History of nutrition/diet dev in Drosophila/rel insects William Ja
WILLIAM JA: Thank you so much.
I’ve changed the title a little bit just
to focus on the organism I’ve worked on
for the last 10-15 years, Drosophila.
I’d come in with a very naive viewpoint.
I started off my academic life as a chemist
and have only used Drosophila for the last
10 or 15 years.
I’ve always thought very deeply about what
we feed them, what they eat, and so I hope
to share with you a very brief history of
diet development in Drosophila and then, selfishly,
a couple of case studies from my own lab that
will hopefully be informative for the goals
of this workshop.
I find it easy to think about the spectrum
of nutrition or diet development in flies
from both ends, nutrition and dietetics, and
these are very basic applied fields where
the field of nutrition is interested in understanding
all … the role of all the nutrients—all
the molecular players, right, the actual molecules
in the food—that influence the animal physiology
On the other end of the spectrum are dietetics.
You don’t care; you just want the best medium
… the best rearing medium possible that
can maintain the animal, maintain the flies,
for generations and generations.
And obviously for research, for lab medically
relevant research, we need to survive on both
ends of this spectrum; both ends need to be
For Drosophila, you have to maintain the entire
life cycle of flies, and for those who haven’t
seen it, you start off life as an egg.
Over the course of a week there’s a larval
period … a larval stage, the larvae pupae,
and after a total week from egg to pupae these
adults eclose or emerge, and the adults survive
for about a month or two.
They lay eggs to start the cycle over.
Something perhaps underappreciated is that
the medium … the rearing medium for these
flies actually has to be soft enough for the
larvae to eat, to develop, but hard enough
for the adults to stand on, because they will
drown in any swampy-type food; they’ll get
stuck and die.
And so, there’s this picture from about
100 years ago from one of the first fly labs
in Thomas Hunt Morgan’s Fly Room at Columbia
I like this picture because you can see the
glass milk bottles that they used to use to
house flies on the table, but I really like
this photo because on the top you can see
that giant bunch of bananas hanging from the
And it’s because the first medium used for
flies, for Drosophila, in the laboratory was
simply fermented fruit.
They took chunks of rotting banana and threw
it into the bottles.
Now, I’d like to say that over the last
100 years, every couple of decades or so,
the medium has been improved—advances to
the standard medium has been improved.
But there’s never been a single standard,
There’s never been a uniform diet used across
But the two major advances in the last 100
years, in the last century, have been the
addition of yeast because it was figured out
very early on that flies don’t east rotting
They eat the microbes that grow on top of
the rotting fruit, mostly yeast and bacteria.
And the other major advance was the addition
of agar as a gelling agent.
And so, like I said, there’s no single standard
today, but today’s diets include some combination
of yeast, corn meal or corn flour, some form
of sugar, and agar as that gelling agent.
This is a common bottle housing flies today.
This is about the size of a milk bottle made
from plastic now.
You can see pupae on the walls on the bottle
and larvae crawling around.
In this example, the medium is brown because
the sugar used in this one is molasses.
So, I’m going to go through a couple of
case studies from my own lab.
When I first started in Drosophila research,
my first experiment, my first project, was
to understand how dietary and caloric restriction
works in flies.
Caloric restriction was one of the … was
touted as one of the most conserved longevity
interventions that works across species, and
the idea is that you eat less, you live longer
and live healthier.
I don’t want to get into any of the science
I only want to talk about methodology.
If you’re talking about caloric restriction
in people, if you want to do this for yourself,
it’s fairly easy conceptually.
If you eat a lot and you want to be restricted,
you eat less.
The first experiments done in rodents in the
1930s, conceptually, again, the methodology
was pretty simple.
You have access to a lot of chow or limited
chow for the CR group.
With flies this is much more difficult.
Again, they live on an ocean of infinite jello,
right, agar-based food.
And so, how do you implement a restriction
experiment in flies?
And so about 30, 35 years ago, the first experiments
to do this were attempted, and it went by
total food dilution, and Dr. Nielsen—Frosty—already
suggested or implied a problem with diluting
or changing concentrations of nutrients.
In this example, I’ve simply shown 2x and
1x diets, so where all the ingredients in
the food were cut in half, alright?
So, on the left, 10 percent yeast, 10 percent
sugar in this example; on the right, 5 percent
yeast, 5 percent sugar.
And so, this paradigm was started 30, 35 years
ago, and it does work.
You can see an example survival curve; this
is over 30 days or so.
The red curve on the right is the one that’s
food, on the left in black is the 2x food,
and you can see that diluting the food does
extend fly life.
And if you plot this as an average lifespan
or mean lifespan, you can see that diluting
the food increased fly lifespan by about 20
percent in this case.
Okay, so that’s great.
A couple of decades of research have focused
on this paradigm.
The assumption has always been that if you
change the food concentration—you dilute
the food concentration—you’ve reduced
the calories or nutrients going in, and this
leads to lifespan extension.
Now without going into any of the details,
because all of this was published, one of
the first questions we asked, colleagues and
I, about 15 years ago now, was, “Why is
it that, if you cut the nutrients in half,
why don’t the flies just eat twice as much,
You’re not actually restricting the food;
you’re giving them ad lib access all the
And so, this took many years of developing
assays that would actually allow us to measure
how much flies eat, and when you do that,
indeed they eat about twice as much.
So, that suggests that the nutrients or calories
that flies are getting on these different
diets is actually equivalent.
Going back to the drawing board and considering
all the things that change in your experimental
design when you cut the nutrients in half
in the food from total food dilution, and
these are just a few that we thought of.
Food properties certainly change: food hardness,
food pH. buffer capacity is a really underappreciated
component of the diet, of the medium, and
all of these things seem to matter.
Prandial habits are increasingly recognized
as something really important for health.
Even on isocaloric diet … even on a … if
you get the same total number of calories
per day, if you take in those calories at
one giant meal versus lots of small meals,
that’s increasingly recognized to influence
Microbial growth is another obvious parameter,
These media, these fly media, are essentially
rich microbiological culture plates, so it
should be obvious that microbe growth could
be different when these nutrient concentrations
And finally, what I want to focus on next
is water intake.
Flies are one of the few animals in captivity
that I know of where ad libitum water access
is not provided.
Drinking water is not provided.
So, these animals are routinely not allowed
to independently regulate their food and water
intake, and the only reason for that is because
no one has ever provided them with drinking
water; that’s how it started 100 years ago.
If flies in the dietary restriction or caloric
restriction paradigm are eating different
volumes to match nutrients, we ask, since
they’re getting different volumes and different
amounts of water through the jello, through
the food, could water stress or dehydration
stress play any role in these lifespan phenotypes
that people were seeing?
And so, we redesigned this experiment simply
by adding some water, in the form of agar,
for the flies to drink.
And just to summarize these results, because
these were published almost 10 years ago now,
on the left again you see that with food dilution
you get almost a 20 percent increase in lifespan,
but when water is present these lifespan differences
This is true in multiple trials of both sexes
and different genotypes of control flies.
Again, just to summarize, this first case
study is that food dilution does lead to longevity,
but the assumption has always been that this
is through changes or reduction in calories
or nutrients going in.
But through careful measurements of feeding
behavior, careful measurements of actual consumption,
you find that flies are capable of regulating
how much they eat and the actual number of
nutrients, the amounts of nutrients going
in, is about equivalent, and it’s the volume,
or more specifically, the water intake that
And so, I think we’ve essentially ruled
out that calories played a role in those early
Now, I take no credit for this, but the Drosophila
community has quietly shifted away from total
food dilution as a paradigm for implementing
caloric or dietary restriction and really
focused, in the last 10 years especially,
on yeast restriction, which changes the protein-carb
ratio of the diet, and that seems to be a
consistent way to manipulate or regulate mice
What I want to emphasize here is that these
problems in methodology aren’t specific
to the fly model, right?
In mice, food restriction has been used but
also intermittent fasting.
This is where you allow mice access or rodents
access to food for only a certain number of
hours per day, and in some cases, they actually
overeat enough so that they’re equivalent
to control animals, right, in terms of total
number of calories.
Yet there’s still a difference in metabolic
health and physiology, but these all fall
under the umbrella of dietary or caloric restriction.
In C. elegans it’s even arguably more complex.
There are over half-a-dozen methods used,
published to implement caloric restriction
in C. elegans and worms.
And Anne Brunet from Stanford published this
really elegant study of aging cells almost
10 years ago showing that testing half a dozen
different methods of caloric restriction,
or published methods, for caloric restriction
of worms actually all have different—sometimes
overlapping, sometimes not—genetic dependencies;
a different genetic mechanisms underlier.
So, which ones of these methods are the most
relevant to mammalian research?
I can’t answer that question, but I think
what’s important, anecdotally … results
are more comparable across species when we
are very careful in thinking about the methods
that we use, and we find that if we perform
caloric restriction in an animal similarly
to another animal, the results are much easier
to compare and they’re much more consistent—internally
Instead, what happens with each model organism
… we try to pick a method that works for
us, that’s easiest to implement in the lab,
with really very little consideration of how
these methods compare across species.
Okay, so what about that other laundry list
of things that change when you change food
My lab … this is just a small sampling,
but my lab over the last 5, 10 years has started
publishing this series of papers looking at
each of these parameters.
It turns out that flies do like acidic food
over not-acidic food.
High pH food changes food palatability, and
this can affect lifespan.
Mifepristone, RU486, is a common drug additive
for flies used for an inducible gene expression
It turns out this drug tastes bad to them.
It’s bitter to them, and it can change food
palatability, depending on the concentration
of the drug and the base diet, and this is
also something not usually considered in the
design of these experiments.
In fact, perhaps our most impactful work is
the idea that microbes promote added nutrition
for flies, especially protein-based nutrition.
Now, why is that the most impactful on this
list when I just told you that 100 years ago
they argued that microbes served as a food
source for flies?
Now, the reason is because of how hot the
microbiome has been for research—mammalian
research—in the last … especially in the
last 15, 20 years, right?
And Drosophila have joined in this effort.
This is a tiny, tiny sampling of papers, but
many Drosophila labs are now looking at the
influence of microbes on all sorts of phenotypes.
And here you see lifespan, development … I
forget what else is on it … behavior, physiology,
and so forth.
This work is fine.
My problem with it is that many of these papers
actually immediately start pushing the idea
that this work is related to the gut microbiota.
They immediately say microbes affect this
phenotype and therefore the gut microbiota
of the fly is meeting this phenotype, the
fly is a good model for mammalian research,
please continue to fund our work.
And the problem is … the problem is I think
many of these papers haven’t really proven
a gut microbiome interaction in flies yet.
In the laboratory—and this is important
because in the last year we started figuring
out how to make a stable gut microbiota or
find a stable or establish a stable gut microbiota
in flies—but typical laboratory conditions
don’t provide a stable gut microbiota in
It’s a transient gut microbiota.
And at any given moment there’s only about
100 to 100,000 CFUs of bacteria in the fly
gut, and these are rapidly transmitted and
As I just said, flies are a microbivores.
But remember, they’re swimming, they’re
in an environment that is teeming with microbes.
In the laboratory, these are fly vitals, 10
to the 5th and 10 to the 7th on the lower
end per square centimeter of the fly food,
because that’s a rich culture you get.
And so, although all these … well, I shouldn’t
say all … many of these papers tout that
they’re looking at specialized post-microbe
They really ignore the fact that flies are
microbivores, and we all know that nutrition
influences virtually every phenotype we measure
in our animal systems.
This true in worms as well.
In C. elegans, we literally feed them bacteria
as their primary food source.
So, I think the idea of studying specialized
post-microbe interactions in these invertebrate
models, I believe they exist.
I believe these interactions exist.
But I think studying them is a lot harder
in these model systems.
And I’ll leave you with one last example
There have been a couple of high-profile papers
a while ago looking the Drosophila microbiome
influencing fly development.
And these microbes only promote fly development
on low-nutrient conditions, on under-nutrition
And so, like I mentioned, our work focused
on lifespan, not development, and I don’t
know if it’s hard to see, but this curve
in the middle, this orange curve, shows how
the lifespan changes with the protein content
of the diet.
On the high end, you sort of have an over-nutrition
As you reduce the protein concentration, you
have sub-optimal for lifespan, and as you
reduce the protein content of the diet, a
lot more the lifespan plummets because they’re
And so, we found from that paper that adding
microbes to the diet seems to improve impact
of flies most—of fly survival most—on
these under-nutrition diets, and the most
obvious reason is, of course, they’re microbivores.
When they’re starved, providing them with
a protein-rich food source rescues them.
When we looked at these high-profiles papers,
we saw that phenotype was also similar, alright?
And if you look at the mechanisms that they
find undermines microbial-mediated developmental
phenotypes, it’s TOR insulin pathway, a
gene involved with amino acid metabolism,
and upregulation of gut proteases.
And although they continue … a lot less
disagreeing, there’s a lot of conflict in
this field, these mechanisms scream to me
some sort of nutritional phenotype.
But what made these papers so special was
that they found specific strains of bacteria
that mediate these phenotypes.
And so, my lab recently published work looking
at this development, and we obtained these
strains of bacteria, and we simply asked … we
made the hypothesis that if bacteria are simply
serving as food for these papers for promoting
Drosophila development, then our hypothesis
would be that if you found a species of bacteria
that was even better for growth promotion
of the fly—the larvae—that bacteria is
probably just something that just grows better
on fly food.
And that’s a really important distinction.
It’s not just the fact here that it grows
better in general.
It grows better in the lab, on your specific
And so, we obtained these bacterial strains,
and it turns out that is the case.
Here was one of the papers that looked at
On the left is the control Lactobacillus strain,
and on the right is the growth-promoting Lactobacillus
You can see on the bottom that the growth-promoting
strain does indeed promote growth; those are
But under fly-free conditions, on the fly
food, you see on dot plot that the growth-promoting
strain does grow better on fly food, and we
see the same thing for the Acetobacter.
Acetobacter is very good at promoting growth
The strain on the right promotes even better
growth, and that strain does grow better under
fly-free conditions on fly food.
And the differences aren’t big.
But if you do the calculations for biomass
with the micrograms of protein content available
based on how much bacteria are available per
larvae that we place in these vials, we’re
in the ballpark of the nutritional requirements
for larval development.
So, I think this … what I think is the most
parsimonious explanation for how microbes
influence fly physiology and health is the
So, holidic diets, chemically defined diets,
were supposed to solve many of the problems
in the nutritional field.
And, at least for Drosophila, a new holidic
medium comes out every 10 or 20 years, and
the last one was coming from the Linda Partridge
lab in 2014.
This is truly, completely chemically-defined:
all amino acids, micronutrients, minerals,
and so forth.
And the bottom line from these studies is
that they’re close, but they are never perfect.
The last diet … they can sustain multiple
generations, you get good egg-laying, you
get reasonable lifespan, but if you compare
it to a standard stock diet, it’s never
exactly the same; it’s never perfect.
So, they’ve come close, but we’re still
not there, and we still don’t know what
it is that makes an optimal diet despite a
considerable amount of effort—over half
a century to a century.
The other important thing to note that doesn’t
seem to be recognized is that—in the fly
community—is that amino acids are not palatable.
They really don’t like them, and not only
do they not like them, I don’t think they’re
that good for them, either.
They increase a lot of pressure in the gut
and gut desiccation, and so I think there
are a number of problems that we may see as
we continue going toward … if we try to
continue using a chemically defined diet.
And an important thing to note: The base diet
Here was this really big paper just stating
that amino acid imbalance explains all the
differences, the lifespan extensions, the
mechanisms for lifespan extension under dietary
And while the results in this paper I think
are fine, I think one way to think about it—or
it’s a completely different interpretation—if
you say that if you start on an imbalanced
diet, you can rescue that by fixing the imbalance.
That’s a very different interpretation of
the work than the title of this paper.
So, I think I’ve hopefully used up my time.
So, let me just get to a couple of take-home
messages that I’ve just gone through with
some of these case studies.
First, throughout my entire career in the
Drosophila community, measurements of feeding
behavior have been absolutely critical.
Measuring how much flies eat has been absolutely
critical, and it’s not just using any assay.
We had to build the assays.
We had to develop the most high-resolution
assays possible, and high-resolution is critical,
right, because they have to be biologically
and physiologically relevant.
You have to be able to measure physiologically
relevant changes in feeding.
If I eat 10 percent more per day, it doesn’t
seem like a lot, but if I eat 10 percent more
every day for a week or a month or a year
or a lifetime, it has obvious physiological
repercussions on health, right?
And it’s been a fight the whole way with
the Drosophila community: first to convince
the community that feeding measurements are
important, that they should be included in
studies; and then the second fight, more recent
fight, has been to make them use the assays
Take the best assays because, for the most
part, feeding measurements in fly studies
are used as a negative control to show that
changes in diet didn’t result in a change
in feed, or a mutant didn’t eat more or
less than its control, or adding a drug to
the diet didn’t change food palatability.
It is really easy to pick an assay that doesn’t
have the resolution to resolve changes in
feeding, and say that you’ve proven the
negative result, which hopefully we all know
you can’t do, right?
So, I think that’s going to be a tough battle
in the zebrafish community as well.
Ten years ago, I started a collaboration with
a zebrafish lab.
We tried to develop a feeding assay; we only
got past … we didn’t even get past the
So, if this is an effort that the community
wants to push forward, I think it will be
hard, but at least from my work and my experiences,
measuring how much actually goes in has been
And secondly, I’d like to state something
sort of opposite to the goals of this workshop,
but there are some dangers to a standardized
A number of labs periodically ask me, people
starting with the Drosophila community or
starting with Drosophila research, “What
diet should I use?”
And I often say not to worry about it too
If your phenotypes are absolutely dependent
on my brand of cornmeal and your supplier
of yeast, how important is it?
I mean, you’re in trouble if someone needs
to replicate your experiment by buying yeast
from a very specific source.
And of course, the cost-benefit of uniformity
in research is different.
There are thousands of fly labs, fly experiments
are cheap, we can sort of afford this soft
coverage of nutritional space.
I don’t know what that cost-benefit is with
zebrafish, but I think it’s important to
consider that phenotypes should work across
diet if they’re robust feeding types.
But the more important thing is transparency
of research methods, and that’s already
been touted by every speaker here so far.
I don’t know how we do it as a community.
The only thing I do for myself, as a peer
reviewer, I always ask that the diets be explicitly
stated in the methods.
I think that’s all we really can do right
now, and you would be surprised by how many
fly labs still say in the methods, “We used
a standard stock medium.”
But again, everyone’s standard is slightly
So, this is the very last slide.
I just want to say that almost everything
I’ve talked about has been on the palatability
end—food intake, food palatability—but
it’s recognized that the more physiologically
relevant parameter is bioavailability.
How much of the nutrients get absorbed in
What is the actual nutritional value of all
these different ingredients … components
of the food?
And how has the nutritional value changed,
depending on how the diet is formulated, where
the diets come from, where the ingredients
We don’t know.
And the assays for this are even worse or
even less developed than the assays for food
At least in Drosophila we’ve just … my
lab has just started developing radioactive
tracing methods to try to figure out bioavailability,
but they are really hard experiments, and
once they’re done, it’s going to be hard,
again, to convince the community that it’s
something worthwhile, but I think the spectrum
is something important enough to think about.
Anyways, I’ll stop there.
Thank you so much.
STEPHEN WATTS: Any questions for Bill?
UNIDENTIFIED MALE: I think it raises the question
whether it’s good nutrition.
I think extension of lifespan is probably
not what matters to the fly most.
What matters to the fly is, I guess, reproductive
So, when you’re at the end of optimizing
and maximizing lifespan, that is probably
not what’s best for the fly from the perspective
of the fly.
WILLIAM JA: True.
UNIDENTIFIED MALE: So, one needs criteria
of what to achieve with standard nutrition.
WILLIAM JA: True.
I think the diet one uses should be very tied
into your experimental goals, your project
goals, and you’re absolutely right.
There are different optimal diets depending
on what phenotype you’re looking for, so
there’s certainly a different diet for optimal
egg-laying and for reproduction, and it’s
a different protein-carb balance, for example,
that optimizes or maximizes survival or lifespan.
UNIDENTIFIED MALE: You talked about the Drosophila
community’s efforts to develop chemically-defined
diets over the years, but you also mentioned
the important role that the microbial communities
play in nutrition.
And so, can you talk about how those two things
Like, have there been efforts to develop diets
that are able to promote growth in the absence
of microbes, or has there been efforts to
normalize microbial communities across …?
WILLIAM JA: Good question.
It’s been recognized that the microbes were
important on these lab diets since 1914, I
think, or 1917, with some of the first papers
looking at microbes, and some base diets promoted
growth—fly growth—under axenic or germ-free
conditions, and some didn’t.
I’m not sure if it was ever figured out
the specific components, but nowadays the
standard diet is yeast, cornmeal, sugar, is
I think the protein and micronutrient content
is sufficient to sustain development for many
generations without microbes.
Whether that’s a good thing or not, I don’t
know, but I’ll also state that most labs
use propionic acid or phosphoric acid or tegosept
as a preservative, which reduces fungal growth,
but it actually decimates bacterial growth
So, many labs are working under conditions
where the bacteria are constantly being growth-suppressed
by these additives, and this is actually what’s
led to most labs having a transient microbiota.
And again, it’s only been in the last … really
the last year where labs are starting to recognize
that if we change the diet, and we bring in
strains of bacteria from the wild, not the
lab strains anymore, but these strains of
bacteria from the wild actually are able to
stably colonize the fly gut.
We don’t know what they do.
We still have to figure out how to tease or
decouple the influence of nutrition for microbivores
from any specialized host gut microbe interactions,
but at least we’ve figured out how to make
a stable gut microbiome.
So, I think that’s the first step in starting
to dissect that, but few labs are going in
that direction because it’s so much easier
to ignore the microbivore concept.
UNIDENTIFIED MALE: Are zebrafish microbivores?
WILLIAM JA: Um …
UNIDENTIFIED MALE: Well, we can talk about
UNIDENTIFIED MALE: I have a question, just
thinking about the palatability and the bioavailability
of the diet, and really it’s a question
for the previous speaker who showed that high
diets rich in carbohydrates, so high-carbohydrate
diets, don’t work with fish; they don’t
supply the energy required.
And so, is that like a food intake or palatability
thing or absorption thing or metabolizing
Is anything known about that?
UNIDENTIFIED MALE: In terms of Ron’s presentation,
I think the carbohydrate variability among
species of fish is definitely provided to
their natural history.
Some, like the trout and salmon, actually
consume diets that have a very low load of
carbohydrate, so their ability to utilize
soluble carbohydrates for energy is much less
than some more omnivorous species like the
common carp or tilapia, and then we even have
some carnivorous fish like grass carps that
can utilize carbohydrates better.
So, it’s more related to the natural history
of the fish and the presence of the soluble
carbohydrates in the environment, and just
beyond just the basic intake and digestibility,
there are obviously some metabolic effects
that are different amongst the different fish
UNIDENTIFIED MALE: And catfish.
UNIDENTIFIED MALE: So, is that an absorption
UNIDENTIFIED MALE: Yes, but that’s not the
It also has to do with their ability to transport
UNIDENTIFIED MALE: The presence and activity
of carbohydrates and enzymes, too, in the
WILLIAM JA: It’s tough to compare terrestrial
animal nutrition to fish nutrition in this
aspect, and you guys would know more about
this, but it’s very different.