**1. Introduction**

The development of renewable resources is one of the key actions to palliate climate change, which is largely a consequence of the world's dependence on petrol. On the other hand, contamination of the environment is an inevitable consequence of human development. These are global problems that need international agreements [1]. Biotechnology can offer solutions to these challenges, such as the production of bioethanol as a substitute to gasoline [2]. Biotechnology can also offer alternatives to the physical-chemical treatment of contaminating compounds, either by avoiding their production, or by mitigating their impact once it has occurred. The biodegradation of pollutants is, in general, a technology that has good social acceptance [3]. *Pseudomonas pseudoalcaligenes* CECT 5344 was isolated from sludge of Guadalquivir River, and it is able to use cyanide as the only source of nitrogen [4]. Cyanide is an extremely toxic compound used in the synthesis of organic compounds such as nitriles, plastics, paints, adhesives, cosmetics, etc., while mining activities and the jewellery industry are the main source of cyanurated wastes [5–8]. This strain tolerates an unusually high concentration of cyanide (up to 30 mM) [4], but it requires a suitable carbon source for growing. The sequencing of the genome of *P. pseudoalcaligenes* CECT 5344 has made it possible to predict which carbon sources can be used by this

bacterium, such as the assimilation of furanic compounds [9]. Furfurals are aromatic natural compounds formed by the dehydration of sugars during the thermochemical pre-treatment of the lignocellulosic materials to release fermentable sugars. The production of biofuels from lignocellulosic residues, which is part of the so-called second-generation biofuels, constitutes a viable option for reducing the greenhouse effect and for providing an alternative to fossil fuels [10,11]. There are different pre-treatment technologies of lignocellulosic residues. One of the parameters that has to be taken into account to optimize the process is avoiding the formation of potentially inhibitory compounds to the posterior yeas<sup>t</sup> fermentation process [12]. From the food technology perspective, furfurals are potential carcinogenic compounds used as a marker of honey adulteration, generated by acid-catalyzed dehydration of carbohydrates of food-containing sugars [13]. In any case, furfural (F), or fufuralaldehyde, and 5-hydroxymethyl furfural (HMF) are natural products that can be eliminated by using the capacity of some microorganisms to metabolize them [14–22]. Other furanic derivatives are furoic acid (FA) and furfuryl alcohol (FFA), all of them with the common thread of having an aromatic furan ring. The variety of furanic compounds degrading species is limited mostly to Gram-negative aerobic bacteria and some Gram positives [17], with a few exceptions including fungi [14]. In the first degradation route currently proposed, furfural is oxidized to 2-furoic acid (FA), which is subsequently transformed into 2-oxoglutarate, a Krebs cycle intermediate [23]. The complete metabolic pathway for the assimilation of F and HMF, as well as the genetic of the process, was first described in the soil isolate *Cupriavidus basilensis* HMF14 [24] (Figure 1). In this strain, the *hmfABCDE* gene cluster is responsible for the assimilation of furoic acid. The first reaction in the pathway is catalysed by the 2-furoyl-CoA synthetase (HmfD), producing 2-furoyl-CoA from 2-furoic acid. The conversion of 2-furoyl-CoA is into 5-hydroxy-2-furoyl-CoA in *C. basilensis*; HMF14 is catalysed by the molybdenum-dependent 2-furoyl-CoA dehydrogenase (HmfABC). The final steps of the proposed furoic-acid metabolic pathway consist of the transformation of 5-oxo-2-furoyl-CoA into 2-oxoglutarate. No gene has been assigned to the hydrolysis of the lactone, whereas *hmfE* has been proposed to encode a specific 2-oxoglutaroyl-CoA thioesterase [24] (Figure 1). *P. pseudoalcaligenes* contains an *hmfABCDE* gene cluster homologous to the gene cluster shown to be essential for the assimilation of furfural in *C. basilensis* HMF14 (Figure 1). Concretely, the amino acid sequence of HmfA from *C. basilensis* HMF14 (GenBank ADE20399.1) is 64% identical to the homologous protein of *P. pseudoalcaligenes* (BN5\_2298, 76% positives). The % identity/% similarity for the rest of the proteins are: 59%/72%, 77%/83%, 61%/75% and 80%/88%, for HmfB (GenBank ADE20400.1), HmfC (GenBank ADE20401.1), HmfD (GenBank ADE20402.1) and HmfE (GenBank ADE20403.1), respectively. Moreover, this locus also contains downstream *hmfABCDE,* a gene (*benE*) belonging to the Major Facilitator Superfamily (MSF)-family of transporters and two separate genes homologous to genes related to the assimilation of furfural in *Pseudomonas putida* Fu1 [9,25] (Figure 1). One of them belongs to the AraC-family of regulators. AraC from *P. putida* Fu1 (GenBank ACA09742.1) is 75% identical (88% similar) to its orthologous gene product in *P. pseudoalcaligenes* (BN5\_2307, Figure 1). The other upstream gene (*PsfD*) codes for a putative conserved protein usually annotated as maturation factor for molybdenum containing dehydrogenases, like xanthine and CO dehydrogenases [9,25]. In that respect, the furoyl-CoA dehydrogenase was proposed to be a molydo-protein [26]. PsfD protein from *P. putida* Fu1 (GenBank ACA09741.1) is 81% identical (89% similar) to PsfD form *P. pseudoalcaligenes* (BN5\_2306, Figure 1). To date, the architecture of this operon presented in Figure 1B has not been described in this context. Here we show that this operon is functional after an adaptation period ending up with the selection of a punctual mutant in the *araC*-type regulator. Therefore, the locus described here seems to be a hybrid furfural assimilating system containing horizontally transferred genes homologous to the catalytic genes for the assimilation of FA described in *C. basilensis* [24] and the regulatory and accessory genes described in *P. putida* Fu1 [25].

**Figure 1.** Scheme of the genetic organization of the *hmf* operon (**A**–**C**) (adapted from Ref. [9]) and predicted metabolic pathway for the assimilation of furfuryl alcohol in *Pseudomonas pseudoalcaligenes* (**D**). The *hmf* locus in *P. pseudoalcaligenes* CECT 5344 (delimited by a curly bracket, **B**) is located between BN5\_2297 (*osmC*) and BN5\_2308 (*AT*). The corresponding homologous genes in *Pseudomonas mendocina* ymp (*Pmen\_2630-2633*) are consecutive in its genome (**A**) and the syntney of the homologous genes is conserved both upstream and downstream the *hmf* locus (green arrows). The *hmf* locus contains genes homologous to that described in the context of furfural degradation in *Cupriavidus basilensis* [24] (**C1**) and in *Pesudomonas putida* [25] (**C2**). The black arrows (panel **B**) represent genes involved in the transposition of mobile genetic elements. (**D**) Proposed pathway for the assimilation of FFA in *P. pseudoalcaligenes* R1. FFADH and FDH are furfuryl alcohol and furfural dehydrogenase, respectively. \* FFADH and FDH could be the same enzyme, whose coding genes are unknown in *P. pseudoalcaligenes*. Both the oxidation of FFA and HMF in *C. basilensis* (not shown) converge in FA constituting the upper pathway. The lower pathway, encoded by the *hmfABCDE* operon (**C1**), is a series of reactions transforming FA into 2-oxo-glutaric acid. HmfA is a furoyl-CoA synthetase and HmfABC a furoyl CoA dehydrogenase. The transformation of 5-oxo-2-furoyl-CoA into 2-oxo-glutaril-CoA has no assigned gene and can be a spontaneous reaction or could be catalysed by an unspecific lactone hydrolase. Finally, HmfE was proposed to be the thioesterase rendering 2-oxo-glutaric acid from 2-oxo-glutaryl CoA [24].

#### **2. Materials and Methods**

#### *2.1. Bacterial Strains, Media and Growth Conditions*

Strain R1 is a spontaneous mutant of *P. pseudoalcaligenes* CECT5344 [27] resistant to rifampicin (40 μg/mL). *P. pseudoalcaligenes* CECT5344 strain R1D was obtained after four serial transfers of *P. pseudoalcaligenes* CECT5344 R1 to a M9 medium with furfural (10 mM) as the sole carbon source. A dilution 1:100 of the previously grown preculture was used as inoculum. For the rest of the growth curves, unless otherwise stated, the inoculum was the equivalent of an overnight culture diluted 1:10 into fresh medium. Bacterial growth was monitored by measuring the absorbance at 600 nm. Cells were

grown in either minimal medium (M9) adjusted to pH 8.5 [4] or in LB medium [28] adjusted to pH 8.5. Cell cultures were prepared in Erlenmeyer flasks filled with 1/10 (v/v) of their nominal volume in order to ensure aerobic conditions and incubated on a rotatory shaker at 190 rpm and 30 ◦C. For minimal medium, ammonium chloride (5 mM) was used as the nitrogen source and 4 g/<sup>L</sup> of sodium acetate, furfural (10 mM), furfuryl alcohol (10 mM) or furoic acid (10 mM) were added as the sole carbon source. *Escherichia coli* XL1 *blue* MRF cells (Stratagene, Agilent Technologies, Santa Clara, CA, USA) were grown aerobically at 37 ◦C in complex LB medium [28] with ampicillin (100 μg/mL). Where appropriate, the following compounds were added to the media: X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside, 0.2 mM, Appli-Chem (Barcelona, Spain), IPTG (isopropyl-β-d-1-thiogalactopyranoside, 0.5 mM, Sigma-Aldrich (St. Louis, MO, USA). Electrocompetent cells were prepared by the growth of cultures up to an optical density at 600 nm (OD600) of 0.35 and centrifugation for 20 min at 1000× *g*, followed by three successive washes (4 ◦C) in 1:1, 1:2, 1:50, and 1:500 volumes of 10% glycerol. The last solution, which yields the stock of electrocompetent cells, contained yeas<sup>t</sup> extract (0.125%) and tryptone (0.25%). A mixture of 50 μL of cells (2 × 10<sup>10</sup> to 3 × 10<sup>10</sup> CFU/mL) and 1 to 10 ng of DNA was electroporated in 2-mm cuvettes with a Bio-Rad Gene Pulser II apparatus, (Bio-rad, Hercules, CA, USA) operated at 2.5 kV, 25 μF, and 200 Ω (4- to 5-ms time constants).

#### *2.2. Preparation of Cell-free Extracts*

The cells of *P. pseudoalcaligenes* CECT 5344 R1D grown with furfuryl alcohol as a carbon source (500 mL culture) were collected by centrifugation at the end of the logarithmic phase and resuspended in 50 mM Tris-HCl (pH 8), containing a complete protease inhibitor cocktail, as recommended by the supplier (Roche, Penzberg, Germany) and glycerol (10%). The cells were disrupted by two passages through a French pressure cell operated at 130 MPa and the cell debris removed by centrifugation at 18,000× *g* for 15 min.

## *2.3. Enzymatic Assays*

Furfural dehydrogenase (FDH) and furfuryl alcohol dehydrogenase (FFADH) were assayed as previously described for *C. basilensis* [24], but optimizing the assay for *P. pseudoalcaligenes* (pH and temperature). FFADH (E.C.1.1.1.–) was assayed spectophotometrically at 65 ◦C and pH 9.5 (50 mM Tris/phosphate/carbonate) following the increment of absorbance at 340 nm due to the production of NADH (ε = 6220 <sup>M</sup>−1·cm<sup>−</sup>1). The reaction mixture contained NAD+ (1.5 mM), FFA (5 mM) and the appropriate amount of cell-free extract (50–100 μL cell-free extract, 0.5–1 mg protein, approximately) in a final volume of 1 mL. FDH (E.C. 1.2.3.1) was measured spectrophotometrically at 65 ◦C and pH 6.5 (50 mM Tris/phosphate/carbonate) following the increment of absorbance at 522 nm due to the reduction of the artificial electron acceptor DCPIP (ε = 21,000 <sup>M</sup>−1·cm<sup>−</sup>1). The reaction mixture contained, in a final volume of 1 mL, 0.33 mM PMS, 0.1 mM DCPIP, 5 mM F, and the appropriate amount of enzyme (50–100 μL cell-free extract, 0.5–1 mg protein, approximately).

#### *2.4. Chromatographic Separation of FDH and FFADH*

Cell-free extract from FFA-grown cells was loaded into an anion exchange chromatography (mono Q 5/50 GL, GE Healthcare (Chicago, IL, USA) attached to an Akta Purifier, GE. All the chormatographies were carried out at 4 ◦C. The column was equilibrated in bu ffer A (Tris/HCl 50 mM pH 8, 2 mM DTT and glycerol (2%)). The cell-free extract (2 mL) was loaded into the column at a flow rate of 1 mL/min. The unbound protein was washed with 12.5 mL of bu ffer A. Then, it was applied at a gradient of 1 mL from 0 to 0.1 M NaCl, always in the same bu ffer, which was maintained as isocratic during 5 mL, followed by a 20 mL gradient from 0.1 to 0.5 M NaCl and 0.5 to 1 M during 10 mL. Finally, the column was regenerated with 5 mL of bu ffer A containing 2 M NaCl. The fractions (0.5 mL) were analyzed for the presence FFADH and FDH activities.

## *2.5. Analytical Methods*
