One point five billion years ago, that's billion, with a B, the very first multicellular organisms developed. At this time, eukaryotic organisms developed the endosymbiotic relationships with various types of bacteria to form such important organelles as the mitochondria and plastids (Ochoa de Alda et al., 2014). Plastids originated as cyanobacteria, and conferred on their new hosts the ability use photosynthesis to generate sugars from environmental carbon dioxide, the first choroplasts. Eukaryotes which could integrate these bacteria into their own cell physiology as biochemical factories were given a leg up in the endless evolutionary arms race, and so all the various eukaryotic groups diversified and thrived. Three major groupings served as the roots of all future photosynthetic eukaryotes: the Rhodophytes, or 'red algae', which utilize phycobiliproteins as accessory pigments, giving them their red color (and it is the absence of these phycobiliproteins which makes the green chlorophyll of modern green algae and plants so visible); the Chlorophytes and Charophytes, the latter of which all organisms we commonly think of plants evolved from, and is commonly grouped paraphyletically (excluding 'plants') with chlorphyta as the 'green algae'; finally there are the Glaucophytes, which are poorly studied, and may possibly be true origin point from which all these modern groups have evolved. These three clusters together are a group called Archaeplastida (or in the same sense, Plantae)(Ball et al., 2011).
But this is not the end of the story of endosymbionts. These single-celled ancestors of these three algal group in turn have been integrated into additional groups, many of which have no common names, resulting in various groups of so-called secondary endosymbionts, wherein the original host is consumed, and develops a symbiotic relationship with its host in turn. The most notable example of this in a group we typically picture as 'plants' (which, if we include the Charo- and Chlorophyta, we can name Viridiplantae). Another major group are the brown algae, which includes the large kelp of the forests of the Sargassum sea, and in fact is not particularly closely related to the Viridiplantae; it acquired its chloroplasts long before it achieved multicellularity, from an ancestral rhodophyte, so plants as we know them (Viridiplantae, the green algaes and land plants) weren't even vaguely involved in this group of macroscopic photosynthesizers (McFadden, 2001). It continues on like this, with numerous other single-celled eukaryote lineages having arisen where red or green algae acquired through endosymbiosis. In contrast to the Archaeplastida, there is a group simpled called SAR (short for its three constituent groups, Stramenopiles, Alveolates, and Rhizaria), which contains a number of secondary endosymbiote events, including the brown algae in Stramenopiles, Dinoflagellates in Alveolates, which includes members responsible for the 'Red Tides' that make seafood toxic in certain seasons and conditions, and various amoeboid organisms in Rhizaria, including Foraminiferans, which will come up again later (Sina et al., 2012).
So we find, over the past one point five billion years, a huge host of diverse eukaryotes have arisen which have permanently integrated ancestral cyanobacteria into their physiology, to the extent that neither organism can survive without the other, and to the extent that they have ceased to be distinct organisms.
...but there are other ways to skin a cat. All that time ago, eukaryotic cells consumed cyanobacteria, and eventually, through the complexities of the numerous components that compose the evolutionary process, they transitioned, over millions of years, from simply consuming the cyanobacteria, to storing them to using them. As Rome was not built in a day, endosymbiosis was not simply something that happened in a flick of a switch.
And it is here where we come across the concept of Kleptoplasty
Trench (1975) best described Kleptoplasty as “the phenomenon where, under natural circumstances an animal acquires intracellularly and retains undamaged plant chloroplasts, free from other associated plant organelles and cytoplasm. Such chloroplasts show sustained active photosynthesis, and photosynthetic products become available to and are utilized by the animal host.” The only modification I would make to this wonderfully concise definition is to expand it to include non-animal organisms. Kleptoplasty has been observed in ciliates (Alveolates)(Stoecker, 1991), foraminiferans (Rhizaria) (Lee, 1998; Chai and Lee, 2000; Bernhard, 2003), and members of the sacoglossan sea slug family Plankobranchidae (Williams and Walker, 1999; Händeler et al., 2009).
In Kleptoplasty, the association between the chloroplasts and the host are not multi-generational. When a foraminiferan, or a ciliate, or a sacoglossan slug reproduces, their offspring do not inherit the chloroplasts of the mother. They must reacquire chloroplasts, and in many instances must reacquire them over and over again, as they can lack the machinery needed to keep them maintained for their lifespans. Their source of chloroplasts, or kleptoplasts, as they're called once stolen, varies from group to group. Foraminiferans consume diatoms (the former from the R, the latter from the S in SAR), ciliates use cryptophytes (which belong to a group whose relationship to other eukaryotes is currently unclear) (Johnson et al., 2004), and the Plankobranhidae, a group of sea slugs, use various green algae, and in some instances brown algae (which frankly I found to be a pleasant if unexpected surprise). Meanwhile, in something parallel to normal kleptoplasty Nudibranches even consume and store certain dinoflaggellates (zooxanthellae) which have endosymbiotic plastids, and use the wholesale organisms as supplementary sources of metabolic energy (Sutton and Hoegh-Guldberg, 1990).
Sacoglossan sea slugs, colloquially called sap-sucking sea slugs, feed on algae by using their radula, a tooth surface that in sacoglossans is used like simple drill, to penetrate the membranes, and then siphon out the cytoplasm of the algal cells. In Plankobranchidae, various species extract chloroplasts from species-specific prey algae, and rather than simply digest them along with the other cytoplasmic contents, sequester them within their gut diverticulae for various lengths of time. Of these, Elysia chlorotica has the most robust displays of kleptoplasty and is able to survive its entire 9-11 month lifespan under starved conditions using energy generated solely by their sequestered plastids (Pierce et al., 1996, Green et al., 2000; Rumpho et al., 2001). It is worth noting that this behavior is only undertaken under starvation conditions; if provided with food, the slugs will preferably continue to feed than depend on their kleptoplasts (Middlebrooks et al., 2011). E. chlorotica is the most thoroughly studied of the various kleptoplast sacoglosssans, along with its sole food source, Vaucheria litorea, a stramenopiles. Their chloroplasts, as secondary endosymbionts, were originally attained through intake of primary endosymbiont red algae (Cavalier-Smith, 2000). V. litorea chloroplasts are contained in four membranes, the chloroplast double envelope, the periplastid membrane and chloroplast endoplasmic reticulum; at some point during their internalization in E. chlorotica, the outermost two membranes are removed.
Chloroplasts only carry a small fraction of their ancestral genetic code which would be needed to maintain them (1-5%), and the required maintenance genes are now carried within the genome of their hosts (Martin and Herrmann, 1998). Most chloroplasts can only perform prolonged photosynthetic activity for several hours following isolation from their hosts(Green et al., 2005), including their isolation that occurs when sequestered in the gut cells of Plankobranchids. However, chloroplasts taken from algae (as opposed to land plants) such as V. litorea are unusually robust and continue to function for several days (Green et al., 2005), which could contribute to the longevity of kleptoplasts within E. chlorotica (Trench et al., 1973).
It remains disputed whether or not kleptoplasty is truly occuring in sacoglossans; some studies have shown that slugs reared with or without sunlight both can live for long periods of times without food, which may suggest that kleptoplasts may not be playing a major metabolic roll. If the chloroplasts are being maintained, however, the means by which E. chlorotica maintains the acquired plastids without the necessary maintenance genes found in the host alga continues to be a major topic of study. Horizontally transferred genes from the original host plant to the slug genome has been proposed and disputed (Wägele et al., 2011, Pelletreau et al., 2011), however a growing number of algal genes have been found expressed from the E. chlorotica genome (Pierce et al., 2007, 2009; Rumpho et al., 2008, 2009; Schwartz et al., 2010) which would help to explain E. chlorotica's success in kleptoplasty compared to less robust kleptoplasty in other Plankobranchids. Future work continues to further identify evidence of horizontal gene transfer from V. litorea to E. chlorotica and elaborate the nature of plastid maintenance.