Thrombospondin-1 is a potent suppressor of T cell activation via its receptor CD47. However, the precise mechanism for this inhibition remains unclear. Because H2S is an endogenous potentiator of T cell activation and is necessary for full T cell activation, we hypothesized that thrombospondin-1 signaling through CD47 inhibits T cell activation by antagonizing H2S signaling. Primary T cells from thrombospondin-1 null mice were more sensitive to H2S-dependent activation assessed by proliferation and induction of interleukin-2 and CD69 mRNAs. Exogenous thrombospondin-1 inhibited H2S responses in wild type and thrombospondin-1 null T cells but enhanced the same responses in CD47 null T cells. Fibronectin, which shares integrin and glycosaminoglycan binding properties with thrombospondin-1 but not CD47 binding, did not inhibit H2S signaling. A CD47-binding peptide derived from thrombospondin-1 inhibited H2S-induced activation, whereas two other functional sequences from thrombospondin-1 enhanced H2S signaling. Therefore, engaging CD47 is necessary and sufficient for thrombospondin-1 to inhibit H2S-dependent T cell activation. H2S stimulated T cell activation by potentiating MEK-dependent ERK phosphorylation, and thrombospondin-1 inhibited this signaling in a CD47-dependent manner. Thrombospondin-1 also limited activation-dependent T cell expression of the H2S biosynthetic enzymes cystathionine beta-synthase and cystathionine gamma-lyase, thereby limiting the autocrine role of H2S in T cell activation. Thus, thrombospondin-1 signaling through CD47 is the first identified endogenous inhibitor of H2S signaling and constitutes a novel mechanism that negatively regulates T cell activation. Autophagy (macroautophagy), a cellular process of "self-eating", segregates damaged/aged organelles into vesicles, fuses with lysosomes, and enables recycling of the digested materials. The precise origin(s) of the autophagosome membrane is unclear and remains a critical but unanswered question. Endoplasmic reticulum, mitochondria, Golgi complex, and the plasma membrane have been proposed as the source of autophagosomal membranes. Using electron microscopy, immunogold labeling techniques, confocal microscopy, and flow cytometry we show that mitochondria can directly donate their membrane material to form autophagosomes. We expand upon earlier studies to show that mitochondria donate their membranes to form autophagosomes during basal and drug-induced autophagy. Moreover, electron microscopy and immunogold labeling studies show the first physical evidence of mitochondria forming continuous structures with LC3-labeled autophagosomes. The mitochondria forming these structures also stain positive for parkin, indicating that these mitochondrial-formed autophagosomes represent a novel mechanism of parkin-associated mitophagy. With the on-going debate regarding autophagosomal membrane origin, this report demonstrates that mitochondria can donate membrane materials to form autophagosomes. These structures may also represent a novel form of mitophagy where the mitochondria contribute to the formation of autophagosomes. This novel form of parkin-associated mitophagy may be a more efficient bio-energetic process compared with de novo biosynthesis of a new membrane, particularly if the membrane is obtained, at least partly, from the organelle being targeted for later degradation in the mature autolysosome.