Production of Phloroglucinol, a Platform Chemical, in Arabidopsis using a Bacterial Gene

Abstract

Phloroglucinol (1,3,5-trihydroxybenzene; PG) and its derivatives are phenolic compounds that are used for various industrial applications. Current methods to synthesize PG are not sustainable due to the requirement for carbon-based precursors and co-production of toxic byproducts. Here, we describe a more sustainable production of PG using plants expressing a native bacterial or a codon-optimized synthetic PhlD targeted to either the cytosol or chloroplasts. Transgenic lines were analyzed for the production of PG using gas and liquid chromatography coupled to mass spectroscopy. Phloroglucinol was produced in all transgenic lines and the line with the highest PhlD transcript level showed the most accumulation of PG. Over 80% of the produced PG was glycosylated to phlorin. Arabidopsis leaves have the machinery to glycosylate PG to form phlorin, which can be hydrolyzed enzymatically to produce PG. Furthermore, the metabolic profile of plants with PhlD in either the cytosol or chloroplasts was altered. Our results provide evidence that plants can be engineered to produce PG using a bacterial gene. Phytoproduction of PG using a bacterial gene paves the way for further genetic manipulations to enhance the level of PG with implications for the commercial production of this important platform chemical in plants.

Figure 1. Schematic representation of gene constructs used for plant transformation.

(a) Binary vector construct containing either bacterial (PhlDbact) or synthetic (PhlDsyn) PhlD gene under the control of Cauliflower Mosaic Virus (35S CaMV) promoter. (b) A similar construct like in (a) but the PhlD gene was fused to the chloroplast signal sequence (TP) of Arabidopsis NifS-like protein (AtCpNifS) that targets protein to chloroplasts. Restriction sites are explained in the text. OCS, octopine synthase terminator; Ω TMV, 5′ leader sequence of tobacco mosaic virus.

Figure 2. Expression analysis and quantification of PhlD transcript in transgenic lines.

Total RNA was isolated from wild-type (WT) and transgenic lines and used for RT-PCR (left panel, a,b) and qRT-PCR (right panel, c,d). Wild-type and selection marker Basta resistance gene (BAR) were used as negative and positive controls for the transgene, respectively. Primers specific to the chloroplast transit sequence (TP) were used to confirm the presence of chloroplast targeting signal sequence. Ubiquitin (UBQ5) primers were used as an internal control for normalization. Right, (c) and (d) show quantification of PhlD transcripts in transgenic lines as relative expression. The chloroplastic 19–2 transgenic line showed the highest expression. At 40 cycles no amplification product was detected in wild-type (WT) by qRT-PCR. Gels shown are cropped. All un-cropped gels are presented as “Un-cropped Fig. 2” in “Supplementary Figures and Tables file”.

Figure 3. Detection of PG in transgenic lines by GC-MS.

(a–c). Extracted ion chromatogram (EIC) of WT, cytosolic line (4–2), and chloroplastic line (19–2) extracts for m/z 342 ([M]+ for TMS3-phloroglucinol). (d) EIC of an authentic PG standard showing co-elution with the peak detected in transgenic lines: m/z 342 ([M]+ for TMS3-phloroglucinol). (e) Mass spectral match of PG standard and chloroplastic line 19–2 at 8.86 minutes demonstrating demonstrating near identical spectra between the authentic phloroglucinol and the peak produced in the chloroplastic (19–2) line.

Figure 4. Identification of the PG conjugate in the transgenic extract by GC-MS (left) and UPLC-QTOF-MS (right).

(a–c) A prominent m/z 342 signal was detected at 14.78 min in transgenic lines (b,c) and not in wild-type (a). (d) EIC of m/z 342 for orange peel extract supporting that orange peel has the same metabolite as that observed in the transgenic lines. (e) Mass spectral match for the metabolite detected at 14.78 minutes for the transgenic line 19–2 and orange peel. (f) UPLC-QTOF analysis of plant extracts showing a prominent signal eluted at 0.83 min with mass  m/z 287.076 in the transgenic line (upper panel) and not detected in the wild-type (bottom panel). (g) In-source fragmentation of the 287.076 compound showing a fragment at 125.0239, which indicates a hexose conjugate of PG. MS/MS analysis confirms that 125.0239 is a product ion of 287.0764, a pattern consistent with orange peel extract. (h) In source MS/MS fragmentation of putative PG conjugate from transgenic line 19–2 and PG standard in positive ionization mode showing that the in source 127.0375 fragment of PG conjugate matches that of PG standard and strongly indicate that the novel compound in transgenic plants represents a hexose conjugate of PG.

Figure 5. Validation of phlorin as the novel PG conjugate.

(a,b) Phlorin extract from orange peel (a) and the transgenic plant (b) extracts were treated with β-glycosidase enzyme mix (rapidase) and resolved in GC-MS before and after digestion showing that the phlorin peak disappears and the PG peak accumulates after digestion. (c) Arabidopsis leaf discs were incubated with PG in the presence or absence of D-glucose, extracted and analyzed by GC-MS. Orange peel extract was used as a control to indicate phlorin formation.

Figure 6. Quantification of PG and phlorin in transgenic lines.

(a,b) Relative quantification of free PG (a) and phlorin (b) expressed as arbitrary units (AU).

Figure 7. Phenotypes of chloroplastic transgenic lines.

Lines 3–2 and 19–2 grown in soil showed toxicity symptoms and reduced growth as compared to wild-type (WT). Arrows indicate the zoomed in seedlings (left) and leaves showing symptoms (right).

Figure 8. Metabolite variation in transgenic lines.

(a) Heat map combined with hierarchical clustering of the modified GC-MS dataset for wild-type (WT), cytosolic (3–2) and chloroplastic (9–2) transgenic lines represented as a heat map. Red color indicates a higher amount of a metabolite and green color indicates a lower amount of a metabolite. (b) Example metabolites that are significantly altered in transgenic lines.