Virus Structures

Transcript of Part 1: Virus Structures: General Principles

00:00:02.08	Hello, I'm Stephen Harrison of Harvard Medical School,
00:00:05.25	Children's Hospital Boston, and the Howard Hughes
00:00:08.05	Medical Institute. This is the first of three lectures on virus
00:00:14.13	structures. This first lecture will be about general features
00:00:18.28	of the molecular organization of virus particles. The
00:00:22.15	second two will be about specific properties of virus
00:00:26.07	particles relevant to the molecular mechanism of infecting
00:00:30.22	a cell. Viruses are carriers of genetic information from one
00:00:37.04	cell to another; in that sense, they're effectively
00:00:41.16	extracellular organelles. The infectious virus particle,
00:00:46.18	sometimes called a "virion," is a molecular machine that
00:00:50.01	packages viral genomes, escapes from the infected cell,
00:00:54.22	survives transfer from one cell to another, and attaches,
00:00:59.01	penetrates, and initiates replication in the new host cell.
00:01:03.21	It's thus not just a passive package, but rather an active
00:01:12.15	payload deliverer. Now most people know viruses as
00:01:17.17	pathogens because the virus bears the genetic
00:01:21.12	information needed to usurp the cellular biosynthetic
00:01:25.24	machinery and replicate itself. The selective advantage
00:01:31.10	for evolution of the virus may be a selective disadvantage
00:01:34.21	for the host, and as a result, hosts evolve defense
00:01:39.25	mechanisms, the immune system in the case of humans
00:01:45.08	and other higher vertebrates. Now, viruses come in two
00:01:52.17	major flavors: enveloped viruses, in which the infectious
00:01:57.22	virus particle is surrounded by a lipid bilayer membrane
00:02:02.20	derived from host cell membrane; and non-enveloped
00:02:06.27	viruses, rather unimaginatively, that have no lipid bilayer
00:02:11.12	membrane, and the protective coat is just protein. These
00:02:15.01	two structural modes correspond to different modes of exit
00:02:21.09	and entry into cells, different mechanisms of assembly,
00:02:25.13	and different mechanisms of infection, as we'll see in the
00:02:31.18	course of this lecture and the next two. Now, just as
00:02:36.17	quick examples, on the left is an example of a non-
00:02:44.12	enveloped virus particle, a rotovirus particle. This image is
00:02:49.17	based on reconstructions from many electron
00:02:55.00	micrographs, and we'll go into some of the details of that
00:02:59.12	in the third lecture. And on the other side is an example of
00:03:05.25	an enveloped virus particle, also studied by electron
00:03:11.11	microscopy, and in the cross section of the image that
00:03:16.07	you see at the bottom, you can clearly see evidence of
00:03:19.22	the lipid bilayer with α-helical segments of the protein on
00:03:25.25	the outside traversing it. Just to remind you of sizes and
00:03:33.07	distances, both the rotovirus particle and the Sindbis virus
00:03:37.24	particle, as the right-hand one was called, have outer
00:03:42.23	shells that are about 700 angstroms in diameter, or 70
00:03:46.15	nanometers. That's about a millionth the size of a tennis
00:03:50.11	ball. Recall that chemical bonds that are between one
00:03:54.23	and two angstroms in length, and that's why chemists use
00:04:01.06	angstroms rather than nanometers; it's the natural unit of a
00:04:04.26	chemical bond. So when I say 700 angstroms, you can
00:04:08.04	think of that as 500-700 atoms across. Of course it's a
00:04:13.11	volume, and so the molecular mass of these particles is
00:04:17.08	some tens of millions of daltons. So bear in mind during
00:04:23.10	this lecture the following three questions. We'll talk about
00:04:28.15	more than just these three, but the main point of the
00:04:34.00	lecture will be to try to introduce you to the following
00:04:37.26	issues. First, why most non-enveloped viruses and a
00:04:41.29	number of smaller enveloped viruses have highly
00:04:45.05	symmetric structures. Second, what do the building
00:04:49.05	blocks of these particles look like? Turns out that the
00:04:52.14	same kinds of building blocks have been used over and
00:04:54.27	over again in the evolution of different viruses, even
00:04:58.02	viruses with very different replication strategies. And
00:05:01.08	finally, what the outer proteins of some enveloped viruses
00:05:04.27	look like. So let's begin with symmetry. What does
00:05:09.20	symmetry mean? Symmetry, as suggested by the image
00:05:14.00	on the left, means that there's some operation, in the case
00:05:18.12	of physical objects some physical operation like a rotation,
00:05:22.07	that brings the object into self-coincidence. In this case, if
00:05:26.16	you rotated this figure by 120° about the axis represented
00:05:31.29	by that triangle, and you closed your eyes while you did it,
00:05:36.02	you wouldn't realize that you had done the rotation. That's
00:05:39.22	called a threefold axis, and as you can imagine, symmetry
00:05:42.27	of more complicated objects can have other symmetry
00:05:46.14	axes, and so the icosahedron (we'll come back to that in
00:05:50.09	a minute) represented on the right has fivefold axes,
00:05:53.23	threefold axes, and twofold axes of symmetry. Some
00:05:59.06	viruses have helical symmetry. Helical symmetry is
00:06:05.11	represented by a screw axis, and so tobacco mosaic
00:06:09.12	virus, which was studied historically as one of the very
00:06:13.01	first viruses for which detailed biochemistry and detailed
00:06:17.13	structure became available, is a helical array in which the
00:06:22.23	nucleic acid, the RNA, is wound into a groove on the
00:06:27.17	protein subunit and winds up with the protein, which forms
00:06:35.22	this helical array. There are a number of other helical
00:06:39.23	organizations in virus particles: Vesicular stomatitis virus is
00:06:44.16	a much more complicated enveloped virus with an outer
00:06:47.18	glycoprotein, that's what this "G" is, on the right. But as
00:06:52.13	you can imagine, helical symmetry yields elongated
00:06:57.25	particles that get unwieldy. And so far more common is
00:07:04.01	the isometric, that is, roughly spherical characteristics, of
00:07:09.19	virus particles with icosahedral symmetry. The
00:07:13.06	icosahedron, one of the Platonic solids, the fanciest one,
00:07:16.28	so to speak, with 20 triangular faces, is simply a
00:07:20.18	representation of icosahedral symmetry. An object needn't
00:07:25.17	have icosahedral shape in order to have icosahedral
00:07:29.26	symmetry. And likewise, I could destroy the symmetry of
00:07:33.13	this object by painting an asymmetric object on each
00:07:37.17	face, rather than an object with threefold symmetry. The
00:07:44.09	icosahedral symmetry is represented or characterized by,
00:07:53.09	as I said, twofold axes, fivefold axes, and threefold axes.
00:07:58.27	And if you place a single asymmetric subunit into a space
00:08:07.24	governed by icosahedral symmetry and then operate on it
00:08:11.28	with the symmetry axes, you get 59 others, that is, there
00:08:16.21	are 60 locations in all that are related to each other by
00:08:23.27	these various symmetry axes, by these various symmetry
00:08:27.00	operations. And so a particle with icosahedral symmetry
00:08:33.27	will have 60 subunits. What I've flashed in here is a sort
00:08:39.24	of schematic representation of what might be a protein
00:08:43.22	subunit, to suggest that a small particle with 60 protein
00:08:49.28	subunits appropriately interfaced with each other can form
00:08:55.02	an icosahedral structure. Now this is indeed a schematic
00:09:00.21	representation of an actual virus particle. Parvovirus is
00:09:05.14	one of the very smallest and simplest of the viruses, have
00:09:10.04	a single kind of protein subunit that forms a small shell,
00:09:18.25	and 60 of them decorate or assemble into that shell, as
00:09:24.02	suggested here. And so, if we take a slightly closer look
00:09:30.14	at that protein subunit in a traditional ribbon diagram
00:09:35.19	representing the fold of the polypeptide chain, you see
00:09:39.26	that it's based on a quite simple, compact domain
00:09:44.21	represented here in red, with large loops emanating from
00:09:49.23	it. That compact domain has this sort of fold, it's called a
00:09:55.03	"jelly roll β-barrel," or sometimes a "cupin fold," and this
00:09:59.04	particular representation comes from the structure of
00:10:03.10	canine parvovirus, a virus of dogs, as you can imagine.
00:10:08.09	But the parvovirus family includes viruses, such as adeno-
00:10:14.26	associated viruses, now being used as vectors for efforts
00:10:21.04	at gene therapy. So in this case, the simple jelly roll β-
00:10:28.16	barrel structure has been elaborated with loops in order to
00:10:33.09	make a particle of adequate size, as shown here. And
00:10:39.00	that particle can package about a 5 kilobase single-
00:10:45.22	stranded DNA genome. Since the molecular mass of the
00:10:51.14	coat protein is about 50 kilodaltons, there's just enough
00:10:55.23	volume inside to package that genome, of which about a
00:11:00.19	third is actually given over to encoding for the coat
00:11:04.11	protein. So that's a relatively expensive way of spending
00:11:08.29	your genetic information. You've got to dedicate a full
00:11:12.04	one-third just to specifying the cardboard box, if you wish,
00:11:18.18	with which you're going to deliver the payload that you
00:11:22.24	actually wish to deliver. FedEx wouldn't like a system in
00:11:27.26	which a third of the weight were in the box. So what
00:11:38.05	about trying to package larger genomes with larger
00:11:41.24	coding capacity? One example, although still by no
00:11:46.29	means as efficient as the viruses we'll start to talk about,
00:11:55.03	are the so-called picornaviruses. These are small positive-
00:11:59.04	strand RNA viruses, of which poliovirus and the human
00:12:03.10	common cold virus (human rhinoviruses) are good
00:12:06.27	examples. In this case, there are three different protein
00:12:11.26	subunits, each with one of these β jelly roll designs, very,
00:12:17.11	very similar to that red β jelly roll in the canine parvovirus,
00:12:22.21	that assemble as shown into an icosahedral structure.
00:12:27.06	And so one-sixtieth of this structure has three jelly rolls, a
00:12:32.06	red one, a blue one, and a green one, designated in that
00:12:37.10	order, VP3, VP1, and VP2, for the colors as I named
00:12:42.03	them, forming the sort of assembly that you see here.
00:12:48.03	Now, those three subunits, as I said, look strikingly like
00:12:54.22	that same β jelly roll we saw in the parvovirus subunit, but
00:12:59.29	the loops are a little less extensive, because in this case,
00:13:03.05	with three subunits, the size of the particle doesn't need
00:13:09.22	to be additionally augmented by taking up space with
00:13:13.20	those loops, and one can still package adequate
00:13:16.14	amounts of RNA. One other feature of the architecture of
00:13:21.13	this particle that's noteworthy, and we're going to see in
00:13:27.01	various forms as we look at even more complicated
00:13:30.24	viruses particles, is the nature of the interaction among
00:13:38.19	the subunits, which not only involves interfaces between
00:13:46.09	pre-folded, rigid domains of subunits, but an elaborate
00:13:53.07	inner scaffold and a little bit of an outer scaffold made by
00:13:58.17	parts of this protein's subunit that fold up only when then
00:14:04.17	particle assembles. And so on the right, you can see
00:14:06.29	some hint of this in a blow-up of VP1, VP2, and VP3,
00:14:13.02	where you can see that, in addition to the jelly roll
00:14:18.15	domains, there are extended arms, they happen to be N-
00:14:24.06	terminal and extend inward in the particle, that fold
00:14:28.10	together when the particle assembles. Now these viruses
00:14:34.15	manage to package a nine kilobase single-stranded RNA
00:14:38.07	genome, but still use about a third of the genome to
00:14:42.16	encode the coat. The same size of package can be
00:14:50.23	achieved with a single kind of subunit, if that subunit can
00:14:55.13	have multiple conformers. Why does it need to have
00:14:58.22	multiple conformers? I told you that icosahedral symmetry
00:15:05.25	requires that there be 60 and only 60 identical structures
00:15:15.11	that form an icosahedrally symmetric shell, so that if you
00:15:19.21	want to use 180 protein subunits, as in the
00:15:23.14	picornaviruses, either they have to come in three chemically distinct
00:15:29.07	kinds, three colors, if you wish, or they need to have three
00:15:34.18	distinct conformers. In this example, from a simple plant
00:15:39.08	virus called tomato bushy stunt virus, shows that indeed
00:15:44.29	one can make a very similar package with the jelly roll β-
00:15:49.17	barrels packed essentially in the same orientation and the
00:15:54.13	same packing style, so to speak, as in the picornaviruses,
00:16:01.11	but where there's only one kind of subunit, and blue, red,
00:16:05.20	and green correspond to three different conformations of
00:16:09.21	that subunit. Those conformers can be achieved by
00:16:14.11	alternate hinges between rigid domains, as shown here,
00:16:19.18	between the two different major conformations. The red
00:16:25.10	and the blue in the previous slide are actually extremely
00:16:29.10	similar to each other and would be represented by what
00:16:33.08	you see here on the right. And a second conformation not
00:16:38.18	only with a somewhat different hinge but also with an N-
00:16:43.00	terminal arm folded up in an ordered way, whereas it's
00:16:47.04	disordered and hangs into the center of the virus particle
00:16:50.15	on the other conformation. So then again one sees here
00:16:56.29	that there is an elaborate inner scaffold that dictates the
00:17:01.27	assembly formed by parts of the protein subunit that are
00:17:08.17	not rigidly folded, are not ordered, until the assembly
00:17:13.24	comes together. We can have a look at that scaffold, it's
00:17:18.00	instructive, in this blow-up of the particle by just focusing
00:17:23.24	on those 60 of the 180 subunits that have an ordered
00:17:28.11	arm. And if we now focus in on that array of protein
00:17:36.10	subunits, 60 of the 180, and look over here at where
00:17:41.23	three of them interact at a threefold axis of the
00:17:45.27	icosahedral symmetry, you see that there's an inner
00:17:49.10	scaffold formed by the N-terminal arms of the protein
00:17:52.22	subunit that dictates the size and characteristic of the
00:17:59.25	whole assembly. These arm-like extensions, which fold
00:18:05.02	together to form an inner scaffold, also form flexible links
00:18:08.19	to the RNA. This is a good example of the undemanding
00:18:14.07	packaging of a genome, as I like to call it. If the package
00:18:20.04	required either a specific nucleotide sequence for a lot of
00:18:25.14	the RNA, or a defined structure in three dimensions, let's
00:18:30.22	say, for the RNA, then the RNA could not evolve to
00:18:34.19	encode the other functions that matter: an RNA-
00:18:39.21	dependent RNA polymerase, for example. And so
00:18:44.03	packaging of nucleic acids in viruses like these involve
00:18:49.25	both a short packaging sequence or packaging signal
00:18:57.20	that is recognized by a few copies of the protein subunit
00:19:03.14	and that can act as an assembly origin; and then a large
00:19:07.21	of nonspecific, charge-neutralizing interactions to
00:19:11.07	condense the RNA into the center of the particle. And so
00:19:15.19	in the case of tomato bushy stunt virus, at the tip of the
00:19:19.00	arm is a very positively charged polypeptide segment that
00:19:25.14	condenses the RNA and neutralizes the strong negative
00:19:30.02	charge on the phosphates. The specific recognition
00:19:36.15	interactions in the case of bushy stunt virus, we don't
00:19:40.05	have a picture of, but we do have a picture of how that
00:19:44.05	same part of the arm recognizes a specific packaging
00:19:49.24	sequence in the case of a related plant virus called alfalfa
00:19:54.03	mosaic virus. And in that case, a positively charged, 25-
00:19:59.13	residue-or-so, N-terminal segment that is not well ordered
00:20:07.15	on the protein subunit as it folds on its own, co-folds with
00:20:14.05	a short packaging sequence represented here by a
00:20:20.07	standard two-dimensional sequence representation of two
00:20:24.02	stem-loops. And the three-dimensional structure
00:20:28.28	represented here leads to a specific recognition, because
00:20:34.28	the stem-loops form a defined three-dimensional
00:20:38.29	interaction stabilized by their folding together with the N-
00:20:48.04	terminal arm of a small number of subunits. Probably one
00:20:52.22	dimer is responsible for recognizing this pair of stem-loops,
00:20:57.18	and the full packaging sequence might have three such
00:21:00.26	pairs and three dimers, but the protein shell is composed
00:21:07.12	of a much larger number, and the all remaining protein
00:21:09.28	subunits will have nonspecific, positively charged,
00:21:16.12	charge-neutralizing interactions with the RNA. Now, let's
00:21:22.14	go on and talk about still larger and more complicated
00:21:25.11	virus particles. Here's a representation, a surface
00:21:28.20	representation, of a papillomavirus. Papillomaviruses
00:21:34.09	cause warts and, in some cases, cancer in humans and
00:21:41.12	many other animals. The recently introduced vaccine
00:21:46.22	against human papillomavirus 16, 18, and one or two
00:21:50.28	other types, is a vaccine that prevents transmission of the
00:21:56.03	virus, which causes cervical cancer. So this surface
00:22:01.05	representation shows you that these viruses, which
00:22:05.00	package a double-strand DNA genome, are based on an
00:22:13.07	assembly of pentameric building blocks. In this case, the
00:22:20.11	pentameric building blocks are positioned not only at
00:22:26.22	positions of fivefold symmetry in this icosahedral shell, but
00:22:31.29	also at a general nonsymmetrical position so that this
00:22:37.09	pentamer is actually surrounded by six other pentamers. A
00:22:42.02	fivefold peg in a sixfold hole, so to speak. This sort of
00:22:48.24	assembly can nonetheless be stabilized by the same sorts
00:22:53.09	of principles that we've seen in the simpler viruses,
00:22:57.04	namely, the tying together of rigid or relatively rigid building
00:23:02.20	blocks by flexible, and hence potentially multidirectional,
00:23:11.00	arms. So here the pentameric assembly of the protein L1
00:23:17.22	that forms this structure is represented here, and as you
00:23:22.27	see, there are loops coming out of it with dotted lines
00:23:26.22	here, that form the interactions between the pentamers
00:23:33.08	shown here. And of course, since this is a fivefold peg in
00:23:36.19	a sixfold hole, its arms have to be directed in different
00:23:41.18	ways, but the pentamer itself is a rigid, fivefold-symmetric
00:23:47.01	object, just like its chemically identical mate here on a
00:23:53.20	fivefold position. Now, this subunit, the L1 subunit, is also
00:24:01.27	based on the same sort of β jelly roll building block that
00:24:06.15	we saw in the positive strand RNA viruses that we were
00:24:11.05	just talking about, and it's elaborated by various loops that
00:24:17.19	vary from virus type to virus type, just one of the reasons
00:24:24.08	that these viruses come in a great variety of serotypes, of
00:24:31.25	immunologically distinct types, because these loops,
00:24:36.19	which are on the outside of the virus particle (that is, this
00:24:39.24	is the part of the pentamer that faces outward, and this is
00:24:46.11	the part that would face inward, this would be the inside
00:24:49.13	of the virus, this the outside of the virus)... these loops are
00:24:53.25	free to vary evolutionarily because they're not so critical
00:24:59.07	for the formation of the stable assembly or for forming the
00:25:05.04	rigid pentamer, and hence can respond, if you wish, to
00:25:10.05	the pressures of their coevolution with the human immune
00:25:16.17	response, or the immune response of the particular animal
00:25:19.28	that they infect. Now, a similar principle, if you wish,
00:25:29.14	namely, reuse of the same kind of building block but in
00:25:35.01	environments that don't have a simple symmetry, is
00:25:43.02	exemplified by the adenoviruses. These are even much
00:25:46.27	larger structures, and I will try to make a few points by
00:25:53.06	talking about the adenovirus structure. The particle has a
00:25:58.28	strikingly icosahedral shape with fibers coming out of the
00:26:04.13	fivefold positions that are responsible for cell attachment.
00:26:08.20	The main part of the coat is represented by a protein
00:26:12.29	called "hexon" because it forms these sorts of
00:26:16.14	hexagonally packed arrays, but in fact the hexon is not a
00:26:19.20	hexamer, it's a trimer. It's a trimer, however, with two of
00:26:24.05	these β jelly roll domains, rather similar in their overall
00:26:28.00	shape, next to each other, so it has a kind of hexagonal
00:26:31.14	outline. As a result, the face of the icosahedron does
00:26:39.13	have threefold symmetry, and the whole structure has
00:26:42.20	threefold symmetry, but the hexon itself actually is only a
00:26:49.15	threefold, and not a sixfold, symmetric entity. Now one
00:26:54.25	quite interesting aspect of the structure here is that there
00:27:01.26	is a bacteriophage called PRD1 (and indeed several
00:27:05.13	other bacteriophages now known) that has essentially
00:27:09.19	exactly the same design. Adenoviruses are viruses of
00:27:13.11	humans and vertebrates and actually a large number of
00:27:17.06	other animal species, so with this structure, one can make
00:27:25.08	the point that even viruses of bacteria have strong
00:27:32.23	resemblances in their design to those of humans and
00:27:41.20	plants, for that matter. Indeed you saw similarity between
00:27:45.19	the plant viruses, like tomato bushy stunt virus, and the
00:27:51.08	picornaviruses, such as polio and the human common
00:27:54.02	cold virus. This doesn't mean, in my own view, that these
00:27:59.10	viruses are so ancient, if you wish, in their design, in their
00:28:05.12	structure, that they antedated the divergence of bacteria
00:28:10.17	and animals, or animals and plants. Rather, we know that
00:28:15.11	viruses can jump species. They can jump from insects...
00:28:19.20	indeed, there are viruses that infect both insects and
00:28:22.28	people, and there viruses that infect both insects and
00:28:26.29	plants. And so the transfer of genetic information that I
00:28:31.17	alluded to at the very beginning of the talk, the notion that
00:28:35.06	a virus particle is package that gets genetic material from
00:28:41.09	one kind of cell to another, may well be true not just for
00:28:45.13	the cells within you or between you and another individual
00:28:50.23	of the same species, but across species. We know that
00:28:55.05	flu jumps from swine to people, as we all learned from the
00:28:59.09	2009 pandemic, or from birds to people. But also,
00:29:05.09	ultimately, through eons of time, from one kingdom to
00:29:10.24	another. At any rate, it does means that the structures
00:29:15.20	we're talking about show a striking similarity and a striking
00:29:19.11	unity, whatever the evolutionary details. In the case of the
00:29:24.04	adenoviruses, the subunit on the fivefold axis is a different
00:29:31.27	protein subunit from the hexon. It's got one β jelly roll
00:29:37.10	domain instead of two, so that again there's a kind of
00:29:42.21	duplication and elaboration as this structure develops into
00:29:48.29	a much larger shall to package, in this case, a 35
00:29:54.01	kilobase-pair, double-strand DNA genome, much, much
00:29:58.04	larger genome. And indeed there are viruses based on
00:30:02.09	very similar kinds of protein subunits, the same double jelly
00:30:07.08	roll structure with a separate, related but genetically and
00:30:13.02	chemically distinct, single jelly roll pentamer on the fivefold
00:30:17.20	axes. There are even much larger viruses based on this
00:30:23.25	kind of subunit. Now, an interesting point in relating the
00:30:31.02	adenovirus structure to the bacteriophage that I
00:30:36.21	mentioned is based on or has a similar kind of major outer
00:30:44.00	shell subunit, are the mechanisms by which the virus
00:30:53.16	forms a defined and specific structure. As you can
00:30:56.27	imagine, in this sort of structure, how in the course of
00:31:00.26	assembly is the relationship between one fivefold position
00:31:06.03	and another fivefold position determined? How is the size
00:31:09.22	of this structure determined, rather than allowing, let us
00:31:14.19	say, multiple hexons to start forming much bigger and
00:31:18.11	bigger triangles? In the case of the phage, the answer is
00:31:26.15	particularly simple. One stripped off the outer shell of
00:31:31.15	hexon-like subunits and penton-like subunits, and
00:31:34.27	discovered that, from the x-ray crystal structure of this
00:31:38.22	particle, that there is an extended protein called a "tape
00:31:44.11	measure" protein by the investigators who discovered
00:31:47.14	this, that in effect stretches from a fivefold position here to
00:31:54.17	a twofold position here and then meets another one,
00:31:57.28	twofold symmetric to the next fivefold, and that
00:32:01.15	organization, think of the scaffolds that we talked about
00:32:06.28	before, governs the fixed size of the particle. In this case,
00:32:12.27	the scaffold protein is not an arm of the same protein
00:32:17.15	subunit, it's a separate protein, but the same principle
00:32:21.25	applies, and likewise, in the adenovirus particle, there are
00:32:26.12	several different so-called "glue" or "cement" proteins
00:32:30.19	that form, in effect, a scaffold that knits together the
00:32:34.18	structure in a way that leaves no ambiguity for the size
00:32:44.07	and characteristic of the final particle. In all of these
00:32:50.24	structures, the papillomaviruses, the adenoviruses, the
00:32:55.15	picornaviruses, the plant viruses such as tomato bushy
00:33:02.21	stunt, we see a simple construction principle at work that
00:33:09.16	is a little bit like an assembly line, like a factory assembly
00:33:13.23	line. There is in all cases a fixed assembly unit, happens
00:33:19.03	to be a dimer in the case of the coat protein of TBSV.
00:33:24.05	You saw that it was a pentamer in the case of the L1
00:33:26.21	protein of the papillomaviruses. The same of the
00:33:30.14	polyomaviruses like SV40. And you saw that the
00:33:35.04	adenovirus hexon, the trimeric adenovirus hexon, is
00:33:38.24	likewise a mass-produced assembly unit. But in order to
00:33:44.18	determine how that mass-produced assembly unit fits into
00:33:49.16	a defined structure of larger size, how the positioning of
00:33:57.18	that subunit doesn't simply lead to errors in the building of
00:34:03.09	a larger or smaller particle, there's a framework or scaffold
00:34:07.07	just as in the construction of a building, let's say, that
00:34:10.10	ensures accurate placement of these mass-produced
00:34:15.15	assembly units. And we've also seen that, interestingly
00:34:19.07	enough, there's a recurring architectural motif that has
00:34:23.17	appeared in the evolution of these structures (and it's a
00:34:26.09	complicated one, so it probably evolved only once) over
00:34:30.21	and over again. Now you might well ask, is this the only
00:34:34.22	architectural motif? Why are all viruses based on a so-
00:34:41.12	similar building block, and the answer is, that isn't the
00:34:44.00	case. There's at least one other, and that sometimes is
00:34:47.09	called the "HK97" fold, after the bacteriophage HK97 in
00:34:55.05	which it was discovered. You can see that this protein
00:34:57.13	subunit looks quite different, it's got some α-helices, it's a
00:35:00.20	somewhat irregular-looking structure, and it's found in the
00:35:07.12	bacteriophage P22 and a large number of other double-
00:35:11.01	strand DNA bacteriophage, where it forms a shell with a
00:35:22.10	number of these subunits forming both hexamers and
00:35:25.15	pentamers, so that there are 60 hexamers and 12
00:35:30.14	pentamers (there are always 12 pentamers in any
00:35:33.12	icosahedral structure), as suggested here. These viruses
00:35:39.03	assemble with an inner scaffold, but the scaffold in this
00:35:43.16	case is discarded by proteolytic digestion in some cases.
00:35:48.24	In this case, it's actually reused; it exits from the particle
00:35:53.08	and gets reused in the case of P22. And the particle then
00:36:00.24	changes some details of its organization as the scaffold
00:36:04.27	exits, as part of the process by which the double-strand
00:36:09.18	DNA is injected, actually pumped, if you wish, into the
00:36:14.20	particle at the next stage in assembly. So these are cases
00:36:19.02	in which the shell preassembles around a scaffold. The
00:36:23.20	scaffold is ejected, either chewed up or literally ejected
00:36:31.17	and reused, and a series of events involving motor
00:36:36.10	proteins are responsible for inserting DNA into these
00:36:42.22	structures. Now you could ask whether this is true only of
00:36:48.16	bacteriophage, answer: "no." You might anticipate that
00:36:52.09	the answer would be no from what I told you about
00:36:55.04	adenovirus and PRD1 for example. Here are two
00:37:01.07	bacteriophage protein subunits that have this sort of
00:37:05.04	structure, but the herpesviruses, of which the herpes
00:37:09.05	simplex 1, the cold sore virus, is one example, are based
00:37:15.24	on a much more elaborately looped, elaborately
00:37:21.13	decorated, version of the same fundamental fold. The
00:37:26.04	structure that we have at the moment is from electron
00:37:28.25	microscopy and not yet at the same resolution that the x-
00:37:33.04	ray structures of these subunits have yielded, but you can
00:37:37.29	probably see in this relatively low resolution representation
00:37:43.03	of the herpesvirus that this part, for example, there's a
00:37:47.20	long α-helix, corresponds to the much simpler,
00:37:53.03	undecorated fold you see here. And then these are loopy
00:37:56.04	structures that stick out and make the protein subunit
00:37:59.10	much larger and have to do with other interactions that
00:38:03.11	the protein subunit of the herpes particle makes. The
00:38:07.12	herpesvirus particle is more complicated, it's both larger
00:38:09.27	and more complicated than the phage particles, and so
00:38:13.25	there are other interactions of those surface loops that are
00:38:18.18	important. Herpesviruses like the phage have very tightly
00:38:25.20	coiled DNA that is inside, that's pumped into them in this
00:38:31.23	reconstruction from electron cryomicroscopy, you can
00:38:34.23	actually see the coiling of the DNA. The DNA is actually
00:38:39.20	coiled this way, that is, circumferentially about the axis of
00:38:45.00	the particle, it's injected through one vertex, and as you
00:38:48.09	see, there's a specialized internal structure here, to which
00:38:53.19	then the tail of the phage that ultimately injects it back
00:38:58.04	into a new host cell, is attached. The cross section here
00:39:04.21	looks as if you have circumferential layers of density in the
00:39:12.12	other direction because, as you can see from this
00:39:15.05	diagram, DNA coiled about a vertical axis, if the order is
00:39:25.23	such that, from particle to particle, there isn't exactly a
00:39:29.16	piece of DNA here, but this one might be here or here,
00:39:34.10	then on average, you will get radial shells of density as
00:39:40.07	you see here, fitting tightly into the interior of the particle, I
00:39:46.03	like to say, with a gardener's analogy (it might not be
00:39:50.10	relevant for all people listening), like winding a hose into a
00:39:54.06	hose pot, or rope into a bucket. Now finally, let's talk a
00:40:01.26	little bit about enveloped viruses. Enveloped viruses
00:40:08.25	acquire their envelope, in general, their membrane (this is
00:40:12.04	not true of all enveloped viruses, but true of almost all of
00:40:19.08	them), by budding out of the cell, either out of the cell
00:40:22.22	surface, or into an intracellular compartment such as the
00:40:29.03	endoplasmic reticulum or the Golgi apparatus, and then
00:40:33.16	being transported out. And in that budding process, wrap
00:40:38.19	themselves, if you wish, in a membrane that's derived from
00:40:42.23	host cell lipids, although host cell proteins are in general
00:40:47.21	excluded. Some of the smaller enveloped viruses have
00:40:53.21	icosahedral symmetry, and their structure and assembly is
00:40:58.15	determined by regular interactions within an icosahedral
00:41:03.08	shell, just as the ones you've seen in the non-enveloped
00:41:06.26	viruses. But larger and less regular enveloped viruses are
00:41:12.08	also seen, such as HIV or influenza, in which the protein
00:41:21.12	interactions are less perfect, but that doesn't matter for
00:41:26.11	protecting the nucleic acid nearly so much, because the
00:41:30.03	lipid bilayer in effect is an impermeable barrier against
00:41:35.07	agents that might get in and degrade or damage or
00:41:41.00	cleave the nucleic acid. So the budding process that I
00:41:48.06	mentioned can either involve, as in the case of the so-
00:41:52.02	called alphaviruses, or which Sindbis virus is one of the
00:41:56.18	prototypes and well studied... and a recent human
00:42:02.01	outbreak of an alphavirus is the chikungunya virus, which
00:42:07.03	had a major outbreak in the French island of Réunion,
00:42:12.22	and led to considerable interest and publicity about the
00:42:16.21	properties of that virus. The alphaviruses have a core that
00:42:22.25	preassembles in the cytoplasm and then two species of
00:42:28.05	glycoprotein that are synthesized on the rough ER,
00:42:33.29	exported to the cell surface, and then the particle buds
00:42:39.03	out through a process by which the inward-directed C-
00:42:47.01	terminal tips of the glycoprotein, which stick through the
00:42:52.09	membrane in a single α-helical segment (you might
00:42:55.04	remember very early on, I showed you a cross section
00:42:58.05	that showed that), interact one-to-one with the
00:43:05.04	icosahedrally symmetric core that's assembled rather like
00:43:09.21	a non-enveloped virus in the cytoplasm, and buds out. In
00:43:16.07	other cases, such as influenza, there's no preassembled
00:43:21.03	inner particle, but rather, the assembly occurs at the
00:43:24.20	membrane, as you see here, where the inner structures
00:43:28.18	and the glycoproteins that incorporate in the membrane
00:43:33.20	come together as part of the elaborate budding event.
00:43:38.17	Separate cellular machinery, in some cases, is then
00:43:42.05	needed to the finish the pinching off, whereas these
00:43:44.27	viruses don't seem to need a separate pinching off
00:43:47.23	mechanism. In the case of HIV, these micrographs show
00:43:54.15	particularly dramatic examples of HIV budding. It's
00:43:58.27	directed in this case by the interaction of the N-terminal
00:44:04.21	domain of the inner protein, the so-called "Gag" gene
00:44:09.10	product, and that protein has a myristoyl group at its N
00:44:17.06	terminus and a very positively charged surface, and
00:44:22.10	interacts with the membrane to drive budding, as shown
00:44:25.20	here. In these micrographs, you can see that the HIV
00:44:33.01	particle is rather sparsely decorated with an envelope
00:44:38.25	glycoprotein that has the function of attaching the virus
00:44:44.17	particle to a new host cell and mediating viral entry. In the
00:44:51.14	case of the smaller icosahedrally symmetric enveloped
00:44:57.11	viruses, like dengue virus for example, the outer coat is
00:45:03.29	much more tightly packed, it forms a very regular array, in
00:45:08.14	case with 180 subunits of the protein whose structure is
00:45:14.21	shown up here, forming a perfect icosahedral array, and it
00:45:18.26	is an assembly of that array that drives particle budding. In
00:45:27.23	all cases of enveloped viruses, the entry process (and
00:45:33.21	that will be the topic of the next part of this set of lectures)
00:45:39.28	involves fusion of the viral membrane with a membrane of
00:45:44.17	the host cell. So just as the assembly process, the
00:45:48.24	maturation process, the exit process, involved budding
00:45:53.19	out and pinching off, so entry involves the reverse
00:45:58.15	process: attachment and fusion of the two membranes.
00:46:03.17	We'll talk about fusion in great detail in the next part of
00:46:07.02	this series, but just to give you a hint of what's to come,
00:46:10.20	an important of all of these viral envelope proteins is that,
00:46:16.17	under suitable circumstances, they can be triggered to
00:46:20.24	undergo a major conformational rearrangement. It's that
00:46:25.26	rearrangement that drives the fusion event, so that in the
00:46:31.14	case of the dengue virus particle, there is a
00:46:33.24	rearrangement from the dimeric structure shown here, a rather plate-like
00:46:42.14	organization of two somewhat elongated protein subunits,
00:46:48.27	into a trimer in which hydrophobic residues at the tip of
00:46:54.05	one of the domains, this yellow domain, so-called "domain
00:46:57.05	II," cluster together at one end of the trimer and interact
00:47:02.18	with the target cell membrane in order to begin the
00:47:06.03	process by which the two membranes are brought
00:47:09.03	together. In the case of dengue virus, this conformational
00:47:13.10	change is triggered by proton-binding, a signal that the
00:47:17.26	virus has arrived in the low pH compartment of an
00:47:21.19	endosome. In other cases, other signals are read out, so
00:47:27.00	to speak, by the fusion mechanism. We can look at this in
00:47:36.00	one more slide, where the interaction with the target cell
00:47:40.13	membrane is shown, and there is a zipping-up process of
00:47:44.08	the C-terminal part of the subunit that actually is part of
00:47:49.28	the pinching together of the two membranes, and leading
00:47:54.10	to an elaborate bit of molecular machinery. Now not all
00:48:04.06	enveloped glycoproteins form such a regular array. In the
00:48:08.11	case of the influenza virus particle, the proteins on the
00:48:13.03	surface of the virus particle sticking out from the
00:48:17.15	membrane are rather spike-like. There are two of them, as
00:48:22.14	you probably know, or two species, the hemagglutinin
00:48:26.06	and neuraminidase, the "H" and "N" of H1N1 or H5N1,
00:48:32.18	that you read about when pandemics threaten. The
00:48:38.12	hemagglutinin is the protein that undergoes a low pH-
00:48:44.16	triggered conformational rearrangement to drive fusion.
00:48:50.05	We'll be hearing quite a lot about that in the next part.
00:48:54.17	The hemagglutinin shown here is a spike-like structure as
00:48:58.13	I mentioned, its molecular design doesn't look anything
00:49:02.10	like that of the envelope protein of dengue virus, it's a
00:49:08.14	stalk-like structure, and long α-helices project the
00:49:19.04	receptor-binding site at the top about 120 or 130
00:49:24.23	angstroms away from the membrane. We'll use that
00:49:30.13	structure to discuss fusion mechanisms in much more
00:49:33.22	detail in Part 2 of this series. See you then.
00:49:40.17