New Insights into the Genetic Organization of the FK228 Biosynthetic Gene Cluster in Chromobacterium violaceum No. 968

Abstract

The biosynthetic gene cluster of FK228, an FDA-approved anticancer natural product, was identified and sequenced previously. The genetic organization of this gene cluster has now been delineated through systematic gene deletion and transcriptional analysis. As a result, the gene cluster is redefined to contain 12 genes: depA through depJ, depM, and a newly identified pathway regulatory gene, depR.

FK228 produced by Chromobacterium violaceum no. 968 is a potent histone deacetylase (HDAC) inhibitor recently approved under the commercial name Istodax by the FDA for the treatment of cutaneous T-cell lymphoma (9, 13). FK228 represents a small family of natural products that also includes spiruchostatins (12), FR901375 (11), and our newly discovered thailandepsins (Y.-Q. Cheng and C. Wang, U.S. patent application 61/235,253). All members of this family of natural products are produced by rare Gram-negative bacterial species, and each of them contains a signature disulfide bond that is known or presumed to mediate a novel mode of anticancer action in which a reduced thiol group “warhead” chelates a Zn²⁺ in the catalytic center of HDACs, thereby inhibiting the enzyme activities (2, 6).

We previously reported the cloning and preliminary characterization of the FK228 biosynthetic (dep) gene cluster in C. violaceum (2). Based upon limited genetic evidence and mostly bioinformatic analyses, we predicted 14 genes to be included in the dep gene cluster (depA through depN) (Fig. 1 A). We also proposed a hybrid nonribosomal peptide synthetase-polyketide synthase pathway for FK228 biosynthesis. In this study, we redefined the genetic organization of the dep gene cluster; our studies also revealed an unexpected gene that regulates FK228 biosynthesis.

FIG. 1.

Redefined FK228 biosynthetic (dep) gene cluster and putative operon organization. (A) All redefined dep genes are shown by solid black bars, open reading frames (ORFs) beyond the dep gene cluster are shown by white bars with solid outlines, and a pseudogene (“depN”) is shown by a white bar with a dotted outline. Previously annotated gene (orf) names that have been revised are in parentheses. Wavy lines with arrowheads indicate putative operons. (B) Relative levels of FK228 produced by the wild-type strain (CvWT), 14 gene deletion mutant strains, a complementation strain, and an overproduction strain of Chromobacterium violaceum no. 968, detected and quantified by LC-MS. Data are mean values of results from duplicate experiments, with error bars indicating standard deviation.

To precisely determine the upstream and downstream boundaries of the dep gene cluster, we employed a multiple cloning procedure (2, 7) and later a multiplex PCR method (4, 15) to have 13 genes or open reading frames systematically deleted (Table 1 and Fig. 1A; also see Materials and Methods and Table S2 in the supplemental material). depG could not be mutated despite multiple attempts, and orf19 is a very small gene paired with orf20. Therefore, these two genes were mutated together, and no attempt was made to mutate depB, depC, or depE, because they are obviously cotranscribed with depA (2) (also see below). In addition, mutants of depD and depH were created previously (2, 15). The genotype of each successful mutant was verified by PCR analysis (see Fig. S2 in the supplemental material), and the relative level of FK228 production by mutant strains was quantified by liquid chromatography-mass spectrometry (LC-MS) analysis (Fig. 1B; see also Fig. S3 in the supplemental material).

TABLE 1.

Bacterial strains used in this study

Strainc	Descriptiona	Source or reference
Chromobacterium violaceum strains
968 (FERM BP-1968)	Wild-type strain, FK228 producing, Apr Thior	IPODb (14)
Δorf3::FRT	Mutant strain with an internal part of orf3 replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
Δorf3	Mutant strain derived from the Δorf3::FRT mutant with the FRT cassette excised	This study
Δorf4::FRT	Mutant strain with an internal part of orf4 (depK) replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
Δorf4	Mutant strain derived from the Δorf4::FRT mutant with the FRT cassette excised	This study
Δorf5::FRT	Mutant strain with an internal part of orf5 (depL) replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
Δorf5	Mutant strain derived from the Δorf5::FRT mutant with the FRT cassette excised	This study
ΔdepM::FRT	Mutant strain with an internal part of depM replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
ΔdepM	Mutant strain derived from the ΔdepM::FRT mutant with the FRT cassette excised	This study
ΔdepN::FRT	Mutant strain with an internal part of depN replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
Δ“depN”	Mutant strain derived from the Δ“depN”::FRT mutant with the FRT cassette excised	This study
ΔdepA::FRT	Mutant strain with an internal part of depA replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
ΔdepA	Mutant strain derived from the ΔdepA::FRT mutant with the FRT cassette excised	This study
ΔdepD	Mutant strain with an internal part of depD deleted	2
ΔdepF::FRT	Mutant strain with an internal part of depF replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
ΔdepF	Mutant strain derived from the ΔdepF::FRT mutant with the FRT cassette excised	This study
ΔdepH	Mutant strain with an internal part of depH deleted	15
ΔdepI::FRT	Mutant strain with an internal part of depI replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
ΔdepI	Mutant strain derived from the ΔdepI::FRT mutant with the FRT cassette excised	This study
ΔdepJ::FRT	Mutant strain with an internal part of depJ replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
ΔdepJ	Mutant strain derived from the ΔdepJ::FRT mutant with the FRT cassette excised	This study
ΔdepR::FRT	Mutant strain with an internal part of depR (orf18) replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
ΔdepR	Mutant strain derived from the ΔdepR::FRT mutant with the FRT cassette excised	This study
Δorf19/20::FRT	Mutant strain with most part of orf19/orf20 (small genes manipulated together) replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
Δorf19/20	Mutant strain derived from the Δorf19/20::FRT mutant with the FRT cassette excised	This study
Δorf21::FRT	Mutant strain with an internal part of orf21 replaced with an FRT cassette (Gmr GFP+) from pPS858	This study
Δorf21	Mutant strain derived from the Δorf21::FRT mutant with the FRT cassette excised	This study
ΔdepR/pVP01-52b	The ΔdepR mutant complemented with a depR (orf18) gene carried on broad-host-range expression vector pVP01-52b based on pBMTL-3	This study
WT/pVP01-52b	Wild-type strain harboring pVP01-52b, as an overproduction strain	This study
Escherichia coli strains
DH5α	General cloning host	Lab stock
S17-1	Host strain for interspecies conjugation	Lab stock

^aAp^r, ampicillin resistance; GFP, green fluorescence protein; Gm^r, gentamicin resistance; Thio^r, thiostrepton resistance.

^bIPOD, International Patent Organism Depository, Tsukuba, Japan.

^c“depN” was previously annotated as a gene of unknown function. The current work revised it as a pseudogene that likely occupies part of the regulatory region of depA and its associated large operon.

The upstream or left boundary of the dep gene cluster was determined to begin with depM according to the following observations (Fig. 1). First, the deletion of either orf3, orf4 (previously annotated as depK), or orf5 (previously annotated as depL) had no obvious effect on FK228 production, indicating that these three putative genes play no role in FK228 biosynthesis and thus are not part of the dep gene cluster. Second, deletion of depM or “depN” resulted in a 35% to 55% decrease in FK228 production, suggesting a positive involvement of these two “genes” in FK228 biosynthesis. The deduced product of depM is an aminotransferase presumably catalyzing the removal of an amine group from a cysteinyl-S-peptidyl carrier protein (PCP) intermediate in trans to form 4-mercaptobutanyl-S-PCP in the initiation module of the FK228 biosynthetic pathway (2). “depN,” originally annotated to encode a small hypothetical protein of unknown function (2), is now believed to be a pseudogene, because no depN transcript has ever been detected under the conditions in which other dep genes are expressed (see below for details). The stretch of DNA occupied by “depN” likely contains regulatory elements upstream of depA. The observed reduction of FK228 production by the C. violaceum Δ“depN” mutant might be an unintended result of the deletion of some of those regulatory elements of depA and its associated operon. Third, deletion of further downstream genes, depA, depD, or depF, resulted in complete loss of the FK228 production, confirming the critical biosynthetic roles of those genes as predicted in the originally proposed biosynthetic model (2). Fourth, orf5 and depM share a 74-bp (A+T)-rich intergenic region but transcribe divergently, indicating that they are transcriptionally unrelated.

The downstream or right boundary of the dep gene cluster was determined to end with depR according to the following observations (Fig. 1). First, deletion of depR (previously annotated orf18) completely abolished FK228 production, and complementation of depR restored ca. 70% of the FK228 production capacity, indicating a critical role of depR in FK228 biosynthesis. Second, deletion of orf19-orf20 or orf21 did not abolish FK228 production but resulted in a slight (16 to 20%) decrease of yield, presumably due to a reduced fitness of mutants. orf19 to orf21 are bacterial housekeeping genes encoding MinE, MinD, and MinC, respectively, which are components of the conserved bacterial Min system involved in cell division and septum formation (1, 5, 8). orf19-orf20 or orf21 mutants were found to be more sensitive to antibiotics and grew much slower than the wild-type strain, and these produced variable amounts of minicells. Third, deletion of other downstream genes, depH, depI, or depJ, resulted in various degrees of reduction of FK228 production, most likely due to the accessory roles of their deduced products involved in either FK228 biosynthesis or self-resistance (2, 15).

Furthermore, the role of depR as a pathway regulatory gene governing FK228 biosynthesis was ascertained with the following evidence. First, depR appears to encode an LysR-type transcriptional activator that contains an N-terminal helix-turn-helix DNA binding domain (pfam00126) and a C-terminal LysR-type sensor domain (pfam03466) (10); within this sensor domain there are two conserved catalytic cysteine residues (C199 and C208) that could form a disulfide bridge upon sensing oxidative stress (3). Second, deletion of depR abolished FK228 production, and complementation of the ΔdepR mutant restored much of the FK228 production capacity (see above). In addition, an overproduction strain, C. violaceum wild type (WT)/pVP01-52b, containing an ectopic copy of depR on an expression vector, produced ca. 10% more FK228 than the wild type (Fig. 1B; see Fig. S3 in the supplemental material), suggesting that DepR is a rate-limiting factor in FK228 biosynthesis. This 10% increase of FK228 production also represents a moderate success of yield improvement through metabolic engineering. Third, semiquantitative reverse transcription (RT)-PCR experiments (Fig. 2) clearly indicate that the expression of most redefined dep genes, except for depJ and depR itself (mutation of depR was designed as such that a 520-bp internal part of depR was deleted, leaving a 260-bp 5′ region of depR intact for RT-PCR detection), was drastically downregulated or turned off in the ΔdepR mutant under a condition in which all genes were expressed in the WT strain. The expression of genes outside the redefined dep gene cluster was not affected by the depR deletion. Because depR was detected to cotranscribe with orf19 to orf21 and is the last gene in the operon (Fig. 1A; see below for details), depR thus does not self-regulate the operon where it resides. Considering that orf19 to orf21 are critical min genes for bacterial physiology, it is surprising to find that depR as part of a housekeeping gene operon turns out to be a gene regulating a secondary metabolite biosynthetic gene cluster. Perhaps there is an underlying link between bacterial septum formation and cell division, and between oxidative stress sensing and FK228 production, which is worth further investigation.

FIG. 2.

Transcriptional regulation of dep genes by a newly identified pathway regulatory gene, depR. (A) Detection of individual gene expression by semiquantitative RT-PCR from an RNA sample of the wild-type strain. (B) Amplification of individual gene expression by semiquantitative RT-PCR from an RNA sample of the ΔdepR mutant strain. 16S rRNA gene amplification was included as a control reaction.

Finally, the operon organization of the dep gene cluster was assessed by overlap RT-PCR. As shown in Fig. 3, transcriptional coupling between two adjacent gene pairs, depA and depB, depB and depC, depC and depD, depD and depE, depE and depF, depG and depH, and depI and depJ, but not between depM and “depN,” “depN” and depA, or depH and depI, could be detected by RT-PCR. Transcriptional coupling between gene pairs orf21 and orf20, orf20 and orf19, and orf19 and depR was also detected. Therefore, the putative operon organization of the dep gene cluster can be assigned as depM alone, depABCDEFGH, depIJ, and depR cotranscribed with orf21, orf20, and orf19 (Fig. 1A). depJ appears to be able to transcribe from two alternative transcriptional start sites; one is located upstream of depI so that depI and depJ are cotranscribed, and this operon is under the control of depR; another one is located just upstream of depJ itself, independent of depR regulation.

FIG. 3.

Detection of putative operon organization of the dep gene cluster and neighboring genes. (A) Detection of transcriptional coupling between two adjacent genes by overlap RT-PCR from an RNA sample of the wild-type strain. (B) Control reactions showing PCR amplification of an intergenic region of two adjacent genes from a genomic DNA sample of the wild-type strain. The unusually large amplicon of depD-depE was due to a special consideration in primer design to avoid two highly homologous intragenic regions toward the 3′ end of depD.

In conclusion, the dep gene cluster is now redefined to contain 12 genes: depA through depJ, depM, and a newly identified pathway regulatory gene, depR. Those genes are organized into four putative operons, depABCDEFGH, depIJ, depM, and depR cotranscribed with orf21, orf20, and orf19 in opposite orientation. Precise determination of the gene organization of the dep gene cluster and identification of a pathway regulatory gene should facilitate complete elucidation of the mechanism by which FK228 is biosynthesized, production of FK228 in heterologous hosts, yield improvement of FK228 through metabolic engineering, and discovery and engineering of FK228 analogs as anticancer lead compounds.