David Roos (U Penn) Part 2: The apicomplexan plastid

Uploaded by ibioseminars on 26.03.2010

Hello, my name is Davis Roos and I'm a Professor at the University of Pennsylvania in Philadelphia,
and in the second segment of this iBioLecture,
I'd like to talk to you about the discovery of the apicoplast and I think you'll see what I mean
by this somewhat whimsical title "Something old, something new, something borrowed,
and not blue, but something green." In the last segment we talked about
Apicomplexa parasites, a group of five thousand eukaryotic protozoa,
which are obligate intracellular parasites living inside the host cells
including humans and a wide variety of other organisms. These parasites include
the malaria parasites responsible for hundreds of millions of cases of disease globally
and on the order of two million deaths every year they include Toxoplasma,
a parasite that is even more widespread, infecting approximately a third of the world’s
population, normally without adverse effects, but with several significant exceptions
to that rule including the importance of Toxoplasma as an opportunistic pathogen
associated with AIDS and other immunosuppressive disorders, and particularly as a
congenital pathogen, the leading source of congenital neurological birth defects
in many parts of the world. This is a malaria parasite that we are looking at,
and we discussed the various organelles inside this parasite last time through.
We looked at the nucleus, the secretory organelles including the golgi apparatus here,
and the specialized organelles involved in host cell attachment and invasion.
We talked a great deal about the inner membrane complex, this double membrane
that is required for the assembly of daughter parasites inside the mother,
in a fascinating process known as schizogony. These are eukaryotic cells
and as such they harbor in addition to a nucleus, the mitochondrion, an endosymbiotic organelle
and we'll say a little bit more about endosymbiosis in general in a moment,
but the focus of this talk will actually be on the apicoplast, or Apicomplexa plastid,
an organelle with, I think you'll agree, a remarkable biological history, which lends
insights into the evolution of the eukaryotic cells in general and potential targets
for drug development. Indeed, this story got its start through stories
not on the cell biology of organelles, but on the mechanism of action of drugs
and the identification of candidate drug targets. It began a decade ago through
the work of graduate student Maria Fichera, who was interested in asking the following question:
Why do these drugs, drugs like chloramphenicol, clindamycin, azithromycin,
all well known as antibacterial antibiotics, all well known as effective antibiotics
because they inhibit protein synthesis on bacterial ribosomes, but not on human ribosomes.
Why is it that these compounds are effective against malaria parasites and Toxoplasma parasites?
Clindamycin is regularly used clinically to treat patients. Now the initial suggestion
was that perhaps there is something bacterial like about the synthesis of proteins
in Toxoplasma and malaria parasites, but Maria was quickly able to show that that is not the case.
Cytoplasmic protein synthesis doesn't appear to be bacteria-like,
it's certainly not sensitive to these antibiotics although it was difficult to look at
mitochondrial protein synthesis, certainly mitochondrial function
is unimpaired by these drugs. And yet, these compounds kill parasites and in a very peculiar way.
Let me describe this phenomenon which Maria defined as the delayed death phenotype.
Here's the way it works. We can treat parasites with up to
10,000 times the lethal dose of drug, here using two different antibiotics,
one a protein synthesis inhibitor, one a fluoroquinolone
that blocks the replication of bacterial DNA,
and they grow like there is no tomorrow, through 48 hours,
or 6-8 cell cycles in tens of thousands of times the lethal dose of drug.
They escape from the host cell, survive extracellularly, and invade a new host cell without difficulty,
and there, inside that new host cell, they die or more precisely, they don't actually die.
These parasites grow, but they grow more slowly, and they grow more slowly
to an extent that is determined by the duration and concentration
of drug treatment that they saw way back 6 cell cycles earlier.
Now I must confess I can't give you an explanation for the reason for this,
but this very peculiar phenomenon is a characteristic of all of these compounds,
structurally unrelated compounds, that have in common
only the fact that the inhibit transpeptidation on bacterial ribosomes.
And so that made it seem very unlikely that these drugs would have unusual
different off-target activities in malaria parasites, suggesting that they must be
inhibiting protein synthesis, but not the protein synthesis that we knew of
and that drew our attention to a mysterious set of ribosomes that was identified
many years ago, many decades ago, in both Plasmodium and Toxoplasma parasites.
On an episomal DNA, a 35,000 nucleotide circular DNA originally thought to be
the parasites mitochondrial genome, but when the true mitochondrial genome
was discovered in malaria parasites and later in other Apicomplexa parasites as well
that left this 35 kb circle as a molecular biological mystery,
a mystery without a function, without a home, but the one thing that we knew about it
from work chiefly done by Ian Wilson's laboratory in the UK, was that these DNAs
contain ribosomal genes, and what’s more those ribosomal genes
had sequence characteristics that suggested that they might be susceptible to macrolide antibiotics
drugs like those that we've seen before. So to investigate this further we cloned
and sequenced the entire sequence from these parasite genomes
and taking advantage of predicted open reading frames,
particularly that for the elongation factor Tu gene, a very widely sampled gene
that's know from throughout life, we can begin to get
an idea of what this mysterious episomal DNA might be. And those studies
yielded several findings, the first of which was that all of the
Apicomplexa, Plasmodium, Toxoplasma, the poultry pathogen Eimeria are monophyletic,
they form a single group and that of course was no surprise it was no surprise at all.
The second was that among the bacterial world, the region from about here downwards,
the sequences look not at all like mitochondrial genomes indicated in yellow,
down at the bottom of your screen, and in fact they look more closely related
to cyanobacteria, blue-green algae, in fact the ancestor of chloroplasts in plants and algae,
and this wasn't much of a surprise either because we and others had previously noticed
that this episomal DNA bore many similarities to chloroplast DNA
and so we suspected that just as the ancestor of all plants and algae had acquired
an organelle from a blue-green algae, cyanobacterium, giving rise to modern day
chloroplasts, similarly these parasites might have acquired another cyanobacterial endosymbiont,
giving rise to a novel organelle. What was actually a much greater surprise though
was that these parasites look not just a little bit like the chloroplasts associated
with plants and algae, but a whole lot like them, as if we hadn't cloned a bit of
parasite DNA at all, but had inadvertently contaminated our cultures
with some pond scum from the pond outside of our laboratory. But no, we didn't
contaminate the cultures, this is indeed genuine parasite DNA
that looks for all the world like it came from a plant. So this raises a bit of a paradox,
because we know a great deal about the evolution of these parasites,
we know for sure that they are not plants and algae, they diverged off of the common
eukaryotic lineages shown to the left here, prior to the divergence of animals and fungi,
probably prior to the divergence of animals, fungi and plants, they are most closely related
to ciliates like paramecium and to dinoflagellates like the organisms that cause
red tide and poison the shellfish industry. And yet they harbor what appears to be a plant chloroplast.
There's really only one possible resolution to this paradox,
and that comes from considering the phenomenon of endosymbiosis.
It's now well-known universally accepted that virtually all, perhaps all eukaryotes
harbor a mitochondrion or at one point harbored a mitochondrion
subsequently lost, and that that mitochondrion was acquired
by a horizontal transfer event when the ancestor of all eukaryotes ate a bacterium,
an alphaproteobacterium in fact, giving rise to the ancestor of mitochondria,
surrounded by a double membrane and containing that mitochondrial genome
and similarly we know that the ancestor of all plants and algae acquired a cyanobacterium
by a similar sort of horizontal transfer event, an invasion of the ancestor
or an engulfment of the bacterium depending on your perspective
giving rise to the chloroplast, a distinctive organelle, again, surrounded by a
double membrane and harboring the DNA that was acquired from that cyanobacterial ancestor.
So either everything else we think we know about the origin of these parasites
is wrong, and in fact the Apicomplexa should branch off of the plant lineage
or instead maybe they just picked up a bit of plant lineage by horizontal transfer,
as indicated by this diagram. The process of secondary endosymbiosis
argues that an ancestral parasite ate a eukaryotic algae, which had previously acquired a
chloroplast from the engulfment of a cyanobacterium. These parasites have
maintained that plastid organelle despite having dispensed
with most of the functions we think of as associated with chloroplasts, photosynthesis for example
does not take place in the parasites we work with. Here's another cartoon
version of what you might think of old video game aficionados
as the Pacman model of organelle evolution in which an ancestral eukaryote
ate a unicellular algae giving rise to the chloroplast surrounded by a double membrane
and along comes the ancestor of these parasites which then engulfed
that eukaryotic plant or algae giving rise to modern day Apicomplexa
parasites harboring an endosymbiotic organelle, which we know is essential
as the target for these various drugs. Consistent with this model,
the organelle that we now know as the Apicomplexa plastid,
or apicoplast is surrounded by four membranes, which we can see in these Toxoplasma parasites,
organelles distinct from the golgi apparatus, the mitochondrion, the nucleus.
Now you might wonder how an organelle as striking as this could have been missed
by cell biologists for all these years, and the answer of course is that it wasn't missed
at all, this organelle had been seen many times and had been the subject of much debate,
but given a variety of uninformative names, the spherical body,
the golgi adjunct, or depending on your linguistic affinity
in France called the organelle plurimembranaire, or in Germany, the Hohlzylinder
but what we can now say is that the answer to this cell biological mystery,
what is this organelle, is it a distinctive organelle, is the same as the answer
to our molecular biology mystery, what is this episomal DNA,
and the answer to our pharmacological mystery, how is it that these drugs,
normally active only against bacterial species, are active against organisms such as Toxoplasma and plasmodium.
So the apicoplast or Apicomplexa plastid is a novel organelle acquired
by secondary endosymbiosis, harboring its own genome and essential for parasite survival,
quite an exciting find. It's not every day that we discover a new organelle,
but the big question of course is what does the apicoplast do? Now we've already
cloned its genome, sequenced it in its entirety, we know every gene associated
with that organelle or genome on the order of thirty protein coding genes,
another thirty RNA genes encoding ribosomal RNAs and transfer RNAs, but unfortunately
that organellar genome which was so informative for telling us about the origin
of the apicoplast, about the mechanism of action of drugs such as
fluoroquinolones and macrolides and rifampicins against these parasites
is completely uninformative in terms of what we presume must be
the metabolic functions that make it essential for parasite survival and that probably
shouldn't be much of a surprise because we know that all endosymbiotic organelles
chloroplasts of plants, mitochondria of eukaryotes in general,
encodes some proteins in their own genome but most of the proteins associated
with their function are actually have been transferred to the nuclear genome
and are translated on cytoplasmic ribosomes, and post-translationally
imported into the organelles. And so if we are going to gain some insight
into what the metabolic pathways might be and whether those perhaps could be targeted
as the target for antibiotic treatment, we are going to need to look into the
nuclear genome of these organisms themselves.
We can do so and we can find those genes, we can screen through available genome
sequences, at that time a fairly limited number of genes, we can identify
proteins associated with the ribosomes, a ribosomal protein S9 for example
is a protein that is essential for ribosomal protein function which is readily
identifiable as bacterial type, or eukaryotic type, and which in this case this particular gene
possesses a long amino terminal extension in the predicted gene, suggesting
that there might be targeting information responsible for translocation
across those multiple membranes of the apicoplast. And indeed, in work done
in my colleagues laboratory in Australia, Ross Waller, a graduate student
of Jeff McFadden in Melbourne, raised antibodies to these proteins and showed
on western blots that in fact they bind to proteins of a molecular weight consistent
with a processing of this large precursor into a smaller version,
which would be the mature ribosomal protein. And the same is true for other proteins
for Toxoplasma and Plasmodium. Now this gives us the tools that we need
to be able to explore not only what these proteins might be doing, certainly we
already knew that there were ribosomes associated with the apicoplast, it's not that surprising
to find ribosomes encoded in the nucleus and imported into the apicoplast,
but this gives us tools which we can use to explore how proteins might traffic
across those many membranes of the organelle. We know that those proteins
are in fact targeted to the organelle, and we know that from in situ hybridization experiments,
binding studies done on fixed samples in Dr. McFadden's laboratory,
where we can see in Toxoplasma and in Plasmodium in living parasites fused
to a fluorescent protein reporter that proteins are targeted to the apicoplast
at the apical end of the parasite nucleus. In Plasmodium we can watch the process
of replication as a ring stage parasite proceeds to this segregating schizont
that will then burst out as sixteen daughter merozoites via this
remarkable structure that mediates the segregation of organelles among the
various daughters, and an even more extreme example in work recently carried out
by Boris Striepen at the University of Georgia, we can see the segregation
of the apicoplast in another Apicomplexa parasite Sarcocystis neurona
an important pathogen of horses as it is segregated among the scores of daughters
that are developing in those parasites. We can map out the targeting signals
that are responsible for this in much greater detail by cut and paste molecular genetics.
When we look at that long N-terminal extension, suspected to mediate targeting
to the apicoplast, we can see that indeed it does if we take the apical extension from one protein
and cut off the mature part of the protein indicated in yellow here,
and fuse that to a fluorescent protein reporter, it goes into the apicoplast.
But the N-terminal signal itself doesn't look very much like a classical chloroplast
targeting signal, in fact its extreme N-terminus doesn't look anything like
a chloroplast targeting signal, it looks for all the world like a secretory signal sequence,
the kind of signal responsible for secreting pancreatic enzymes for example,
into the small intestine. Indeed, if we take that small N-terminal region,
the hydrophobic region indicated here in blue, and fuse it to a fluorescent protein reporter,
it behaves like a secretory signal sequence, it secretes a fluorescent protein
outside of the parasite, you can see four Toxoplasma parasites here, silhouetted in black.
The rest of this long N-terminal extension on its own mediates no targeting at all,
as we see in these parasites, nicely stained within the cytoplasm. But that region
can be replaced with a genuine chloroplast targeting signal from a pea plant
or from the experimental organism, Arabidopsis, indeed we can make a completely
synthetic nuclear encoded plastid protein by taking a pancreatic signal sequence,
fusing that to a chloroplast targeting domain from pea plants,
and fusing that to a fluorescent protein derived from a jellyfish and bingo,
it goes directly into the apicoplast in both Toxoplasma and Plasmodium
so faced with the daunting problem of how to target proteins into these organelles,
these parasites have evolved a remarkable mechanism. Fusing what we normally
think of as completely distinct targeting pathways, the blue bit,
the secretory signal sequence, with the green bit, a plastid targeting domain,
and we can map out the nature of those plastid targeting signals in great detail.
Experiments here carried out by post-doctoral fellow Omar Harb,
in which we can distinguish between proteins that target to the apicoplast,
and those that target to the membrane of the apicoplast as opposed to the lumen, shown here in red,
and we can follow the processing of those proteins on western blots as the targeting signal is cleaved off.
Those that have lost the ability to target apicoplast but which still harbor
secretory signal sequences, and those that have lost all targeting information as well.
We know the targeting to the chloroplast of plants and targeting to the apicoplast
of these organisms is quite complex and involves a variety of redundant signals,
for this particular targeting signal there are in fact one, two, three,
completely distinct plastid targeting signals, any one of which is sufficient
to mediate targeting to the organelle, so in quite an unusual story,
very different from what you may have learned in introductory cell biology,
we take two targeting signals, normally thought of as quite distinct,
the secretory signal sequence, which mediates co-translational
translocation across the endoplasmic reticulum where upon
signal peptidase processes off the signal sequence, the pink bit,
to expose a sub terminal domain, the yellow bit, normally thought of
as a post-translational targeting signal mediating translocation
across the membrane of the chloroplast, now this yellow bit is exposed
within the lumen of the endoplasmic reticulum and as that protein
winds its way through the secretory pathway, mediates translocation across
the remaining membranes into the apicoplast where upon a pitrilysin protease
cleaves that to expose the mature protein indicated in green.
We know that this process proceeds through the classical secretory pathway
by a number of studies carried out by Manami Nishi in this one experiment,
we use a fluorescence photobleaching study to follow the timing of targeting
to the apicoplast, in this case two parasites labeled in red
with a marker for the inner membrane complex, the apicoplast labeled
in green, we've photobleached one apicoplast just prior to parasite division
and you'll notice that it recovers through the production of new protein
very rapidly within a few minutes. We'll then take the same cells
and photobleach the other apicoplast here, and now this has been done
just at the time as parasite replication begins, subsequent to the division
of the golgi apparatus, here you can see the developing daughter parasites
within the mother as they replicate now and extend over a subsequent 80 minutes,
no recovery of protein trafficked into the apicoplast indeed, five hours later
we still see no protein targeted to the apicoplast and we won't until those parasites
start to divide again. Trafficking to the apicoplast occurs only over this very
narrow window of time indicated here by the white triangles as the apicoplast
is itself dividing, and as the golgi apparatus is dividing
and most active as well. We can see this process and the organelles
that are involved in an electron micrograph with distinctive large vesicles,
here the dark electron dense vesicles, distinct from the COPI and COPII vesicles
associated with trafficking to and from the golgi apparatus
and indeed in these larger vesicles we can stain protein destined for the apicoplast.
But I raise this problem not as a cell biological question of distinctive
targeting of the apicoplast, but as a problem of really
fundamental importance if we are interested in identifying targets for parasite survival.
Remember, this is an essential organelle, it’s likely to harbor a variety
of metabolic functions, what are those functions and can we target that
with new drugs. Now you might imagine a variety of strategies
that one could take to explore the functions of a novel organelle.
We know quite a lot about the function of chloroplasts for example,
and we know that through, for example, enzymological studies on purified
chloroplasts, but here we suffer from the difficulty in obtaining
large numbers of parasites. If we worked on the chloroplasts of pea plants,
or spinach plants, we could go down to the market or field and purchase
a truckload of spinach and isolate grams or kilograms of plant chloroplasts
for enzymological studies. Similarly, if we were interested in taking more
modern proteomics approaches, we would want to have large amounts of material,
and yet for these parasites, we can obtain at best a gram of parasites
and of course the plastid organelle is going to be only a small fraction of that gram.
We can imagine other approaches as well, we could devise genetic screens,
saturating the parasite genome with insertional transgenes and perhaps
selecting for genes that have acquired plastid targeting signal by virtue
of screens for targeting to the apicoplast, and indeed, we've pursued all
of those approaches, enzymological studies, proteomic studies, genetic studies,
in my laboratory and in other laboratories as well, and in sum total
we've learned a little bit about the function of the apicoplast,
we've identified some viable candidates, but the most successful approach
by far has been an informatics approach, and I'd like to try to describe
that to you just briefly. So this illustration indicates what you might imagine
as a conceptual schema, a strategy for identifying nuclear encoded
apicoplast proteins through genome database mining. We'll go to the
genome sequences that are available, any sequences, complete,
incomplete, genome sequences, EST sequences from any organism that has an apicoplast
and we will troll through those sequences asking a variety of
pretty simple minded questions. Asking for example, for any genes
that shows some level of similarity with proteins known to be associated
with plants, or with chloroplasts. Now there are going to be many many
false positives in this kind of search, after all, many proteins are associated
with plant chloroplasts that won't necessarily be associated with the apicoplast.
There will be false negatives there will be proteins that we simply
can't recognize because they're too divergent. Similarly, we can look
for proteins that have either one or the other of those distinctive targeting signals,
we can look for proteins with a secretory signal sequence for example.
But once again, we may not be able to identify all signal sequences
particularly in organisms for which the structure of the genes is not well characterized.
Moreover, by searching for signal sequences, we are undoubtedly
going to identify proteins destined for the endoplasmic reticulum
or the golgi apparatus, or the secretory organelles important for invasion
or the plasma membrane or the parasitophorous vacuole or the host cell beyond,
so false positives, false negatives, and similarly for all of the various
screens that we imagine carrying out. But all of the screens listed on this chart,
I would suggest harbor two features in common, and those that
they hold in common with all successful computational approaches.
The first is that they are hopelessly non-specific, but the second is that they
are all computationally tractable, and therefore easy to do, and if we do
a search like that we identify many proteins with secretory signal sequences
or similarity to plants or cyanobacteria or with N-terminal extensions,
but what we are looking for are those candidates that harbor two or ideally three
of those and in our first set of screens we identified several
hundred proteins that might be associated with the apicoplast.
A single approach, an informatics approach, undoubtedly many
false examples, but certainly hypotheses that we can test experimentally at the lab bench.
Here's what one such protein looks like, a protein with a N-terminal
hydrophobic signal sequence a subterminal region rich in charged
amino acids, one of the hallmarks of plastid targeting,
at least in Plasmodium parasites, although Toxoplasma parasites look
a little bit different, and most importantly in this case, a N-terminal region
that shows unequivocal similarity to ferredoxin the terminal electron acceptor
in photosynthesis and while these parasites are
not photosynthetically active, they've retained two proteins,
vestigal relics of photosynthesis in the ancestor of the apicoplast.
Here's the other such protein, ferredoxin NADP reductase
and in this case we've taken this protein, fused it to a
yellow fluorescent protein reporter, transfected it into Toxoplasma parasites
where we can readily carry out transient transfection studies in an overnight experiment
with proteins derived from either Plasmodium or from Toxoplasma,
demonstrate by co-localization with a validated apicoplast protein
that indeed this particular protein is targeted to the apicoplast.
The net result of these studies is what we believe to be a complete metabolic pathway
map for the apicoplast. We know that this organelle harbors its own DNA
and the machinery necessary to replicate it. It harbors its own transcriptional machinery
and those transcripts are translated into proteins using proteins encoded
on the apicoplast genome and additional proteins that are imported from the nucleus
And we now know through nuclear encoded proteins found to be associated
with the apicoplast, that a variety of other metabolic processes
take place within the four membranes of the apicoplast as well.
We know that this organelle is involved in lipid biosynthesis, using a fatty acyl synthase
that is significantly different from the fatty acyl synthase
of animal cells and humans in particular. The type II fatty acyl synthase
consists of multiple subunits which are assembled together in a process
that is the target of effective anti-microbial antibiotics
that might be candidates for treatment for Toxoplasmosis or for malaria.
We know that the apicoplast carries out isoprenoid biosynthesis,
the precursors to making cholesterol, although in this case not converting
into cholesterol isoprenoid units that are used certainly for modification of
transfer RNAs and possibly for other functions as well.
We know that the heme biosynthetic pathway involves not only the
cytoplasm and the mitochondria associated in the standard C-4 heme pathway
in for example, mammalian cells but also the apicoplast in an interesting process
that we don't understand yet in great detail. And we can dive down
into further detail to identify all the various components associated
with these pathways including several that are particularly attractive
as drug targets, some of which have compounds that are in clinical trials
as anti-malarials as we speak. But the point that I'd like to make today
is not just that malaria parasites stole a plastid from an ancestral plant,
and maintained that plasmid as a non-photosynthetic organelle
that's essential for parasite survival and therefore attractive as a drug target.
True though that may be, the real point that I wanted to make
and that I would like you to take from this segment of the iBioLecture
is that its these computational bioinformatics data mining approaches
that have been most successful and are really transforming the way we think
about doing biological experiments. In this area of parasitology
as in all areas of biomedical research. And in the next segment of this iBioLecture
we'll talk a little bit further about that, taking advantage of the genome sequences
now available for malaria parasites, as well as the mosquito vectors
and the human hosts and posing the challenge of whether we can design
and mine genome databases to take those genomes, identify the genes,
and develop diagnostics and therapeutics, drugs and vaccines
that might be effective against these organisms. And I look forward
to being able to tell you more about that in the next segment of this lecture series.