The wastefulness of photorespiration is probably a consequence of just two factors. Early in the evolution of photosynthesis there was a higher carbon dioxide to oxygen gas ratio in the ancient atmosphere. Indeed the atmosphere was likely anaerobic in the earliest times on Earth. Rubisco evolved its active site when oxygen was rare and carbon dioxide was common. Since ancient times, rubisco has not yet evolved a mechanism to discriminate between the two similar substrates (O=C=O and O=O). The reactions are similar too; the substrate is attached at a point along RuBP resulting in its splitting into organo-monophosphates. 3-phosphoglycerate is a common product of both reactions. So the difficulty of a protein to distinguish such similar molecules and to catalyze one reaction but not the other just has not happened yet. Photorespiration losses have not been intolerable either; the selection pressure is probably not severe in most environments. But in hot, dry, environments that are heavily-populated with plants, where the local carbon-dioxide content of the atmosphere is greatly reduced, selection should have resulted in a few adaptations to overcome photorespiration. Indeed it has!
It is no surprise that if O=C=O and O=O compete for the active site of rubisco, then mechanisms that would concentrate carbon dioxide around chloroplasts would evolve among plants competing for low carbon dioxide supplies in hot, dry climates (where it is already depleted!). While some unicellular algae and cyanobacteria apparently use some membrane pumps to concentrate carbon dioxide in their cells (at an ATP cost), terrestrial plants evolved a Calvin-cycle add-on cycle. While the Calvin cycle produces 3-phosphoglycerate (a three-carbon sugar-phosphate) as its fixation product, this add-on cycle produces oxaloacetate (a four-carbon sugar-phosphate) as its fixation product. Thus the Calvin cycle is sometimes called the C3 cycle; the add-on is usually called the C4 cycle. Other names for the add-on include C4 photosynthesis, and the Hatch-Slack cycle (in honor of its discoverers). This pathway is found in 16 monocot and dicot families and has apparently evolved as three different specific pathways and two fundamentally-different separations.
The C4 cycle is depicted here in an overview as occurring in two adjacent cells with material passing symplastically between them through plasmodesmata.
You can observe four basic stages of the C4 cycle: carbon fixation (carboxylation), transport, decarboxylation, and regeneration. The basic steps are now outlined.
In the C4 cycle, atmospheric carbon dioxide is taken into the fluid environment of the cells and enters into the typical bicarbonate equilibrium. The carbon is fixed (attached) to phosphoenolpyruvate by the enzyme phosphoenolpyruvate carboxylase. The product of this reaction is a four-carbon acid. PEP carboxylase is a relatively-recently evolved cytosolic enzyme. The product of this fixation reaction is oxaloacetate, a four-carbon acid that gives the cycle its C4 name.
Typically oxaloacetate is converted to a different 4-carbon acid for transport. In some species, NADP:malate dehydrogenase reduces the oxaloacetate to malate by using NADPH as its reducing power. In other species, aspartate aminotransferase converts oxaloacetate to aspartate (with the glutamate/α-ketoglutarate support shuttle). Whichever is produced, the 4-carbon acid is transported somewhere...in C4 plants this goes to an adjacent cell via plasmodesmata.
In CAM plants the C4 acid is transported into a vacuole.
Where the C4 acid goes specifically determines what happens next. In some C4 plants the acid ends up in the chloroplast where NADP:malic enzyme decarboxylates the acid. In other C4 plants the acid ends up in the mitochondrion where NAD:malic enzyme decarboxylates the acid. In yet other C4 plants, the acid ends up in the cytosol where phosphoenolpyruvate carboxykinase decarboxylates the acid. These three options are shown below:
You can also see that the product of the decarboxylation is usually pyruvate or phosphoenolpyruvate. This C3 acid is returned back through plasmodesmata, perhaps with some alterations. In some species alanine aminotransferase uses the glutamate/α-ketoglutarate shuttle to produce alanine as a return intermediate.
Finally, the returned C3 acid is converted back to PEP by means of pyruvate-orthophosphate dikinase. This step requires much energy input in the form of ATP. The PEP can now accept another carbon dioxide and shuttle it in again; the C4 cycle is complete.
The effect of all of this processing is to pump carbon dioxide from the atmosphere into the place where the Calvin cycle is trying to use Rubisco. This pump uses a lot of ATP, but it accomplishes the goal of dumping CO2 into the Calvin cycle area to "swamp out" photorespiration and favor the Calvin cycle. The enzymes in the C4 cycle are mostly common respiratory or glycolytic enzymes except for PEPcarboxylase. Apparently this enzyme is the one true innovation represented in the evolution of C4 biochemistry. One critical feature of PEPcarboxylase is that it uses bicarbonate rather than carbon dioxide in its active site...bicarbonate is sufficiently different from oxygen gas, that oxygen is not a competitive inhibitor (or worse!). A second feature of PEPcarboxylase is its high affinity for bicarbonate; this allows the enzyme to draw carbon dioxide into the cell through the bicarbonate equilibrium even at depeleted carbon dioxide levels in modern atmospheres in dry, hot habitats.
Other features of plants with C4 photosynthesis is the identity of the two cells involved. Just to remind us, here is the cross section of a "typical" C3 leaf.
As you can see mesophyll cells carry out photosynthesis and possess chloroplasts for that purpose. As you recall from our discussions of translocation, these mesophyll cells provide materials for translocation in the phloem. The bundle sheath cells are responsible for loading the phloem by active transport. Sugars move from mesophyll to bundle sheath and then into phloem.
In addition to evolution of PEP carboxylase, C4 plants have evolved the separation of the C4 carboxylation from the oxygen sensitive C3 carboxylation step into this pathway. The mesophyll cells carry out the C4 steps, and shuttle those organic acids through plasmodesmata into the bundle sheath cells, where the Calvin cycle is therefore improved in its competition with photorespiration by the extra carbon dioxide. The bundle sheath cells in C4 plants are enlarged and their chloroplasts have much higher chlorophyll levels (looking darker!) to drive the Calvin cycle...compared to their mesophyll cells that are using limited light reactions to drive mostly the PEP carboxylase and regeneration cycles. Here are views of a typical C4 plants: Zea mays, Flaveria, and Amaranthus tricolor.
In the top panel you see a light micrograph of corn leaf. This shows the typical Kranz (wreath) anatomy. Each vein is surrounded by bundle sheath cells with the intensely chlorophyll-rich chloroplasts. The mesophyll cells around those have fainter-looking chloroplasts and contact with gas spaces that communicate via stomata with the external atmosphere.
In the lower panel you see a on the left a leaf of flaveria that has been probed with an enzyme-linked nucleic acid to produce blue pigmentation where PEP carboxylase mRNA is present. Obviously this is hybridizing in the mesophyll. On the right is a leaf of amaranthus that has been probed with a radioactive nucleic acid to produce an autoradiogram where RubisCO mRNA is present. This is shown to be primarily in the bubble-like cells of bundle-sheath surrounding each vein.
Shortly after the discovery of the C4 cycle the crassulacean acid metabolism (CAM) pathway was dissected and elucidated as a C4 variant. This pathway is found in desert succulents and epiphytes. In the typical C4 cycle, the fixation reaction occurred in a mesophyll cell and the decarboxylation reaction occurred in a bundle sheath cell, in CAM plants the two reactions are separated temporally rather than spatially. The CAM cycle is shown below...notice how all the reactions occur in the same cell but at different times of the day.
As you can easily observe, the fixation step occurs at night with the guard cells open to receive carbon dioxide during the cool night. This process is driven by use of starch to make the PEP required for PEPcarboxylase activity. The malic acid is transported to the vauole and accumulates there at night. There is a strong pH change in the vacuole at night!
The next morning, the malic acid is transported back to the cytosol for the decarboxylation reaction. This floods the Calvin cycle with carbon dioxide while the guard cells are closed. The energy of course comes from the light reactions and accumulated starch is used the next night for another round of carbon fixation.
It is immediately obvious that CAM plants will be slow-growing. They do photosynthesis in the daytime (remember the light reactions and Calvin cycle can ONLY operate in the light!). Much of this gain is lost the next night by use of the starch. In a very clear way, CAM plants are "spinning their wheels" using a lot of energy to go almost nowhere. Desert plants indeed are slow growing, but they are very competitive in hot, dry environments. The advantage in water savings by having guard cells open at night and closed in the daytime should not be under-estimated! It is apparently worth the trade-off in slow growth rate.
Obviously, with every cell having both the Calvin cycle and the CAM pump, one might expect the CAM plant to be able to use the Calvin cycle alone if conditions allow. Indeed many CAM plants are facultative in that they function as CAM plants under hot, dry conditions but as C3 plants if well-watered and kept in cooler temperatures. Such plants grow at rapid rates under cooler, moister conditions, and grow dramatically slower under warmer, drier conditions. But they do not succumb to desiccation!
The other interesting observation one must make, then, is that in CAM plants the activity of PEP carboxylase must be controllable to shut down the wasteful C4 style fixation in the daytime. How this happens is shown below, illustrating the role of phosphorylation of proteins in terms of activation/inactivation cycles.