Non-Thermal Equilibrium Transport by Dynein Molecular Motors in Live Neurons, and Breakthroughs in Linear and Non-Linear Ultrasound Imaging
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 340 | |
| Author | ||
| License | CC Attribution - NonCommercial - NoDerivatives 4.0 International: You are free to use, copy, distribute and transmit the work or content in unchanged form for any legal and non-commercial purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/45084 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
00:00
Electronic mediaDC motorClimate changeHot workingSpare partMedical ultrasonographyGentlemanEffects unitMorningLaser coolingMusical developmentAngeregter ZustandCardboard (paper product)AfternoonSecretaryGroup delay and phase delayCooling towerAtomismLecture/Conference
01:17
Local Interconnect NetworkHalo (optical phenomenon)Particle physicsSynthesizerSpeckle patternNoise reductionSpeckle imagingThermodynamic equilibriumAfternoonShip classGroup delay and phase delayQuantum fluctuationMedical ultrasonographySynthesizerHot workingCell (biology)Continuous trackLecture/ConferenceMeeting/Interview
02:08
AtomhülleBrightnessMagnetic coreOpticsAngeregter ZustandElectron energy loss spectroscopySynthesizerSEEDParticleCrystallizationTransfer functionGruppensteuerungQuality (business)TeilchenrapiditätSemiconductorPhotonPotentiometerBrightnessMaterialDotierungAngeregter ZustandOLEDSpace probeConcentratorSingle (music)Resonance (chemistry)Relaxation (physics)Energy levelInfraredLightSynthesizerAtomhülleRulerMinuteScanning acoustic microscopeBestrahlungsstärkeCrystal structureComputer animation
04:45
SEEDSynthesizerParticleBrightnessMagnetic coreAtomhülleEmission nebulaQuantumQuality (business)Magnetic coreInertiaConcentratorCrystallizationDotierungBrightnessAtomhülleSizingSpace probeHourLecture/Conference
05:32
QuantumEmission nebulaBrightnessQuantumQuality (business)HourDaySpare partDyeingColorfulnessSpace probeLecture/ConferenceComputer animation
06:21
SynthesizerAxionAntiparticleCell (biology)CoatingContinuous trackModel buildingAngeregter ZustandDiffusionFahrgeschwindigkeitEngineAnalemmaQuantum fluctuationRotating radio transientContinuous trackRestkernSpace probeFahrgeschwindigkeitFormation flyingChemical substanceBill of materialsMeasurementVoltageAxionSpare partBeta particleCell (biology)Signal (electrical engineering)ClockQuantum dotMetreBicycleCosmic microwave background radiationCosmic distance ladderLightNoise (electronics)ScoutingDC motorSeparation processFocus (optics)ChannelingGround (electricity)ColorfulnessColor codeFlow measurementFinger protocolDiffusionSpeckle imagingYearField-effect transistorRoll formingHot workingBusFusion powerBarrelSensorMechanicMikrofluidikElectric power distributionReceiver (radio)Pulp (paper)WoodturningRückezugMitsubishi A6M ZeroSecretaryGreen politicsMeeting/Interview
13:24
SynthesizerThermodynamic equilibriumProzessleittechnikGenerationAngeregter ZustandElectric power distributionQuantum fluctuationMeasurementSingle (music)Engine displacementMeasurementThrust reversalOrbital periodElectric power distributionClockAngeregter ZustandEngine displacementCell (biology)Direct currentDyeingVideoPresspassungSigmaSingle (music)TrajectoryEngineFahrgeschwindigkeitEnergy levelQuantization (physics)Tape recorderQuantum fluctuationBrightnessThermodynamic equilibriumTemperatureHeatField-effect transistorContinuous trackRegentropfenDC motorVolumetric flow rateSpeckle imagingWater vaporInkAlcohol proofSunriseShort circuitMinuteOrder and disorder (physics)GasCoatingMusical ensembleAutumnVanRailroad carKilogramLine segmentForgingBending (metalworking)PolradPhotographySpeise <Technik>ConcentratorTransportHyperbelnavigationLecture/Conference
21:01
CamshaftElectric power distributionBookbindingModel buildingEngineElectric currentCell (biology)Hauptsatz der Thermodynamik 2Quantum fluctuationOpticsHorn antennaForceSensorBell LaboratoriesWasserwiderstandRadioactive decayHauptsatz der Thermodynamik 2Series and parallel circuitsTemperatureScale (map)Model buildingSunriseTemperaturabhängiger WiderstandViscosityNatürliche RadioaktivitätMeasurementGeokoronaThermodynamic equilibriumConcentratorOrder and disorder (physics)Reaction (physics)VideoCell (biology)Bending (metalworking)Quantum fluctuationAxionEngineRoll formingKopfstützeFahrgeschwindigkeitVoltageWireElectrical conductorThermalYearSigmaTiefgangKelvinBrownGearAlcohol proofPropeller (marine)Stock (firearms)Blood vesselAmplitudeTurningContinuous trackGround stationHot workingNoise (electronics)MonthLimiterSummer (George Winston album)MagnetSpeise <Technik>Computer animation
28:33
Speckle imagingSpeckle patternAngle of attackAudio frequencyMeasurementSpeckle patternMedical ultrasonographyStandard cellQuality (business)YearDensityMeasurementAngle of attackAudio frequencyWeek
29:19
Angle of attackSuperheterodyne receiverCash registerWatercraftEGPRSAlcohol proofAudio frequencySpeckle imagingSpace probeDistortionRefractive indexBrechungPressureNoise reductionSpeckle patternStandard cellMedical ultrasonographyOrder and disorder (physics)Speckle imagingForgingBlood vesselSpeckle patternNoise reductionController (control theory)MeasurementCommercial vehicleDiffractionMeeting/Interview
29:54
TransceiverSensorSuperheterodyne receiverWatercraftSteeringAutomatic watchSynthesizerSpeckle imagingNoise reductionSpeckle patternThermalAudio frequencyLimiterAngle of attackMinuteSoundAudio frequencyNear field communicationSimplified Chinese charactersFood storageComputer animationMeeting/InterviewLecture/Conference
30:39
OpticsDimmerTransverse modeAudio frequencyTissue paperLinear motorSoundNoise reductionMechanicMaterialAudio frequencyContrast (vision)FahrgeschwindigkeitSource (album)Spare partMode of transportAlcohol proofKontraktionSpeckle imagingOrder and disorder (physics)Magnetic resonance imagingWind waveField-effect transistorWater vaporMitsubishi A6M ZeroFlatcarComputer animation
32:53
Speckle imagingTissue paperMode of transportMobile phoneSensorHose couplingYearScanning probe microscopySummer (George Winston album)MachineCell (biology)Magnetic resonance imaging
33:56
Local Interconnect NetworkSpeckle imagingSurface acoustic waveLecture/ConferenceMeeting/Interview
34:32
Continuous trackDC motorAxionCell (biology)Field-effect transistorRail profileLecture/ConferenceMeeting/Interview
35:29
Local Interconnect NetworkDoorbellThermodynamic equilibriumSpeckle imagingNoise reductionSpeckle patternLinear motorGravitational singularityDC motorDirect currentContinuous trackFinger protocolTypesettingBrightnessTissue paperMechanicFACTS (newspaper)MachineSpare partSizingController (control theory)Hose couplingTool bitNew YearMetreSensorNightYearFiberLecture/Conference
39:54
PinkLocal Interconnect NetworkSpeckle imagingAtmosphere of EarthSizingMint-made errorsSignal (electrical engineering)SensorOptischer SchalterLecture/Conference
40:37
Speckle imagingThermodynamic equilibriumThermodynamic equilibriumEngineYearStagecoachQuantum fluctuationTemperatureCartridge (firearms)Water vaporDestroyerDC motorLecture/Conference
42:02
Speckle imagingThermalLocal Interconnect NetworkEngineThermodynamic equilibriumNoise reductionMobile phoneAxionWhiteRulerAudio frequencyMotion captureInterference (wave propagation)PaperPlane (tool)Focus (optics)Cell (biology)ProzessleittechnikDie proof (philately)Communications satelliteSemiconductorForceLastOpticsFACTS (newspaper)Field-effect transistorToolMinuteAngle of attackCocktail party effectFundamental frequencyMembrane potentialHot workingMidnightLecture/ConferenceMeeting/Interview
45:46
VideoNanotechnologyElectric power distributionGround stationCommercial vehicleMaterialComputer animation
Transcript: English(auto-generated)
00:14
OK, good morning. Fantastic. The third gong we can begin. Hello, my name's Adam Smith and it's my great pleasure to welcome you to this session with Steve Chu.
00:23
As you all know, Steve was awarded the 1997 Nobel Prize in Physics for his work with cooling and laser trapping of atoms, but Steve is a man of many parts and since then he's been Director of the Lawrence Berkeley National Laboratory, he's been US Secretary of Energy, he's now Professor of Physics and Professor of
00:42
Cellular and Molecular Physiology at Stanford. And I for one have long since ceased to be surprised by where Steve's inquisitiveness takes him. So today he's producing yet another new piece of work for those in Lindau. He's going to be talking about molecular motors in living neurons and new developments in ultrasound imaging.
01:02
And after that we're going to have a free Q&A with Steve in which he tells me it's anywhere safe to go. So you can talk about what he talks about now or you can ask about energy or climate change or sustainability or politics, I dare say. Anyway, over to Steve. Thank you. Actually there's a session this afternoon for a much more free ranging discussion.
01:25
But today I want to concentrate on some new things that several people in my lab are doing. It's a very small group of people now. And the outline of the talk is I'm going to tell you about a new class of particles that have emerged over the last decade
01:42
and our ability to improve them and apply those new particles to live cell tracking in neurons. And I'm going to go and apply some concepts in physics, things like the fluctuation theorem, and I'll explain what that is, to show how we can analyze the data. And finally I'll go to ultrasound imaging.
02:02
So the first is just the synthesis of rare earth upconverting nanoparticles. So the team that did the work of the synthesis, the functionalization, the characterization, and the biology experiments are shown here, Chen Liu, Yun-Hsuan Zhang, and Sam Peng. The idea of rare earth nanoparticle probes is that you have in a crystal of sodium yttrium fluoride
02:26
embedded in purities YB and erbium. And the idea is that the YB is a half, only one vacancy unfilled shell has a very simple optical spectra. And so when you're coming with light in the infrared at 980 nanometers, you can excite the YB.
02:46
There's very rapid resonant energy transfer among the YB within the crystal, and you can transfer that energy to the erbium. And so it's because the erbium energy level over here is very closely resonant with the energy level of YB.
03:01
Of course you can transfer back, but you dope more YBs and erbiums, maybe ten times as many, so another photon will excite the YB and transfer, and now you get a double excitation. So in addition to transferring one step, you transfer another step. Very rapid non-radiative relaxation down here, so you no longer transfer energy back to the YB.
03:25
So what happens is out comes green and red light when you put in infrared light. Now you look at this and say, well, wait a minute, your doping is 20% YB, 2% erbium. Why don't you go to 90% YB and 10% erbium? And indeed people tried, but they were unable to grow any really decent crystals with higher doping.
03:48
So we started this by actually having nanoparticle synthesizers give us the materials, and we got very discouraged by the quality of the brightness.
04:01
And so in an act of desperation, we decided to try it ourselves, but we had one little twist. We're going to borrow something from the semiconductor industry. We know in semiconductors that if you start with a single perfect crystal of gallium arsenide, for example, you can put alloys of materials on it, gallium aluminum arsenide,
04:23
indium gallium aluminum arsenide, all sorts of stuff. And if you do it carefully, you can molecularly epitaxially grow a perfect crystal. So we said, okay, we know how to make a perfect crystal of sodium yttrium fluoride, so we'll start with a little C core, five, six nanometers in diameter.
04:40
We harvest those little particles, we put it back into a wet pot with the soles that allow us to grow any concentration we want. And what we found is we can grow anything we wanted with high quality crystals, just like epitaxial growth. And so that was a start with a core of a perfect crystal, put it on a shell of anything we wanted,
05:04
and then cap it with an inert shell so you don't transfer the energy to the environment. So when we did this, our particles were about 150 times brighter than previously published particles. And we can grow them any size we wanted, any doping we wanted, and now several different impurities.
05:24
So the difference between these particles and other fluorescent probes is shown here. In the red, you see the luminescence from these upconverting nanoparticles, very constant luminescence, and the time goes to six hours.
05:42
It's actually, it's the time of the patients of the postdocs doing the experiment. This is the highest intensity, six hours, nothing's changing. If you come back the next day, they're still there at that brightness. The gray is a quantum, a high quality quantum dot, and it blinks and blinks and blinks, and finally it actually photobleaches.
06:04
And in the orange and whatever that color that is, purple, you see organic dyes or fluorescent proteins. And so these are much more stable than the other probes, especially the non-blinking part is very stable. So what can we do with it?
06:22
So I'm going to show you one first application, and the application is the study of neurons. And in this picture of a neuron, you have the cell body where, let's see, does that show up? Does it show? Okay, great. This is the cell by the nucleus of the cell in this cartoon, and these are dendrites that receive chemical inputs from other neurons.
06:47
If there's a threshold, the neuron fires, there are voltage spikes that propagate down this long axon, finally getting to here where neurotransmitters are released. Now if you stare at this cell long enough and realize that this length could be not only microns,
07:03
but tens of microns, hundreds of microns, up to about a half a meter long, and you're a physicist, so you say, hmm, what's the diffusion of a molecule to get a half a meter down here through a few microns? And the answer is never going to get down there. And so what happens is the neuron has little buses, cargos, where they put molecules and things that you want to get down to the end point here,
07:31
on little motors, and these little motors chug along on railroad tracks or microtubules, and bring molecules down here.
07:40
But there's only, there's a two-way traffic. Neurons on the other side of this neuron also emit molecules, and they are absorbed by these little fingers of the axons, and they too are trafficked back the opposite way back to the cell. So the neuron actually engages in two-way trafficking. For example, a molecule called nerve growth factor, if it doesn't receive nerve growth factor molecules from this end and traffic back,
08:09
the cell will kill itself. Very, you know, sensitive cell. And so anyway, we're going to look at this trafficking, and the way we're going to look at it is we'll take embryonic rat neurons,
08:25
and so Sam and Yunshang learned how to actually dissect embryos of rats with tweezers, put it out, put it into a microfluidic cell in a proper broth, and the, in this mixture here, the neurons would actually grow their axons into this other chamber,
08:44
and this distance is 900 microns. Okay, so it's nearly a millimeter. The reason we want to do this is we can sprinkle our probes over here, and we can look at the trafficking along a very clean channel, unencumbered by a lot of other fluorescent dots.
09:01
So this is what we do. It turns out that if you put on nerve growth factor and drop the particles on, the cell grabs it and traffic its back, traffics it back. This was a project that was started by one of my postdocs 15 years ago, and now a professor in chemistry at Stanford, using quantum dots.
09:23
But it was a little bit frustrating because these quantum dots would blink, and we weren't really sure what was going on. And so we wanted to do something better than that. So when I got out of becoming secretary of energy, I said, we need different probes. And so can we get these different probes? Now we have the non-blinking probes.
09:42
And so here we have our microfluidic cell with the cell bodies down here, and we sprinkle our particles up here. And this is a image, a movie of a nanoparticle. It's a slightly accelerated real time. Their signal to noise is very good.
10:01
There's essentially no background because you're illuminating the cell with 980 nanometer light, and you're detecting in the red and the green. So this is the raw signal. You see that sometimes the cell gets, or this trafficking gets a little confused and backtracks. And if you blow up this section where it got confused, you find that indeed it changed tracks.
10:23
So in so doing, we can actually follow these trafficking of the neurons from the distant axon all the way back to the cell body. This is a color code in time. If you blow up this section, you find that sometimes it stalls. It does a little diffusive thing.
10:40
It goes backward. Sometimes it clearly goes into different tracks. The resolution here is about five nanometers. And so you know that it's gone into different tracks. So we can follow this tracking. Furthermore, we can follow the tracking over the 900 microns, nearly one millimeter.
11:01
And again, this is color coded in time. So where you see black and purple going to red, you see the different tracks. So what happens is this cargo gets transported over here. We move the stage, refocus. It gets transported over here. And you see that there are some pauses. There are motions going forward. The motions going forward don't have a constant velocity.
11:22
And there's a very complicated motion, sometimes a little bit going backwards. So if you form a distribution of velocities, you find that for the most part it's going forward. Zero minus velocity means it's going backward. And this is what we find. And when you look in vitro studies of systems, you find similar sorts of behavior, a wide dispersion of velocities.
11:46
So I'm going to skip that. The ultimate reason why we're doing this besides the biophysical measurements and biology measurements are that we want to actually study neural transport in cells coming from humans who have APOE2,3, and 4 alleles.
12:05
So if those of you may know that if you have APOE4 allele in you, you've inherited that, your chances of getting Alzheimer's are increased. If you get old, your chances of getting Alzheimer's are very much increased. By the time you're 85 or 90, you might have a 50% chance.
12:23
But you have an even higher chance when you have APOE4 alleles. There is no known cause for the disease. There are associations with it. But the smart money is now on not the formation of these amyloid plaques. That's an end point. By the time you form an amyloid plaques, large parts of your brain are already dead.
12:44
But it's the clearing of how the cells traffic molecules, for example, A-beta. So that's where the focus is shifting. And so we're collaborating with Tom Sudoff, who got a Nobel Prize in medicine physiology several years ago
13:01
for his work identifying protons in vesicle fusion. But he's now become interested in understanding this mechanism. And so to that end, he's developed ways of taking human cells from subjects and inducing them to grow into neural cells that we can and he can use in his experiments.
13:23
So now I'll turn to some physics and describe to you what the fluctuation theorem is. So the fluctuation theorem is a statement about entropy. Just to remind you, entropy delta S is equal to the change in heat over temperature.
13:41
And the theorem states that the probability of entropy flow to increase the entropy, that's P of sigma, to the time-reversed thing where in time-reversal picture, your entropy is either going to remain constant or become negative. So the ratio of entropy flowing the right way in the correct time order,
14:04
where you can ask me lots of questions about what time means, but that ratio grows exponentially large as time increases. So that's a statement of statistical mechanics that says the probability of entropy increasing grows exponentially large to the probability of entropy shrinking.
14:24
Entropy shrinking is, imagine a drop of red ink in water and it's a certain diameter. Entropy growing means the drop dissipates and goes to pink water. Entropy shrinking is a drop actually concentrated and gets to redder drop. For a very, very short period of time that's possible,
14:41
but over any finite amount of time because the laws of large numbers, it goes the other way. And so this is a formal statement of that. But what's important about this statement is it's true in equilibrium and it's true in non-equilibrium. All right, so how do we talk about and measure the properties of entropy flow?
15:04
And so again, this is what I just said, that this says that this ratio of entropy going the right way becomes exponentially large. So this probability of flowing entropy running the clock forward is running the motor forward. The motor is hydrolyzing units of energy, ATP,
15:23
and so that's the probability of the motor advancing this cargo forward by a delta x. And the entropy going backward is simply the reverse direction. The probability of going backward is equal to this, okay? So that's a measurable quantity. So we say at time t equals zero, you have this red dotted line
15:42
and you look at the correlation in this long track and a little time later, 10 milliseconds later, where is the particle? Where is that cargo? Most of the time it's gone forward, a little bit of the time it goes backward. After 40 milliseconds, it's mostly gone forward. After 100 milliseconds, it's moved forward.
16:01
So this is a measured quantity, these probability distributions. The solid line is a fit to a Gaussian. So this is just e to the minus quantity x squared minus mu squared over 2 sigma squared. So that becomes a measured quantity. I look at this expression and say, great, I can take the log of this measured quantity,
16:22
p of delta x over p of minus delta x, and the log of that is equal to sigma t over Boltzmann's constant. And what I get is this. I'm now looking at the entropy per unit length, I've divided by delta x, and I find that this measured quantity phi is equal to the displacement mu is the mean displacement,
16:44
the moving of this blob of Gaussian probability, divided by the dispersion, the width. Now, if you stare at the single particle as it traverses this long neuron, it turns out that on certain trajectories, it asymptotes to a certain level,
17:02
and another trajectory then pauses, and another trajectory asymptotes to another level, and yet another level. And these seem to be discrete levels, 0.1, 0.2, 0.3. All right, now, we're going to interpret this as one motors, two pairs of motors, three pairs of motors,
17:20
because remember, in the formula for this, it's really just saying that phi is the entropy per unit length that reaches some steady state. And so you say, well, if it's entropy per unit length, and you've got three motors burning ATP versus one motor burning ATP, you should have three times more entropy, or possibly, so that's the idea.
17:45
But let's not look at just one cargo, let's look at a whole bunch of cargos. So here are 12 cargos, and in all the cargos, the same level of quantization, so this is the histogram of all those cargos, and yet the velocity for 0.1, 0.2, 0.3, 0.4 is very dispersed.
18:05
Oops, okay, that's a puzzle. But this is very exciting, because now you know this quantization in the fluctuation theorem, the quantization in entropy per unit length is very discrete, and it's true for all the cargos, no matter which, for all these 12 cargos.
18:24
Okay, so hold that thought. We think that this is the number of motors, because you just look at the algebra and says, it really is the amount of entropy and, you know, four motors burning up ATP will generate that much more. Now we go back and we put in bigger particles, brighter particles,
18:45
and we are now able to look at single molecular steps in this neuron, this live neuron. These are now particles giving out five million counts a second, we're recording five million counts a second, which means we can take video rates of one millisecond,
19:03
0.8 milliseconds to get the image and 0.2 milliseconds to read it out. And if you blow up these, this is the system at room temperature, 30 degrees centigrade through seven degrees, you see these single steps. So for the first time, you see very clear single steps in a live neuron
19:22
for many, many minutes until it runs its course, you come back the next day, the neuron hasn't curled up and died. This is good. It's just infrared heat. It warms up the water, but it doesn't induce phototoxic damage. And so we can make a histogram at room temperature of the steps.
19:41
This is point, rather eight nanometers, 1624, partially resolved. And so you see, and this is zero steps. And you compare that to, again, the in vitro studies and you find that there's similar distributions. These are in vitro studies where it's different, it's not mammalian,
20:01
it's purified in yeast, but at least it's a very similar distribution of these so-called dynein molecules. Now, what else can we do? We can have these distributions, but we have them at higher temperatures, going to body temperature for a mammalian cell. And we find that the distribution narrows up, it looks a little bit better.
20:22
And if you plot, let's say it makes a step and you plot the time before it makes another step, the so-called dwell time, what you find in this histogram is there's a rise and a fall, a rise and a fall. Once you see a rise and a fall, you immediately think that this dwell time
20:41
is described by two rate constants, one rate, k1, but until you satisfy this first step of k1 probability, then you do the other rate, k2. And when you fit for k1 and k2, what we find, for example, it was described by a single rate constant, for example, like the lifetime of an excited state of a dye molecule,
21:02
it's just a single exponential decay. So when you see a rise and a fall, you know immediately there are multiple rate constants. So what we find, remarkably, is that the k1 and k2 are equal to within a few percent, 5%. And they're thermally activated because as you go from room temperature to series 7 degrees,
21:24
they actually increase. So here we have something that could not have been seen before, because we now have the time resolution to see these two rate constants. And so in so doing, what we've done is we've created a new model for how the dyeing steps.
21:42
I'm not going to talk you through the model because you can ask me questions later, but it's a new model that previously people couldn't see, develop this model because they weren't able to measure in vitro, in vivo rather, with these very high ATP concentrations.
22:01
So when they had to slow up the reaction in order to see single steps, and when they slowed up the reaction, they hid a lot of stuff. So that's the bottom line of this. I should also say something about non-equilibrium thermodynamics. This, starting back from the fluctuation theorem,
22:22
the probability of entropy increasing versus entropy decreasing can also be cast in a form where the probability of doing some more positive w over minus w, the time reversed one, and you can write that in terms of an exponential of change in entropy
22:41
over kT where the entropy is given by the change in the Helmholtz free energy, delta A, and the amount of work, delta W. And so if you consider the system, the vesicle and the motor, and the reservoir, this reservoir of ATP and the rest of the cell, what you find is this fluctuation theorem then becomes a statement about the force over kT.
23:06
So again, you just take the log of both sides per unit length x, that's our old friend phi again, equals sigma T, that was the first version of the fluctuation theorem, and now you find that you have a force over kT.
23:21
Now, the remarkable thing about this force is that it doesn't seem to have the right value because phi, remember, was 0.1, 0.2, 0.3, and you say, ah, well, the cells are 37 degrees, I stick in 37 degrees, and I get as a force 0.41 pN for the motor system, as the minimum force.
23:45
However, there's a problem because in vitro, where in vitro would always measure a weaker force because you don't know if it's got the full molecular components, but as a lower limit, very careful in vitro studies have found that the force is around 4.3 pN,
24:04
about an order of magnitude stronger. And so this is a problem because you're saying, hmm, if it's 37 degrees, I'm getting an order of magnitude smaller, so what about T? Is T a real temperature? Well, it is the temperature of the cell. I can guarantee you that we're not 10 times 310 Kelvin.
24:25
And so you can say, well, okay, but it's really the drag force over T, and so is there another way of measuring the drag force? The drag force is a cargo, it's a sphere, and so a Stokes drag, 6 pi eta, the kinetic viscosity times the radius of the sphere times the velocity.
24:47
But you look up in the literature what the viscosity is in cells and in axons in cells, and they're all over the map. And they end up giving you a force anywhere from 0.2 pN to 20 pN. 20 pN makes no sense because that means the viscous drag force
25:03
is actually greater than the force of the motor. And my faith in nature is that biology will never design a motor that cannot overcome viscous force. It's just a waste of life. So now, you can try to make better in vitro measurements,
25:26
not in vitro, in vivo measurements of viscosity, but there's another way. And the other way is something we're gearing up to do, which is a much more straightforward way, and it's to appeal to another theorem called the fluctuation-dissipation theorem.
25:42
You're all familiar with fluctuation-dissipation theorem. It was conjectured by people at Bell Laboratories when Bell Labs engineer discovered that a resistor or wire had thermal noise, voltage fluctuations, and Nyquist said,
26:04
hmm, that's interesting, you have these thermal fluctuations, but these fluctuations seem to be related to the resistance of the conductor. So is there a universal relationship between the fluctuations in voltage and the dissipation, the resistance? So he proposed a thing called the fluctuation-dissipation theorem.
26:23
The first, most famous example of the fluctuation-dissipation theorem is the relationship between Brownian motion of a particle in a fluid and Stokes' drag. But if you look back at how Einstein did that, he was such a genius, he didn't have to propose a theorem that he had to prove. He finessed it.
26:42
Now, it turns out the fluctuation-dissipation theorem is a theorem, and it applies to Brownian motion of particles, it applies to resistors and cells, it applies to many, many things. It's deeply connected, fluctuations and dissipation. Along comes Lars Onsager in 1930, and says, you know, well, that's in equilibrium.
27:01
What if you put a system slightly out of equilibrium? Is the path back to equilibrium also related to the fluctuations in equilibrium? So this was the so-called Onsager regression. I think about 30 years later, he gets a Nobel Prize for this, because it's the first statement in non-equilibrium thermodynamics
27:21
where you begin to try to understand non-thermal thermodynamics and statistical mechanics. It too later became a theorem. And it not only became a theorem slightly out of equilibrium, it can be significantly out of equilibrium. But what fluctuation-dissipation theorem out of equilibrium,
27:40
it requires one other caveat. You have to reach some steady state. And when you reach a steady state, that means time invariant. Equilibrium is time invariant. Once you reach a steady state, and what you can see is the fluctuation in this vesicle, and the dissipation, the amount of resistance it's feeling as it's moving forward,
28:03
are going to be related, but that means you have to measure the fluctuations and the motion at the proper time scale. Unfortunately, the time scale is going to be about 1 to 10 microseconds instead of a millisecond. So we're gearing to do that experiment. We think we know how to do it. We can measure nanometer steps with one microsecond resolution.
28:24
We'll let you know if it's successful, perhaps in a half a year. Anyway, that's what's happening there. This is just math, so that, I'm not going to tell you about that. So let me continue, two minutes. Okay, two minutes, I'm just going to show you some end results of ultrasound.
28:42
This is your standard high-quality ultrasound imaging without computer cartoon renditions. This is the raw data of a little fetus, 3 centimeters long. We figured out how to get rid of speckle. I'm not going to be able to tell you how to do that. It's very straightforward physics.
29:00
This is a demonstration of some corn starch, of higher density corn starch. This is an experimental measurement. When we angle average and frequency average, you can see what we have here. This is a simulation. It agrees with theory. We've taken pictures of wrists.
29:24
We didn't know you had to get an R.R.B. in order to take an ultrasound image of your own wrist, but we're now wiser for that. And we've done other things. And so with this standard device used to image small animals, this is a blood vessel of the wrist, this is a tendon, this is the edge of the bone.
29:42
This is what we see now. With getting rid of the speckle. This is only a factor of six reduction because we couldn't have total control of the electronic pulses in this commercial unit. But it's achieving the fraction limit now. And we think we can get 10 times better.
30:02
I just want to say that we're going to other ways to make it a practical thing. So as you scan over your abdomen and your liver, we're now developing a way of auto-pointing so you can get very different angles. In one fell swoop. Finally, in the last minus one half minute, I'll get two more minutes.
30:20
I'm going to tell you a little bit about frequency, non-linear ultrasound coming with two frequencies, F1 and F2 at five and six megahertz. And you're going to detect a difference frequency at F1 minus F2. Okay. Why are we doing this?
30:41
Well, we thought we were going to get rid of speckle, but we, so we go to a Chinese grocery store, buy a pig kidney. And this is the vast picture, this is the part you eat or some people eat. And anyway, if you have a standard ultrasound image, you can't actually tell, see this with different from water.
31:02
But if you do non-linear, it really shows up. You go to Safeway, buy a piece of salmon and these little flecks of fat, you can see the flat. And you can't see much. So I said, hmm, nature is talking to us, trying to tell us something. And so I'm going to go to this. And so we went to a surgeon, a brain surgeon at Stanford who does brain surgery of glioblastoma tumor, very deadly brain tumor.
31:28
And she also does experiments with mice of human glioblastoma tumors. So this is an MRI scan of a mouse, and this is the tumor. And if you go to a slice towards the nose of the mouse, you don't see the tumor.
31:41
In normal ultrasound, you do see the tumor, but you see a lot of other stuff. And in difference frequency, you only see the tumor. We understand why this is. It's, if you just look at the terms of the nonlinear equation, this term is very similar to linear term.
32:01
And this, there's another term that is orthogonal to this linear term. In the sense, orthogonal, you can show analytically, this is the wave equation, the nonlinear wave equation, this is the source term. This is velocity of material shaking around, and this is some higher order correction. But the two terms, if you come in at 90 degrees, they scale as two over the cosine theta, but cosine theta is near zero.
32:28
And so what we found, kind of by accident, is you come in at 90 degrees, you get a totally different contrast mechanism. And contrast mechanism is everything in biology. Remarkably, not only was speckle reduction overlooked in this way, but this was also overlooked.
32:45
And so we're very excited that this could also be a new way of measuring cancer tumors. And let me end by saying that ultrasound is the cheapest imaging modality we have in clinical medicine, looking deep inside the body.
33:03
It's now entering into cell phones. We think everything I've showed you can actually be used in a cell phone. So suppose you're in Sub-Saharan Africa, you can do the cell phone, you can upload the data, have it analyzed over the internet and get it downloaded again. So you need no hospital, no x-ray shielding, no nothing.
33:23
And so we think, and this is actually we think going to be comparable in resolution to MRI, we think we can get a couple hundred micron resolution. You might ask, why was I interested in ultrasound? Well, I'll end by, you know, inspiration comes from strange places.
33:42
I advised Siemens, I was advising Siemens for four years. Siemens' biggest moneymaker is their medical imaging, MRI and CT scans. They make the best MRI machines in the world. The chief technology officer at Siemens says he's scared to death that Black Swan will emerge and take away all our business from CT and MRI.
34:03
And I said, like, for example, what? And he said, like, ultrasound. So I'll end there. Thank you.
34:20
That's a beautiful talk encompassing chemistry, physics, biology. It is possible in 2019 to be a Renaissance scientist. There you have it. Okay, we don't have long, but let's have some questions. Start with, yep, go.
34:45
Okay, so the question is in these tracks, they're actually dynein. Dynein goes from the axon to the cell body and kinesin goes the other way. But in the dynein motors, the question was if you see a bunch of them, do they interact, do they bump them? And they get in each other's way, more or less.
35:03
As near as we can tell, they go and, like, they can stall. We're, but we need to look at much more data to see, for example, when they change tracks, are they changing tracks in the same place? Because maybe the tracks run out of rail and then they have to jump. Or is it something like that? We can also put in dyes.
35:20
So there's much to be discovered on that. But they seem to interact in a non-cooperative way. Okay, go, please.
35:45
That's a very important question. Just to remind you all, there are two sets of motors. Kinesin, the microtubules are polarized, so there's a plus and minus end. Kinesin walks along one direction, dynein walks along the other direction.
36:00
These cargos seem to have both sets of motors. And so what we see is it's going down and then it reverses a little bit. Sometimes it changes tracks, but sometimes it's on the same track. So it's on the same track within a few nanometers. We know they can't have reversed polarity, so it engages. The mechanism for which motors get turned on and turned off is unknown.
36:22
Okay, but if you add in the distant axon, the instructions are put on the dynein motors and use them. But the fact that they reverse means Kinesin had to been there and similarly going the other way. So we don't know. Thank you.
36:50
Okay, good questions. The size is tunable. We can make it as small as 10 nanometers. But remember, the brightness goes as the number of impurities, which is the volume.
37:02
So for 10-20 nanometers, you have to take data every 100 milliseconds. For this once every millisecond to get nanometer resolution, we went up to 150 nanometers. But the wonderful thing about these parts is you can dial in what size you want.
37:23
So we can actually dial it in. We now understand the growth mechanism enough that we can dial it in to 5% with a 2% dispersion in size and 5% in this is where we want it. So it's kind of under control now.
37:43
And we're, Stanford's patenting it. There's a company interested and maybe in a couple of years, many people can have these particles. Question here.
38:03
Yeah, very good questions. Let me tell you how surgeon, when they operate, they go and they poke their finger. I'm not kidding. The tumor feels different. That's actually when they do a palpation. And I had a friend who sadly had a corrective cancer that spread to the lungs.
38:22
They opened up her entire chest. The surgeon went into the whole lungs and went like this, feeling for tumor. Okay, so what they're really doing is they have sensitive fingers. They feel the nonlinearity. And sadly, the sad thing about this nonlinearity is if they are taking more successive slices,
38:44
if they take, you know, say it feels normal again, but if they leave a little margin of a fraction of a millimeter, it's going to feel normal. But then they take the last final slice, they give it to pathology, they solve the patient, and then 40% of the time, it comes back, the margins weren't clear.
39:02
So the doctor has to call the patient, says, sorry, you have to go back on the knife, but if you don't and we don't get it out, it could metastasize. Okay, and I had a friend, two breast cancers, within a year, both margins not clear by the finger.
39:21
So if you had any way of testing with a much more sensitive than finger to a fraction of a millimeter, that would be very, very good. Okay, so these could readily have direct clinical applications. Why does it go nonlinear? Tissue is very nonlinear. You push and it starts to push back.
39:41
And so the surgeon's finger has become pretty good, but it's all intuitive. It's like machine learning, they don't tell you what the answer is, but they charge a lot of money.
40:00
Yeah, the step size was eight nanometers. And so if you look at the error and the mean, it was about a half a nanometer, the standard deviation was about two, two and a half nanometers. How do you achieve it? Oh, it's a centroid. It's an isolated particle and there's this big blob and you find,
40:20
the signal is so good you can find the center of the dot. It's just like super resolution, except bazillion more counts. I think, I don't think it's failed yet.
40:40
I think my suspicion is that it's not in equilibrium. And it's been shown actually physically with optical tweezers set up. If you shake it and you know exactly what you're doing and you move it along, you can actually have enough and you reach a steady state non-equilibrium.
41:01
It's very well described by this Onsager non-equilibrium stuff and you can dial in what the temperature is. So what does effective temperature mean? It is really that the fluctuations are fluctuating locally much more than Brownian motion at three seven degrees or in the case of this other experiment done, I think, three years ago
41:22
at room temperature with an optical tweezer shaking it. So we think that if the motor is working very fast, it's actually inducing fluctuations. Think of the cargos as stagecoach and you got teams of horses and they're pulling on against each other, they're not fully coordinated and they're shaking each other and so this is what we see.
41:42
It's not boiling water, it's just induced fluctuations. So we think that's it but we want to actually use the fluctuation dissipation theorem to really prove that that's really what's happening. The strong conjecture is it's really out of thermal equilibrium. So this would be the first motor that's in nature
42:01
that really seems to be out of equilibrium. Question here. Optical forces and fluorescence dimension, no. Because first of all, we're not focusing on lasers, we're wide field. So there is no gradient.
42:21
And this is, remember, this is infrared light bathing the axon of the cell, okay. And then we're not heating up the chromium and killing the cell.
42:40
No, I don't, there is no special secret. What actually, I'll tell you what it is. When you get stuck, instead of reading papers, I hate to read papers, I say, okay, what are the fundamentals? What's really going on? In ultrasound, what's really going on is what's the origin of speckle?
43:01
And go back to freshman interference physics and say if you come in at different frequencies and different angles, you can get rid of it but what's not realized is there were in fact debates in the ultrasound literature, is it better to use different frequencies or different angles? But then you say, wait a minute, if you come at different frequencies, you have planes of interference and different frequencies may have different planes.
43:23
You come in different angle, the planes are over here. So the proof is, of course, they're orthogonal. Just, this is a picture proof. But it wasn't realized in the literature and it wasn't realized when we talked to experts. It's no, no, no, it's one or the other, okay. So, and in epitaxial, we're just borrowing from the semiconductor industry
43:44
for epitaxial growth. It was so, that was it. Okay, one more. So, in your, what is the secret to science communication across the spectrum?
44:04
I can't say I know the secret to communicating with politicians. With some, you can. Because if you explain it very clearly, you're not talking down to them, you're talking up to them in a sense that, you know, every good scientist
44:21
passes the grandfather-grandmother test. If they're not scientists, can you explain it to a lay person in very simple terms without jargon? If you can do that, you yourself understand it like the back of your hand. And it's when you can explain it like that, that people on the other side can understand.
44:41
And then you look for body clues, just like when you're dealing with people in general, you're looking for body clues as to whether their eyes are glazing over or if you say something and they're getting angry, or they're going like this and they're looking really excited, you know. So, I think it's things like that. But it's always simple language, try not to use jargon.
45:05
We unfortunately have to stop. But one last question from me, which is an extension of that question. So much captures your attention, do you have a rule for deciding what you're actually going to focus on? No, I'm terrible that way.
45:21
I'm the most probably scientifically promiscuous person you've ever met. But, and in fact, it leads to certain inefficiencies where I'm in the process of writing paper or doing something, I find, oh, that's interesting. And then go tailing off and then pretty soon it's midnight.
45:40
Yeah, but you seem to get somewhere with each of these problems. That's luck. Okay, thank you all very much. Thank you for your lovely questions.