Walk inside a cell: animal, plant and bacterium

What you are standing in

A cell is too small to picture. The usual drawing in a school book puts the nucleus, a few mitochondria and some wavy lines of ER inside a circle, all about the same size and all floating in empty space. This exhibit takes one animal cell, enlarges it a million times and lets you walk around inside it at an ordinary eye height. Everything is drawn to the sizes measured in real cells, so a mitochondrion is a capsule you could sit on and a ribosome is a dark bead the size of a large marble.

The cell is shaped like one spread out on a glass dish: a low dome 30 µm long, 24 µm wide and 8 µm tall in the middle. You start on its floor, which is the part of the plasma membrane stuck to the glass. The nucleus fills the middle, with rough ER wrapped around its left side and the Golgi stack on its right. Mitochondria and lysosomes are scattered around, microtubules fan out from a point beside the nucleus, and small vesicles ride along them towards the edge of the cell.

Why the model is a million times larger

The obvious choice would be to make the cell large enough that your body fits inside it. A magnification of 100,000 is not enough. At that scale a 20 µm cell becomes a 2 m bubble, and the nucleus would be smaller than your head. A person 1.6 m tall would stand 16 µm high, taller than the whole cell.

At a magnification of 1,000,000 the numbers become easy to read. One metre in the model is one micrometre in the cell, and one millimetre is one nanometre. Your eye height of 1.6 m stands for 1.6 µm. The scale box in the top left corner shows this, and its bar is redrawn for whatever is straight ahead of you. If you look at a wall 3 m away, the bar shows how long 1 µm looks at that distance.

Sizes in the model

These are the main parts and the sizes they are built to. The real figures are ranges, because cells differ from one type to the next.

The cell's volume works out at about 3,000 µm³. Cell Biology by the Numbers uses 2,000 µm³ as a rule of thumb for a human cell, with most falling between 500 and 4,000 µm³, and it uses 3,000 µm³ for a characteristic mammalian cell in its worked examples. HeLa cells spread on a slide reach about 40 µm across while staying only a few micrometres tall, which is why the model is a low dome rather than a ball.

Checking a few numbers yourself

Stand next to the cut-open mitochondrion near the start. It is 1 m wide, so its top is a little below your eye level. In the real cell that is 1 µm, about a hundredth of the width of a human hair.

The membrane is the hardest part to see. At 4 to 10 nm thick it is 4 to 10 mm in the model, about as thick as a pencil, wrapped around a room 30 m long. Cell Biology by the Numbers compares it to the skin of an aircraft against the width of its body. That is why one patch on the right-hand wall is drawn 20 times thicker than the rest. In that patch each lipid head is 16 mm across instead of 0.8 mm, and you can count the two layers.

The DNA in the nucleus is the number that does not fit at all. A human cell carries about 2 m of DNA (two copies of a genome of about 3 billion base pairs, 0.3 nm per base pair). In the model that would be 2,000 km of thread packed into a ball 9 m long. The purple threads you see inside the window are a small sample drawn to show the texture.

The membrane in cross-section

A phospholipid has a head that mixes with water and two tails that do not. In water the tails hide from it by lining up tail to tail, which gives two layers with heads on both outer faces. Each layer is about 2 nm deep, so the core of the bilayer is about 4 nm thick. Proteins sit in this film. The blue barrel in the magnified patch is a channel that spans both layers, and the protein with green beads on the outside carries a chain of sugars.

The membrane is fluid. Lipids drift sideways past their neighbours all the time, but they very rarely flip from one layer to the other, so the two layers can keep different mixes of lipids. The model holds them still.

The nucleus and its pores

The nuclear envelope is two membranes, an inner and an outer, with a gap of a few tens of nanometres between them. The window cut into the front of the nucleus shows both rims, about 4 cm apart in the model. The outer membrane carries on into the rough ER, which is why the ER sheets in the model start right next to the nucleus.

Nuclear pores pass through both membranes at once. Each is a ring of proteins about 120 nm across with a channel through the middle, and a vertebrate nucleus has about 2,000 of them. Messenger RNA and ribosome subunits go out through the pores, and proteins needed inside the nucleus come in. The model shows 300 to 1,100 pores depending on your device.

Inside, the nucleolus is a dense region with no membrane around it, drawn 2 µm across in the model. It forms around the genes for ribosomal RNA, and the two halves of the ribosome are put together there before they are exported.

Where proteins are made and packed

A ribosome reads messenger RNA and joins amino acids into a chain. Ribosomes floating in the cytosol make proteins that stay in the cytosol. When the start of a new protein carries a signal for the ER, the ribosome docks onto the ER and feeds the growing chain through the membrane into the sac, which is what makes the ER look rough. Smooth ER has no ribosomes. It is a network of tubes 30 to 100 nm wide that makes lipids and stores calcium.

From the ER, proteins travel in small vesicles to the Golgi apparatus. A mammalian cell has 40 to 100 Golgi stacks, usually gathered near the centrosome, and each stack has four to eight cisternae. The model shows one stack of six. Proteins enter on the cis side, nearest the ER, have sugars trimmed and added as they move through, and leave from the trans side in vesicles headed for the membrane, the lysosomes or other compartments.

Kinesin walking on a microtubule

Microtubules are hollow tubes about 25 nm wide, usually built from 13 rows of tubulin. Each row is a chain of alpha and beta tubulin pairs that repeats every 8 nm. Most microtubules start at the centrosome beside the nucleus and run out towards the edge of the cell. They are stiff and nearly straight over the distances inside a cell.

Kinesin-1 has two motor heads joined by a long coiled stalk, and a tail that holds the cargo. It walks hand over hand, as PDB-101 describes it. One head stays bound to the microtubule while the other lets go, swings forward past it and binds 16 nm further on. Each step moves the cargo 8 nm and uses one ATP. Cell Biology by the Numbers gives about 800 nm per second for conventional kinesin in the test tube, with about 100 steps before it falls off.

At a million times life size, 800 nm per second becomes 0.8 m per second, a slow walking pace. But the heads take about 100 steps a second, far too fast to follow on screen. The guide step about kinesin therefore switches on slow motion at 1/20 of real speed, so you can see each head swing past the other.

Mitochondria and lysosomes

Mitochondria have two membranes. The outer one is smooth. The inner one is folded into cristae, which give it a large surface for the proteins that make ATP. Cell Biology by the Numbers describes a typical mitochondrion as a capsule roughly 2 µm long and 1 µm across, though in many cells they join into branching networks tens of micrometres long. The model uses ten separate capsules and cuts one open so you can see the cristae.

Lysosomes are bags of more than 60 kinds of digestive enzyme, and a cell normally has hundreds of them. The model shows six. Their inside is kept at pH 4.5 to 5, and the enzymes work best at that acidity. The cytosol around them is at about pH 7.2, where the enzymes do not work properly, so a leaking lysosome does not digest the rest of the cell.

Shrinking ten and a hundred times

At a million times life size the cell is easy to walk through, but anything smaller than about 20 nm is a crumb at your feet. A ribosome is the size of a marble and a potassium channel is 8 mm across. To see molecules you need a second step of magnification, and the Size buttons in the top right give you two more: ×10 and ×100.

The buttons do not rebuild the cell. They shrink you. At ×10 the scale is 10,000,000 to 1, one metre in front of you is 100 nm, and your eyes are 160 nm above the floor. At ×100 the scale is 100,000,000 to 1, one metre is 10 nm, and you stand 16 nm tall, about the height of a proteasome. The box in the top left corner always shows the current magnification, what one metre stands for, and your real height, so the numbers stay honest whichever step you are on. The cell itself does not change. A mitochondrion that was the size of a sofa becomes a building 100 m long, and walking past it at ×100 would take minutes.

Molecules at their true size

Molecules are drawn as bead models, with one bead for each folded part of a protein, in the style of the illustrations on RCSB PDB-101. They are built at their true size and placed where they work, then hidden until you are close enough for them to show. From across the cell they cost nothing to draw.

There are four places to find them in the animal cell:

• On the floor membrane, near the front right. Potassium channels, sodium-potassium pumps, receptors and aquaporins sit in a patch of membrane paved with lipid heads 0.8 nm apart. A clathrin-coated pit, a cage about 100 nm wide, is pulling a bud of membrane inwards. Actin filaments 7 nm wide run across the floor, with a myosin V motor walking on one of them, and a rope-like intermediate filament lies nearby.
• In the crowd beside them. A box 100 nm on each side is filled with proteins at the density of real cytosol, about 3 million per cubic micrometre. At ×100 it is a wall of molecules you can push into. A proteasome, an Hsp90 chaperone, a glycogen granule and a string of ribosomes reading one messenger RNA sit at its edge.
• In the nucleus. A stretch of DNA wound round nucleosomes, with RNA polymerase II making a strand of RNA, and a nuclear pore complex in full detail on the envelope.
• On organelles. A ribosome docked on the rough ER through its Sec61 channel, a row of ATP synthases at the cut edge of the open mitochondrion, and a dynein motor under the kinesin track.

The densities are real where the guide book says so (the crowd box, the lipid floor) and thinned out elsewhere, because a real membrane is so packed with protein that the channels would be hard to pick out.

A plant cell at the same scale

The Cell button switches to a leaf cell built to the same scale. It is 40 µm long, 24 µm wide and 20 µm tall, inside a wall of cellulose. Cell Biology by the Numbers describes plant cells of this kind with most of their volume taken up by a vacuole, and that is the first thing you notice: a translucent block fills the middle, and the rest of the cell is squeezed into a thin layer around it and a room at one end where the nucleus sits.

Chloroplasts lie on the floor under the vacuole and against one wall. They are drawn 5 µm long, inside the range of 4 to 6 µm given for vascular plants, and one is cut open to show its grana, stacks of flat thylakoid sacs, and a grain of starch. On the far wall a pit field of plasmodesmata, channels about 50 nm wide, connects this cell to its neighbour. Along the near wall, organelles stream round a loop of actin cables at 4 µm per second, inside the range of 1 to 100 µm per second measured for cytoplasmic streaming. At ×100 inside the cut chloroplast you can find Rubisco, the enzyme that fixes carbon dioxide, and photosystem II, which splits water and releases oxygen.

A bacterium at the same scale

The third cell is E. coli. Cell Biology by the Numbers uses about 1 µm wide, 2 µm long and 1 µm³ in volume as the rule of thumb, so at a million times life size the whole organism is a capsule 2 m long lying on the dish, the size of one of the animal cell's mitochondria. That is not a coincidence of the model. Mitochondria descend from bacteria that were taken in by another cell in the distant past, and they still keep a small genome of their own. The grid on the dish is drawn in 1 µm squares.

At ×1 you are too big to go inside, so the walk starts at ×10. Inside, ribosomes fill most of the space around the nucleoid, the tangle of the chromosome in the middle with no membrane around it. The envelope overhead has three layers: the inner membrane, a thin net of peptidoglycan and the outer membrane. Flagellar motors about 45 nm across sit in the envelope, each turning a corkscrew filament about 100 times a second, and a lattice of chemoreceptors at one end tells the motors when to reverse. Because there is no nucleus in the way, ribosomes start reading messenger RNA while it is still being made, which the model shows at two places near the start.

The guide book

The Guide book button opens a list of 78 things to visit: 49 in the animal cell, 14 in the plant cell and 15 in the bacterium. You can search by name, filter by cell and by kind (organelles, molecules and processes) and sort by size. Choosing one switches cell and size if needed and walks you to a spot in front of it. A card then shows its real size, how many a cell has, a short description, its sources, and a bar comparing its size with a protein, a ribosome, E. coli, an animal cell and your own height at the current size.

Seven processes, the stages of mitosis, the cell cycle, the cell plate of a dividing plant cell and binary fission in bacteria, are not played out in the 3D cells. Their cards show a small drawing instead.

What a million times does to speeds and distances

Because 1 m in the model is 1 µm in the cell, every length can be converted by moving the unit up by six steps. A cell 20 µm across is 20 m across, about the length of a bus and a half laid end to end. A ribosome 25 nm across is 25 mm, 2.5 cm, the size of a large marble. A protein of 5 nm is 5 mm, a grain of rice seen end on. A membrane 5 nm thick is 5 mm thick, a layer you could hold between two fingers around a hall 30 m long.

Speeds scale the same way, because time is not stretched. Kinesin walks at about 800 nm per second, which is 0.8 m per second in the model. A protein that drifts a few micrometres in a second covers a few metres. You walk at about 1.4 m per second, which is 1.4 µm per second in the real cell, a little faster than a kinesin. Crossing the 30 m length of the cell takes you about 21 seconds. At ×100 the same walk is 14 nm per second, and the Walk control has a fast setting for that reason.

The units hide one trap. Areas scale with the square and volumes with the cube. A mitochondrion of 1 µm³ is 1 m³ in the model, a cube a metre on each side. The whole cell, about 3,000 µm³, is 3,000 m³, a hall larger than an Olympic swimming pool, which holds about 2,500 m³. The numbers in the table above are lengths. When this guide gives a count per cell, it means per 3,000 of those cubic metres.

How many molecules, and how long they take to cross

Cell Biology by the Numbers gives about 3 million proteins in every cubic micrometre of a cell, so a human cell of 2,000 to 4,000 µm³ holds on the order of 10 billion (10¹⁰) proteins. For the 3,000 µm³ of this model that works out at 3,000 × 3 million, about 9 billion. In the model, one cubic metre of space around you holds 3 million of them. The average gap between neighbouring proteins is the cube root of 1 µm³ divided by 3 million, about 0.007 µm or 7 nm, which is 7 mm on screen. A typical protein is about 5 nm across, so the crowd is nearly touching. This is the density the crowd box reproduces. Everything else in the model is much emptier than that.

Molecules in a cell do not travel in straight lines. They are knocked about by water molecules, a motion called diffusion, and the average time to cover a distance x grows with the square of that distance. For a particle moving in one dimension it is t ≈ x² / 2D, where D is the diffusion constant. In three dimensions the 2 becomes a 6. The book gives D as about 100 µm²/s for a protein of average size in water and about 10 µm²/s inside a cell. Putting in the numbers:

• Across an E. coli cell, x = 1 µm: t ≈ 1 / 60, about 0.02 s. The book quotes roughly 0.01 s.
• Across a HeLa cell, x = 20 µm: t ≈ 400 / 60, about 7 s. The book quotes about 10 s.
• Along an axon 1 cm long, x = 10,000 µm: t ≈ 10⁸ / 60, about 1.7 million s, roughly three weeks. The book says about two weeks, since it rounds D and the factor.

A motor moving at 1 µm per second would take 2 to 3 hours to cover 1 cm, which is why nerve cells carry cargo on microtubules rather than waiting for it to diffuse. At the scale of a cell, though, the two methods are close. At 20 µm a motor needs 20 s and diffusion needs under 10 s, so a small cell can let molecules diffuse and keeps motors for cargo that is large or must go to a particular place.

Common misunderstandings

• "Organelles float far apart in empty water." In a real cell the cytosol is a crowded gel and the organelles are packed close, with the ER stretching almost everywhere. The open room you walk through in the model is a simplification so that you can move.
• "Molecules know where to go." Most reach their partner by chance, by diffusion, and react only when they happen to meet in the right orientation. Motors like kinesin are the exception, since they follow tracks, but the cargo still has to find the track.
• "A kinesin walks like a person." It looks like it from the animation, but the heads are small enough to be pushed about by the surrounding water all the time. The walk is a ratchet in which each ATP makes one head swing forward and bind, so the jiggling is turned into steps in one direction.
• "The nucleus is the cell's brain, so it controls everything directly." The nucleus stores and copies the genome and makes RNA, but proteins are made in the cytosol and on the ER, and many of the cell's decisions are made by signals arriving at the membrane. Mitochondria even keep a small genome of their own.
• "All cells look like the drawing in the textbook." The model shows one cell type spread on glass, with the sizes of the ones commonly studied. A red blood cell has no nucleus and a muscle fibre has many, and cell volumes range over several orders of magnitude.

How the model differs from a real cell

• The cytosol is nearly empty here. A real cell holds about 3 million proteins in each cubic micrometre. The floating specks around you are a few thousand at most, and they drift slowly instead of jostling at high speed.
• Ribosomes are thinned out. The model draws a few thousand. A real cell has far more, packed so densely on the rough ER that the sheets look studded.
• Actin and intermediate filaments are sparse. Bundles of actin cross the floor and a few intermediate filaments wrap the nucleus, but the dense actin mesh under the membrane is left out so the paths stay clear.
• Molecules are hidden from afar. The molecule models appear only when you are close enough to see them, and they sit still instead of tumbling.
• Most organelles keep still. In a living cell, mitochondria move, split and join, and the ER changes shape from minute to minute.
• The nucleus has a window. A real envelope is closed apart from its pores.
• Colours are invented. Organelles have no colour of their own at this size. Electron microscopes show them in grey, and fluorescence images use dyes.

Moving around

On a computer, use WASD or the arrow keys to walk and drag with the mouse to look around. R lifts you up and F lets you sink back down. Shift makes you move faster. Clicking the floor walks you there. On a phone, the left half of the screen is a joystick and the right half turns your view. The two arrows on the right lift you up or down.

The panel on the right of the stage has three rows. Cell switches between the animal cell, the plant cell and the bacterium. Size shrinks you to ×10 or ×100 and back. Walk sets your pace to slow, normal or fast, which helps at ×100, where normal walking covers only 13 nm a second. The small map in the top left shows where you are from above; click it to walk to that spot.

Name tags appear when you come close to something, and at ×100 they label single molecules. The All name tags control shows more of them at a time. The Parts list and the Guide book both take you straight to whatever you choose.