Algal chloroplast pyrenoids: Evidence for convergent evolution

Almost all aquatic eukaryotic algae have pyrenoids, microcompartments within the chloroplast containing the enzyme ribulose 1, 5-bisphosphate carboxylase/oxygenase (Rubisco). However, the shapes and locations of the pyrenoids within chloroplasts vary widely in different algal taxa (1, 2), as the chloroplasts in the different alga arose from different endosymbiotic associations. In the green alga Chlamydomonas reinhardtii, it has been shown that intact pyrenoids are essential for high rates of photosynthesis (3). However, beyond the well-studied Chlamydomonas system and some significant work in diatoms, pyrenoids in other algal lineages have not been investigated. In this issue, Moromizato et al. present a proteomics study of the pyrenoid from the chlorarachinophyte Amorphochlora amoebiformis (4). This is a welcome study as it significantly expands our understanding of the pyrenoid in eukaryotic algae beyond the few model systems that have been studied to date. The work of Moromizato et al. adds another algal lineage, notably, one that went through two endosymbiotic events.

Up to one-third of global CO₂ assimilation is done by eukaryotic algae containing pyrenoids (2). This contrasts with terrestrial plants where Rubisco is dispersed throughout the chloroplast stroma. Aquatic eukaryotic algae concentrate carbon dioxide using a CO₂ concentrating mechanism (CCM) internally to enhance CO₂ assimilation. From work in Chlamydomonas, the packaging of Rubisco within the pyrenoid is thought to be essential for efficient CO₂ assimilation. Therefore, the pyrenoid is thought to be required for efficient photosynthesis and the success of the organism. However, the morphology of pyrenoids varies greatly between algal lineages. One reason pyrenoid structure varies so much is that the chloroplasts of different algal lineages arose independently (5, 6). Many algae such as Rhodophytes, Chlorophytes (includes Chlamydomonas), and Glaucophytes acquired their chloroplasts directly from cyanobacterium endosymbiosis. However, most photoautorophic eukaryotes other than the above groups acquired their chloroplasts by a secondary endosymbiosis of Rhodophyte and Chlorophyte algae. The former includes a large group of algae: the Stramenopiles (diatoms belong to this group), Cryptophytes, Haptophytes, and a part of Alveolata; and the latter includes some groups of Excavata and Rhizaria (chlorarachinophytes belong to this group). Chloroplasts that arise from primary endosymbiosis have two envelope membranes surrounding the chloroplast stroma. Algae that have undergone secondary endosymbiosis, like diatoms and chlorarachinophytes, have four membranes surrounding the stroma (2). In addition, the position of the pyrenoid and thylakoid membranes varies between algae. In Chlamydomonas, the pyrenoid is in the center of the chloroplast surrounded by a starch sheath, and modified thylakoid membranes, known as tubules, penetrate the pyrenoid (Fig. 1). In diatoms, in contrast, there is no starch sheath, and the layers of specific thylakoid membranes dissect the axis of the disc-shaped pyrenoid: These membranes are denoted as the pyrenoid-penetrating thylakoid (Fig. 1). In A. amoebiformis, the pyrenoid is pinched off to one end of the chloroplast and a set of interplastidic membranes bisect it (Fig. 1). Are these diverse types of pyrenoids all essential to the operation of a CCM?

Fig. 1.

Pyrenoids of Chlamydomonas, Phaeodactylum, and Amorphochlora. Top row—sketches of the algal cells showing pyrenoid location, shape, and position within chloroplasts. The cell nucleus is indicated (N) and the membrane complexity highlights secondary endosymbiotic events giving rise to Phaeodactylum and Amorphochlora. In all cases, pyrenoids interact closely with membranes, while that of Chlamydomonas is surrounded by a starch sheath and Amorphophora carries a capping vesicle. Bottom row—magnification showing details of pyrenoid structure and elements that are common to all three pyrenoids. Rubisco is depicted in blue. The linker proteins are shown as string and balls connecting Rubisco. The linker proteins are EPYC1 in Chlamydomonas, PYCO1 in Phaeodactylum and PPAP28 in Amorphochlora. Each pyrenoid is penetrated by a set of membranes, each containing a carbonic anhydrase indicated as either α-CA or θ-CA. The carbonic anhydrase converts bicarbonate (HCO₃⁻) to carbon dioxide (CO₂) for Rubisco. The bicarbonate is brought into the membrane by a bestrophin-like protein (BST) in Chlamydomonas and Phaeodactylum. In Amorphochlora, the protein that is responsible for bicarbonate transport is not known, although a bestrophin-like protein (PPAP30) was found in the pyrenoid proteome. CV stands for capping vesicle.

In PNAS, Moromizato et al. present a proteomics study of the pyrenoid from the chlorarachinophyte Amorphochlora amoebiformis.

A second reason pyrenoid structures are so varied is that the algal lineages likely developed pyrenoids independently (2, 5). It is thought that the endosymbiotic events that led to the algal lineages occurred over a billion years ago when the CO₂ concentration in Earth’s atmosphere was much higher than today, high enough to saturate Rubisco and allowing for good rates of photosynthesis (3, 6, 7). When the atmospheric CO₂ level dropped over the past 300 My, algae had to adapt and most developed CCMs to increase the CO₂ concentration for Rubisco. This means that each algal lineage developed its CCM independently. This new work by Moromizato et al. supports the hypothesis that pyrenoids are physiologically accomplishing the same task of concentrating CO₂ for Rubisco even if they are morphologically distinct and are using different components.

Moromizato et al. (4) started their research by taking advantage of the structure of the A. amoebiformis pyrenoid and a fluorescently tagged Rubisco small subunit (RbcS), to develop a simple yet effective method to isolate clean pyrenoids. While the proteome of Chlamydomonas has been studied using isolated pyrenoids, Moromizato et al. perhaps facilitated by the ultrastructure of pyrenoids in this species, have described a very successful isolation of pyrenoids (4). They then did a proteomic analysis of the isolated pyrenoids, identifying over 150 proteins. These were ranked according to relative abundance, and the sub-cellular location of up to 30 of these was assessed through a Green Fluorescent Protein-tagging and microscopy approach. Further, they then raised antibodies to a number of these proteins, enabling them to obtain even greater resolution for the placement of the proteins within the pyrenoid using immunogold localization.

Their findings support the hypothesis that different algal lineages adapted to decreasing CO₂ levels independently but often ended up with similar CCMs. In Chlamydomonas, transiently high concentrations of CO₂ are generated by conversion of bicarbonate (HCO₃⁻) to CO₂ in the acidic thylakoid lumen. The carbonic anhydrase that catalyzes this HCO₃⁻ to CO₂ conversion, CAH3 (an α-carbonic anhydrase), is found in the thylakoid tubules within the pyrenoid (Fig. 1) (8, 9). Thus, the CO₂ generated leaks out of the thylakoid tubules right where Rubisco is packaged. In diatoms, a θ-carbonic anhydrase is located in the lumen of the pyrenoid-penetrating thylakoids. In A. amoebiformis, the pyrenoid is invaginated by an inner pair of plastid membranes, and Moromizato et al. found that the θ-carbonic anhydrase PPAP12 is localized to inside the pyrenoid-penetrating periplastidal compartment. Thus, PPAP12 is positioned well to generate CO₂ for Rubisco in the pyrenoid. In Chlamydomonas and diatoms, the thylakoid lumen is acidic due to the transport of H⁺ into the thylakoid by the electron transport chain, directing the carbonic anhydrase reaction equilibrium heavily toward formation of CO₂ in this compartment. It is not clear whether the periplastidal compartment in A. amoebiformis is acidic, but recently it has been reported that the periplastidal compartment in diatoms is slightly acidic (10, 11) supporting the hypothesis that this system has evolved to elevate CO₂.

Other similarities include the fact that Rubisco is gathered into the pyrenoid through its association with a linker protein. In Chlamydomonas, this protein is EPYC1 (12), while in Phaeodactylum tricornutum, the linker protein PYCO1 carries out this role (13). In A. amoebiformis, the linker protein may be PPAP28, an intrinsically disordered protein with 90 short repeats, typical of linker proteins involved in Rubisco condensation (14). All three of these proteins appear to have different phylogenetic origins, which fits with the different origins of the chloroplasts of these algae. Thus, the condensation of Rubisco and the localization of carbonic anhydrase are similar in all three taxa.

In addition to differences across species in location and shape of the pyrenoid, the pyrenoids feature different carbonic anhydrases. The carbonic anhydrase found in the thylakoids of diatoms (15) and in the perpilastidal space of A. amoebiformis are θ-type, while the carbonic anhydrase used in Chlamydomonas is an α-type. The transporters used to bring HCO₃⁻ to the lumenal carbonic anhydrase may also be different. In Chlamydomonas and Phaeodactylum, bestrophins are responsible for HCO₃⁻ transport into the thylakoid lumen (Fig. 1) (16, 17). Moromizato et al. also detected bestrophin-like proteins in the A. amoebiformis pyrenoid proteome. However, it is not clear whether the bestrophins in A. amoebiformis are on the same plastid membranes that invaginate the pyrenoid or not. Another surprising finding by Moromizato et al. was the presence of a number of GTPases in the A. amoebiformis pyrenoid. GTPases have not been detected in other pyrenoid proteomes. It will be interesting to learn about their function.

These disparate traits emphasize a growing consensus for the convergent evolution of pyrenoids in diverse microalgae, underpinned by the common observation of Rubisco condensation and the association with membranes having an acidic interior and associated carbonic anhydrase. This recognition raises intriguing questions about the drivers of CCM evolution, which may arise through a catalytic advantage obtained through Rubisco condensation, acidification, and colocalized carbonic anhydrase activity (18). Moromizato et al. add to a growing understanding of the variety of CCM solutions in algal systems through high-quality analysis of pyrenoid proteins (19).