How designing brand-new enzymes could change the world

Growing up in central Wisconsin,
I spent a lot of time outside.

In the spring, I'd smell
the heady fragrance of lilacs.

In the summer, I loved
the electric glow of fireflies

as they would zip around on muggy nights.

In the fall, the bogs were brimming
with the bright red of cranberries.

Even winter had its charms,

with the Christmassy bouquet
emanating from pine trees.

For me, nature has always been
a sense of wonder and inspiration.

As I went on to graduate school
in chemistry, and in later years,

I came to better understand
the natural world in molecular detail.

All the things that I just mentioned,

from the scents of lilacs and pines

to the bright red of cranberries
and the glow of fireflies,

have at least one thing in common.

They're manufactured by enzymes.

As I said, I grew up in Wisconsin,
so of course, I like cheese,

and the Green Bay Packers.

But let's talk about cheese for a minute.

For at least the last 7,000 years,

humans have extracted a mixture of enzymes

from the stomachs of cows
and sheep and goats,

and added it to milk.

This causes the milk to curdle --
it's part of the cheese-making process.

The key enzyme in this mixture
is called chymosin.

I want to show you how that works.

Right here, I've got two tubes,

and I'm going to add chymosin
to one of these.

Just a second here.

Now my son Anthony,
who is eight years old,

was very interested in helping me
figure out a demo for the TED talk,

and so we were in the kitchen,
we were slicing up pineapples,

extracting enzymes from red potatoes

and doing all kinds of demos
in the kitchen.

And in the end, though,

we thought the chymosin demo
was pretty cool.

And so what's happening here

is the chymosin
is swimming around in the milk

and it's binding to a protein
there, called casein.

What it does then
is it clips the casein --

it's like a molecular scissors.

It's that clipping action
that causes the milk to curdle.

So here we are in the kitchen,
working on this.

OK.

So let me give this a quick zip.

And then we'll set these to the side
and let these simmer for a minute.

OK.

If DNA is the blueprint of life,

enzymes are the laborers
that carry out its instructions.

An enzyme is a protein that's a catalyst,

it speeds up or accelerates
a chemical reaction,

just as the chymosin over here
is accelerating the curdling of the milk.

But it's not just about cheese.

While enzymes do play an important role
in the foods that we eat,

they also are involved in everything
from the health of an infant

to attacking the biggest
environmental challenges

we have today.

The basic building blocks of enzymes
are called amino acids.

There are 20 common amino acids

and we typically designate them
with single-letter abbreviations,

so it's really an alphabet of amino acids.

In an enzyme, these amino acids
are strung together,

like pearls on a necklace.

And it's really the identity
of the amino acids,

which letters are in that necklace,

and in what order they are,
what they spell out,

that gives an enzyme its unique properties
and differentiates it from other enzymes.

Now, this string of amino acids,

this necklace,

folds up into a higher-order structure.

And if you were to zoom in
at the molecular level

and take a look at chymosin,
which is the enzyme working over here,

you would see it looks like this.

It's all these strands and loops
and helices and twists and turns,

and it has to be in just
this conformation to work properly.

Nowadays, we can make enzymes in microbes,

and that can be like a bacteria
or a yeast, for example.

And the way we do this
is we get a piece of DNA

that codes for an enzyme
that we're interested in,

we insert that into the microbe,

and we let the microbe use
its own machinery, its own wherewithal,

to produce that enzyme for us.

So if you wanted chymosin,
you wouldn't need a calf, nowadays --

you could get this from a microbe.

And what's even cooler, I think,

is we can now dial in
completely custom DNA sequences

to make whatever enzymes we want,

stuff that's not out there in nature.

And, to me, what's really the fun part

is trying to design an enzyme
for a new application,

arranging the atoms just so.

The act of taking an enzyme from nature
and playing with those amino acids,

tinkering with those letters,

putting some letters in,
taking some letters out,

maybe rearranging them a little bit,

is a little bit like finding a book

and editing a few chapters
or changing the ending.

In 2018, the Nobel prize in chemistry

was given for the development
of this approach,

which is known as directed evolution.

Nowadays, we can harness
the powers of directed evolution

to design enzymes for custom purposes,

and one of these is designing enzymes
for doing applications in new areas,

like laundry.

So just as enzymes in your body

can help you to break down
the food that you eat,

enzymes in your laundry detergent

can help you to break down
the stains on your clothes.

It turns out that about
90 percent of the energy

that goes into doing the wash

is from water heating.

And that's for good reason --

The warmer water
helps to get your clothes clean.

But what if you were able
to do the wash in cold water instead?

You certainly would save some money,

and in addition to that,

according to some calculations
done by Procter and Gamble,

if all households in the US
were to do the laundry in cold water,

we would save the emissions
of 32 metric tons of CO2 each year.

That's a lot,

that's about the equivalent

of the carbon dioxide
emitted by 6.3 million cars.

So, how would we go
about designing an enzyme

to realize these changes?

Enzymes didn't evolve
to clean dirty laundry,

much less in cold water.

But we can go to nature
and we can find a starting point.

We can find an enzyme
that has some starting activity,

some clay that we can work with.

So this is an example of such an enzyme,
right here on the screen.

And we can start playing
with those amino acids, as I said,

putting some letters in,
taking some letters out,

rearranging those.

And in doing so, we can generate
thousands of enzymes.

And we can take those enzymes

and we can test them
in little plates like this.

So this plate that I'm holding in my hands

contains 96 wells,

and in each well is a piece of fabric
with a stain on it.

And we can measure
how well each of these enzymes

are able to remove the stains
from the pieces of fabric,

and in that way see how well it's working.

And we can do this using robotics,

like you'll see
in just a second on the screen.

OK, so we do this and it turns out

that some of the enzymes
are sort of in the ballpark

of the starting enzyme.

That's nothing to write home about.

Some are worse, so we get rid of those.

And then some are better.

Those improved ones
become our version 1.0s.

Those are the enzymes
that we want to carry forward

and we can repeat this cycle
again and again.

And it's the repetition of this cycle
that lets us come up with a new enzyme,

something that can do what we want.

And after several cycles of this,

we did come up with something new.

So you can go to the supermarket today
and you can buy a laundry detergent

that lets you do the wash in cold water
because of enzymes like this here.

And I want to show you
how this one works too.

So I've got two more tubes here,

and these are both milk again.

And let me show you,

I've got one that I'm going
to add this enzyme to

and one that I'm going
to add some water to.

And that's the control,

so nothing should happen in that tube.

You might find it curious
that I'm doing this with milk.

But the reason that I'm doing this

is because milk
is just loaded with proteins,

and it's very easy to see
this enzyme working in a protein solution,

because it's a master protein chopper,

that's its job.

So let me get this in here.

And you know, as I said,
it's a master protein chopper

and what you can do is you can extrapolate
what it's doing in this milk

to what it would be doing in your laundry.

So this is kind of a way to visualize
what would be happening.

OK, so those both went in.

And I'm going to give this
a quick zip as well.

OK, so we'll let these sit over here
with the chymosin sample,

so I'm going to come back
for those toward the end.

Well, what's on the horizon
for enzyme design?

Certainly, it will get it faster --

there are now approaches
for evolving enzymes

that allow researchers to go
through far more samples

[unclear] like I just showed you.

And in addition to tinkering
with natural enzymes,

like we've been talking about,

some scientists are now trying to design
enzymes from scratch,

using machine learning,
an approach from artificial intelligence,

to inform their enzyme designs.

Still others are adding
unnatural amino acids to the mix.

We talked about
the 20 natural amino acids,

the common amino acids, before --

they're adding unnatural amino acids

to make enzymes with properties unlike
those that could be found in nature.

That's a pretty neat area.

How will designed enzymes affect you
in years to come?

Well, I want to focus on two areas,

human health and the environment.

Some pharmaceutical companies

now have teams that are dedicated
to designing enzymes

to make drugs more efficiently
and with fewer toxic catalysts.

For example, Januvia,

which is a medication to treat
type 2 diabetes,

is made partially with enzymes.

The number of drugs made with enzymes
is sure to grow in the future.

In another area,

there are certain disorders

in which a single enzyme
in a person's body doesn't work properly.

An example of this
is called phenylketonuria,

or PKU for short.

People with PKU are unable to properly
metabolize or digest phenylalanine,

which is one of the 20 common amino acids
that we've been talking about.

The consequence of ingesting phenylalanine
for people with PKU

is that they are subject
to permanent intellectual disabilities,

so it's a scary thing to have.

Now, those of you with kids --

do you guys have kids, here,
which ones have kids?

A lot of you.

So may be familiar with PKUs,

because all infants in the US
are required to be tested for PKU.

I remember when Anthony, my son,
had his heel pricked to test for it.

The big challenge with this
is what do you eat?

Phenylalanine is in so many foods,
it's incredibly hard to avoid.

Now, Anthony has a nut allergy,
and I thought that was tough,

but PKU's on another level of toughness.

However, new enzymes
may soon enable PKU patients

to eat whatever they want.

Recently, the FDA approved an enzyme
designed to treat PKU.

This is big news for patients,

and it's actually very big news

for the field of enzyme-replacement
therapy more generally,

because there are other targets out there
where this would be a good approach.

So that was a little bit about health.

Now I'm going to move to the environment.

When I read about
the Great Pacific Garbage Patch --

by the way, that's, like,
this huge island of plastic,

somewhere between California and Hawaii --

and about microplastics
pretty much everywhere,

it's upsetting.

Plastics aren't going away anytime soon.

But enzymes may help us
in this area as well.

Recently, bacteria producing
plastic-degrading enzymes were discovered.

Efforts are already underway
to design improved versions

of these enzymes.

At the same time, there are enzymes
that have been discovered

and that are being optimized

to make non-petroleum-derived
biodegradable plastics.

Enzymes may also offer some help
in capturing greenhouse gases,

such as carbon dioxide, methane
and nitrous oxide.

Now, there is no doubt,
these are major challenges,

and none of them are easy.

But our ability to harness enzymes
may help us to tackle these in the future,

so I think that's another area
to be looking forward.

So now I'm going to get back
to the demo --

this is the fun part.

So we'll start with the chymosin samples.

So let me get these over here.

And you can see here,

this is the one that got the water,

so nothing should happen to this milk.

This is the one that got the chymosin.

So you can see that it totally
clarified up here.

There's all this curdled stuff,
that's cheese,

we just made cheese
in the last few minutes.

So this is that reaction

that people have been doing
for thousands and thousands of years.

I'm thinking about doing this one
at our next Kids to Work Day demo

but they can be
a tough crowd, so we'll see.

(Laughter)

And then the other one
I want to look at is this one.

So this is the enzyme
for doing your laundry.

And you can see that it's different
than the one that has the water added.

It's kind of clarifying,

and that's just what you want
for an enzyme in your laundry,

because you want to be able
to have an enzyme

that can be a protein chowhound,
just chew them up,

because you're going to get
different protein stains on your clothes,

like chocolate milk
or grass stains, for example,

and something like this
is going to help you get them off.

And this is also going to be
the thing that allows you

to do the wash in cold water,
reduce your carbon footprint

and save you some money.

Well, we've come a long way,

considering the 7,000-year journey
from enzymes in cheese making

to the present day and enzyme design.

We're really at a creative crossroads,

and with enzymes,
can edit what nature wrote,

or write our stories with amino acids.

So next time you're outdoors
on a muggy night

and you see a firefly,

I hope you think of enzymes.

They're doing amazing things for us today.

And by design,

they could be doing
even more amazing things tomorrow.

Thank you.

(Applause)