This is an old video so it may not be monitored by HMMI, but I have to ask this question. I understand that a stem cell will begin to differentiate by selectively transcribing certain genes and not others, through the use of external proteins called transcription factors (TF). I understand now that these TF's are themselves, combinations of a library of perhaps 2,000 basic TF proteins that are produced from 3% of the genes in DNA. Dr. Tjian explained that by combining these 2,000 proteins into much bigger assemblies, you can get hundreds of thousands of combinations that can and will bind to your DNA genome at different places, thereby activating some genes and turning off others. What he did not explain was the following: 1. How does a stem cell "know' which of the 2,000 or so base TF's to combine to produce a specific Activation complex? For example, if the stem cell needs to turn on transcription of gene 3452 on chromosome 22 so that it becomes a fat cell, let's assume the activation protein complex to do that must be a combination of base TF proteins 3,2455, 62, 344, etc. What signal does the stem cell get that tells it which of the 2,000 TF proteins in the nucleus need to combine? How does it know which of the proteins to select? They are all in the nucleus at the same time. Seems like the stem cell needs a supervisor with a master blueprint ... some "one" to call out the directions for assembling the proteins needed to produce that bigger gene specific transcription factor?? 2. How does that bigger assembly of gene specific transcription factors know how to find gene 3452 on chromosome 22? Does it actually randomly cruise down the entire 3 billion base pairs looking for a gene specific site? How does it unwind all that DNA from its many many histones, so it can look for that specific site? If there is an answer to the above in any other video on UA-cam, please post a link to it here. Thanks
I like your Shoshin. The formation of both complexes you mention in 2 (complexes of DNA and protein(s)) and 1 (complexes of multiple proteins) are relatively energetically favorable, compared to all other possible arrangements of these components. en.wikipedia.org/wiki/Gibbs_free_energy This means that eventually (at equilibrium), they should be formed more than other possible arrangements (picture all of these components diffusing semi-randomly in some space, and partly by chance coming close to each other in different combinations, then the ones that are energetically favourable are more likely to interact for longer periods of time). However, this doesn't imply that these complexes will be formed in a relevant timescale. Given the size of the space, the number and diffusion coefficients of the molecules, and other important parameters, having all of these components come together into a complex could take an amount of time that approaches infinity. If we use biologically relevant parameters, then you can simulate how long it should take for these complexes to form. The results of the simulations may not match with observations about how fast these complexes form (in reality), which implies that you are missing a part of the picture. So regarding Q2 (forget about it unwinding histones for now, assume that the site it is looking for is available, otherwise it will not bind it): if a transcription factor had to 'search' (move past) all of the available DNA in 3D space, this would take longer than the amount of time that people observe for transcription factors to find their sites on DNA. So a concept called 1D diffusion, where transcription factors can also 'track' along the DNA rather than just diffusing in 3D space (reducing the dimensionality of the search, here in space, decreases the amount of time required to find what you're looking for) is believed to occur, and this may reconcile the mismatch between the observed and simulated time required. I'm not sure about the coherence of 'expected time intervals' between simulations and observations for Q1 (protein complex formation), but it's an interesting point, so thanks.
Very remarkable questions!. To establish and maintain the cell identity, some mysterious "control-tower' is mandatory. And genes should be regulated as a cluster or a group. Fore example, lets assume that to establish the identity of fat cells, a set of genes such as A, B, E, K, Y should be expressed as a cluster, whereas to establish myocytes, a different set of genes such as A, B, R, X, Z should be expressed as a cluster. In that case, A, B need to be expressed in both cells, but E, K, Y should be exclusively expressed in fat cells. Just combination of TFs can do this job, but will be very complex, error-prone, not robust. Something more may be required.
If not evolution then what? This is hardly perfect machinery or we wouldn't have cancer. Any alternative you suggest must account for that fact. This is a really interesting vid, points to epigenetics.
Fascinating, maybe the best talk on cell-biology I've seen. Thank you so much!
This is an old video so it may not be monitored by HMMI, but I have to ask this question. I understand that a stem cell will begin to differentiate by selectively transcribing certain genes and not others, through the use of external proteins called transcription factors (TF). I understand now that these TF's are themselves, combinations of a library of perhaps 2,000 basic TF proteins that are produced from 3% of the genes in DNA. Dr. Tjian explained that by combining these 2,000 proteins into much bigger assemblies, you can get hundreds of thousands of combinations that can and will bind to your DNA genome at different places, thereby activating some genes and turning off others.
What he did not explain was the following:
1. How does a stem cell "know' which of the 2,000 or so base TF's to combine to produce a specific Activation complex? For example, if the stem cell needs to turn on transcription of gene 3452 on chromosome 22 so that it becomes a fat cell, let's assume the activation protein complex to do that must be a combination of base TF proteins 3,2455, 62, 344, etc. What signal does the stem cell get that tells it which of the 2,000 TF proteins in the nucleus need to combine? How does it know which of the proteins to select? They are all in the nucleus at the same time. Seems like the stem cell needs a supervisor with a master blueprint ... some "one" to call out the directions for assembling the proteins needed to produce that bigger gene specific transcription factor??
2. How does that bigger assembly of gene specific transcription factors know how to find gene 3452 on chromosome 22? Does it actually randomly cruise down the entire 3 billion base pairs looking for a gene specific site? How does it unwind all that DNA from its many many histones, so it can look for that specific site?
If there is an answer to the above in any other video on UA-cam, please post a link to it here. Thanks
Frank Gregorio
It is caused by magic
I like your Shoshin.
The formation of both complexes you mention in 2 (complexes of DNA and protein(s)) and 1 (complexes of multiple proteins) are relatively energetically favorable, compared to all other possible arrangements of these components. en.wikipedia.org/wiki/Gibbs_free_energy
This means that eventually (at equilibrium), they should be formed more than other possible arrangements (picture all of these components diffusing semi-randomly in some space, and partly by chance coming close to each other in different combinations, then the ones that are energetically favourable are more likely to interact for longer periods of time). However, this doesn't imply that these complexes will be formed in a relevant timescale. Given the size of the space, the number and diffusion coefficients of the molecules, and other important parameters, having all of these components come together into a complex could take an amount of time that approaches infinity. If we use biologically relevant parameters, then you can simulate how long it should take for these complexes to form. The results of the simulations may not match with observations about how fast these complexes form (in reality), which implies that you are missing a part of the picture. So regarding Q2 (forget about it unwinding histones for now, assume that the site it is looking for is available, otherwise it will not bind it): if a transcription factor had to 'search' (move past) all of the available DNA in 3D space, this would take longer than the amount of time that people observe for transcription factors to find their sites on DNA. So a concept called 1D diffusion, where transcription factors can also 'track' along the DNA rather than just diffusing in 3D space (reducing the dimensionality of the search, here in space, decreases the amount of time required to find what you're looking for) is believed to occur, and this may reconcile the mismatch between the observed and simulated time required. I'm not sure about the coherence of 'expected time intervals' between simulations and observations for Q1 (protein complex formation), but it's an interesting point, so thanks.
Very remarkable questions!. To establish and maintain the cell identity, some mysterious "control-tower' is mandatory. And genes should be regulated as a cluster or a group. Fore example, lets assume that to establish the identity of fat cells, a set of genes such as A, B, E, K, Y should be expressed as a cluster, whereas to establish myocytes, a different set of genes such as A, B, R, X, Z should be expressed as a cluster. In that case, A, B need to be expressed in both cells, but E, K, Y should be exclusively expressed in fat cells. Just combination of TFs can do this job, but will be very complex, error-prone, not robust. Something more may be required.
You have to take into consideration that stem cells can divide symetrically and asymetrically, making their proteomes different
Priceless. Thanks for the intriguing insights. Loved it
Fantastic Lecture! Thank you!
Love you sir. Salute sir.
Superb !
Mind blowing !
Hello 👋 sir
If not evolution then what? This is hardly perfect machinery or we wouldn't have cancer. Any alternative you suggest must account for that fact. This is a really interesting vid, points to epigenetics.
To me, evolution is more powerful than saying, it just got put here..
If not « adding of atoms in pure hazard » until we find a combination this is surely intelligent design