An Introduction to Polyketide Assembly Lines

Transcript of Part 3: Vectorial Specificity of Assembly Lines

00:00:08.03	Hello again.
00:00:09.20	I'm Chaitan Khosla
00:00:11.11	from Stanford University,
00:00:13.10	and welcome to the third part
00:00:16.24	of this trilogy of lectures
00:00:18.29	on polyketide antibiotic biosynthesis
00:00:22.05	and assembly lines.
00:00:25.29	So, in the first lecture
00:00:27.21	I introduced you to the basics
00:00:30.02	of what an assembly line is,
00:00:33.13	what's the chemistry that happens on an assembly line,
00:00:37.03	and what an assembly line looks like.
00:00:41.15	We looked at the basic architecture
00:00:45.06	of the 6-Deoxyerythronolide B synthase,
00:00:48.19	or DEBS,
00:00:50.18	which I have used subsequently
00:00:52.27	to make most of the points
00:00:55.08	that I have made.
00:00:57.12	I introduced to you
00:00:59.16	the architecture of this assembly line,
00:01:01.25	or at least as best as we can gauge today
00:01:05.12	based on data we have.
00:01:07.24	In the second lecture
00:01:10.14	I introduced you to the tools,
00:01:13.09	and how we've used those tools
00:01:15.24	to study different kinds of questions.
00:01:18.23	So, we talked about refactoring
00:01:21.18	a metabolic pathway
00:01:23.24	like this assembly line of DEBS
00:01:26.03	in E. coli.
00:01:27.09	We talked about how to reconstitute that pathway
00:01:30.07	in a test tube.
00:01:32.02	I pointed out that you can
00:01:33.25	isolate individual reactions
00:01:36.12	on the assembly line
00:01:38.05	and look at the kinetics,
00:01:39.22	whether you're looking at extender unit selection,
00:01:41.24	translocation,
00:01:43.08	elongation,
00:01:44.27	modification events.
00:01:46.24	We also talked about how these tools
00:01:48.23	can be used to study specificity
00:01:51.03	and, in particular,
00:01:52.29	we talked about two types of specificities:
00:01:56.00	stereospecificity,
00:01:58.24	I mentioned that stereochemistry
00:02:01.06	is controlled by reductive domains,
00:02:03.23	as well as extender unit specificity,
00:02:07.17	and I mentioned, over there,
00:02:09.25	you not only have enzyme-substrate recognition
00:02:13.15	but also protein-protein recognition.
00:02:18.04	This final lecture focuses on what, in my mind,
00:02:22.10	is the grandest challenge of all
00:02:25.03	in this field:
00:02:27.05	understanding how an assembly line
00:02:30.17	actually comes about.
00:02:33.20	What do I mean by that?
00:02:37.05	I've been showing you,
00:02:39.06	quite freely I might add,
00:02:41.00	schemes like this
00:02:43.08	throughout my earlier two lectures.
00:02:47.20	Every time I show you one of these schemes
00:02:50.08	or, for that matter,
00:02:52.04	you could open your favorite journal of biochemistry today,
00:02:55.26	and chances are in any given issue,
00:02:58.07	you'll find somebody having published
00:03:00.14	yet another one of these
00:03:02.13	kinds of assembly lines
00:03:04.04	that they have cloned, sequenced,
00:03:05.21	and started to characterize.
00:03:07.07	And in all cases you'll find
00:03:09.01	cartoons like the kind you see in this scheme.
00:03:15.26	The question that lies ahead of us is,
00:03:18.11	how do we know that
00:03:21.02	this is how the assembly line works?
00:03:26.27	And one way to recognize
00:03:29.12	why that question
00:03:31.25	is not an obvious question
00:03:34.15	is by thinking back
00:03:36.08	to all those orphan polyketide assembly lines
00:03:39.23	that I talked about to you
00:03:41.27	at the beginning of my first lecture.
00:03:43.29	So, I mentioned to you,
00:03:45.29	there's these large numbers of orphan assembly lines
00:03:49.03	that make polyketides
00:03:51.10	whose structure we don't know.
00:03:54.14	If I gave you the sequence
00:03:57.02	of one of those assembly lines
00:03:59.17	and asked you
00:04:01.17	to draw me an equivalent scheme
00:04:03.06	like the kind you see in front of you
00:04:05.02	in this slide,
00:04:07.10	chances are you won't be able to do it
00:04:09.15	with any confidence.
00:04:11.26	And the reason why I can show
00:04:15.07	this scheme in this slide,
00:04:16.24	and hold my head up high,
00:04:19.03	is because I know the answer
00:04:23.07	for this assembly line.
00:04:25.26	I know the structure
00:04:27.20	of 6-Deoxyerythronolide B,
00:04:30.13	and using that knowledge I am essentially
00:04:34.14	back-programming on this assembly line
00:04:37.08	what must happen at each stage
00:04:40.07	in order to get that final product.
00:04:43.20	But that begs the question,
00:04:45.27	how does nature know
00:04:48.01	that this is how the assembly line must work?
00:04:50.27	Why does nature start here,
00:04:54.01	and move here?
00:04:55.29	Or putting it differently,
00:04:57.22	why can't nature take a product
00:05:00.09	from this module
00:05:02.09	and instead of passing it on to this module,
00:05:05.12	why can't it just skip these modules altogether
00:05:08.21	and pass it on to this?
00:05:10.29	Or, why can't it take this precursor
00:05:15.13	and pass it to some other polyketide synthase
00:05:18.05	that's floating around in the same cell?
00:05:21.20	Where are the rules
00:05:24.17	that dictate how this vectorial translocation
00:05:28.17	of polyketides
00:05:30.15	must occur in the manner
00:05:32.11	that I tell you it's occurring
00:05:34.20	in this scheme?
00:05:36.20	That is the fundamental challenge
00:05:39.21	associated with this field.
00:05:43.27	The way we think about this challenge
00:05:46.23	goes back to the work
00:05:50.08	of an inspirational enzymologist
00:05:54.26	from a previous generation,
00:05:56.16	whose name you may have heard:
00:05:58.25	Bill Jencks, or William P. Jencks.
00:06:02.29	Jencks made huge contributions
00:06:06.07	to our understanding
00:06:08.08	of how enzymes work.
00:06:10.15	One of the problems he talked about
00:06:12.29	and gave a lot of thought to,
00:06:15.09	as is discussed in this article
00:06:17.07	that I would encourage any student
00:06:19.08	who's interested in machines in biology
00:06:24.03	to read,
00:06:25.21	was the question of how do you get
00:06:28.24	vectorial processes occurring in nature,
00:06:32.19	off the back of proteins like enzymes?
00:06:37.13	This article is chock-full
00:06:40.05	of many really interesting insights
00:06:42.21	that were way ahead of their time.
00:06:46.01	I certainly don't have the time
00:06:47.27	to go into all of them.
00:06:50.03	But there's one insight
00:06:51.27	that I'm going to focus on
00:06:53.21	because I think it does a very nice job
00:06:55.27	of framing the understanding
00:06:58.11	of the vectorial chemistry
00:07:00.05	that I'm here to tell you about.
00:07:02.26	What Jencks says in this article
00:07:07.03	is that proteins that do coupled vectorial processes,
00:07:10.27	that do vectorial processes,
00:07:13.06	differ from your garden variety enzyme
00:07:16.13	in the sense that they undergo large changes
00:07:21.08	in substrate specificity
00:07:23.14	at different points in their overall catalytic cycle.
00:07:29.26	And a really good example
00:07:32.14	that he uses in his quintessential
00:07:35.01	chicken-scratch cartoon manner
00:07:37.27	that's shown over here
00:07:40.08	is the example of how
00:07:43.03	myosin moves on actin fibers.
00:07:46.15	That you probably know about
00:07:49.00	because it's what textbooks are made up...
00:07:51.09	biochemistry textbooks are made of.
00:07:53.18	What you know is that myosin
00:07:56.07	moves along an actin highway
00:07:59.16	that is made up of actin monomers
00:08:02.15	by essentially going from one monomer of actin
00:08:07.15	to the next and so on.
00:08:11.07	And that movement
00:08:13.18	is powered by the hydrolysis of ATP,
00:08:17.19	by myosin.
00:08:19.28	And what Jencks postulated back then,
00:08:22.14	or what he articulated in this paper,
00:08:25.28	was that the reason you get this vectorial movement
00:08:32.07	is because there are different states of myosin
00:08:35.15	at different points in the overall cycle,
00:08:39.15	and some of these states
00:08:42.07	bind tightly to myosin,
00:08:44.22	the unbound states,
00:08:47.03	whereas ATP- or ADP-bound states,
00:08:50.19	the myosin does not bind very tightly
00:08:54.00	to actin.
00:08:56.21	The key point that Jencks made
00:08:59.14	is the reason this is a machine,
00:09:01.22	and it's not like myosin
00:09:03.23	dances back and forth on this actin highway,
00:09:07.09	the reason it motors along this actin highway
00:09:10.25	is because there is a distinct change
00:09:14.27	in specificity of myosin for actin
00:09:18.14	at different points in the catalytic cycle.
00:09:21.21	When myosin binds to ATP,
00:09:26.01	it exchanges with an unbound state
00:09:32.03	that binds to one monomer of actin,
00:09:36.02	whereas when that ATP is hydrolyzed,
00:09:39.23	that myosin sets up the binding
00:09:44.09	of a distinct conformational state
00:09:46.29	to the next.
00:09:49.03	It can't...
00:09:51.11	this exchange between the bound and the unbound state
00:09:54.29	is prohibited, because if this happened
00:09:57.03	you would be wasting this ATP hydrolysis,
00:10:00.21	and you'd be back to where you started from.
00:10:04.01	But the reason this moves forward
00:10:06.20	is because this state
00:10:09.00	has a distinct specificity
00:10:12.23	than this state of the enzyme.
00:10:15.27	So the big changes
00:10:19.02	in the specificity of myosin
00:10:21.13	are what Jencks proposed
00:10:23.12	is the basis for this vectorial process.
00:10:28.02	We think about our challenge
00:10:32.06	very much analogous
00:10:34.17	to how Jencks proposed
00:10:37.13	how myosin works on actin.
00:10:39.29	And more specifically,
00:10:42.10	there are two postulates
00:10:44.20	that we draw upon
00:10:46.28	from Jencks' analogy
00:10:49.05	to actin and myosin
00:10:51.07	to explain how this vectorial process happens.
00:10:55.08	The first one
00:10:57.19	is that the specificity change
00:11:00.14	that Jencks talked about,
00:11:03.05	in our context,
00:11:05.23	happens at the level of protein-protein interactions.
00:11:09.26	I gave you an example
00:11:12.04	in my previous lecture
00:11:14.04	of an acyltransferase
00:11:16.21	that has protein-protein specificity
00:11:19.14	in addition to its specificity
00:11:21.24	for its normal substrate.
00:11:23.18	I'll be elaborating on this theme
00:11:25.19	much more in this lecture.
00:11:28.07	And the second hypothesis
00:11:31.09	that helps us think about assembly line
00:11:34.20	vectorial biochemistry
00:11:37.03	is that this core of the protein
00:11:39.22	that exists in each module,
00:11:41.20	this KS and AT,
00:11:44.26	has specificity for the acyl carrier protein,
00:11:50.15	in an interesting way,
00:11:52.20	that leads to establishing directionality.
00:11:55.26	Bear with me if that's not clear.
00:11:58.10	Let me first start by giving you
00:12:00.14	some information, some background
00:12:02.18	on the relevance of protein-protein interactions.
00:12:05.17	So, about 15 years ago,
00:12:09.07	when Rajesh Gokhale was in my lab,
00:12:12.08	he came up with an interesting postulate
00:12:15.27	about how assembly line directionality happens.
00:12:21.02	He hypothesized
00:12:22.27	that specificity that leads to this chain
00:12:26.07	moving from this module to this module
00:12:29.08	is not in the active sites
00:12:32.12	for different substrates,
00:12:35.04	but lies at the interface
00:12:37.18	between these modules.
00:12:39.22	So you have these docking sites
00:12:43.08	that exist between modules,
00:12:46.02	that are orthogonal
00:12:48.13	to the core modules
00:12:50.22	that are doing the actual enzymology,
00:12:53.19	that allow two modules
00:12:56.05	to come together
00:12:58.04	so that you have this directionality
00:13:01.05	that is established.
00:13:03.03	How do we know that?
00:13:04.28	We know that through experiments like this.
00:13:07.19	This is one example of an experiment
00:13:09.14	that Stuart Tsuji did in my lab
00:13:11.26	some years ago
00:13:14.08	where he kinetically isolated,
00:13:16.25	from the entire DEBS assembly line
00:13:19.28	you see over here,
00:13:22.03	these two modules,
00:13:24.00	Module 2 and Module 3.
00:13:25.21	He pulled them out of the overall system,
00:13:28.09	put them into a test tube,
00:13:30.09	and reconstituted their full enzymology
00:13:33.16	that's shown in this scheme,
00:13:36.06	measuring the appropriate parameters
00:13:38.21	that one gets from these kinds of experiments.
00:13:43.05	Now, what he does is
00:13:45.19	he switches these short peptide linkers
00:13:48.11	that flank these modules
00:13:50.25	for their counterparts from this module.
00:13:54.29	And if he switches just one of the two linkers,
00:13:58.25	what he gets are two modules
00:14:01.18	that by themselves
00:14:03.25	are just as good of catalysts as they were here,
00:14:08.02	but when you put them together,
00:14:10.04	they are simply unable to talk to each other.
00:14:15.07	If, however,
00:14:16.26	you take this module that has this linker
00:14:19.08	instead of this linker,
00:14:21.25	and couple it to this module
00:14:26.23	that has this linker
00:14:28.29	instead of this linker,
00:14:31.12	this pair over here
00:14:33.18	is indistinguishable from this pair.
00:14:36.23	And so now you see
00:14:39.06	how these matching docking sites
00:14:43.22	provide for the ability for Module 2
00:14:47.02	to talk to Module 3,
00:14:49.20	as opposed to Module 5
00:14:51.23	or any of the other modules.
00:14:53.25	So you're seeing the establishing of directionality
00:14:57.14	off the back of protein-protein interactions.
00:15:02.27	Now, what the Gokhale/Tsuji experiments do
00:15:05.18	is they tell us why Module talks to Module 3.
00:15:09.21	What they don't tell us is,
00:15:12.01	why don't chains go back and forth?
00:15:15.00	It's analogous to the question
00:15:17.09	that Jencks was asking:
00:15:18.28	why is it that myosin moves
00:15:20.21	in only one direction on actin,
00:15:23.12	as opposed to just going back and forth,
00:15:25.19	back and forth?
00:15:27.20	That question has an exact analogy over here,
00:15:30.27	because as you'll remember from the first lecture
00:15:33.11	of my three lectures,
00:15:35.06	I introduced to you that chain translocation
00:15:38.09	-- when a chain moves from one module
00:15:40.02	to the next --
00:15:41.24	it does so by a thiol
00:15:43.27	to a thioester exchange reaction.
00:15:47.05	And that reaction
00:15:48.26	can just as easily run in the backward direction
00:15:51.03	as it can run in the forward direction.
00:15:53.23	And so what precludes chains
00:15:55.22	from just going back and forth
00:15:57.15	and messing up the chemistry
00:15:59.06	so you'll never get an antibiotic
00:16:00.24	out of these assembly lines?
00:16:02.22	What's written in this code
00:16:04.17	that allows that to happen?
00:16:09.05	So, the question is,
00:16:12.06	how does this event
00:16:14.25	lead to a movement here,
00:16:17.01	but when it comes time to move the chain forward,
00:16:20.14	it doesn't just go back
00:16:22.15	into the same module,
00:16:24.09	but rather goes to the next module.
00:16:29.12	Again, over here,
00:16:31.17	protein-protein interactions
00:16:33.19	play a very important role.
00:16:35.13	And in particular,
00:16:37.24	the interactions,
00:16:40.04	not just in these docking sites
00:16:42.00	that Gokhale and Tsuji identified,
00:16:43.29	but also between this protein
00:16:46.22	and this protein,
00:16:48.21	play a very critical role.
00:16:50.23	How do we know that?
00:16:52.17	Through experiments like this
00:16:54.11	that Nick Wu did,
00:16:56.06	where he isolated, kinetically,
00:16:58.02	this chain translocation event,
00:17:01.26	exposed this module over here
00:17:06.08	to the same substrate,
00:17:08.07	same docking partners,
00:17:10.15	same receiving module,
00:17:12.22	but now he changes
00:17:14.19	the identity of the donor carrier protein.
00:17:17.20	So all that's being changed over here
00:17:19.19	is the identity of this protein.
00:17:22.05	And what you see is
00:17:24.11	you have a big difference
00:17:26.19	in specificity constants
00:17:28.24	just based upon who this player is
00:17:32.19	in the overall chain translocation event,
00:17:35.11	suggesting that these protein-protein interactions
00:17:39.08	are very important.
00:17:41.17	We now know,
00:17:43.06	through work of Shiven Kapur
00:17:45.07	and others in my lab,
00:17:47.01	what the basis for this is.
00:17:49.11	So, it turns out that these carrier proteins,
00:17:51.29	which are helical proteins,
00:17:54.11	have different helices
00:17:56.10	-- different portions of this helical protein --
00:17:59.12	that dictate specificity at different events.
00:18:02.22	How do we know that?
00:18:05.02	So, you can take this red carrier protein
00:18:07.23	that talks to this red acceptor module,
00:18:12.06	replace it with the green carrier protein,
00:18:15.05	it doesn't do a good job of talking
00:18:17.14	to this acceptor module,
00:18:19.12	but now if you just put a portion of this helix
00:18:22.14	into this protein,
00:18:24.14	now this chimeric protein
00:18:26.19	is just as good,
00:18:28.12	or almost just as good,
00:18:30.06	as this red carrier protein
00:18:32.23	in delivering cargo
00:18:34.21	through this chain translocation event,
00:18:37.10	suggesting that it's this helix over here
00:18:41.02	that plays a very important role
00:18:43.08	in dictating the specificity
00:18:45.02	during chain translocation.
00:18:47.28	You can do exactly analogous experiments
00:18:50.05	in that other theoretical reaction,
00:18:53.22	the chain elongation reaction
00:18:55.23	where that carbon-carbon bond is formed
00:18:58.00	between the extender unit and the growing polyketide chain
00:19:01.23	to give you this newly elongated chain.
00:19:04.25	You can kinetically isolate this event,
00:19:07.10	you can use the same substrates,
00:19:10.20	same catalysts,
00:19:12.28	just change the identity of this carrier protein.
00:19:16.06	And what you see is,
00:19:17.28	depending on which module you use,
00:19:19.24	you can have big effects
00:19:21.23	on the specificity constant over here,
00:19:24.11	as Alice Chen showed in these experiments.
00:19:27.22	You can use the same kind of approaches
00:19:30.10	that Shiven Kapur used
00:19:32.05	to make chimeric carrier proteins
00:19:34.11	and show which portions
00:19:36.16	of these carrier proteins
00:19:38.12	play a critical role in dictating the specificity
00:19:42.22	that is measured in these parameters.
00:19:45.11	And so the take-home message
00:19:47.23	from experiments like this
00:19:51.03	is that even though chain elongation
00:19:55.07	and chain translocation
00:19:57.19	are sequential events
00:20:00.19	along the assembly line,
00:20:02.16	and even though all of those events
00:20:04.24	involve ketosynthase-acyl carrier protein interactions,
00:20:09.14	the specificity of those events
00:20:12.04	is being driven by different faces
00:20:14.23	at this carrier protein.
00:20:17.21	So this blue face of this carrier protein
00:20:20.26	is what's driving the translocation reaction,
00:20:24.17	whereas the yellow face of the carrier protein
00:20:27.19	is what's driving the elongation reaction.
00:20:31.20	So you have this Janus-like biochemistry
00:20:36.01	in these carrier proteins.
00:20:38.19	And the reason that allows you to do things
00:20:42.19	in a vectorial manner
00:20:45.13	is because now you can set up
00:20:48.18	so that a carrier protein
00:20:51.06	can only recognize the next module.
00:20:54.06	How do we know that?
00:20:56.01	We know that through experiments
00:20:57.23	where we engineer these carrier proteins
00:21:00.06	to be confused carrier proteins.
00:21:03.07	This is one example of that experiment,
00:21:05.16	where we engineer a module
00:21:07.29	that stutters.
00:21:10.09	So if you take a natural module,
00:21:12.12	like this one shown over here,
00:21:14.09	and its own carrier protein over here,
00:21:16.18	you give it a nucleophilic substrate
00:21:18.23	and electrophilic growing polyketide chain,
00:21:22.10	and you ask it to catalyze chain elongation,
00:21:25.00	it'll do so,
00:21:27.06	it'll do one round of chain elongation,
00:21:31.01	and it'll stop at this point.
00:21:34.07	And the reason it stops at this point
00:21:36.28	and doesn't go back
00:21:39.09	to the same carrier protein
00:21:41.08	is that this carrier protein
00:21:43.27	is a partner of this ketosynthase
00:21:46.17	in this chain elongation reaction
00:21:49.10	that forms this critical carbon-carbon bond,
00:21:52.22	but it is not a partner
00:21:55.16	for chain translocation.
00:21:57.16	Its chain translocation partner
00:21:59.13	is elsewhere on that assembly line.
00:22:02.14	And so the way you can test that
00:22:05.22	is you can take the recognition features
00:22:08.14	of the next carrier protein,
00:22:11.29	build them into this system
00:22:14.16	to generate a chimeric carrier protein
00:22:17.26	that can not only do a chain elongation event,
00:22:21.13	but this chimeric thing
00:22:23.12	can then donate it back
00:22:25.10	to this ketosynthase
00:22:27.19	and go through yet another round of chain elongation
00:22:30.11	to give you a longer module.
00:22:32.15	So here you've got a module
00:22:34.19	that is stuttering;
00:22:37.18	that's doing two rounds of elongation
00:22:40.07	instead of one and only one round,
00:22:42.10	because we've kind of scrambled up its recognition features.
00:22:47.02	And so what that allows me to do
00:22:49.10	is to leave you with a basic picture
00:22:52.17	of how you might wanna think about this assembly line.
00:22:56.12	So, I showed you this cartoon.
00:22:59.27	The question we're asking is,
00:23:01.29	how does the growing polyketide chain
00:23:05.10	pick one and only one path?
00:23:07.17	It's going through at least
00:23:09.17	300 Ångstroms of distance
00:23:11.19	as it travels through these modules
00:23:14.21	from one active site to another,
00:23:17.08	and comes out as a 6-Deoxyerythronolide B.
00:23:21.11	How does it pick one and only one path
00:23:24.08	to take through this entire assembly line?
00:23:27.27	And so, if we isolate just these two modules,
00:23:31.05	Modules 2 and 3,
00:23:33.24	I can show you what I think
00:23:36.05	might be a model for how this works.
00:23:39.18	So we start off
00:23:42.00	with the upstream module,
00:23:44.12	this pink module,
00:23:46.26	bound to its carrier protein
00:23:49.16	in the elongation state
00:23:51.19	-- what Jencks would call...
00:23:53.28	has specificity for the elongation state --
00:23:57.06	and that specificity
00:23:59.06	is derived from this yellow portion
00:24:01.07	of the carrier protein
00:24:03.11	that docks into this active site
00:24:05.14	and puts its substrate
00:24:07.17	where it belongs.
00:24:09.25	Meanwhile, the green carrier protein
00:24:11.24	of this green module
00:24:13.28	is just sitting in neutral mode.
00:24:16.25	So this carrier protein
00:24:19.02	then catalyzes chain elongation,
00:24:22.26	it yanks out that chain,
00:24:25.01	it can't return it back to the same site
00:24:27.05	because it doesn't have the appropriate recognition features,
00:24:30.27	instead its got recognition features, now,
00:24:34.00	using that panhandle linker,
00:24:36.06	that docking site that was identified as well as this helix
00:24:41.14	that Kapur identified.
00:24:44.02	It docks into this module,
00:24:46.11	but in a different orientation,
00:24:48.11	so now it can translocate
00:24:50.20	the chain to the next module.
00:24:54.14	When that happens,
00:24:56.13	that carrier protein goes back
00:24:58.13	into the neutral mode
00:25:00.15	and this carrier protein,
00:25:02.03	which is charged with the extender unit,
00:25:04.09	comes in to do the elongation reaction.
00:25:07.15	And so this three-stroke process
00:25:10.11	goes through again and again
00:25:12.19	to give you the final product,
00:25:15.19	to allow chains
00:25:17.13	to move forward through the assembly line.
00:25:21.10	And so I'll leave you with this picture.
00:25:24.19	I started with the picture to the left,
00:25:27.13	to motivate these remarkable antibiotic machines.
00:25:31.16	I've added one more picture
00:25:33.11	that I suspect most of you
00:25:34.26	who've had a class in biochemistry recognize.
00:25:38.10	It is the picture that we often use
00:25:40.18	to describe how proteins are made
00:25:43.06	on ribosomes.
00:25:47.24	Let me ask you the following question:
00:25:51.03	what do these pictures mean to you?
00:25:54.27	Chances are if I just flashed
00:25:57.11	both these pictures in front of you,
00:26:00.10	you would immediately recognize
00:26:02.11	what I'm saying.
00:26:04.27	If, on the other hand,
00:26:07.08	you had just arrived today from Mars,
00:26:10.08	and I showed you these pictures,
00:26:13.02	they would mean absolutely nothing to you.
00:26:16.29	So why do these pictures mean so much to you?
00:26:21.00	I would submit to you they mean a lot to you
00:26:24.03	because you have seen the movie in each case,
00:26:28.04	and that is where I will leave you
00:26:30.21	as the future of my field.
00:26:35.12	What I showed you in the past three slides
00:26:39.13	were essentially the first three frames
00:26:42.22	of a movie
00:26:45.15	that will be the DEBS movie
00:26:47.26	over the next decade.
00:26:50.09	There's no Hollywood,
00:26:51.26	no Disney as yet,
00:26:54.06	but it's backed up by
00:26:58.01	data based upon...
00:27:00.11	through a combination of biochemistry,
00:27:03.14	structural biology,
00:27:05.09	and similar types of chemical/biological approaches.
00:27:10.01	Our challenge for this field
00:27:11.27	is to take those static pictures
00:27:15.06	that I've been showing you so far
00:27:17.15	and convert them into a movie,
00:27:19.22	so once you've seen that movie,
00:27:21.29	you will recognize the engineering implications
00:27:25.01	of every state
00:27:27.07	in every one of those thousands of assembly lines
00:27:30.20	that nature has give you,
00:27:32.21	and be able to rationally think
00:27:34.23	about how to recombine them,
00:27:36.23	or how to extract interesting antibiotics
00:27:39.23	from the right kind of assembly line in nature.
00:27:43.06	I can't do that today,
00:27:46.17	nor can any of my coworkers.
00:27:49.24	But I hope this inspires you
00:27:52.16	to think about how that might be done in the future.
00:27:55.03	This brings me to the end of the lecture.
00:27:58.16	I'd like to thank you for your attention.
00:28:02.24	I'd like to thank all of my students
00:28:05.16	and postdoctoral fellows
00:28:07.21	who, over the years,
00:28:09.16	have worked with me on DEBS
00:28:11.26	and related polyketide assembly lines.
00:28:14.29	Their names show up to the left of the slide,
00:28:18.27	and I've been acknowledging their papers
00:28:21.18	as we've gone along through
00:28:23.20	these different modules.
00:28:26.27	And I'd like to also thank my collaborators:
00:28:30.20	David Cane at Brown University,
00:28:33.05	with whom I've had a long-standing collaboration
00:28:36.19	on the study of DEBS,
00:28:38.23	as well as my collaborators
00:28:41.00	at the synchrotron
00:28:42.23	-- SLAC synchrotron --
00:28:44.21	Matthews Irimpan,
00:28:46.02	Tsutomu Matsui,
00:28:47.25	and Thomas Weiss,
00:28:49.11	with whom we've had very productive collaborations
00:28:51.25	associated with the
00:28:54.05	crystallographic and SAXS characterization
00:28:56.29	of DEBS.
00:28:58.12	And last, but not least,
00:29:00.03	I'd like to thank the NIH
00:29:02.01	for having supported all our work on DEBS
00:29:05.13	over the past 15 years.
00:29:07.27	Thank you.