of the alkyne-terminated derivative. Deprotection of the pivaloyl-protected metalloporphyrin with the use of diisobutylaluminum hydride was performed as previously published. Briefly, the product of pivaloyl-group deprotection was dissolved in dry N,N-dimethylformamide (DMF) at a concentration of 0.001 M under an argon atmosphere, and the solution was cooled to 0°C with a water/ice bath. A large excess of sodium hydride (60% dispersion in mineral oil), enough to cover the tip of a small metal spatula, was scooped into the solution, and the reaction mixture was stirred for about 15 min. A large excess of propargyl bromide (80% in toluene), amounting to 5% of the volume of DMF solvent, was added slowly (dropwise by a syringe), and the reaction was allowed to warm up to room temperature and was left to react overnight.

The progress of the reaction was checked periodically with matrix-assisted laser desorption/ionization–time-of-flight (MALDI-TOF) mass spectrometry (MS). If the reaction was not completed, additional sodium hydride and propargyl bromide were added, and the mixture was left to react for an extra day. Removal of the solvent and chromatographic purification were performed as previously described. This, as well as the chromatographic purifications in the earlier synthetic steps, ensured the removal of any reactants and impurities, yielding a pure alkyne-terminated palladium-porphyrin derivative as confirmed by proton nuclear magnetic resonance ( 1 H–NMR) nuclear magnetic resonance and MALDI-TOF MS. The synthesis of the oxygen-sensing, ethylglutamate metalloporphyrin dendrimer was performed as previously published, via a copper-catalyzed click-type reaction of the alkyne-terminated palladium-porphyrin with a second-generation azido-ethylglutamate dendron subunit (42). Purification of the porphyrin-dendrimer was modified from the published protocol. After removal of the solvents (DMF and water) via rotary evaporation, the residue was dissolved in a small volume of ethanol, and the porphyrin-dendrimer was precipitated via addition of ultrapure water followed by centrifugation. The supernatant was removed, and the ethanol dissolution and precipitation/centrifugation cycles were repeated twice more, followed by drying under high vacuum to afford the product as red solid. MALDI-TOF MS and LC–MS analysis of the final product showed that no unreacted dendron monomers were left from the alkyne-azide click reaction. The mass of the metalloporphyrin (structure shown in Fig. 3B) was determined by MALDI-TOF MS to be approximately 5412.11 Da.

Top: Liquid bandage formulation components: (A) New-Skin brand ethanol-nitrocellulose matrix, (B) in-house synthesized oxygen-sensing metalloporphyrin, and (C) fluorescein reference dye. Bottom: Protocol for liquid bandage application. Bandages are painted on to the inner area of the flaps, allowed to dry for 1 min, and sealed with Tegaderm to prevent interference from room air. After 48 hours of postsurgical measurements, the bandage is removed along with the Tegaderm during redressing. Photo credit: Emmanuel Roussakis, MGH. Schematic adapted with permission from reference (50).

The paint-on bandage material was formulated from three components: the commercially available New-Skin liquid bandage (an ethanol-based nitrocellulose solution), an oxygen-sensing palladium-porphyrin Clickaphor dendrimer, and the oxygen-insensitive reference dye fluorescein (Fig. 3). The oxygen sensor and the reference dye are first codissolved in ethanol (200 proof) at concentrations of 180 and 10 μM, respectively. This stock solution is diluted with the commercial New-Skin liquid bandage at a ratio of 1:1 to achieve final porphyrin and fluorescein concentrations of ∼90 and ∼5 μM, respectively. Concentrations were chosen so that both dyes’ emission was visible under room lighting when excited by the same source. The mass per volume percentage ratio of the porphyrin sensor in the final formulation is no more than 0.4% (w/v).

A concern with any bandage or device that comes in contact with human skin, especially for an extended period of time, is that it may leave residual material on the skin after removal. We wished to explore whether, upon removal of the dried bandage formulation along with the Tegaderm film, any oxygen-sensing metalloporphyrin remained on the tissue. To test for residual deposited metalloporphyrin, a 1 cm by 1 cm portion of excised human abdominal tissue was painted with the liquid bandage solution containing a commercial porphyrin sensor, allowed to dry for 1 min, sealed with Tegaderm, and left for 24 hours before removal. This procedure mimics the removal of the Tegaderm and oxygen-sensing bandage from the human subjects. Following tissue digestion, inductively coupled plasma (ICP)–MS was performed at the Harvard School of Public Health. This experiment was then repeated using the Pd-porphyrin dendrimer synthesized for this study and kept under much harsher conditions in 100% humidity at 37°C for 3 days. Following this incubation period, the Tegaderm/bandage was removed, and the area was wiped with an alcohol pad. External metals analysis of the digested samples looking for traces of palladium was conducted at Brooks Applied Labs (ISO/IEC 17025 certified) using an ICP-MS, an order of magnitude more sensitive than the system at the Harvard School of Public Health.

Calibration of the paintable bandage formulation

The liquid bandage formulation was first validated spectroscopically by measuring the porphyrin and fluorescein emission throughout deoxygenation with nitrogen, as shown in Fig. 4A. The liquid formulation’s red-to-green response when placed in a cuvette and excited with an ultraviolet (UV) flashlight is also shown by the unfiltered cell phone images in Fig. 4B. Note that the green emission of the fluorescein reference dye, with a broad peak near 532 nm, does not change in response to oxygen. This unresponsiveness to oxygen permits fluorescein to act as an internal reference against which the oxygen-sensitive phosphorescence of the porphyrin may be measured. Without the glutamate dendrimer, one or both dyes are subject to aggregation within the New-Skin nirtocellulose matrix when drying as a very thin paint on film. Thus, it is the combination of the inherent photophysical properties of Clickaphor Red porphyrin alkyne core, along with its potential to be easily converted into derivatives that leads to its optimal oxygen-sensing performance within polymeric matrices and formulations. A comparison of Clickaphor Red to its alkyne-terminated metalloporphyrin core precursor to widely used, commercial porphyrins such as Pd(II)meso-tetrakis(pentafluorophenyl)porphyrin (PdTPFPP) (fig. S4C) was performed to demonstrate the need for the ethylglutamate porphyrin-dendrimer to prevent aggregation and self-quenching of the dyes within the paintable New-Skin nitrocellulose formulation. While commercial porphyrins and alkyne cores may be used for preformed thin films, the Clickaphor Red formulation is specifically formulated for painting directly onto human skin.

(A) Spectra of the liquid bandage formulation in ethanol demonstrating that only the porphyrin’s emission is oxygen dependent (660-nm peak) and (B) images taken with a cell phone camera of the ethanol formulation under different oxygenation conditions when excited by a 405-nm light-emitting diode flashlight, demonstrating that the red-to-green color change is visible to the naked eye under room lighting. (C) Mean normalized phosphorescence intensity from camera images shown in (E), demonstrating sensitivity to oxygen from 0 to 160 mmHg (gray circle), and that the modified Stern-Volmer relation (I0/I) fits the calibration data (red square). (D) Stern-Volmer calibration with a rolling average applied in both x and y to account for gas flow hysteresis during repeated cycles and (F) demonstration of ex vivo calibrations performed on human breast tissue with different levels of melanin, which had no apparent effect on the sensor performance. Photo credit: Haley Marks, MGH.

Next, the oxygen response in the presence of an autofluorescent background was explored by painting the liquid formulation onto ex vivo breast and abdominal tissue. Discarded human skin tissue was collected under Massachusetts General Hospital (MGH)–discarded tissue IRB protocol 2015P001267. The dried bandage was then removed from the skin using Tegaderm and placed faceup over the skin to expose it directly to the chamber environment. The sample chamber consisted of a phosphate-buffered saline–soaked sponge inside a petri dish, which is sealed with a poly(dimethylsiloxane) (PDMS) lid. Nitrogen and air supplies were run through a gas proportioner, whose humidified output is inserted into the chamber’s lid with a needle. The pO2 in the chamber was measured using a Clark-type electrode as a reference standard, as shown in the photo in fig. S5. The dish was heated to ∼32°C, and 100% humidity was maintained throughout imaging to approximate skin temperature and hydration conditions. For all calibrations, the chamber was first purged with nitrogen and allowed to equilibrate for ∼30 min before slowly introducing known levels of oxygen every 5 min or until the Clark electrode gave a stable reading, with images acquired at each step and processed in accordance with the following imaging section.

Multiple calibrations were performed throughout the duration of the 6-month-long study to account for potential aging of the formulation, to determine the role of melanin in calibrations, and to compare the effects of freezing tissue. As four of the patients in this study did not donate their discarded tissue, ex vivo breast tissue harvested from various different donors was used to carry out calibrations throughout the study duration. These tissues were collected and used for calibrations throughout the study and were also frozen so that they could be thawed for skin pigmentation matching. To determine the effect of skin pigmentation of the ex vivo calibrations, frozen samples with various pigmentations were thawed and matched to the intrinsic R R + G background autofluorescence baseline of each subject. The mean normalized phosphorescence intensity R R + G from each bandage calibration image (shown with a false color map in Fig. 4E, bottom) was converted into pO2 values using the following modified Stern-Volmer relation I 0 I = R 0 R 0 + G 0 R R + G = 1 + G 0 R 0 + G 0 K sv [ p O 2 ] (1)

One fresh discarded abdominal tissue sample used for calibration was collected from an enrolled subject, shown in the example lookup table in Fig. 4. Since nitrogen and air levels were adjusted manually with a gas proportioner, some minor degree of hysteresis occurred throughout deoxygenation and reoxygenation. To compensate for this, a rolling average was applied to both x and y, and a York linear fit for this corrected calibration is shown in Fig. 4D. Ultimately, it was found that background contributions from varying levels of melanin had little effect on calibrations (Fig. 4F) but that using fresh versus frozen tissue caused a change in the tissue structure, which affected its ability to retain moisture and its breathability, as well as the baseline autofluorescence, as is evident in fig. S6. In addition, it is known that a skin pH > 6 could potentially cause measurement error (fig. S4A) as the fluorescein has a pKa (where Ka is the acid dissociation constant) = 6.4. Skin pH typically ranges from 4 to 6 in normal, healthy humans, and pH higher than 6 would indicate bacterial infection (55).

Photography and image analysis

Commercially available Nikon D70s DSLR cameras were modified for collection of the chromophore’s full emission spectrum by removing the infrared (IR) rejection filter and attaching a custom 3D-printed filter slider containing two bandpass filters in the green (525/30 nm, Chroma Technologies) and red (660/40 nm, Chroma Technologies) spectral regions. For simultaneous excitation of the dyes near the porphyrin’s Soret band, blue/UV bandpass filters (385/70 nm, Chroma Technologies) were mounted in front of two bilaterally mounted Vivitar flash units set at 1/16 of their maximum power. A 1/16 level flash excitation corresponds to 2-mW total irradiance over the course of the 48-hour monitoring period. According to Mitra and Foster (56), this is far below the wattage required to induce enough reactive oxygen species (ROS) to negatively affect the tissue oxygen consumption readings. The spectra of the dyes overlaid with the camera filters are shown in fig. S1. The flash units were mounted to the camera body on a triangular arm, which allowed the clinician to hold the camera with one arm while maintaining one arm free for pressing the trigger button, adjusting camera focus, or adjusting patients’ monitors or dressings. A photograph of the custom DSLR camera setup on a tripod alongside the excitation and emission spectra of the dyes is also shown in fig. S1.

For postsurgical measurements, photographs were taken in sets of six at 0 and 20 min after the application of the oxygen-sensing bandage and then hourly at 1 to 6 hours, followed by an additional photograph set with acquisitions every 6 hours between 12 and 48 hours. Each set of six photographs contains a red, green, and no-filter image, with and without the flash on. The “flash off” images account for any background signal from room lighting. The “no filter” images were used as a quality control measure to ensure proper camera orientation and flash intensity. Taking photographs at the 20-min postapplication mark allowed the bandage to reach oxygen tension equilibrium with the tissue (51). A monitoring duration of 48 hours was chosen to match the time period over which flap failure is most likely to occur (6). During photography, a black sheet with a 25 mm hole was used to expose only the liquid bandage while blocking any interfering fluorescence signal from surrounding medical supplies such as bed sheets, gowns, tubing, and the ViOptix probe itself. While the sheet was not necessary from a technical standpoint, it provided a straightforward means for ensuring deidentification of the images while also allowing for fully automated image processing by standardizing the analyzed region of interest (ROI). Images were converted from .nef (RAW) to 16-bit .tiff RGB images using Nikon’s View-Nx software for analysis. Converted images were then processed in MATLAB using the following abbreviated algorithm: categorize as red- or green-filtered image and as flash on or off, align corresponding background and signal images, subtract background from signal image for each color to correct for interfering lighting, align corrected red and green images, perform matrix algebra of aligned images to get map of the phosphorescence intensity normalized to total luminescence ( R R + G ), and export raw and processed data. Inverted logic masks were then used to normalize data to the surrounding autofluorescent tissue. The developed MATLAB function “tif2phos.m” can be found in the Supplementary Materials and is shown graphically in fig. S2.


All statistical analyses were performed using the R language (57) in the RStudio environment. A .cvs and .Rmd file containing the complete raw dataset and statistical analysis, respectively, can be found in the Supplementary Materials. The linear mixed-effects regression (LMER) model for predicting a continuous outcome (changes in phosphorescence or pO2) based on continuous predictors (changes in stO2 and changes in time) and accounting for random effects was constructed before data analysis as follows Δ p O 2 , ij = β 0 + β 1 Δ % st O 2 , j + β 2 t 0 , j + β 3 t * Δ % st O 2 , j + b 0 , j + ϵ ij (2)

The fixed effects are defined as follows: Δ % stO2, the change in blood oxygen delivery t, the time (in hours) since the Tegaderm was applied over the bandage t * Δ % stO2, the interaction term between time and blood oxygen delivery, which accounts for changes in oxygen saturation experienced during flap uptake and ϵ, the residual error. The null hypothesis is that β1, the coefficient describing the relationship between blood oxygen delivery and tissue oxygen consumption, is equal to zero, and the alternative hypothesis is that β1 is nonzero. As some cases were bilateral, to account for the correlation between two bandages worn by the same subject, we include the subject specific random intercept b0, j along with the residual error ϵ, which are assumed to have a normal distribution with mean equal to 0 and an unknown SD, where j indexes subjects and i indexes the bandage within a subject. This analysis was also repeated using the raw % phosphorescence (where %phos = R R + G ) in place of bandage pO2 to confirm that the raw data inversely correlate with stO2 regardless of the quality of the calibration.


Here, we show definitive evidence of dispersed light sensing in octopus skin and document the expression of a candidate light sensor in skin of the same species, Octopus bimaculoides. Two previous studies have speculated that cephalopod skin may be intrinsically sensitive to light, noting that chromatophores in both squid and octopus skin seem to expand when the skin is illuminated, but neither study provided more than preliminary observations (Florey, 1966 Packard and Brancato, 1993). We found that chromatophores in the skin of O. bimaculoides expand significantly and repeatedly when exposed to bright white light, a behavior we call light-activated chromatophore expansion, or LACE. We attribute LACE to light, as we minimized heat reaching the samples by using fiber optics to illuminate the skin, which itself was submerged underwater. LACE responses clearly show that O. bimaculoides skin can detect light by itself, independent of eyes.

While octopus LACE is a robust behavior, we found that some of the parameters of LACE differ from those noted by Packard and Brancato (1993). For instance, they report that chromatophores in denervated Octopus vulgaris skin expand 1 s after a flash of bright white light, which differs from the average 6 s (adults) or 15 s (hatchling) latency for LACE we found in O. bimaculoides. This incongruence in latency may be attributable to differences between the species and/or the preparation itself, as it seems that Packard and Brancato did not isolate skin samples, but denervated portions of skin still attached to the whole animal. We observed a high degree of variation in both the latency of LACE and the time to full expansion of the chromatophores in our preparations and attribute at least some of this variation to differences in the time between dissecting the tissue and running LACE experiments. We also observed differences in LACE between the hatchlings and adults, where adult skin responded more consistently and robustly than the skin from younger animals. We speculate that this could be caused by the presence of more light sensors in adult versus hatchling skin. However, despite these differences from preliminary reports, our data are the clearest demonstration to date that Octopus bimaculoides skin is intrinsically light sensitive, and that light detected by the skin causes the chromatophores to expand.

We hypothesized that r-opsin, a key light sensing protein in the eyes of octopuses and other animals, may also detect light in octopus skin and underlie LACE. To support this hypothesis, we looked for evidence of opsin expression in the skin and determined the action spectrum for LACE. Consistent with our hypothesis, we found that r-opsin is expressed in the skin of O. bimaculoides. This result is similar to Mäthger et al. (2010), who detected r-opsin mRNA from one PCR trial of skin from another cephalopod, the cuttlefish Sepia officinalis, although it is not yet known whether cuttlefish have LACE. Additionally, opsin expression by itself is weak evidence for the ability of skin to detect light. Other essential r-opsin cascade genes, including G-protein α (q) and phospholipase C, are also expressed in the skin of O. bimaculoides, suggesting that the necessary genes for functional opsin-based phototransduction are expressed in octopus skin (Speiser et al., 2014). Finally, the LACE action spectrum is also consistent with our hypothesis. Spectral sensitivity analysis of the opsin from the eyes of another octopus O. vulgaris shows a λmax of 474 nm (Brown and Brown, 1958). If the same opsin found in octopus eyes underlies octopus LACE, then LACE activity should peak close to the known spectral sensitivity of octopus opsin. Indeed, we found that the latency to LACE is shortest in blue light, and fitting the Govordovskii curve to the action spectrum data gives a λmax of 480 nm. Taken together, these data strongly support our hypothesis that opsin phototransduction underlies LACE. Future work should continue to test this hypothesis by manipulating the function of opsin phototransduction proteins and observing how they affect LACE.

Because r-opsin is known to function in light sensing, cells in octopus skin that express opsin are excellent candidates for dispersed light sensors that could underlie LACE. We identified ciliated peripheral sensory neurons in the skin of hatchling O. bimaculoides using α- and β-tubulin antibodies. These cells were similar in morphology and position (Sundermann-Meister, 1978 Sundermann, 1983 Mackie, 2008 Buresi et al., 2014) to cells described as mechanoreceptors in both squid and cuttlefish (Budelmann and Bleckmann, 1988 Bleckmann et al., 1991). It is not yet known whether these peripheral sensory neurons act as mechanoreceptors in the skin of O. bimaculoides. Intriguingly, we localized r-opsin expression to these same peripheral sensory neurons in hatchling skin, raising the possibility that aside from a mechanoreceptive function, these sensory cells may also be dispersed light receptors in octopus and other cephalopods. Unfortunately, the precise connections between candidate dispersed light sensors in octopus skin, the chromatophores and the CNS remain unclear, as does their relationships with LACE and merits further investigation to test the hypothesis that the r-opsin-expressing neurons detect light.

Our finding of opsin expressed in known mechanoreceptors raises the question of whether opsin has a role in mechanoreception, in addition to its well-established role in light detection. While our work is the first description of this opsin expression pattern in mollusks, opsin-expressing mechanoreceptors have been recently described in the annelid Platynereis, zebrafish and Drosophila (Backfisch et al., 2013 Senthilan et al., 2012). From work on mechanoreception in Drosophila antennae, we now know that opsin is required for anntenal mechanoreceptors to detect vibrations, highlighting a previously unknown role for opsin in senses besides light detection (Senthilan et al., 2012). We do not yet know whether the opsin-expressing cells we found in hatchling O. bimaculoides skin function as mechanoreceptors, light sensors or both, or the extent to which opsin is required for detecting either of these stimuli. Still, our results compel future research into the role of opsins in senses other than photoreception. We believe that the phylogenetic spread of opsin expression in mechanoreceptors among vertebrates, annelids, arthropods and now mollusks, suggests that such mechano-sensory roles for opsin could be ancient in animals.

Finally, uncovering dispersed light sensitivity in octopus skin raises the question of how it evolved to underlie LACE in octopuses. Our study is the best evidence so far for light-sensitive skin in cephalopods and we hypothesize that LACE may play a role in modulating body patterning for camouflage, alongside the canonical control exerted by the CNS. However, while cephalopods are unique among mollusks for their body-patterning abilities, we know that most other mollusks, especially bivalves, gastropods and chitons, are able to sense light with their skin. There is rich literature describing behaviors like phototaxis or shadow responses and physiology linked to light sensing in the skin of other mollusks (Ramirez et al., 2011). We do not yet know if or how cephalopods use their light-sensing skin for these other more typical molluscan behaviors. However, the widespread distribution of dispersed light sensing and associated behaviors throughout the phylum suggests that dispersed light sensitivity could be an ancestral molluscan trait that has been co-opted in the cephalopod lineage to mediate novel body-patterning behaviors in response to light. Understanding the underlying molecular mechanisms for dispersed light sensing across the mollusk classes would help clarify the evolutionary history of dispersed light sensing and associated behaviors. Our study provides a framework for future comparative work that can integrate already known behavioral data with molecular data for light-detecting components in various mollusks. This work could address the question of whether diverse mollusk behaviors that rely on dispersed light sensing share a common molecular mechanism for light detection, and thus whether dispersed light sensing was present in ancestral mollusks.


PDT is an effective form of treatment for an increasing number of human conditions, ranging from cancer to several skin conditions. Benefit deriving from the use of light is known since ancient time, but only in the last decades of twentieth century, we witnessed the rapid expansion of knowledge and techniques. Recent improvements of the therapy are related especially to photosensitizer’s development and delivery systems. Nowadays, the use of LED-based devices represents the emerging and safest tool for the treatment of many conditions such as skin inflammatory conditions, aging, and disorders linked to hair growth. Although the use of LED in the treatment of hair disorders has now entered common practice, better controlled studies are still needed to corroborate its efficacy.