Though another chapter already deals with Section E(iii) of the 2017 CICM Primary Syllabus ("describe ... cellular organelles and their properties"), the mitochondria deserve some VIP treatment, including a chapter all of their own. Apart from their actual physiological importance, these organelles also call for increased attention because Question 18 from the first paper of 2016 specifically asked the trainees to "describe the structure and function of the mitochondrion". The pass rate, predictably, was 19%. Nobody saw this coming. Reflecting on that paper, otherwise sane people began to question their approach to exam preparation; surely if they can ask you this, then they can ask you anything.
Important structural features of the mitochondrion include:
- Two membranes
- Outer membrane with pores
- Inner membrane without pores
- These membranes meet at contact sites (area where membrane lipids and proteins may be exchanged with other organelles)
- Two separate compartments
- Intermembrane space or "outer compartment"
- Matrix space or inner compartment
- Organisation of the inner membrane into cristae (a comb of many folds) which increases its surface area
- Molecular components
- ATP synthase molecules on the inner surface of the inner membrane
- Mitochondrial ribosomes in the inner compartment
- Mitochondrial DNA in the inner compartment
- Electron dense-granules in the inner compartment (composed of RNA)
Mitochondrial function includes:
- ATP synthesis functions, including:
- Citric acid cycle
- Electron transport chain
- Beta-oxidation of long chain fatty acids
- Regulatory and synthetic functions
- Haem synthesis
- Calcium ion storage
- Urea cycle
- Haem synthesis
- Steroid synthesis
- As a byproduct of these functions
- Heat production
- CO2 production
- Production of reactive oxygen species
- Non-metabolic roles
It is hard to call out any single best reference for the time-poor trainee to read before the exams. Remarkably for a subject with little direct clinical relevance, UpTodate have an excellent summary article on mitochondrial function which covers most of the bases and would make a good single reference for this topic, but some of the stuff included in the college comments (eg. haem synthesis, urea cycle) is not mentioned there. McCarron et al (2013) give a good rundown of the structure (and motility) of this organelle but has minimal information on their metabolic function. Keith Garlid's chapter on mitochondria from Nick Sperelakis' Cell Physiology Sourcebook (3rd ed, 2001, p. 139-151) covers energy production comprehensively, but completely omits other aspects, and therefore cannot be recommended. Ganong have some mitochondrial material in Chapter 2 (p. 34 f the 23rd edition) but it is only half a page and not enough to answer Question 18 from the first paper of 2016. Nor does Guyton & Hall contain enough detail to cover a good response, no matter how one reconstructs the marking rubric from the college comments. Immo E. Scheffler's Mitochondria (2007, 2nd ed.) covers every possible facet of the subject over about 450 pages, but has no summary and is therefore completely useless to the exam candidate. In short, there is no single best reference, and this chapter has been cobbled together out of multiple sources.
Without repeating oneself, the most important (examinable) features of mitochondrial structure are as follows:
If one wanted to use a diagram to illustrate these features within the timeframe of a 10-minute SAQ, one would be lucky to produce something like this (from Cellular Organelles by Bittar):
It is probably useless to expand much further on this subject because realistically that's all you can expect to be able to cram into your head for the CICM Part I exams. However, this site has a rich history of chasing topics into deep dark rabbit holes. And so:
A mitochondrion is generally bacteria-sized, usually about 3-4 μm in length and about 1 μm in diameter. They do come in all shapes and sizes, and the idealised bean-shaped mitochondrion seen in textbooks is really an unfair stereotype for mitochondrial body shape, which bears about as much morphological relationship to a real mitochondrion as a glossy magazine cover does to a normal female body. With the help of fluorescent biolabels, one can celebrate the remarkable diversity of mitochondrial shapes and sizes. Here's an image of MitoTracker Green-labelled mitochondria from McCarron et al (2013); the scale bar is 10 μm.
These were just mitochondria from vascular smooth muscle cells. Given the diversity of cell types and energy demands, one may expect to see a vast diversity of mitochondrial phenotypes. This extends out of the strictly medical context and into broader biology. Far from the normalities of human cells, in the anoxic goop at the bottom of ponds there are single-celled eukaryotes which have no use at all for oxygen, and therefore no metabolic/oxidative purpose for mitochondria. However, they have mitochondrial remnants (mitosomes and hydrogenosomes) which perform vital metabolic functions and which obviously have an even greater diversity of physical structures.
The outer membrane of the mitochondria is a relatively porous lipid bilayer. It is full of porin proteins which permit a relatively unobstructed movement of molecules in either direction. The porins which allow this are called VDACs (voltage-gated anion channels), though their "voltage-gated" properties are not normally seen in vivo. In vitro, at a membrane potential of -40 mV or so, these gates shut, allowing some anions to cross (but not ATP). According to Colombine (2004) most of the time these are open in living cells, allowing passage for molecules up to 5-6 kDa in size. So porous is this membrane that ions exchange relatively freely between the cytosol and the intramembranous space, and there is no electrochemical gradient or potential difference across this membrane.
The intermembranous space is a narrow 20 nm gap between the inner and the outer membrane. Because the outer membrane is full of abovementioned porins and promiscuously allows any small molecule to get in and out of the intermembranous space, its ionic content closely resembles the cytosol. There is a difference in protein content, however - there are various unique proteins in this space which complete the final steps of porphyrin and urea metabolism, contribute to protein logistics (eg. transport, folding, sorting), and to apoptosis. Particularly, the apoptosis-triggering cytochrome C protein lurks in the intermembranous space.
The intermembranous space was thought to be folded up to increase the surface area of the inner membrane, and these comb-like folds were called the cristae mitochondriales. As Johannes Herrmann (2010) describes, these cristae were thought to be continuous invaginations of the intermembranous space. Turns out, they are actually separated from the rest of the intermembranous space by semipermeable junctions, which turn the intercristal space into a functionally distinct compartment.
The intercristal space is something of a toxic waste dump for the mitochondria. Because of the electron transport chain this compartment ends up being filled with a fluid which is rich in hydrogen ions. That makes sense, because the hydrogen ion gradient between the intermembranous space and the mitochondrial matrix is used by the electron transport chain to produce ATP. As a consequence, the intermembranous fluid is more acidic than the cytosol. This has been confirmed experimentally by investigators like Cortese et al (1992), who shoved sensitive fluorescent probes up in there, and measured a pH difference of 0.4-0.5 points. Additionally, this is where all the reactive oxygen species end up. The cristae junctions are thought to prevent these contents from spilling into the cytosol.
The inner membrane is also a lipid bilayer, but it contains a substantial amount of cardiolipin which makes it even more ion-impermeable than normal. Ion and protein entry in and out of the mitochondrial matrix is tightly controlled by transport proteins in this membrane. Additionally, this membrane is the site of embedded electron transport chain enzymes. Because of the highly selective controlled permeability of this membrane, electrochemical gradients of a significant scale are established between the intercristal space and the matrix. Authoritative sources (eg. Lane, 2018) give −200 mV as the value for the transmembrane potential of the inner mitochondrial membrane.
The mitochondrial matrix is a viscous jelly-like mass filled with closely-packed protein molecules, which appears as a heterogeneous reticular electron-dense compartment when examined by an electron microscope. The protein content has been estimated as 560g/L (Srere et al, 1980), far in excess of the protein content of the cytosol. One of the benefits of multiple cristae is to decrease the distance between inner matrix components and the inner membrane surface; in such a densely packed material the rate of diffusion even for small molecules will be slow.
Though in the pre-genetic era it would have been mad to say so in polite society, these days it is widely accepted that the origins of mitochondria are probably extracellular. These organelles essentially represent a highly modified domesticated form of previously independent wild bacteria, which - somewhere at the dawn of time, and for reasons unknown - developed a co-dependent relationship with protists, and thereby permitted multicellular eukaryotic life as we know it.
A better justice to this topic is done by Gray, Burger & Lang (2001). Without expanding overmuch on what is clearly a fascinating topic, it will suffice to say that the circumstantial evidence for this is abundant, even though we have no fossil record of pre-symbiotic mitochondria-less protozoa. There are some modern species of protozoa which are genetically close to the "root" of the Eukaria tree, which have no mitochondria and possess a karyomastigont, a unique group of cytoskeletal structures thought to be a developmental remnant carried over from pre-Eukarya motility systems. It is not clear when the endosymbiosis occurred, but it must have been no earlier than about 2 billion years ago, because prior to that there would not have been enough oxygen around to justify the existence of aerobes.
So, what sort of bacteria would mitochondria have been, when they were "free-range"? According to Gray et al (1999), on the basis of mitochondrial DNA analysis, the closest living relative of the these "proto-mitochondria" (let's call them Mitochondromonas) would have been from the family of α-proteobacteria, a group which includes a number of obligate intracellular parasites such as Rickettsia, Anaplasma and Ehrlichia. They would have been weakly Gram-negative or Gram-intermediate like spirochetes, and obviously capable of (or dependent on) aerobic metabolism. You would have used macrolides or tetracyclines to kill them.
It actually serves the mind better to treat these organelles as something apart from the general milieu of the cell's interior, as housepets rather than furniture. They certainly have more personality than the boring Golgi apparatus or endoplasmic reticulum, which behave more like toasters or couch cushions. Mitochondria are motile and self-replicating; they roam the cell are like grazing cows, and undergo binary fission when they need to expand their numbers. They collect in herds near areas of highest ATP demand, undergo rapid shape changes in response to stress, and are capable of rapid purposeful movement in one direction (McCarron et al report short bursts of motion up to 1000 nm/s).
Mitochondria are home to the main oxygen-dependent processes of ATP synthesis from glucose and lipid metabolism. In summary:
A detailed discussion of mitochondrial metabolic enzymes falls outside of the remit of this chapter. It's covered thoroughly by Vakifahmetoglu-Norberg et al (2017) and Rolfe & Brown (1997). For the sake of keeping things interesting, it is worth noting that at a normal resting state, 90% of the oxygen used by mammalian cells is used by the mitochondria, but this use is relatively inefficient.
Without digressing extensively into the territory of metabolic biochemistry, it is important to point out that not all of the energy produced by the metabolic combustion of substrates such as glucose ends up stored chemically in ATP. The oxidation of 1 mol of glucose yields 686 kcal of energy, whereas the energy stored in the first phosphoanhydride bonds of 36 moles of ATP is 262.8 kcal. The process of "combusting" metabolic substrates is therefore inherently inefficient, and the remaining energy (61.7% of the total) is released as heat. Thus, the complete oxidation of 1 mol of glucose produces around 423 kcal of heat. This heat is liberated before any actual ATP is created: the oxidation of metabolic fuel first produces reducing agents (adenine dinucleotides NADH and FADH), which are in turn used to donate electrons to the electron transport chain.
The electron transport chain uses these electrons to pump protons out of the matrix, and then lets protons leak back into the matrix through ATP synthase in order to produce ATP. Thus, if all proton movement back into the matrix was via these ATP-producing processes, the electron transport chain would theoretically be 100% efficient, i.e. oxygen use would be stoichiometrically 100% coupled to ATP synthesis. However, in mammals, 80% of the consumed oxygen is used in ATP synthesis, and 20% is "uncoupled" from ATP synthesis by what most authors describe as a "mitochondrial proton leak" (Jastroch et al, 2010; Brand et al, 1994).
This "proton leak" is due to the no-zero permeability of the inner membrane to hydrogen ions which must be constantly counteracted by the actions of the electron transport chain. It is demonstrated by elegant experiments where oxygen consumption of cells is measured at rest, then with ATP synthesis disabled by oligomycin, and then with the electron transport chain disabled by cyanide (Challoner, 1968). Basically, if the inner membrane was totally proton-impermeable and the oxygen use of a mitochondrion was 100% coupled to ATP synthesis, then all proton pumping should have a linear relationship with the rate of ATP synthesis, and should cease completely when you disable ATP synthase with oligomycin. But it doesn't. A slow leak of protons back into the matrix persists, which is clearly unrelated to the intentional ATP-producing proton leak.
This is important. The leak of protons back into the matrix, when it bypasses ATP synthase, uncouples ATP production from the metabolic combustion of glucose and oxygen. You can combust as much glucose and oxygen as you like, but if all the protons end up leaking back into the matrix, all of that combustion will not produce any ATP: its results will be only heat, water and CO2. This is the key feature which permits mitochondria to regulate heat production. By introducing more porin proteins to facilitate a greater proton leak, one may increase thermogenesis and burn a ton of metabolic substrates without producing very much ATP. One such porin protein (thermogenin, also called "uncoupling protein 1") is found in the mitochondria of the brown adipose tissue of mammals.
In summary, mitochondrial metabolism produces heat because the reactions of substrate oxidation are exothermic reactions. By regulating the amount of proton leak, one can keep the rate of ATP production stable, but increase or decrease the rate of substrate oxidation, thereby changing the rate of heat production. By some accounts ( Chrétien et al, 2018) the interior of mitochondria may get as hot as 50°C, i.e up to 12 degrees hotter than the surrounding cytosol. This is thought to be the main cause of mammalian endothermy; we keep warm by having about five times as many mitochondria per cell as compared to cold-blooded animals, so we can waste our metabolic fuel through uncoupled oxidation and still have enough mitochondria to produce ATP.
Of course, we mammals do not own the intellectual property rights to mitochondrial thermogenesis. By working hard enough and burning enough glucose, ordinarily cold-blooded species can produce a surprising amount of heat. Church et al (1959) inserted temperature probes into the flight muscles of locusts and found that they heat up rapidly during flight (up to 35-40°C!) and ultimately produce so much heat that the insect has to stop and rest, lacking any specific mechanisms to cool itself. Nor are we a unique class for uncoupling our oxidative phosphorylation: even plants can do it provided they express the appropriate uncoupling proteins, eg. the skunk cabbage Symplocarpus foetidus which heats itself so that its volatile aromatic molecules can stink more hideously even in ambient cold conditions (Ito, 1999).
The electron transport chain, predictably, is in the business of shuttling electrons around. One of the natural byproducts of this industrial process is the production of reactive oxygen species. Murphy (2009) and Andreyev et al (2005) offer excellent overviews of this subject. In summary, the production of reactive oxygen species is mainly due to the escape of electrons from various metabolic processes, which in turn causes a reduction of molecular O2 to O2−. Under normal conditions (i.e. when mitochondria are busily producing ATP and there is plenty of metabolic substrate) the rate of O2− production is probably quite low, but under certain experimental or pathological conditions, up to 1-2% of total O2 consumed will be converted to O2−. These conditions include situations where the electron transport chain has been damaged by some sort of chemical adversary, during apoptosis, or in conditions of ischaemia. The H2O2 which is produced from O2− then goes on to act not only as a biochemical terrorist (vandalising cell membranes by peroxidation of lipids, etc) but also as a secondary messenger (Starkov, 2008).
This topic is complex; it is covered very well by Wang & Youle (2009) and to discuss it in any great detail would probably exceed the needs of this summary. In short, mitochondria participate in the final stages of apoptosis (beyond the "point of no return"). The intermembrane space is full of proapoptotic proteins like cytochrome C, but these are too large to get through the outer membrane pores, and therefore are safely sequestered. When an apoptotic trigger (and there are several possible triggers) causes an increase in the permeability of the outer membrane to proteins, cytochrome C spills into the cytosol and activates caspases- a group of cysteine proteases that dismantle cell contents in a rampage of proteolysis. This can be a cascade of events triggered by the mitochondria themselves, or it can be triggered by other factors (including extracellular signals) in which case the mitochondria act as amplifiers.
The cells use calcium as a secondary messenger, and so it would make sense the cytosolic calcium concentrations be kept intentionally low. It is present therein nanomolar concentrations, usually. Apart from keeping calcium out of the cell altogether using membrane permeability, the other aspect of this homeostasis is the rapid sequestration of intracellular calcium in mitochondria. Vandecasteele (2001) describes these processes very well. In summary, even at very low cytosolic calcium concentrations (300 nM) mitochondria suck up calcium and store it in their matrix. They then contribute to the shape of intracellular calcium concentration curves and play a role in cellular events which rely on increases in intracellular calcium - for example. smooth muscle contraction (Roux, 2009)
Mitochondria are the source of a lot of membrane lipids, particularly phosphatidylethanolamine and phosphatidylcholine. The total importance of this lipid synthesis for the rest of the cell is unclear (i.e. it is not like they make most of it) but they certainly manage their own membrane repair and synthesis internally. They also make their own cardiolipin locally, for their inner membranes. According to Mesmin (2016), they get their lipid substrate directly from the endoplasmic reticulum via outer membrane contact sites. If they do not get enough phosphatidic acid from external sources, lipid synthesis is crippled, and this is demonstrated by morphological changes in the mitochondria (i.e. they have fewer cristae). If phosphatidylethanolamine synthesis is completely knocked out genetically, the knockout mice never make it past day 10 of embryonic development (Steenbergen et al, 2005).
Synthesis of haem (heme?) is a long multi-stage process which takes place throughout the mitochondria and the cytosol. Chung et al (2012) expand on the details which exceed this short summary. In short, the role of mitochondria is to start the process by condensing glycine with succinyl-CoA, exporting the product (δ-aminolevulinic acid) out into the cytosol. A few more steps later, another intermediate (coproporphyrinogen III) is imported back into the mitochondria, where haem synthesis is then completed (including the addition of the actual iron). Iron is stored in mitochondria as mitochondrial ferritin, which is separate from cytosolic ferritin, and which (probably) does not represent an important form of body iron stores, because normal iron storage organs have so little mitochondrial ferritin in them (Drysdale et al, 2002)
Without going on a long digression about the urea cycle, it will suffice to say that it passes through the mitochondria but starts and finishes in the cytosol. The one main reaction in the mitochondria is a middle step, where carbamoyl phosphate (a product of ammonia and L-arginine) imported from the cytosol and converted into L-citrulline by ornithine transcarbamoylase. A deficiency of this enzyme gives rise to the sort of high ammonia that might require sodium benzoate therapy, and is a genetic disorder - but not a mitochondrial disease per se, as the enzyme is encoded by the "host" DNA rather than in mitochondrial DNA.
Mitochondria have diverse roles in immune system function (a good reference is the 2018 article by Lim et al). They contribute something little to most aspects of it. They produce reactive oxygen species both as an offensive capability or for use as secondary messengers; they participate in the innate immune response against viral infection; their role in cell injury extends from activating signalling pathways for repair to triggering apoptosis. It would be far-fetched to describe their role in immunity as central, but it is worth mentioning.
Mitochondria possess some DNA of their own. This encodes only some of the normal mitochondrial proteins, revealing the extent to which they are "domesticated" by eukaryotic cells. For instance, only some of the electron chain enzymes (and in some cases, only some subunits from a given enzyme) are encoded in the mitochondrial DNA, and the others come from the "host" nucleus. Stewart & Chinnery (2015) give a decent overview of this topic.
The mitochondrial genome is polyploid: instead of one copy of a chromosome from mum and another from dad like normal God-fearing nuclear DNA, the mitochondria contain multiple 10 identical copies of the same DNA molecule. These molecules are circular, and suffer from a much faster mutation rate than nuclear DNA (there is something like 10 times greater rate of point mutations), probably due to their proximity nearmost the boiling cauldrons of reactive oxygen species. This leads to heteroplasmy: of the ten or so copies of the mitochondrial genome, all will probably end up with completely different mutations, and so each copy of the DNA will be slightly different.
The college comments to Question 18 from the first paper of 2016 suggest vaguely that answers of candidates who mentioned the maternal inheritance were somehow superior to the others. It is worth noting that, though the spermatozoa have mitochondria of their own, the total mitochondrial contribution of the oocyte completely dwarfs the sperm (by a factor of two) and so the inheritance of mitochondrial DNA is mainly maternal.
They don't just make ATP and haem- these things also constantly produce proteins; or if not whole proteins, then integral subunits of proteins. Fox (2012) covers this nicely. Basically, the mitochondrial DNA encodes mainly for electron transport chain proteins (and not actually for all of them- some subunits are derived from nuclear DNA). Eight of the subunit protein complexes are produced in the mitochondrial matrix, by intramitochondrial ribosomes. Everything else (Fox quotes about 1000 different proteins) is imported from the cytosol.