The N Terminus of Phosphodiesterase TbrPDEB1 of Trypanosoma brucei Contains the Signal for Integration into the Flagellar Skeleton

The precise subcellular localization of the components of the cyclic AMP (cAMP) signaling pathways is a crucial aspect of eukaryotic intracellular signaling. In the human pathogen Trypanosoma brucei, the strict control of cAMP levels by cAMP-specific phosphodiesterases is essential for parasite survival, both in cell culture and in the infected host. Among the five cyclic nucleotide phosphodiesterases identified in this organism, two closely related isoenzymes, T. brucei PDEB1 (TbrPDEB1) (PDEB1) and TbrPDEB2 (PDEB2) are predominantly responsible for the maintenance of cAMP levels. Despite their close sequence similarity, they are distinctly localized in the cell. PDEB1 is mostly located in the flagellum, where it forms an integral part of the flagellar skeleton. PDEB2 is mainly located in the cell body, and only a minor part of the protein localizes to the flagellum. The current study, using transfection of procyclic trypanosomes with green fluorescent protein (GFP) reporters, demonstrates that the N termini of the two enzymes are essential for determining their final subcellular localization. The first 70 amino acids of PDEB1 are sufficient to specifically direct a GFP reporter to the flagellum and to lead to its detergent-resistant integration into the flagellar skeleton. In contrast, the analogous region of PDEB2 causes the GFP reporter to reside predominantly in the cell body. Mutagenesis of selected residues in the N-terminal region of PDEB2 demonstrated that single amino acid changes are sufficient to redirect the reporter from a cell body location to stable integration into the flagellar skeleton.In eukaryotes, the ubiquitous second messenger cyclic AMP (cAMP) is generated from ATP by membrane-integral or by cytoplasmic, CO₂-regulated cyclases (35, 44). The cAMP signal is processed by a small group of receiver proteins, including the regulatory subunit of protein kinase A (28), cAMP-gated ion channels (4), and the guanine-nucleotide-exchange proteins EPAC1 and EPAC2 (39). The cAMP signal is terminated by the action of a family of cyclic nucleotide-specific phosphodiesterases (PDEs) (9). This paradigm is rather straightforward, involves a limited number of players, and is generally well understood, at least in mammalian cells. However, much less is known about how individual cAMP signals are temporally and spatially controlled. Since most eukaryotic adenylyl cyclases are integral membrane proteins, often restricted to specific membrane subdomains (10), cAMP signaling is usually initiated at the cell membrane (40). However, diffusion of cAMP away from its site of generation is rapid, with diffusion coefficients being about 400 μm²/s (8, 15, 29), translating into diffusion velocities of 30 to 40 μm/s. As a consequence, the signal would reach the center of the cell with a diameter of 3 μm within less than 50 ms and would rapidly saturate the entire cell. While regulation through fluctuating cellular levels of cAMP represents a valid paradigm of cAMP signaling, it has become clear that other, more localized modes of cAMP signaling must also exist. Several groups have shown that the cAMP response of a given cell can differ depending on what set of receptors activates the cyclase response (14, 30, 41, 42). Similarly, the cAMP response of endothelial cells depends on the subcellular site where the cAMP is produced. They tighten their barrier function when cAMP is produced by membrane-bound adenylyl cyclases but become more permeable when cAMP is produced in the cytoplasm (17, 45). The distinct subcellular localization of cAMP signals was experimentally demonstrated using an array of techniques (29, 40, 55, 56).Physically tethered PDEs might serve to confine newly synthesized cAMP to defined microdomains. Only cAMP-binding proteins that are localized within or extend into such microdomains would be able to receive the cAMP signal (17, 49). cAMP concentrations within such domains might rise and fall rapidly, reaching peak concentrations much more rapidly and locally far beyond the steady-state cAMP levels measured in whole-cell extracts. Such spatially organized, tethered PDEs can generate local sinks into which cAMP disappears (1, 23). This paradigm would allow the simultaneous presence of numerous local cAMP concentration gradients within a single cell, allowing great flexibility in signal generation and intracellular signal transmission. This concept is based on the distinct subcellular localization and physical association of PDEs with subcellular structures and on the existence of localized subcellular cAMP pools, for which there is extensive experimental support (3, 5, 13, 50, 52). Interestingly, PDEs localized in different subcellular regions may still be able to compensate for each other. Ablation of the cilium-specific PDE1C from the olfactory neurons in the mouse did not prolong response termination, as long as the cytoplasmic PDE4 in the cell body was still present (11).The unicellular eukaryote Trypanosoma brucei is the causative agent of human sleeping sickness in sub-Saharan Africa. It belongs to the large order of the kinetoplastida, which includes many medically and economically important pathogens of humans, their livestock, and their crops worldwide (27). Trypanosomes are very small cells (about 15 by 3 μm in diameter) that carry a single flagellum (10 by 0.5 μm). The volume of a procyclic trypanosome of strain 427 is (9.6 ± 0.8) × 10⁻¹⁴ liter (Markus Engstler, personal communication), with the flagellum representing about 15% of this. A signaling threshold concentration of 1 μM cAMP corresponds to just about 30,000 molecules of cAMP per cell. Given a diffusion coefficient of 400 μm²/s (29), unrestricted diffusion of cAMP would swamp the cell within 50 ms. Obviously, temporal and spatial control of cAMP signaling is crucial for T. brucei. Strategically located, physically tethered PDEs might thus play an important role in the architecture of the cAMP signaling pathways in T. brucei.The genomes of T. brucei and of other kinetoplastids, such as T. vivax, T. cruzi, Leishmania major, L. infantum, and L. braziliensis, all code for the same set of five cyclic nucleotide-specific PDEs (25, 53). In T. brucei, the genes for T. brucei PDEB1 (TbrPDEB1; subsequently termed PDEB1) and TbrPDEB2 (PDEB2) are tandemly arranged on chromosome 9 and code for two very similar cAMP-specific PDEs, each with two GAF (mammalian cyclic GMP-dependent PDEs, Anabaena adenylyl cyclases, Escherichia coli FhlA) domains (21) in their N-terminal regions (38, 57). These two PDEs were also studied experimentally in T. cruzi (12) and L. major (24, 52), and orthologues are present in all kinetoplastid genomes available so far. Despite their high overall sequence similarity, PDEB1 and PDEB2 exhibit distinct subcellular localizations (31). PDEB1 is predominantly found in the flagellum, where it is stably associated with cytoskeletal components that are resistant to detergent extraction. In contrast, PDEB2 is mostly localized in the cell body, from where it is fully extractable by nonionic detergents. However, a minor fraction of PDEB2 also associates with the flagellar skeleton in a Triton-resistant manner, most likely through interaction with PDEB1. Earlier work has shown that both PDEB1and PDEB2 are essential enzymes in bloodstream-form T. brucei (31), while TbPDEA, TbPDEC, and TbPDED play minor roles (20; S. Kunz, unpublished data).