Perquinolines A-C: Unprecedented Bacterial Tetrahydroisoquinolines Involving an Intriguing Biosynthesis.

Metabolic profiling of Streptomyces sp. IB2014/016-6 led to the identification of three new tetrahydroisoquinoline natural products, perquinolines A-C (1-3). Labelled precursor feeding studies and the cloning of the pqr biosynthetic gene cluster revealed that 1-3 are assembled by the action of several unusual enzymes. The biosynthesis starts with the condensation of succinyl-CoA and L-phenylalanine catalysed by the amino-7-oxononanoate synthase-like enzyme PqrA, representing rare chemistry in natural product assembly. The second condensation and cyclization events are conducted by PqrG resembling an acyl-CoA ligase. Last, ATP-grasp RimK-type ligase PqrI completes the biosynthesis by transferring a γ-aminobutyric acid or β-alanine moiety. The discovered pathway represents a new paradigm for assembly of the tetrahydroisoquinoline cores of natural products.

concentration of 0.20-0.30 nM and a protein concentration of 3-6 µg/test tube. Receptor density was determined in the range of B max = 2500-3000 fmol/mg protein. Opioid receptor binding was determined using the radioligand [ 3 H]diprenorphine (specific activity of 31 Ci/mmol, PerkinElmer, Rodgau, Germany) with a K D value of 0.068±0.018 nM±SEM at a final concentration of 0.20-0.30 nM, a protein concentration of 3-10 µg/test tube and a B max = 1400-3000 fmol/mg protein.
Compounds were tested in the range of 1 nM to 100 µM as triplicates. Unspecific binding was determined in the presence of 10 µM CGP12177 (2AR) or naloxone (µOR). Protein concentration was established by the method of Lowry using bovine serum albumin as standard. [4] Resulting competition curves were analyzed by nonlinear regression using the algorithms for one-site competition of PRISM 6.0 (GraphPad, San Diego, CA). EC 50 values derived from the resulting dose response curves were transformed into the corresponding K i values utilizing the equation of Cheng and Prusoff. [5] Preparation and manipulation of DNA.
DNA extraction and manipulation, E. coli transformation and RP4-based E. coli -Streptomyces conjugation were performed according to standard procedures. [1,6] The Phusion DNA polymerase (Thermo Fisher Scientific, USA) was used to amplify fragments used for gene expression. Dream Taq polymerase (Thermo Fisher Scientific, USA) was used for PCR for the verification of gene deletions in cosmids or chromosome of Streptomyces strains. DNA fragments were purified from agarose gels using the QIAquick Gel Extraction Kit (Qiagen, Germany). Plasmid and chromosomal DNA were purified with QIAprep Spin miniprep kit and DNeasy Blood and Tissue Kit (Qiagen, Germany). DNA processing enzymes used in this work were obtained from New England Biolabs (USA). All constructs were verified by sequencing. Sanger sequencing and oligonucleotides synthesis was provided by Eurofins Genomics, Germany. Oligonucleotides used in this study are listed in Table S2.

Sequencing and analysis of Streptomyces sp. IB2014/016-6 genome.
Genomic DNA of Streptomyces sp. IB2014/016-6 was isolated from 30 mL culture grown in TSB at 28°C for 48 hours. Total DNA isolation was performed according to the salting out procedure followed by RNase treatment. [1] Isolated DNA was sequenced by Illumina sequencing. The purity and concentration of the genomic DNA was determined using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). For sequencing an Illumina paired-end sequencing library (TruSeq sample preparation kit; Illumina, USA) was constructed according to the manufacturer's protocol. The Streptomyces sp. IB2014/016-6 draft genome sequence was established on an Illumina MiSeq system in rapid run mode (2x250 nt) with a pair distance of about 500 bp. Upon sequencing and processing of the obtained data, a de novo assembly was performed using the SPAdes Assembler (version 3.8.1) with default settings. [7] Annotation of the genome was performed by means of NCBI Prokaryotic Annotation Pipeline. [8] For the identification of secondary metabolites clusters antiSMASH 3.0 was used. [9] The assembled and annotated draft sequence of the Streptomyces sp. IB2014/016-6 was deposited in the GenBank database under accession number SGBF00000000.

Generation of the construct for pqrA gene inactivation.
Two DNA fragments, D21-1 and D21-2, flanking gene pqrA have been amplified using the primer pairs DpqrA1BamHI/DpqrA1EcoRV and DpqrA2EcoRV/DpqrA2XbaI (Table S2). The obtained 2 kb fragments were cloned into a pJET1.2 vector using CloneJET PCR cloning kit (Thermo Fisher Scientific, USA). The construct containing fragment C21-1 was digested with EcoRV (primer) and XbaI (vector MCS) and ligated with the C21-2 fragment, which was retrieved with the same restriction enzymes. The resulting plasmid was digested with EcoRV and ligated with the spectinomycin-resistance cassette aadA from patt-saadA. [10] The final construct was cloned into pKGLP2 [11] as BamHI-XbaI giving pKGLPpqrA::aadA.
This construct was transferred into Streptomyces sp. IB2014/016-6 via an intergeneric conjugation as described. [6c] The exconjugants were grown under non-selective conditions and screened for white spectinomycin-resistant colonies when grown on MS supplemented with 70 µg mL -1 X-gluc. The deletion of pqrA gene was confirmed by PCR using the 16-6-28315ChF/16-6-28315ChR primer pair.
Cloning of the pqr-gene clusters.  [12] Transformants were selected on YNB medium supplemented with the Yeast Synthetic Drop-out Medium Supplements without leucine (Sigma-Aldrich, USA).
Colonies were plated by patches of 100, washed of and analysed by PCR for the presence of clones harbouring the desired construct using primers pCLY10HindV and pCLY10NotV annealing to opposite sides of vector and annealing to the cloned region outside of the homology fragment (Table S2). Also, primers to the central part of both clusters were used to verify the cloned fragments 16-6Cl21ChF/16-6Cl9ChR and B24891ChF/B24891ChR, respectively. Positive clones were further pooled out, the total DNA was purified using standard protocol [13] and transformed into E. coli XL1Blue to give p10-86-16-6 and p3-47B24891 clones carrying the desired region of Streptomyces sp. IB2014/016-6 and S. odonnellii NRRL B24891 chromosome. Constructs were introduced into S. lividans TK24 by the intergeneric conjugation. Resulting strains were grown in NL19 medium and the production of 1-3 was analysed as described above.
Cloning, expression, purification and characterization of PqrA.
pqrA gene was amplified with the primers 16-6-pqrAnFNdeI and 16-6-pqrAnRxho using Streptomycetes sp. IB2014/016-6 chromosomal DNA as a template. The obtained fragment was digested with NdeI and XhoI and inserted into a similarly digested pET28b vector to generate pET28b-pqrA. PqrA structure was modeled using SWISS-MODEL Workspace using E.
coli 2-amino-3-ketobutyrate coenzyme A ligase (1fc4.1.A) as template. [14] Point mutations in pqrA were introduced using QuikChange II kit (Agilent Technologies, USA) and primers listed in Table S2. 5% of material after PDA Detector was re-directed to SQ Detector 2 for mass spectrometry analysis (Waters, Germany).
Fractions containing 4 and 5 were collected by 2767 Sample Manager based on mass detection, evaporated and dissolved in methanol.

Construction of the mutant strains.
The constructs with deletion of the individual pqr genes were generated using the PCR-based λ-Red recombination technique [15] and a hygromycin resistance marker from the IMES. [16] Gene deletions were confirmed by PCR. Primers used to amplify the hygromycin cassette and to verify the mutation are listed in Table S1. The mutagenized constructs were introduced into S. lividans TK24 J1074 by RP4-based conjugation from E. coli ET12567 containing pUB307 [17] and selected for with apramycin resistance. Integration of constructs into genome of recipient strain was proved by PCR. Strains were cultured as described above for 1-3 production, metabolites were extracted with butanol from cultural liquid and aceton:methanol (1:1) from biomass and analyses by high resolution LC-MS with m/z range from 50 to 1000 [M+H] + . When needed ms2 experiments were performed.
85 ± 12 µM 62 ± 15 µM Figure S4. Incorporation of isotopic labeled precursors into 1.   Figure S5. Nucleotide alignment of pqr gene clusters from Streptomyces sp. IB2014/016-6 and S. odonnellii NRRL B24891. Alignment was performed with Geneious v. 8.1.7 software (Biomatters Ltd., New Zealand) using Clustal W algorithm. Red correspond to less than 30% identity, green-brownmore than 30% but less than 100%, green corresponds to 100% identity.          The amount of product of reaction (when detected) was estimated from the corresponding peak area and peak of 4 was taken as 100%.