Article pubs.acs.org/biochemistry
Tryptophan-Rich Sensory Protein/Translocator Protein (TSPO) from Cyanobacterium Fremyella diplosiphon Binds a Broad Range of Functionally Relevant Tetrapyrroles Andrea W. U. Busch,† Zachary WareJoncas,† and Beronda L. Montgomery*,†,‡,§ †
Plant Research Laboratory, Department of Energy, Michigan State University, East Lansing, Michigan 48824, United States Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, Michigan 48824, United States § Department of Microbiology & Molecular Genetics, Michigan State University, East Lansing, Michigan 48824, United States ‡
S Supporting Information *
ABSTRACT: Tryptophan-rich sensory protein/translocator protein (TSPO) is a membrane protein involved in stress adaptation in the cyanobacterium Fremyella diplosiphon. Characterized mammalian and proteobacterial TSPO homologues bind tetrapyrroles and cholesterol ligands. We investigated the ligand binding properties of TSPO from F. diplosiphon (FdTSPO1), which was functionally characterized in prior genetic studies. Two additional TSPO proteins (FdTSPO2 and FdTSPO3) are present in F. diplosiphon; they are similar in size to reported bacterial TSPOs and smaller than FdTSPO1. The longer cyanobacterial TSPO1 is found almost exclusively in filamentous cyanobacteria and has a relatively low degree of homology to bacterial and mammalian TSPO homologues with confirmed tetrapyrrole binding. To probe distinctions of longform TSPOs, we tested the binding of porphyrin and bilin to FdTSPO1 and measured binding affinities in the low micromolar range, with the highest binding affinity detected for heme. Although tetrapyrrole ligands bound FdTSPO1 with affinities similar to those previously reported for proteobacterial TSPO, binding of cholesterol to FdTSPO1 was particularly poor and was not improved by introducing an amino acid motif known to enhance cholesterol binding in other bacterial TSPO homologues. Additionally, we detected limited binding of bacterial hopanoids to FdTSPO1. Cyanobacterial TSPO1 from the oxygenic photosynthetic F. diplosiphon, thus, binds a range of tetrapyrroles of functional relevance with efficiencies similar to those of mammalian and proteobacterial homologues, but the level of cholesterol binding is greatly reduced compared to that of mammalian TSPO. Furthermore, the ΔFdTSPO1 mutant exhibits altered growth in the presence of biliverdin compared to that of wild-type cells under green light. Together, these results suggest that TSPO molecules may play roles in bilin homeostasis or trafficking in cyanobacteria.
T
Rhodobacter sphaeroides (RsTSPO) bind to various ligands, including tetrapyrroles.20,24−26 Analyses of a Rhodobacter TSPO knockout mutant, which exhibited accumulation of photosynthetic pigments in the growth medium, led to the suggestion that TSPO is involved in controlling the export of porphyrin from cells.27 A physiologically relevant correlation of ligand binding with TSPO function has been reported for A. thaliana.26 In these studies, the binding of the cyclic tetrapyrrole heme facilitated AtTSPO degradation.26 Hemebound TSPO interacts with the aquaporin PIP2;7 that is correlated with a downregulation of the intracellular levels of both TSPO and PIP2;7 through an autophagic pathway.28 Expression of a stable mutant of TSPO that was deficient in
ryptophan-rich sensory protein/translocator protein (TSPO), previously known as the peripheral benzodiazepine receptor in mammals,1 is nearly ubiquitously found across organisms and has been studied in animals, plants, and bacteria.2−5 TSPO is a membrane-spanning protein,6 whose complete function has not been fully elucidated.7−12 Studies in mammalian, plant, algal, and bacterial systems assigned functions for TSPO homologues in stress adaptation, oxygen sensing, photosynthetic pigment biosynthesis, and porphyrin transport,3,13−20 whereas a function in translocation of steroids, specifically cholesterol,21 across the mitochondrial membrane as well as intracellular porphyrin transport and conversion in mammalian cells19 are subjects of current debate.22,23 Common to many TSPOs studied to date is their ligand binding activity. TSPO proteins from the model plant Arabidopsis thaliana (AtTSPO), mammals (HsTSPO for human and MmTSPO for mouse), red alga Cyanidioschyzon merolae, and the bacterium © XXXX American Chemical Society
Received: October 5, 2016 Revised: December 5, 2016 Published: December 6, 2016 A
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
conservation of a conserved histidine associated with heme binding in TSPO from A. thaliana,26 an alignment of TSPO sequences was performed with Clustal OMEGA.36 Cloning and Purification of TSPO from F. diplosiphon. TSPO from WT F. diplosiphon (also Tolypothrix sp. PCC 7601 37 ), which encodes FdTSPO1 (GenBank entry AAT36314.1), was amplified from genomic DNA with primers containing EcoRI and XhoI restriction sites (forward, 5′-ccaaa GAA TTC ATG ACT CAA TCA AAT AAT ACC-3′; reverse, 5′-gcaaa CTC GAG TTA TTT TTC CAC TGG-3′; restriction sites are underlined, and bases added for increased cleavage efficiency by restriction enzymes are shown in lowercase). The amplified fragment was restricted with EcoRI and XhoI and cloned into pGEX-6P-1 (GE Healthcare, Uppsala, Sweden) carrying an N-terminal glutathione S-transferase (GST) tag,38 which had been restricted with the same enzymes. A version of FdTSPO1 containing a LAF mutation (hereafter the LAF mutant), to be tested for its impact on cholesterol binding, was constructed using the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA) and LAF mutation primers (forward, 5′-CTT GTA TCT GGC TAG CTT TCG CTG CTG TAT TAG-3′; reverse, 5′-CTA ATA CAG CAG CGA AAG CTA AGA TAC AAG-3′) according to the manufacturer’s instructions. The GSTFdTSPO1 fusion was expressed in BL21 cells grown in super broth [35 g/L tryptone, 20 g/L yeast extract, and 5 g/L sodium chloride (pH 7)] at 37 °C to an optical density at 578 nm (OD578) of approximately 0.5−0.6. Expression was induced with 100 μM isopropyl β-D-1-thiogalactopyranoside (IPTG) and grown for ∼16 h at 17 °C. Cells were pelleted at 5500g for 15 min and lysed in purification buffer [20 mM Tris and 200 mM NaCl (pH 7.5)], with either a French press at an ∼16K psi working pressure or a Constant Systems (Kennesaw, Georgia) Cell Disruptor at 20K−30K psi with three passages each. Lysed cells were subjected to a low-spin centrifugation at 12000g for 10 min at 4 °C, and the obtained supernatant was centrifuged at 174587g for 2 h to collect the membrane fraction. The harvested membrane fraction with a total protein concentration of ∼10 mg/mL was solubilized with 1% n-dodecyl β-D-maltopyranoside (DDM) (Affymetrix, Maumee, OH) for 16 h while being stirred. The insoluble portion was removed by centrifugation at 174587g for 1−2 h. Solubilized membrane protein was incubated for at least 1 h while being stirred with Glutathione Sepharose 4B resin (GE Healthcare) equilibrated with purification buffer. The resin was loaded on a gravity column and washed with at least 10 column volumes of purification buffer containing 0.2% DDM. Bound FdTSPO1 was eluted from the column by overnight incubation with PreScission Protease (GE Healthcare) or alternatively by elution with 15 mM glutathione-containing purification buffer, followed by overnight dialysis with PreScission Protease and subsequent removal of GST and protease by a second affinity purification. Buffer exchange into purification buffer containing 0.02% DDM was achieved by dialysis in SnakeSkin Dialysis Tubing with a 10000 molecular weight cutoff (MWCO) (Thermo Scientific, Rockford, IL) or by using a 10 mL Amicon stirred cell and ultrafiltration disk membranes with a 10000 MWCO (Millipore, Billerica, MA). The latter was also used for concentrating the protein, or Amicon ultra-4 centrifugal filter units (10 kDa) were used instead. All steps were performed at 4 °C or on ice. TSPO2 and TSPO3 were amplified via polymerase chain reaction (PCR) from genomic DNA with primers containing
heme binding did not negatively affect PIP2;7 protein levels as observed for overexpression of the wild-type (WT) form of TSPO.28 This observed effect of TSPO on PIP2;7 abundance in A. thaliana established a role for TSPO in drought stress, a condition under which expression of plant TSPO is upregulated.14 We sought to investigate the biochemical properties of TSPO1 (AAT36314.1) from the filamentous freshwater cyanobacterium Fremyella diplosiphon (termed FdTSPO1) to identify ligand binding activities of a cyanobacterial TSPO from an organism with special tetrapyrrole requirements. F. diplosiphon is an oxygenic photosynthetic cyanobacterium and therefore not only employs tetrapyrroles in the respiratory chain, as do other nonphotosynthetic organisms, but also exhibits strong demand for tetrapyrroles in its photosynthetic complexes in the form of chlorophyll in the photosystems and phycobilins in the light-harvesting antenna complexes, or phycobilisomes (PBSs). Notably, PBSs can comprise up to 50% of the total cellular soluble protein.29 Thus, PBSs contribute significantly to cellular tetrapyrrole demands. Additionally, F. diplosiphon exhibits an ability to undergo complementary chromatic acclimation, a process that allows the cell to change the phycobiliproteins contained in its light-harvesting structures, and thereby alter cellular tetrapyrrole content, to sustain light capture upon changes in external light quality.29,30 Given the strong organismal demands for tetrapyrroles and light-dependent changes in the synthesis of bilins, we investigated the ability of FdTSPO1 to bind cyclic tetrapyrroles (porphyrins), open chain tetrapyrroles (bilins), and the known TSPO ligand cholesterol, as well as the structurally similar cyanobacterial hopanoids. We found that FdTSPO1 has a limited ability to bind cholesterol or hopanoids, whereas tetrapyrroles like heme, biliverdin IXα, mesoporphyrin IX, and protoporphyrin IX bind in the low micromolar range. These results support the hypothesis that FdTSPO1 is involved in tetrapyrrole regulation and that it can bind a structurally diverse set of ligands in vivo. The bilins to which FdTSPO1 has been confirmed to bind are physiologically relevant as precursors of the phycobilisomal light-absorbing bilins (biliverdin IXα, heme, and protoporphyrin IX) and derivatives thereof [mesoporphyrin IX, Zn(II)mesoporphyrin IX, and Zn(II)protoporphyrin IX], as well as the enzyme cofactor of the respiratory chain and of detoxification mechanisms (heme).31 To determine whether a TSPO ligand may impact fitness in a TSPO-dependent manner, we tested growth of WT cells compared to that of a ΔFdTSPO1 mutant in the presence of biliverdin and determined that mutant cells exhibited distinct growth responses. In addition to FdTSPO1, we discovered two other putative TSPOs in the genome of F. diplosiphon (FdTSPO2 and FdTSPO3) that are distinct from FdTSPO1 in sequence and size yet also possess the ability to bind some physiologically relevant tetrapyrroles.
■
EXPERIMENTAL PROCEDURES Phylogeny/Genome Analysis. Representative amino acid sequences of putative TSPOs from all kingdoms were retrieved from the NCBI protein database.32 Similarities among three TSPO sequences from F. diplosiphon and proteobacterial, cyanobacterial, and mammalian TSPOs were determined using the pairwise alignment EMBOSS Needle tool.33−35 For computation of a phylogenetic tree representing TSPO sequences from all kingdoms, sequences were aligned with MUSCLE and phylogeny was calculated using ClustalW2 phylogeny. To assess the presence of putative cholesterol binding motifs and B
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry EcoRI and XhoI restriction sites (forward, 5′-ccaaa GAA TTC ATG ATT AAA TCT TGG ATG G-3′; reverse, 5′-gcaaa CTC GAG TTA TGC TTC TTG-3′; and forward, 5′-ccaaa GAA TTC ATG ATT GAA TCT TGG TGA TAG GG-3′; reverse, 5′-gcaaa CTC GAG CTA CGC ATC TTC AG-3′, respectively) and proteins expressed and purified in a manner identical to that described for FdTSPO1 above. Bacterial Strains and Growth Conditions. Growth conditions of F. diplosiphon were essentially as described previously.39 In brief, WT cells were grown at 28 °C in BG-11 medium (Fluka, Buchs, Switzerland) with 20 mM HEPES (BG-11/HEPES) while being shaken at 175 rpm. Cells were grown in either ∼10−15 μmol m−2 s−1 of continuous broadband green light (GL; CVG sleeved Rosco green 89 fluorescent tubes, General Electric, model F20T12/G78) or red light (RL; CVG sleeved Rosco red 24 fluorescent tubes, General Electric, model F20T12/R24), with light intensities determined using a LI-250A light meter in line with a quantum sensor (LI-COR, Lincoln, NE). For growth curves, cells in the exponential phase were diluted to an optical density at 750 nm (OD750) of 0.1 and 20 μM biliverdin IXα (biliverdin IXα hydrochloride, Frontier Scientific) was added from a 10 mM stock solution in 70% DMSO. Growth was assessed using the cell density measured as OD750 over 16 days using a SpectraMax M2 microplate reader (Molecular Devices, Sunnyvale, CA). Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR). Transcript accumulation analyses were conducted essentially as described previously.39 RNA was isolated from WT cells grown under GL or RL as described above using Trizol. RNA treatment and reverse transcription of 500 ng of RNA per sample were as detailed previously.39,40 Primers for FdTSPO1 were those previously reported:39 gtagaacggagattavggtgcg (forward primer) and cagccacvagttagccagatac (reverse primer). For FdTSPO2, the primers were as follows: acctgtaatgttccgcctg (forward primer) and actgcagcccaattagaagtag (reverse primer). For FdTSPO3, the primers were as follows: ccccgtgatgttaaatggttc (forward primer) and aacaatatttgcggaagctgc (reverse primer). ORF10B primers were used to amplify the reference gene as reported previously.40 qRT-PCR was conducted as described using Fast SYBR Green Master Mix, a Microamp fast optical 96-well reaction plate, and an ABI FAST 7500 Real-Time PCR system (Applied Biosystems, Grand Island, NY) in FAST mode according to the manufacturer’s instructions and as previously detailed with a final template dilution of 1:50.39 Experiments were performed with three biological replicates with three technical replicates each. Pull-Down Assays with Tetrapyrrole-Linked Agarose. Covalent binding of tetrapyrrole to low-density aminoethyl agarose beads 6B cross-linked (Goldbio, St. Louis, MO) was performed on the basis of the work of Tsutsui and Mueller.41 Tetrapyrroles used were hemin, biliverdin hydrochloride (BV), protoporphyrin IX (PPIX), mesoporphyrin IX (MPIX), Zn(II)protoporphyrin IX (ZnPPIX), and Zn(II)mesoporphyrin IX (ZnMPIX), which were purchased from Frontier Scientific (Logan, UT). Tetrapyrrole aliquots (final concentration of 13.3 mg/mL) were dissolved in N,N-dimethylformamide (DMF) (Sigma-Aldrich, St. Louis, MO). The tetrapyrroles were derivatized with 1,1′-carbonyldiimidazole (CDI) (SigmaAldrich) at a final concentration of 13.3 mg/mL and heated at 80 °C for 15 min while being rotated. After a cool down period of 30 min at room temperature, the derivatized tetrapyrrole was added to a gravity column with aminoethyl-agarose
from 0.8 mL of slurry/mL of DMF solution, which had been washed with 13.2 mL of distilled water, 3.4 mL of 33% (v/v) DMF, 3.4 mL (v/v) of 66% DMF, and finally with 6.6 mL of 100% DMF per milliliter of DMF solution each. The mixture was incubated for 18 h at room temperature in the dark with rotation. Unbound tetrapyrrole was removed by washing with 25% (v/v) pyridine (Sigma-Aldrich) until the flow through appeared visually clear. The column was equilibrated with water followed by purification buffer with at least 10 column volumes each. Twenty microliters (for FdTSPO1) or 10 μL (for FdTSPO2 or FdTSPO3) of tetrapyrrole-coupled resin or aminoethyl-agarose (negative control) was equilibrated by being washed once with 1 mL of ddH2O and twice with 1 mL of purification buffer with 0.02% (v/v) DDM. To this was added 100 μL of a 2.5 μM solution of purified FdTSPO1 or 50 μL of an ∼4.2 μM solution of purified FdTSPO2 and ∼10.8 μM purified FdTSPO3 in purification buffer with 0.02% DDM, and the mixture was incubated at room temperature for 30 min while being shaken. The resin was washed twice with 1 mL of purification buffer with 0.02% DDM, and a final wash was conducted with 50−100 μL. The resin was pelleted by centrifugation at 500g for 2 min at 4 °C in each step. Protein bound to the column was eluted by incubating the resin with 50−100 μL of 2× sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE) sample buffer at 95 °C for 10 min. The 5× SDS−PAGE sample buffer was added to 10 μL of flow through (after binding) and wash samples (from the last washing step) to a final concentration of 1×. These samples, as well as 10 μL of the elution fraction, were run on a 15% SDS− PAGE and visualized by Coomassie Blue staining or immunoblotting, the latter as described below. Standard SDS−PAGE and staining were performed.42 Ultraviolet/Visible (UV/Vis) Absorption Spectra. UV/vis absorption spectra were recorded on an Agilent HP8453 UV/ visible spectrophotometer (Agilent Technologies) at room temperature in a quartz cuvette with a path length of 1 cm. The measurements were performed in the range of 300−800 nm for interactions of FdTSPO1 with hemin (2 or 4 μM) as indicated. A baseline adjustment was performed at 900 nm. Tryptophan Fluorescence Measurements. Tryptophan fluorescence measurements were performed with purified FdTSPO1 protein and ligand essentially as described by Li et al.;43 2.5 μM protein in 250 μL of purification buffer containing 0.02% DDM with or without 2 mM β-mercaptoethanol was used. Protein was excited at 280 nm and 600 V and emission measured at 290−400 nm with a PTI QuantaMaster spectrofluorimeter (HORIBA Instruments, Edison, NJ). Ligand was added in aliquots until all fluorescence was quenched. The amount of quenching, with the initial signal set at 1, was plotted against ligand concentration. The resulting data points were fitted with the Sigmoidal Fit with Hill function in Origin, i.e., y = Vmax[xn/(kn + xn)], to obtain the binding curve and resulting Kd values. Given that TSPO is not expected to show binding cooperativity, n would equal 1 and the equation would be equivalent to the Michaelis−Menten equation. We assessed parameter n in these analyses as an internal measure of validation. Stock concentrations of tetrapyrroles were prepared in dimethyl sulfoxide (DMSO) and in ethanol for cholesterol. A Hop-22(29)-ene solution was obtained as a 0.1 mg/mL solution in isooctane from Sigma. Control measurements were performed with protein and solvent only. Immunodetection of TSPO. SDS−PAGE was performed42 and blotted at constant values of 25 V and 1 A for 30 min C
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry Table 1. Similarity of TSPOs to FdTSPOa TSPO1 (WP_045870827.1)
TSPO2 (WP_045870064.1)
TSPO3 (WP_045870065.1)
species
accession number
length
gaps
identity
similarity
gaps
identity
similarity
gaps
identity
similarity
Nostoc sp. PCC 7120 Nostoc sp. PCC 7120 Nostoc sp. PCC 7120 Synechococcus sp. PCC 7002 Synechocystis sp. PCC 6803 A. thaliana Solanum tuberosum Physcomitrella patens R. sphaeroides 2.4.1 Synorhizobium meliloti Homo sapiens Rattus norvegicus Schizosaccharomyces pombe
BAB74706.1 WP_010998253.1 WP_010998254.1 WP_012308213.1 BAA18606.1 AEC10886.1 CAH10765.1 EDQ62829.1 ABA79442.2 WP_012881280.1 AAA18228.1 NP_036647.1 NP_595490.2
251 157 157 157 159 196 192 198 159 162 169 169 164
0.0% 42.5% 47.2% 49.4% 44.1% 43.7% 37.7% 36.4% 37.3% 36.1% 45.0% 38.5% 46.4%
85.3% 12.7% 14.2% 15.5% 15.6% 15.7% 16.1% 19.5% 18.3% 19.8% 15.5% 16.2% 15.6%
91.6% 29.0% 27.7% 25.1% 27.8% 25.9% 30.4% 30.3% 30.6% 31.3% 26.2% 29.2% 25.2%
36.0% 0.0% 0.0% 0.0% 1.3% 39.5% 44.2% 21.9% 16.3% 17.7% 37.0% 16.9% 22.7%
15.8% 83.4% 65.0% 51.6% 54.1% 18.2% 14.3% 21.9% 22.7% 26.3% 20.5% 24.2% 21.5%
29.1% 90.4% 79.6% 70.1% 66.7% 29.5% 30.4% 38.5% 39.5% 42.9% 32.5% 40.4% 38.7%
44.3% 0.0% 0.0% 2.5% 3.7% 42.4% 48.3% 16.0% 25.4% 12.4% 40.2% 19.9% 30.2%
17.2% 62.4% 75.2% 42.8% 44.7% 18.3% 15.7% 23.8% 22.1% 25.3% 19.6% 25.4% 20.1%
30.9% 75.8% 86.6% 62.3% 59.0% 27.2% 29.1% 43.6% 37.6% 42.4% 30.4% 42.0% 37.0%
a Sequence similarities and identities for FdTSPO from F. diplosiphon in comparison to other TSPO protein sequences as determined by EMBOSS Needle.33
Figure 1. Genomic context of putative TSPO homologues in F. diplosiphon. FdTSPO1 (top) is adjacent to green light-upregulated genes whose products are involved in phycobilisome biogenesis (cpeYZ and cpeAB) as well as genes involved in phycobilisome degradation upon nitrogen starvation (nblA) and chlorophyll degradation (paO).39 FdTSPO2 and FdTSPO3 (lower) are located next to each other in the genome and are flanked by genes encoding subunits of a K+-transporting ATPase and genes encoding enzymes involved in cell wall turnover (ddpX, D-alanyl-Dalanine dipeptidase) and phytol reutilization (VTE5, phytol-kinase).
recently published whole genome of F. diplosiphon (also known as Tolypothrix sp. PCC 760137) resulted in the identification of two additional putative TSPO-encoding genes. These two putative TSPOs, which are adjacent in the F. diplosiphon genome and denoted FdTSPO2 and FdTSPO3, are neighbored by genes encoding subunits of a K+-transporting ATPase and genes encoding enzymes involved in cell wall turnover (ddpX, 44 D-alanyl-D-alanine dipeptidase) and phytol reutilization from chlorophyll degradation (VTE5, phytol-kinase)45 (Figure 1). FdTSPO2 and FdTSPO3 are equal in length (157 amino acids) and highly similar (63.7% identical and 79% similar). Notably, FdTSPO2 and FdTSPO3 were found to be more similar to TSPO homologues in single-celled cyanobacteria in which only one TSPO homologue is found (Table 1). Additional TSPO homologues were also detected in some filamentous cyanobacterial species (e.g., Nostoc) that are more homologous with FdTSPO2 and FdTSPO3 than with FdTSPO1 (Table 1 and Figure 2). We performed phylogenetic analysis with a variety of putative TSPO protein sequences from prokaryotes and eukaryotes, including higher and lower plants, mammals, photosynthetic and nonphotosynthetic bacteria, algae, archaea, and algal viruses (Figure 2). All three TSPO sequences from F. diplosiphon cluster with sequences from filamentous cyanobacteria, albeit in three distinct clades (Figure 2). FdTSPO2 and FdTSPO3 cluster with other TSPOs from cyanobacteria, whereas FdTSPO1 is present in another distinct cluster with cyanobacterial as well as algal TSPO homologues. On the basis of the same tree, TSPO from the higher plant A. thaliana is more closely related to
with the Trans-Blot Turbo Transfer System (Bio-Rad, Hercules, CA). The blot was incubated in blocking solution (1× Tris-buffered saline, 0.5% Tween 20, and 3% bovine serum albumin) while being rocked for 1 h at room temperature (RT). The anti-TSPO serum was added in a 1:10000 dilution in 10 mL of a blocking solution and incubated while being rocked for 16 h at 4 °C. The blot was washed three times with 10 mL of wash buffer (1× Tris-buffered saline and 0.1% Tween 20) at RT and then incubated with 1:20000 goat anti-rabbit HRP antibody (Pierce, Rockford, IL) in washing buffer for 1 h at RT. The blot was visualized with FemtoGlow Western PLUS chemiluminescent HRP substrate (Michigan Diagnostics, Royal Oak, MI) on a ChemiDoc MP imager (Bio-Rad). The antiTSPO antibody serum was prepared by Pacific Immunology (Ramona, CA) in New Zealand white rabbits, using an excised purified FdTSPO1 protein band separated on a 12.5% SDS−polyacrylamide gel at 100−200 V.
■
RESULTS Sequence Similarities and Phylogeny of Cyanobacterial TSPO Proteins. TSPO1 from freshwater, filamentous F. diplosiphon shows the highest degree of homology with TSPO proteins from other filamentous cyanobacterial species, including Anabaena variabilis and Nostoc sp. PCC 7120 (Table 1). The FdTSPO1 sequence is more similar to eukaryotic and proteobacterial TSPO sequences (i.e., 15−20%) than to sequences from single-celled cyanobacterial species like Synechocystis sp. PCC 6803 (i.e., 15.6%). Analysis of the D
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
Figure 2. Putative TSPOs from the F. diplosiphon group with TSPO homologues from filamentous cyanobacteria. Phylogenetic tree with sequences retrieved from a search of the NCBI protein database.32 The sequences were aligned with MUSCLE, and phylogeny was calculated using ClustalW2 phylogeny. The species name is followed by the protein accession number and the bootstrap value. An excerpt of the tree containing various putative TSPO sequences from all kingdoms is shown.
cyanobacterial sequences than proteobacterial and mammalian TSPO homologues (data not shown), whereas this is not the case for TSPO from the moss Physcomitrella patens (Figure S1). To determine whether the three FdTSPO homologues may be differentially expressed in cells, we examined expression of FdTSPO1, FdTSPO2, and FdTSPO3 genes under distinct light conditions. F. diplosiphon is a chromatically acclimating cyanobacterium that exhibits distinct physiological and morphological changes in response to green light (GL) and red light (RL), an adaptation that allows maximal light absorption in different
depths of the water column where either GL or RL is more abundant. FdTSPO1 is upregulated under GL compared to RL conditions. A 6.5-fold induction under GL versus RL was observed in a microarray analysis,46 an ∼2.5-fold induction in RNA-sequencing-based analysis39 and an ∼2-fold induction by qRT-PCR.39 We compared all three TSPOs by qRT-PCR (Figure 3) and found that FdTSPO2 transcript levels were similar under both light conditions, whereas the FdTSPO3 average transcript level was increased under RL compared to GL, albeit not significantly (Figure 3A). Reproducibly, E
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry FdTSPO1 showed ∼2-fold induction under GL (Figure 3A).39 These data suggest a distinct light-regulated role for FdTSPO1
and Nostoc. Given the distinct clustering of cyanobacterial TSPO sequences and the relatively low degree of homology to bacterial and mammalian homologues, we sought to investigate ligand binding capacities of FdTSPO1 (termed FdTSPO throughout), which has been previously shown to regulate light- and stressdependent responses in vivo39 and is the most distinct from previously characterized bacterial TSPOs. FdTSPO Binds a Range of Functionally Relevant Tetrapyrroles. Tetrapyrrole binding has been established for TSPO from the bacterium R. sphaeroides (i.e., RsTSPO), as well as for plant and mammalian TSPO proteins.6,24,26 Specifically, protoporphyrin IX bound to RsTSPO and HsTSPO (from Homo sapiens) in vitro and to protoplasts with heterologously expressed MmTSPO.6,24,26,43 Heme binding was demonstrated in vitro for RsTSPO and AtTSPO (plant A. thaliana), and for mitochondrial preparations enriched with mammalian TSPO.6,47 Mesoporphyrin IX binding was confirmed for mammalian TSPO mitochondrial preparations, as well.43,47 Given the prior associations of tetrapyrrole binding for TSPO proteins from a range of organisms from bacteria to mammals, we performed qualitative pull-down assays with FdTSPO1 to establish binding activity to known, as well as unknown, TSPO tetrapyrrole ligands (Figure 4). For this assay, we used tetrapyrroles covalently coupled to agarose resin. All of the tested resin-bound tetrapyrroles, including the porphyrins hemin, protoporphyrin IX (PPIX), mesoporphyrin IX (MPIX), Zn(II)protoporphyrin IX (ZnPPIX), and Zn(II)mesoporphyrin IX (ZnMPIX), as well as the open chain tetrapyrrole biliverdin IXα (BV), interacted with purified FdTSPO1 (see Figure S2 for SDS−PAGE of the purified input protein). No interaction was detected with the resin-only control. Given that heme binding, in particular, can be difficult to assess because of the intrinsically sticky nature of heme, we also obtained UV/vis spectra of unbound and bound heme with FdTSPO1. We noted distinct shifts in the UV−vis spectra for the ligand hemin alone, relative to hemin mixed with FdTSPO1, which indicated specific ligand binding (Figure S3). Notably, the two shorter forms of TSPO found in F. diplosiphon, i.e., FdTSPO2 and FdTSPO3, which are more similar to TSPOs from bacterial and eukaryotic systems for which tetrapyrrole binding has been reported,6,24,29 also bound a range of tetrapyrrole-linked resins (Figure S4). These proteins bound tetrapyrroles similar to those bound by FdTSPO1, with the exception of BV, which appeared to show limited interaction with FdTSPO2 and FdTSPO3 proteins. Given the distinction of the longer FdTSPO1 and the fact that its homologues are found primarily in filamentous cyanobacteria, as well as its low degree of homology to TSPO homologues from bacterial and mammalian systems characterized
Figure 3. FdTSPO transcript levels in green light compared to red light. Samples for quantitative real-time reverse transcription PCR (qRT-PCR) were taken for cells acclimated to green light (GL, black bars) or red light (RL, gray bars). The relative expression level compared to that of the ORF10B gene is shown (±SD; n = 3). The relative transcript abundance of FdTSPO1, FdTSPO2, and FdTSPO3 is shown for each TSPO homologue relative to the GL-adapted samples (fold change) for each construct (A) and relative to FdTSPO1 in GL (B). The inset in panel B shows only FdTSPO1 and FdTSPO2. p values determined using an unpaired, two-tailed Student’s t test (*p < 0.01).
under GL, but not for FdTSPO2 and FdTSPO3. When the same data for all three TSPO transcripts were compared to FdTSPO1 levels under GL, overall transcript levels were substantially higher for FdTSPO2 and FdTSPO3 (Figure 3B). Compared to that of FdTSPO1, FdTSPO2 transcript levels were ∼5-fold higher in GL and RL and transcript levels for FdTSPO3 were ∼30-fold higher in GL and ∼56-fold higher in RL (Figure 3B). Taken together, these analyses indicate that two major types of cyanobacterial TSPO homologues exist. A shorter form (represented by FdTSPO2 and FdTSPO3 in F. diplosiphon) is present in filamentous and single-cell cyanobacteria, and a longer form (represented by FdTSPO1) is found in addition to the shorter TSPO homologues. The longer form of TSPO is found nearly exclusively in filamentous cyanobacteria such as Fremyella
Figure 4. F. diplosiphon TSPO binding to tetrapyrrole-bound resin. Pull-down assay for FdTSPO1 with tetrapyrrole-bound agarose resin. Hemin, biliverdin IXα (BV), protoporphyrin IX (PPIX), mesoporphyrin IX (MPIX), Zn(II)protoporphyrin IX (ZnPPIX), and Zn(II)mesoporphyrin IX (ZnMPIX) were covalently attached to aminoethyl-agarose (AE-agarose) resin. Purified FdTSPO1 was incubated with the resin and unbound supernatant collected (FT), washed with buffer (W), and eluted in SDS sample buffer (E). AE-agarose, which was the substrate for the covalent attachment of tetrapyrrole, was used in parallel as a negative control. F
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
Figure 5. Tryptophan quenching of TSPO1 from F. diplosiphon upon binding of tetrapyrrole and cholesterol. Tryptophan fluorescence changes were measured after excitation of 2.5 μM protein at 280 nm and emission measurement at 290−400 nm. Subsequent addition of varying concentrations of ligand (0 μM to ∼3.25 mM) caused quenching of the fluorescence signal. One representative measurement per ligand is shown for wild-type FdTSPO1 [WT (A−E)] and its LAF mutant (F): mesoporphyrin IX (A), protoporphyrin IX (B), hemin (C), biliverdin IXα (D), and cholesterol (E and F). Binding experiments were repeated at least three times. One representative measurement is shown.
Figure 6. Tetrapyrrole and cholesterol binding curves for TSPO1 from F. diplosiphon. Tryptophan fluorescence quenching was plotted against ligand concentration and fitted with the Hill equation to obtain the binding curve and binding affinity. Representative binding curves for wild-type FdTSPO1 [WT (A−E)] and its LAF mutant (F) for mesoporphyrin IX (A), protoporphyrin IX (B), hemin (C), biliverdin IXα (D), and cholesterol (E and F) correspond to the fluorescence quenching graphs from Figure 5. All binding experiments were performed with the wild-type version of FdTSPO1 (A−E) or with a mutant containing the LAF motif (F). Shown are the binding curves of the representative quenching curves in Figure 5.
we performed tryptophan fluorescence quenching assays with FdTSPO1 and putative ligands, based on assays established for Rhodobacter TSPO.6,25 Tryptophan fluorescence is very suitable
for tetrapyrrole binding affinities, we decided to conduct additional studies with FdTSPO1 to determine its quantitative binding of tetrapyrroles. To obtain quantitative binding data, G
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry Table 2. TSPO Ligand Bindingb
a
Numbers in parentheses are standard deviations and numbers (n) of replicates. bBinding affinities of human, R. sphaeroides, and F. diplosiphon TSPO1 for tetrapyrroles are compared.
Figure 7. Cholesterol binding site comparison for TSPO proteins for various organisms. Location of the putative cholesterol binding site compared to the confirmed site of mammalian TSPO, in which the LAF and CRAC motifs have been implicated in cholesterol binding. The CRAC site has a consensus sequence of (L/V)-X1−5-(Y)-X1−5-(K/R). CRAC and LAF sites are shaded and critical residues boxed. The alignment was performed with Clustal OMEGA.36 Abbreviations to the left of the alignment represent species as follows (with numbers after the slash representing the length in amino acids of protein): SynPCC6803, Synechocystis sp. PCC 6803; F.diplosiphon, F. diplosiphon; NostocPCC7120, Nostoc sp. PCC 7120; S.pombe, Sc. pombe; A.thaliana, A. thaliana; S.tuberosum, So. tuberosum; R.sphaeroides, R. sphaeroides; H.sapiens, H. sapiens; M.musculus, M. musculus; S.meliloti, Si. meliloti; P.patens, P. patens.
for TSPO proteins because these proteins have a large number of tryptophans present, with nine tryptophan residues in the case of FdTSPO1. Quenching of tryptophan fluorescence
induced by binding of the respective ligand was measured (Figure 5). Binding curves (Figure 6) were used to calculate binding affinities for the porphyrins mesoporphyrin IX H
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
not yet know the hopanoid molecules found in F. diplosiphon, we tested binding with commercially available Hop-22(29)-ene. This hopanoid is widely distributed among bacteria, including cyanobacteria.56 In our analyses, FdTSPO1 bound Hop-22(29)ene poorly (Figure S6). To determine the potential in vivo physiological implications of the binding of tetrapyrroles to FdTSPO1, we assessed growth of wild-type cells compared to a ΔFdTSPO1 mutant39 in the presence of biliverdin, which binds FdTSPO1 well. These analyses indicated a distinct growth difference for ΔFdTSPO1 under green light conditions compared to WT, though there was no significant difference between the WT and ΔFdTSPO1 mutant in red light (Figure 8). Given that FdTSPO1 is
(Figures 5A and 6A), protoporphyrin IX (Figures 5B and 6B), hemin (Figures 5C and 6C), and the open chain tetrapyrrole biliverdin IXα (Figures 5D and 6D), which were in the low micromolar range. Notably, although FdTSPO1 binds heme, it does not contain the conserved His implicated in heme binding in the Arabidopsis TSPO protein (Figure S5). Binding of cholesterol to FdTSPO1 was also measured, and cholesterol showed a low binding affinity (Figures 5E and 6E). We compared the binding affinities obtained for FdTSPO1 in this study with binding affinities reported for bacterial and mammalian TSPOs from nonphotosynthetic systems (Table 2). Mammalian TSPO binds cholesterol with nanomolar affinity, while FdTSPO1 binds cholesterol poorly at a Kd of 175 μM. TSPO from Rhodobacter (RsTSPO) was reported to have a binding affinity for cholesterol (Kapp) of ≈80 μM.43 Two motifs have been implicated in cholesterol binding, the LAF motif and the cholesterol recognition amino acid consensus (CRAC) site [(L/V)-X1−5-(Y)-X1−5-(K/R)]48−50 RsTSPO lacks the central tyrosine of the CRAC site as well as the LAF motif (Figure 7). The tyrosine residue of the CRAC site was reported to be necessary for cholesterol binding and might be important for stabilizing the ligand interaction through hydrogen bonds.51,52 On the other hand, the predictive value of the CRAC motif has been debated.53 In the case of the human nicotinic acetylcholine receptor (AChR), distinct CARC cholesterol recognition motifs (i.e., with reverse or inverted orientation of the sequence of the CRAC site) were reported. Although nearly all of these proteins also contain a CRAC motif, the CRAC site is found in regions of AChR that are not energetically favored for interacting with cholesterol.54 Using a modeling approach for molecular docking, a favorable interaction of cholesterol with a CRAC-like site that contains a phenylalanine rather than the central tyrosine was found, thereby underlining the difficulty of cholesterol binding predictions based on sequence alone.54 Indeed, the introduction of the LAF motif in RsTSPO, which as described above lacks the central tyrosine in its putative CRAC site, increases the extent of cholesterol binding to levels comparable to that observed for mammalian TSPO49 (Table 2). FdTSPO1 lacks not only the tyrosine of the CRAC site but also the conserved terminal arginine (Figure 7). To test the role of the LAF motif in this context and its effect on binding of cholesterol in FdTSPO1, we introduced the LAF motif creating FdTSPO1-LAF and purified the protein (Figure S2). The presence of this motif in FdTSPO1-LAF did not increase the level of binding of cholesterol relative to FdTSPO1 (Figure 6E,F). In fact, the level of cholesterol binding was slightly lower in the FdTSPO1-LAF mutant instead (Table 2). The level of protoporphyrin IX binding was only slightly reduced in the mutant, suggesting that the introduction of the LAF domain did not have major effects on the protein structure or function (data not shown). Analysis of RsTSPO binding showed that introducing a LAF motif aside a degenerate CRAC motif (lack of central tyrosine) increases the level of cholesterol binding to binding affinities similar to that of mammalian TSPO (LAF and CRAC side), whereas introduction of the LAF site into FdTSPO1 does not increase cholesterol binding affinities in the context of an even more degenerate CRAC side (lack of both tyrosine and arginine residues). Given that cyanobacteria contain hopanoids that are bacterial lipids structurally similar to cholesterol,55 we tested whether FdTSPO1 could bind these molecules. As the hopanoid pathway in cyanobacteria is still being elucidated and we do
Figure 8. Growth of F. diplosiphon wild type and ΔFdTSPO mutant in biliverdin-containing medium. The wild type (WT) and FdTSPO1 knockout mutant (ΔFdTSPO1) were grown in either red light (RL) or green light (GL). Cells in exponential phase that were adapted to the respective light color were diluted to an optical density at 750 nm (OD750) of 0.1 in BG-11/HEPES medium with a final concentration of 20 μM biliverdin IXα, and growth was monitored over a period of 16 days.
upregulated in green light in WT cells,39,46 these data suggest that the TSPO ligand may impact fitness in a TSPO-dependent manner under specific light conditions.
■
DISCUSSION Despite a near ubiquitous presence of TSPO in most organisms, studies exploring the ligands of TSPO and TSPO functions in vivo have yet to fully elucidate the functional roles of TSPO and which of these are impacted by binding of a ligand to TSPO. Many studies employ purified mitochondria or protoplasts to assess TSPO−ligand binding to elucidate TSPO function.57 Recently, there have been several reports that TSPO ligands, whose effects on isolated mitochondria and protoplasts or observations in TSPO-overexpressing mammalian cells have been attributed to TSPO function alone, can produce effects in these systems in a manner independent of TSPO.58,59 Indeed, previously established functions based on TSPO−ligand effects and overexpression of TSPO in cells were later shown not to require the presence of TSPO.7,60 For example, the involvement of mammalian TSPO in heme biosynthesis and porphyrin transport had been proposed. In these studies, overexpression of TSPO in mammalian cells was correlated with an increase in mRNA levels for genes encoding enzymes of the heme biosynthetic pathway.61 Additionally, binding of TSPO ligands to mitochondria resulted in a decrease in the level of porphyrin translocation.19 By contrast, no effect on heme biosynthesis was observed in mammalian TSPO knockout cells in a recent study.22 Also, no direct evidence of TSPO acting as a I
DOI: 10.1021/acs.biochem.6b01019 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry transporter of porphyrins has been reported.10 A role for mammalian TSPO in steroidogenesis had been implied on the basis of the observation that known TSPO ligands stimulated steroidogenesis, while a more recent study with knockout mice contradicted the direct involvement of mammalian TSPO in such steroidogenesis.23,58,62 It is therefore pivotal for fully understanding TSPO function and ligand-dependent TSPO roles in vivo to establish a functional connection to ligand binding. A first step toward accomplishing this goal is to identify physiologically relevant TSPO ligands followed by functional analyses of their roles in the organisms and verification of a direct link between binding of a ligand to TSPO and downstream effects on the cell. In light of the recent controversy surrounding TSPO and its implementation on drug discovery, identification of native TSPO ligands is crucial. Therefore, specifying shared highaffinity ligands, among TSPO homologues, versus speciesspecific ligands is an essential part of solving the TSPO puzzle. We show that tetrapyrrole binding affinities, specifically those of protoporphyrin IX, heme, mesoporphyrin IX, and biliverdin IXα, with the latter two not having been studied with bacterial TSPO before, are generally similar between cyanobacterial and purple bacterial TSPO while the level of cholesterol binding is substantially decreased in FdTSPO (Table 2). The binding affinities of these TSPO proteins are in the range of reported free heme levels in cells, i.e., 0.1 to
