Background

The chloroplast is the organelle that conducts photosynthesis in plants and algae. A major compartment of the chloroplast is the thylakoid lumen, which is enclosed by the thylakoid membrane. The Arabidopsis lumen proteome consists of at least 80 proteins based on mass spectrometry analyses (Peltier et al., 2002; Schubert et al., 2002) and up to 127 proteins according to lumenal targeting peptide prediction (Almagro Armenteros et al., 2019). All lumenal proteins characterized so far are nuclear-encoded and post-translationally transported into chloroplasts [see for reviews on lumenal proteins targeting (Albiniak et al., 2012) and function (Kieselbach and Schröder, 2003; Järvi et al., 2013)]. Lumenal proteins support and modulate photosynthetic activity directly or indirectly (examples of proteins are given in parenthesis), with known function in electron transport (plastocyanin PC, photosystem I subunit PsaN, and photosystem II subunit PsbO), in protein processing (C-terminal processing protease ctpA), assembly (immunophilins), and degradation (Deg proteases), in redox regulation (membrane protein with lumen thioredoxin domain HCF164) and in photoprotection (violaxanthin de-epoxidase and lipocalin in the plastid LCNP). However, the function of several lumenal proteins remains to be elucidated (some of which are Psb-like and pentapeptide repeat-containing proteins) and their regulation is not fully understood. The activity, stability, and distribution in the lumen compartment as soluble or membrane-associated proteins, or protein–protein interactions can be regulated by post-translational modifications such as phosphorylation or N-terminal acetylation (Gollan et al., 2021), or through redox modification or disulfide bond modulation, e.g., by lumen thiol oxidoreductase 1 of the lumenal domain of the kinase STN7 (Wu et al., 2021; for reviews, see Buchanan and Luan, 2005; Kang and Wang, 2016).

The major challenge when performing studies on the lumen proteome is the balance between purity and the quantity of protein needed for further analysis. Indeed, lumenal proteins are lowly abundant compared to thylakoid membrane protein light harvesting complex LHCII and stromal Rubisco [which represent more than 50% of total chloroplast protein content (Hall et al., 2011)]. Fractionation of the chloroplast and isolation of lumenal proteins thus enable their detection and accurate quantification by working within the dynamic range of detection, for example in immunoblot assays (the dynamic range is the lowest to highest concentration of a given protein that can be reliably detected). Then, the accumulation of a protein of interest can be investigated in different growth conditions or stress treatments and in mutants; the lumen proteome has been analyzed from 6–8-week-old plants grown under standard growth conditions (i.e., 120 μmol photons m^-2·s^-1, 21 °C, 8:16 h light/dark) (Peltier et al., 2002; Schubert et al., 2002) and also comparing the end of the dark vs. light period (Granlund et al., 2009) or after cold acclimation (Goulas et al., 2006). In addition, localization and distribution of the proteins, as well as assessment of protein complexes and interaction, can be inferred. Recently, Gollan et al. (2021) refined further localization studies to distinguish free lumenal proteins from membrane-associated ones, using Yeda press to isolate soluble lumenal proteins and subsequent washes with urea and salt to release inner membrane-associated proteins.

To isolate lumen proteins from plants, first the chloroplasts are obtained by homogenization of leaves followed by centrifugation; then, chloroplasts are lysed by osmotic shock, and thylakoid membranes are collected by centrifugation and washed to remove stromal proteins and peripheral membrane proteins. Finally, thylakoid membranes are ruptured to release lumenal proteins with a Yeda press (Kieselbach et al., 1998) or a sonicator (Peltier et al., 2000; Levesque-Tremblay et al., 2009); alternatively, thylakoid membranes are solubilized to release lumenal proteins using a detergent such as Triton X-114 followed by phase partitioning at 30–37 °C (Bricker et al., 2001), 0.04% Triton X-100 (McKinnon et al., 2020), or 0.05% n-Dodecyl β-D-maltoside (Chang et al., 2021).

Here, we report a simple procedure by which a lumenal fraction can be isolated in pure form from the thylakoids in Arabidopsis by sonication, which we adapted from previously described methods (Peltier et al., 2000; Levesque-Tremblay et al., 2009). Lumen isolation by sonication is of interest for several reasons: 1) the Yeda press is no longer commercially available (Hall et al., 2011), so unless already existing in the laboratory, access to one is limiting; 2) operating a sonicator is easier and faster than using the Yeda press for more than two samples (due to faster washing time of instruments parts); and 3) use of detergent can be costly and can disrupt native interactions. Also, the yield with sonication is comparable to using the Yeda press (15–30 μg of lumenal proteins per gram of leaf) and a decreased isolation time (from six to three hours for two samples) is valuable to limit proteolysis and preserve more native states of proteins and complexes during the lumen isolation. All methods present the disadvantage of stromal membrane–associated protein contaminants in the lumenal fraction that are not removed during the washes of thylakoid membranes (Bricker et al., 2001) (for example, see Figure 2, ATPb); the presence of these contaminants could be decreased by using a shorter sonication time and/or decreased power (Peltier et al., 2000). Overall, the sonication method is easy to implement, saves time, plant material, and cost, and is suitable for most studies—unless a large quantity of proteins is required, in which case the Yeda press method should be favored.