Mammalian cells cannot synthesize the amino acid tryptophan and rely on its import as tryptophan or as components of nutrients such as peptides across the plasma membrane for basic cell functioning. Tryptophan is necessary for protein synthesis, and it accounts for ~1.3 % of the amino acids in human proteins. Tryptophan is also the precursor of other vital molecules like serotonin, melatonin, and NAD. Moreover, kynurenines produced from tryptophan appear to play a pivotal role in immunosuppression in inflammatory diseases and cancer [39
The FLIPW nanosensors described in this study allow for noninvasive, real-time, spatio-temporal imaging of intracellular tryptophan levels and flux, offering advantages over conventional analytic methods. The E. coli
transcriptional regulator TrpR was used as the recognition element for the construction of FRET sensors for tryptophan. As noted previously, the use of bacterial proteins for the construction of intracellular sensors reduces the problem of interference with endogenous metabolic and signal transduction pathways in eukaryotic cells [41
]. Genetically encoded nanosensors further offer the advantage of subcellular sensor targeting through judicious choice of leader sequences as demonstrated with nuclear- and ER-targeted glucose nanosensors [42
] and with cell-surface display of glutamate nanosensors [19
]. Most FRET nanosensors are based on the ligand-binding–induced Venus fly trap–like conformational changes of bacterial PBPs [16
], which consist of two well-structured lobes with the ligand-binding site located at the interface. TrpR is about three times smaller than the average PBP and is partially unfolded in the absence of tryptophan [25
]. In the presence of tryptophan, the protein adopts the condensed conformation observed in crystal structures [22
], and the concomitant conformational changes allow for the detection of tryptophan binding by FRET.
FRET has been a successful reporter signal for small molecule sensors [43
]. According to the Förster theory, the efficiency of the energy transfer depends on the distance between the fluorophores and their dipole orientation [45
]. These small molecule nanosensors can be engineered by modification of linker sequences between reporter and sensing domains and/or insertion of fluorophores in surface loops of the sensing domain, resulting in increased and/or reversed signal outputs of FRET nanosensors [29
]. Because TrpR dimerizes to form its ligand-binding sites, we applied a novel approach for engineering of the FRET signal. Insertion of a second TrpR coding sequence in the principal FLIPW-CTY sensor changed the FRET response depending on the position of the insertion site. While insertion before eCFP and between eCFP and Venus almost eliminated the FRET response, a TrpR copy after Venus reversed the FRET response and increased the ratio change. Comparison of structural models of the FLIPW-CTY and FLIPW-CTYT sensors predicted that the fluorophores would be closer together in FLIPW-CTY. Since FRET efficiency is inversely correlated with the distance between the fluorophores as described in the Förster equation [45
], the experimentally determined FRET ratio and the models are consistent.
FLIPW-CTYT was used for monitoring tryptophan uptake in cell cultures grown in 96-well microtiter plates, making the sensor suitable for high-throughput assays in which the effect of drugs or small interfering RNAs (siRNAs) is tested systematically [46
]. The effective KM
for combined tryptophan uptake and metabolism in COS-7 cells in microplate assays and during perfusion was similar with values in the low micromolar range. We found that a LAT1-like transporter is responsible for the observed tryptophan exchange in COS-7 cells. Its transport and exchange characteristics are very similar to that of human LAT1 as expressed in KB cells. The reported affinity of human LAT1 for tryptophan uptake is 21 μM, about 4-fold lower than determined in this study [35
]. However, the values are difficult to compare, because the former uptake experiments were performed under static conditions with LAT1 expressed in oocytes and the obtained affinities relate to the sum of intracellular pools of free, incorporated, and degraded tryptophan in the large oocytes. On the other hand, the affinities obtained using FLIPW-CTYT have been determined for the pool of free tryptophan in the targeted subcellular compartment, i.e., cytosol, under perfusion conditions with histidine as a defined counterexchange substrate. Possibly, the physiological conditions of amino acid exchange in the vascular system are mimicked more accurately by perfusion.
The transporters LAT1 and LAT2 are heteromeric obligatory counterexchangers of large, neutral amino acids with a 1:1 exchange stoichiometry (SLC3 and SLC7) [6
]. As exchangers, they do not change the net intracellular amino acid concentration, but rather modify their relative concentrations. Perfusion of FLIPW-CTYT–expressing COS-7 and KB cells with tryptophan and histidine yielded high-resolution data of the real-time dynamics of free cytosolic tryptophan resulting from system L exchange activity. Importantly, we found that LAT transporters can exchange kynurenines and tryptophan. Since the individual intracellular and extracellular substrate selectivity of the LAT transporters are similar [36
], kynurenine-tryptophan exchange may be bidirectional.
Tryptophan-kynurenine exchange may be part of an endogenous immunosuppressive mechanism during autoimmunity and may support the immune escape of proliferative cell types—like cancer cells—by enhancing the depletion of the local tryptophan pool and increasing the serum kynurenine concentrations (). Kynurenines and hydroxy-kynurenine are natural amino acids that are produced from tryptophan through IDO, whose enzymatic activity is necessary for immune escape [13
]. Macrophages and dendritic cells express IDO for endogenous suppression of the immune system [40
]. Several human cancer cell lines, including KB cells, up-regulate IDO activity in the presence of the proinflammatory cytokine interferon-γ [48
]. In addition, IDO can be up-regulated by the mutation of the tumor suppressor Bin1
], which is lost or attenuated in several cancer types [49
]. Intracellularly produced kynurenines serve as substrates for the exchange for extracellular tryptophan by LAT transporters, which are expressed in many tissues and human primary tumors [10
]. Effectively, tryptophan is sequestered from the local environment and kynurenines accumulate in the serum. The kynurenines contribute to the pool of amino acids that can be taken up in exchange for intracellular tryptophan by surrounding cells expressing LAT transporters, a process which results overall in a tryptophan flux toward the IDO-producing cells. Since resting human T cells express only transporters of system L for the transport of l
], the tryptophan-kynurenine exchange mechanism helps to deplete the intracellular tryptophan as well. Both the accumulation of kynurenines and depletion of tryptophan arrest T cell growth and induce apoptosis [39
]. Thus, tryptophan-kynurenine exchange results in double trouble for T cells (). At the same time, IDO-overproducing cells are protected from the apoptotic effect of kynurenines by the strict counterexchange of tryptophan and its stoichiometric degradation products. Cells, e.g., human macrophages, expressing transport systems with higher affinities for tryptophan than LAT1 will be able to continue to proliferate during ongoing tryptophan sequestration by the combined activity of LAT1 and IDO [50
]. FLIPW sensors can now be used to test whether T cells take up kynurenines using the same pathway leading to a further drain of the essential tryptophan. The sensors can also be used to identify novel drugs and regulatory factors in genomic RNAi screens or screens of chemical libraries.
FRET nanosensors are unique tools for studying intracellular, small molecule steady-state levels and fluxes in vivo and in real time. Ultimately, complete metabolic routes can be monitored by using nanosensors that selectively detect single intermediates. For this means, a set of FRET nanosensors has been constructed that use the ligand-induced conformational changes of PBPs [16
]. As the FLIPW sensors demonstrate, other protein scaffolds that undergo conformational changes upon ligand binding can provide sensing domains for nanosensors with specificities not represented in the PBP family, such as tryptophan. E. coli
tryptophan repressor TrpR is not part of a protein family with different substrate specificities, which could be used for the expansion of the current set of nanosensors. However, the wealth of bacterial transcriptional regulators, which change affinity for operator sequences upon binding of effectors, may provide potential sensing domains for novel FRET metabolite nanosensors.
The FLIPW-CTYT nanosensor has proven to be a robust system with multiple advantages over conventional methods for intracellular tryptophan detection. The new sensor thus provides a complementary tool for monitoring steady state levels, uptake, and counterexchange, and will be an important tool for analyzing the factors that control tryptophan flux in living cells. As the kynurenine/tryptophan exchange demonstrates, such factors might contribute to important cellular processes such as inflammation and immune escape.