Identifying specific protein interactors and spatially or temporally restricted local proteomes contributes significantly to our understanding of cellular processes in virtually all aspects of life. Obtaining such data is challenging, especially when the protein, cell type or event of interest is rare. In recent years, different proximity labeling techniques have been developed that have greatly improved our ability to tackle these questions. However, while effective in mammalian systems, use in plants has been extremely limited due to technical challenges. Recent technological improvements in the form of two highly active versions of the biotin ligase BirA* (TurboID and miniTurboID), prompted us to test this new system on two challenging but widely asked questions in plants: what are interaction partners of low-abundant proteins and what are organellar proteomes in rare and transient cell types. To address the first question, we used the transcription factor FAMA as a test case. FAMA is a master regulator of guard cell development and promotes terminal differentiation of the guard cell precursor by both activating and repressing hundreds of genes. FAMA-expressing young guard cells are rare and restricted to the epidermis of developing aerial tissues, which makes them a good model system to test TurboID applicability under material-limiting conditions. For this experiment, young Arabidopsis seedlings expressing FAMA-TurboID were treated with biotin to label FAMA complexes. Col-0 wild type and seedlings expressing nuclear TurboID under the FAMA promoter were used as controls for unspecific binding of proteins to the beads and stochastic labeling of nuclear proteins, respectively. Biotinylated proteins were isolated by affinity purification with streptavidin-coupled beads and identified by LC-MS/MS. Analysis of proteins labeled by FAMA-TurboID fusions revealed known interactors of this late stomatal lineage specific transcription factor, as well as novel proteins that could explain its dual function as an activator and repressor. Comparison with proteins obtained from classical co-immunoprecipitation approaches with FAMA showed that proximity labeling is superior for identification of meaningful interaction partners of low-abundant proteins.