Biochemical Evidence That phaZ Gene Encodes a Specific Intracellular Medium Chain Length Polyhydroxyalkanoate Depolymerase in Pseudomonas putida KT2442

Polyhydroxyalkanoates (PHAs) can be catabolized by many microorganisms using intra- or extracellular PHA depolymerases. Most of our current knowledge of these intracellular enzyme-coding genes comes from the analysis of short chain length PHA depolymerases, whereas medium chain length PHA (mcl-PHA) intracellular depolymerization systems still remained to be characterized. The phaZ gene of some Pseudomonas putida strains has been identified only by mutagenesis and complementation techniques as putative intracellular mcl-PHA depolymerase. However, none of their corresponding encoded PhaZ enzymes have been characterized in depth. In this study the PhaZ depolymerase from P. putida KT2442 has been purified and biochemically characterized after its overexpression in Escherichia coli. To facilitate these studies we have developed a new and very sensitive radioactive method for detecting PHA hydrolysis in vitro. We have demonstrated that PhaZ is an intracellular depolymerase that is located in PHA granules and that hydrolyzes specifically mcl-PHAs containing aliphatic and aromatic monomers. The enzyme behaves as a serine hydrolase that is inhibited by phenylmethylsulfonyl fluoride. We have modeled the three-dimensional structure of PhaZ complexed with a 3-hydroxyoctanoate dimer. Using this model, we found that the enzyme appears to be built up from a coreα/β hydrolase-type domain capped with a lid structure with an active site containing a catalytic triad buried near the connection between domains. All these data constitute the first biochemical characterization of PhaZ and allow us to propose this enzyme as the paradigmatic representative of intracellular endo/exo-mcl-PHA depolymerases.

Polyhydroxyalkanoates (PHAs) can be catabolized by many microorganisms using intra-or extracellular PHA depolymerases. Most of our current knowledge of these intracellular enzyme-coding genes comes from the analysis of short chain length PHA depolymerases, whereas medium chain length PHA (mcl-PHA) intracellular depolymerization systems still remained to be characterized. The phaZ gene of some Pseudomonas putida strains has been identified only by mutagenesis and complementation techniques as putative intracellular mcl-PHA depolymerase. However, none of their corresponding encoded PhaZ enzymes have been characterized in depth. In this study the PhaZ depolymerase from P. putida KT2442 has been purified and biochemically characterized after its overexpression in Escherichia coli. To facilitate these studies we have developed a new and very sensitive radioactive method for detecting PHA hydrolysis in vitro. We have demonstrated that PhaZ is an intracellular depolymerase that is located in PHA granules and that hydrolyzes specifically mcl-PHAs containing aliphatic and aromatic monomers. The enzyme behaves as a serine hydrolase that is inhibited by phenylmethylsulfonyl fluoride. We have modeled the three-dimensional structure of PhaZ complexed with a 3-hydroxyoctanoate dimer. Using this model, we found that the enzyme appears to be built up from a core ␣/␤ hydrolase-type domain capped with a lid structure with an active site containing a catalytic triad buried near the connection between domains. All these data constitute the first biochemical characterization of PhaZ and allow us to propose this enzyme as the paradigmatic representative of intracellular endo/ exo-mcl-PHA depolymerases.
Polyhydroxyalkanoates (PHAs) 2 are biodegradable polyoxoesters produced by a wide range of bacteria (see Fig. 1). These compounds are accumulated as reserve granules (carbon and energy storage material) in the cytoplasm when the environmental conditions are not optimal for growth due to the limitation of a required nutrient and an excess of a carbon source (1)(2)(3)(4)(5). When growing conditions turn to carbon starvation, accumulated PHA can be biodegraded to monomers and/or oligomers, which are then reutilized by microorganisms as carbon and energy sources (6). PHAs are built from 3-hydroxyfatty acid monomers where the carboxyl group of one monomer forms an ester bond with the hydroxyl group of the neighboring monomer (2). Because the hydroxyl-substituted carbon atoms are of the R configuration conferring chirality to the biopolymer (7), PHAs have been proposed as a source for the production of chiral (R) hydroxyalkanoic acids, scaffolds for the synthesis of value added products (8). The R moiety that can vary from 1 to 11 carbon atoms is also present at this 3 position (Fig.  1), and, thus, short chain length PHAs (scl-PHAs) contain monomers consisting of 4 -5 carbon atoms, whereas medium chain length PHAs (mcl-PHAs) are formed of monomers of 6 -14 carbon atoms (2).
PHAs can be catabolized by many microorganisms through extracellular or intracellular processes depending on the PHA localization (6). In extracellular degradation, exogenous PHA is utilized as a carbon and energy source. Extracellular PHA is released by producer microorganisms after death, and the granules spread into the environment are further hydrolyzed by secreted enzymes into water-soluble oligomers and monomers. The ability to degrade extracellular scl-PHAs is widespread among bacteria in comparison to that of mcl-PHAs. Thus, many extracellular scl-PHA depolymerases have been characterized in depth over the last decade, and Ͼ20 genes have been identified (6, 9 -13). The prototype of extracellular mcl-PHA depolymerases is that of Pseudomonas fluorescens GK13, and only its coding gene and its homologous genes in other strains * This work was supported by the Comisió n Interministerial de Ciencia y Tec-have been isolated and characterized (6,14). On the other hand, when a producer microorganism requires a carbon source, intracellular PHA can be hydrolyzed by intracellular depolymerases, which seem to be anchored to the PHA granule (6,15,16). In contrast to extracellular depolymerases, intracellular degradation of previously accumulated PHAs is poorly understood. However, during the last years, the analysis of the genomes of a paradigmatic polyhydroxybutyrate (PHB) producer bacteria like Ralstonia eutropha strain H16 and other related strains as well as the cloning and characterization of several PHB depolymerases and oligomer hydrolases have revealed the existence of a very complex PHB hydrolytic system for this model microorganism (17)(18)(19)(20)(21)(22)(23). At least four types of enzymes and some enzyme paralogs have been proposed as putative components of this complex hydrolytic system (21). Depending on the depolymerase isoenzymes synthesized by the cell, the end reaction products can be oligomers, monomers, dimers, or a mixture of oligomers, but the function of each depolymerase isoenzyme has not yet been clearly determined (17)(18)(19)(20)(21)(22)(23). Moreover, the expression of these isoenzymes should be tightly regulated to guarantee the hydrolysis of the biopolymer when required for its complete assimilation, but the characterization of such complex regulatory system is just at the preliminary stages (24,25).
The only example of intracellular mcl-PHA degradation described so far is that of the mcl-PHA mobilization in different bacterial species belonging to the genera Pseudomonas. The pha genes involved in mcl-PHA metabolism in these strains have been cloned and sequenced (see Fig. 2). This pha cluster is composed of two synthase-coding genes (phaC1 and phaC2), a putative regulatory gene (phaD), two coding phasin genes having structural and regulatory functions (phaF and phaI) (24,26), and the phaZ gene located between the synthase-coding genes whose product is similar to members of the family V of bacterial lipolytic enzymes containing a potential lipase box (27). Interestingly, the first report describing a process of mcl-PHA mobilization demonstrated the existence of self-hydrolysis of the granules in a strain of Pseudomonas oleovorans (15,28), whereas an mcl-PHA intracellular depolymerase activity was ascribed to the phaZ gene firstly in Pseudomonas putida GPo1 (formerly known as P. oleovorans GPo1) (26) and later in P. putida U only by using mutagenesis/complementation experimental approaches (29,30). Nevertheless, a biochemical demonstration that phaZ encodes a depolymerase enzyme remained to be presented.
In this work the PhaZ from P. putida KT2442 has been purified and characterized after its overexpression in a heterologous host. Therefore, we presented the first formal biochemical evidence that PhaZ is certainly an intracellular mcl-PHA depolymerase, becoming the prototype of this relevant enzyme family.

EXPERIMENTAL PROCEDURES
Materials-n-Phenylalkanoic acids and n-alkanoic acids were supplied by Lancaster Synthesis or by Sigma. [1-14 C]Octanoic acid (50 mCi/mmol) was from Americans Radiolabeled Chemicals. All other products were of analytical quality or high-performance liquid chromatography grade.
DNA Manipulations and Plasmid Constructions-DNA manipulations and other molecular biology techniques were essentially performed as described previously (34). Transformation of E. coli cells was carried out by using the RbCl method or by electroporation (GenePulser, Bio-Rad) (40). Plasmid transference to the target Pseudomonas strains was done by the filter-mating technique (35). DNA fragments were purified by standard procedures using Gene Clean (Bio 101, Inc., Vista, CA). Plasmids and oligonucleotides used in this study are listed in Table 1. To construct pPAZ1, the 904-bp DNA fragment coding for the PhaZ depolymerase was PCR-amplified by using the oligonucleotides PHAZ1 and PHAZ2 using the total DNA of the strain P. putida KT2442 as template. For PCR amplifications, we used 2 units of AmpliTaq DNA polymerase (PerkinElmer Life Sciences), 1 g of DNA template, 1 g of each deoxynucleoside triphosphate, and 2.5 mM MgCl 2 in the buffer recommended by the manufacturer. Conditions for amplification were chosen according to the G ϩ C content of the corresponding oligonucleotides. The PCR product was digested with the engineered endonucleases (Table 1) and cloned in pUC18 or in the corresponding expression vector. Similar procedures were applied to generate the plasmids pPAZ2 and pPAZ3 (Table 1). In the case of pPAZ2, the oligonucleotides applied were PHAZ3 and PHAZ4, and the resulting PCR product was of 891 bp. For pPAZ3 construction, the primers PHAZQE and PHAZ2 were used, and the PCR amplification product was of 884 bp. Nucleotide sequences were determined directly with the same oligonucleotides used for cloning. Standard protocols of the manufacturer for TaqDNA polymeraseinitiated cycle sequencing reactions with fluorescently labeled dideoxynucleotide terminators (Applied Biosystems Inc.) were used. All these constructions were confirmed by sequencing using an ABI Prism 3730 DNA Sequencer.
Purification of PhaZ Depolymerase-A preculture of E. coli M15 (pREP4, pPAZ3) cells was incubated overnight in LB medium plus ampicillin and kanamycin. Culture was diluted until an optical density at 600 nm (A 600 ) of 0.1. When A 600 reached 0.5 the culture was induced with 0.1 mM isopropyl 1-thio-␤-D-galactopyranoside, and cells were further incubated for 4 h. 500 ml of culture was resuspended in 40 ml of lysis buffer (50 mM sodium phosphate, pH 8.0, 5 mM Tris-HCl, 20 mM imidazole, 300 mM NaCl). Cells were broken by a 4-fold passage through a French press (1,000 p.s.i.) and centrifuged at 27,000 ϫ g. The supernatant was loaded onto 2 ml of Ni-NTAagarose column (Qiagen) equilibrated with the washing buffer (50 mM sodium phosphate, pH 8.0, 5 mM Tris-HCl, 300 mM NaCl, 75 mM imidazole) at 0.5 ml/min. After loading, the column was washed with ten volumes of the same washing buffer and then eluted with the elution buffer (50 mM sodium phosphate, pH 8.0, 5 mM Tris-HCl, 300 mM NaCl, 500 mM imidazole) at 0.25 ml/min. All the purification steps were carried out at 4°C.
Electrophoretic and Immunological Techniques-SDS-PAGE was performed routinely as described before (34). Western blot analysis was performed with the ECL Western Blotting Detection Kit (Amersham Biosciences) according to the protocol described by the manufacturer. Rabbit polyclonal antiserum against PhaZ was generated by using SDS-PAGE-separated PhaZ depolymerase material as the antigen. Protein band quantifications were carried out by using a PhosphorImager (Molecular Dynamics).
Characterization of the Biopolyesters PHAs and PHPhAs-PHAs and PHPhAs were isolated following the procedure reported before (41) from cells grown in the described media (see above). The PHPhA and PHA compositions were analyzed by NMR and gas chromatography as indicated previously (39,42).
PHA Depolymerase Radioactive Assay-To prepare the substrate for the radioactive assay P. putida GPo1 strain was cultured in 200 ml of defined medium as described for P(HO-co-HX) production (see above), in the presence of 10 Ci of [ 14 C]octanoic acid. After 24 h of incubation, cells were harvested and resuspended in 3 ml of TE buffer (10 mM Tris⅐HCl, pH 7.5, 1 mM EDTA) plus 1 mg/ml lysozyme (Sigma) and equal volume of glass beads (150 -212 m, Sigma). The mixture was kept at 37°C for 60 min and disrupted by shaking in a vortex mixer. The pellet fraction was isolated by centrifugation at 16,000 ϫ g for 25 min and suspended on 2 ml of saline solution. To eliminate nucleic acids, 2 mg of DNase (2,000 units/mg, Sigma) and 0.1 mg of RNase (100 units/mg, Sigma) were added and incubated 1 h at 37°C. Finally, the mixture was incubated overnight at 60°C in the presence of 1 mg of proteinase K (30 units/mg, Sigma). The solution was centrifuged at 16,000 ϫ g for 25 min, and the pellet fraction was dissolved in 10 ml of chloroform. To extract water-soluble cell components, 2 ml of water was added and the organic and aqueous phases were mixed by vortexing and then separated by centrifugation at 5,000 ϫ g. The organic phase was transferred into a new tube. The procedure was repeated by adding 10 ml of chloroform to the aqueous phase. The chloroform solution was precipitated in ice-cold methanol (10-fold excess) under vigorous stirring. After decanting the methanol/chloroform solvent mixture, the resulting polymer was air-dried overnight and stored at room temperature. The yield was 80 mg of [ 14 C]P(HO-co-HX) (18 nCi/mg). This labeled polymer, hereafter named PHA*, was used to prepare a polymer/water emulsion (PHA* latex) as previously described (14), which was applied as substrate to assay the depolymerase activity of the PhaZ enzyme. To avoid unspe-cific binding of the PhaZ to the polymer substrate, 25 l of PHA* latex (6 g/l) was preincubated in the presence of 0.02 mg/ml lysozyme for 10 min at room temperature. The assay mixture of 250 l contained 0.6 mg/ml lysozyme-preincubated PHA* latex, 0.2 M Tris-HCl, pH 8.0, 0.3 M NaCl, and 5 l of crude extract (10 g/l) or purified enzyme (0.1 g/l). The mixture was incubated at 37°C for 30 min, and the reaction was stopped by adding 10 l of formaldehyde. The samples were centrifuged at 16,000 ϫ g for 25 min at 4°C, and the radioactivity in 200-l portions of the supernatant was determined in a scintillation counter. One unit of depolymerase activity by this method (units*) was defined as the amount of enzyme that catalyzes the hydrolysis (solubilization) of 1 g of PHA* in 1 min. This assay was used to determine optimum pH and temperature as well as to study the thermal stability of PhaZ. For this latter determination, purified PhaZ was incubated from 20 to 65°C at time intervals from 5 to 60 min.
PHA Depolymerase Turbidimetric Assay-Native granules of P(HO-co-HX) (P(HO-co-HX)n) were prepared from crude extracts of P. putida GPo1 strain cultured in 200 ml of defined medium as described for (P(HO-co-HX)) production (see above) by glycerol density gradient centrifugation (37). Artificial cholate-coated granules ((P(HO-co-HX)s) were prepared by emulsifying 300 l of a PHA or PHPhAs solution in chloroform (12.5% (w/v)) in 3 ml of 50 mM sodium cholate by ultrasonication and subsequent evaporation of the solvent following the method described previously (43). These artificial granules were further purified by glycerol density gradient centrifugation. Artificial granules of (P(HO-co-HX) were also prepared in the absence of cholate (P(HO-co-HX)w) as described previously (14). A turbidimetric assay method was performed as described elsewhere (11). A reaction mixture of 500 l contained 1.8 g of PhaZ, 250 g of substrates, 0.2 M Tris-HCl, pH 8.0, 0.3 M NaCl. One unit of depolymerase activity by turbidimetric assay (units A ) is defined as the amount of enzyme that catalyzes the decrease of one A 650 unit per min. Alternatively, one unit of depolymerase activity (units PHA ) is defined as the amount of enzyme that catalyzes the hydrolysis (solubilization) of 1 g of PHA in 1 min.
Lipolytic and Esterase Activity Assays-Rhodamine B-1,2,3tri-(9Z-octadecenoyl)-sn-glycerol and 1,2,3-tributyryl-sn-glycerol plates were used to test PhaZ lipolytic activity as described elsewhere (44). The activity assays against other triacylglycerols were carried out by both rhodamine B plates and turbidimetric methods. Esterase activity against p-nitrophenylhydroxyalkanoates with a chain length of the fatty acid moiety of 2-16 carbon atoms was performed according to the procedure described previously (11).
Identification of the Products Released from PHA after Enzymatic Hydrolysis-For the identification of the PhaZ hydrolysis products, four reaction mixtures were subjected to enzymatic hydrolysis in parallel with 250 g of P(HO-co-HX)w at different reaction times (1,12,30, and 360 min). The first three mixtures were developed in the presence of 1.8 g of PhaZ as a standard turbidimetric assay. Because the enzyme loses most of the activity after incubation at 37°C for 30 min, the last reaction mixture was supplemented every 60 min with an aliquot of 1.8 g of functional PhaZ, to ensure the complete hydrolysis of the polymer. The degradation products were analyzed by analyzing the supernatant of the reaction mixture by mass spectrometry. The LC-MS experiments were carried out on a Finnigan Surveyor (Thermo Electron) pump coupled with a Finnigan LCQ Deca (Thermo Electron) ion trap mass spectrometer. The separation was performed at room temperature on a 100 ϫ 2.1 mm (3-m particle size) Hypersil HyPurity C18 column (Thermo Electron) at a flow rate of 100 l/min and an injection volume of 5 l. The mobile phase was 0.1% ammonium hydroxide in water (A) and 0.1% ammonium hydroxide in methanol (B). The following elution program was used: at the start 95% A and 5% B; after 3 min the percentage of B was linearly increased to 95% in 7 min, then kept constant for 20 min, ramped to the original composition in 1 min, and then equilibrated for 10 min. The detection was monitored by MS-ESI(Ϫ) spectrometry at a source voltage of 4.5 kV and at a capillary heat of 200°C. All spectra were recorded in full scan mode (m/z ϭ 50 -1500).
Homology Modeling of PhaZ-The three-dimensional structure of PhaZ was modeled using Swiss PDB Viewer 3.7 (45). The template used was the hydrolase CumD from P. fluorescens IP01 complexed with isobutyrate (46) (PDB code 1IUP). Raw structures obtained from fitting were subjected to steepest descent energy minimization. A dimer of 3-hydroxyoctanoic acid (HO) was generated using the ChemOffice 8.0 utilities (CambridgeSoft) and was manually docked onto the PhaZ model using RASTOP 2.0.2. (available at www.geneinfinity.org/ rastop/). The figures were rendered with RASMOL 2.7.1 (47).

RESULTS
In Vivo Functionality of PhaZ from P. putida KT2442 by Using P. putida U (PhaZ Ϫ ) as Host-P. putida KT2442 is a model strain in environmental biotechnology characterized by a wide metabolic and physiologic versatility (32,48) that is able to produce mcl-PHA from a broad carbon substrate range (49). The pha gene cluster of this strain is similar to that of other mcl-PHAs producers (45-99% identity) (Fig. 2). Remarkably, the phaZ gene showed a 96% identity with the homologous gene of P. putida GPo1 and U strains (26,29,30). To confirm the functionality of the phaZ gene of P. putida KT2442, we tested its capacity to complement the mutant strain P. putida U (PhaZ Ϫ ) by constructing the plasmid pPAZ2, a phaZ shuttle expression vector derived from pBBR1MCS5, able to replicate in Pseudomonas strains. After 24 h of fermentation in the presence of octanoic acid the P. putida U (PhaZ Ϫ ) (pPAZ2) recombinant cells produced only 12.5% of PHA (w/w of dry cell weight), which is just half of the amount of PHA produced by the parental strain P. putida U (PhaZ Ϫ ) transformed with the control plasmid pBBR1MCS5 (27% w/w of dry cell weight) (Fig.  3). This result demonstrated that the cloned phaZ gene of P. putida KT2442 produces a functional PhaZ that is able to complement the PhaZ Ϫ deficiency reducing the PHA content of the mutant cells. Nevertheless, although the reduction of PHA content confirmed that phaZ gene of P. putida KT2442 was truly functional, this was still an indirect evidence of its putative depolymerase role and a direct biochemical demonstration that this gene certainly encodes such enzyme remained to be investigated.
Cloning and Overexpression of the phaZ Gene from P. putida KT2442 in E. coli-The first approach to biochemically characterize the phaZ gene product was to construct a recombinant E. coli strain producing the PhaZ protein. Fig. 4A (lane 2) shows that E. coli DH5␣ (pPAZ1) overproduced a protein of 33 kDa, which is in agreement with the molecular mass of PhaZ deduced from its amino acid sequence (31.4 kDa). Although most of the overproduced PhaZ protein was deposited with the insoluble fraction of the crude extract (Fig. 4A, lanes 3-6) a Western blot analysis demonstrated that a very low but detect-able amount of soluble PhaZ protein was present in the supernatant of the crude extract after ultracentrifugation (Fig. 4B).
We analyzed the activity of the PhaZ-soluble enzyme using p-nitrophenyl esters and P(HO-co-HX) in its different forms (native granules or latex preparations) as putative substrates. Assays were tested by gas chromatography-MS analysis of the supernatant (50), turbidimetric method (11), or gravimetric analysis (28). Unfortunately, we did not detect any activity of PhaZ by these analyses.
Assuming that these methods could not be sensitive enough to detect the PhaZ activity in crude extracts, we set up a novel and more sensitive assay to detect the PHA hydrolysis in a quantitative way. To this aim, we generated a [ 14 C]P(HOco-HX) latex (PHA* latex) from P. putida GPo1 strain by incorporating 14 C-labeled monomers into the polyester. Cells were cultured under PHA production growth conditions in the presence of 14 C-labeled octanoic acid as carbon source. Remarkably, we were able to detect the solubilization of 35 g of PHA* latex after its incubation at 37°C overnight with the soluble crude extract of E. coli DH5␣ (pPAZ1). As expected, the control sample from E. coli DH5␣ (pUC18) did not release to the supernatant radioactive-soluble products. These results demonstrated that PhaZ produced in E. coli was responsible for the polymer solubilization and, moreover, the new developed method paved the way to detect the activity of the PhaZ depolymerase in a very high sensitive assay extremely useful to deal with its purification.
Purification of PhaZ Depolymerase and Characterization of Its Biochemical Properties-To purify the protein PhaZ by affinity chromatography, we constructed the His-tagged PhaZ overproducer strain E. coli M15 (pREP4, pPAZ3). Although most of the fused PhaZ protein sedimented as insoluble inclusion bodies (Fig. 5A), we were able to detect PHA hydrolysis in the soluble supernatant of crude extracts, suggesting that at least a significant fraction of the protein was produced in a soluble form (Fig. 5B). Thus, the soluble PhaZ enzyme could be purified by metal-chelate chromatography (Fig. 5A) showing a specific activity of 14,100 units*/mg of protein. PhaZ showed the optimal activity at pH 8.8 and 43°C. However, the enzyme was very unstable above 37°C, becoming inactive after 30 min of incubation at this temperature. Depolymerase activity decreases to 25 and 50% at pH 7.0 and 10.5, respectively. For optimal activity, it was extremely important to maintain a high ionic strength (300 mM NaCl) in the reaction mixture (Table 2).  A similar effect was observed using KCl instead of NaCl, suggesting that the effect was not cation-dependent. Particularly, although MgCl 2 did not affect enzyme activity, CaCl 2 caused a significant inhibition (88%). Serine esterase inhibitors, such as phenylmethylsulfonyl fluoride or dodecyl sulfonyl chloride, inhibited the depolymerase activity up to 70% suggesting that the enzyme behaves as a typical serine hydrolase (Table 2). However, reducing agents such as dithiothreitol or 2-mercaptoethanol only reduced 20% the enzymatic activity, suggesting that cysteines are not involved in enzyme activity. Ionic and nonionic detergents, except sodium cholate, clearly inhibited PhaZ activity, even at very low concentrations (Table 2).
Substrate Specificity Range of the PhaZ Depolymerase-Although turbidimetric assays did not provide enough sensitivity to detect PhaZ activity in crude extracts, once PhaZ was purified we reconsidered this method to comparatively and quantitatively measure the enzyme activity on other substrates, because radioactive PHA precursors are not always commercially available. Thus, we compared the activity of PhaZ on nonradioactive P(HO-co-HX)w prepared as PHA* latex but using nonradioactive octanoic acid (see "Experimental Procedures"). Considering an apparent extinction coefficient for P(HO-co-HX)w of 1.44 l g Ϫ1 cm Ϫ1 , we calculated the specific activity of PhaZ as 12,860 units PHA /mg of protein by using the turbidimetric assay. This value is in perfect agreement to the specific activity determined by the radioactive assay (14,100 units*/mg of protein) confirming the reliability of both procedures to detect PhaZ mediated hydrolysis. Nevertheless, it is important to notice that the radioactive method is 10-fold more sensitive than the turbidimetric assay in terms of detection limits for PhaZ activity (see below).
The PhaZ activity was also determined turbidimetrically by using artificial granules of P(HO-co-HX)s prepared in the presence of sodium cholate as surfactant and compared with that of the natural substrate, the native granules of P(HO-co-HX)n (i.e. granules containing associated proteins (GAPs)) ( Table 3). Interestingly, PhaZ hydrolyzes both heteropolymers with similar activity independently of the presence of GAPs.  Because the activity of PhaZ depolymerase was not affected by the addition of sodium cholate, this detergent was utilized to prepare artificial granules of different PHAs to be used as substrates in the turbidimetric assay. Table 3 shows that turbidimetric clearance was detectable when other mcl-PHAs like the heteropolymer P(HN-co-HP) containing 55% of 3-hydroxynonanoic acid and 45% of 3-hydroxyheptanoic acid, and mcl-PHAs containing aromatic monomer such as phenyl-hydroxyalkanoates of 6 (PHPhH), 7 (PHPhh), and 8 (PHPhO) carbon atoms in the aliphatic chain were used as substrates (Table 3). Remarkably, PHB was hardly solubilized demonstrating that PhaZ is specific for mcl-PHAs. Other typical substrates for lipases or esterases such as triacylglycerols or p-nitrophenylalkanoate were not significantly hydrolyzed by PhaZ. All these results taken together suggest that PhaZ is a PHA depolymerase that specifically hydrolyzes mcl-PHAs containing aliphatic and aromatic monomers.
Products of the Enzymatic PHA Hydrolysis-The time course of depolymerase reaction was followed by LC-MS using a C18 column, which allowed us to quantify the amount of released monomer and dimer esters of HO (see "Experimental Procedures"). Fig. 6A shows the LC-MS analysis of commercial HO used as control. This analysis revealed the existence of two chromatographic peaks, a large peak with a retention time of ϳ10 min (peak 1) and a small peak with a retention time of ϳ15 min (peak 2). The ESI(Ϫ) analysis of peak 1 (Fig. 6I) provided two main single charged negative ions that perfectly matched the molecular masses of the deprotonated HO monomer (m/z 159) and of the dimer adduct of HO monomer (m/z 319), whereas the analysis of peak 2 (Fig. 6J) showed two main single charged negative ions corresponding to the molecular masses of the deprotonated HO diester (m/z 301) and of the dimer adduct of HO diester (m/z 603). Interestingly, this result indicates that commercial preparations of HO are slight contaminated with HO dimer providing us an unexpected additional control for our analysis. Fig. 6 (B-F) shows the depolymerase degradation products (monomers and dimers) of P(HO-co-HX) granules analyzed by LC-MS at different times (0 -6 h) of enzymatic hydrolysis. The chromatogram depicted in Fig. 6D (12 min of hydrolysis) clearly illustrates that at short reaction times the amount of dimers released by the depolymerase was higher than that of monomers. It is worth noting that the peak at 15 min shows the expected ESI(Ϫ) mass spectrum of the HO diester (Fig. 6H), whereas the minor peak at 14 min (peak 3) corresponds to the Hx-HO dimer (data not shown). However, when the reaction was completed after 6 h of hydrolysis (Fig.  6F), the dimer disappeared being only able to detect the monomer, which presents the typical ESI(Ϫ) mass spectrum of HO (Fig. 6G). Remarkably, these results indicate that PhaZ depolymerase behaves as endo-and exo-hydrolases being able to hydrolyze both large and small polyester molecules to accomplish the total degradation of mcl-PHA.
Subcellular Localization of the PhaZ in P. putida KT2442-To analyze the cellular localization of PhaZ in P. putida KT2442, we studied by Western blot analyses the presence of the enzyme in the soluble and granule fractions of this strain cultured under a The assays were carried out in the presence of 300 mM NaCl. b Enzyme was preincubated 10 min at room temperature in the presence of the inhibitor. After 10 min the depolymerase and the inhibitor were added to the assay system; the final concentration of the inhibitor and protein was 10 times lower than that of the preincubation solution. c Dissolved in isopropanol; value has been corrected for inhibition by the solvent.

TABLE 3 Substrate specificity of purified PhaZ depolymerase
Values are the average of three independent experiments. Monomer percentages in the composition of the polymers are indicated in parentheses.

Substrate
Relative activity Triacylglycerols a 1,2,3-Tributyryl-sn-glycerol ϩ b 1,2,3-Trioctanoyl-sn-glycerol ND c 1,2,3-Tridecanoyl-sn-glycerol ND 1,2,3-Tridodecanoyl-sn-glycerol ND 1,2,3-Tri-(9Z-octadecenoyl)-sn-glycerol ND a Triacylglycerol lipolytic activity was assayed in rhodamine plates and by clearing halo test. b A very low but detectable activity was observed when the assay was developed overnight. c ND, not detected. d Esterase activity against p-nitrophenylalkanoates was assayed by colorimetry. e One unit (unit A650 ) of turbidimetric specific activity is defined as the decrease of one A 650 unit per min. f Artificial granules were prepared in the absence (dH 2 O) or in the presence of sodium cholate. mcl-PHA production conditions (Fig. 7). These analyses clearly demonstrated that PhaZ forms part of the GAPs. Moreover, we observed that PhaZ is produced in parallel with polyester syn-thesis. Taking into account that the mobilization of mcl-PHA in Pseudomonads is promoted under non-nitrogen limitation (26,50) we tested the influence of PhaZ production in this event. Thus, the de novo synthesis of PhaZ was checked in these cells by transferring them to a non-nitrogen-limited defined medium (Fig. 7, lanes 3 and 4). Interestingly, PhaZ protein content of the granules obtained under non-nitrogen limitation was 4-fold higher than that of the granules isolated from mcl-PHA-producing cells. Furthermore, it should be noticed that also in this case the PhaZ enzyme was exclusively detected in the granule fractions (Fig. 7, lanes 1 and 3). These results agree with those described for a depolymerase activity of P. oleovorans, which is also associated to mcl-PHA granules (51). Surprisingly, some hybridizing bands reactive to the anti-PhaZ antibodies were detected in the supernatant fractions (Fig. 7, lanes 2 and 4), being more intense in the fraction isolated from non-nitrogen-limited cells (Fig. 7, lane 4). Because these bands do not correlate with the molecular mass of PhaZ or putative degradation products we could not ascribe them to known proteins so far.
Modeling of PhaZ Depolymerase Three-dimensional Structure-Taking into account that the primary structure of PhaZ displays a significant similarity with proteins whose three-dimensional structure has been solved, we built a structural model of PhaZ to gain insight into the molecular basis of its enzymatic properties. The highest similarity was found with the meta-cleavage product hydrolase from P. fluorescens IP01 (CumD protein) (46), the C-C bond hydrolase MhpC from E. coli (52), and the aryl esterase PFE from P. fluorescens (53). All of these display similar architectures, with a conserved core domain of the ␣/␤hydrolase type and a helical "lid" domain packed onto the surface of the former. To construct the threedimensional model of PhaZ we chose the CumD enzyme as template (20% amino acid identity and 36% similarity), because its structure complexed with a cleavage product was available, and it might help in modeling the structure of PhaZ complexed with its own substrate. Fig. 8 (A and B) shows the twodomain structure assumed for PhaZ. The core domain is formed by a central 8-stranded ␤-sheet with ␣-helices packed on both sides. Residues 133-192 would be structured as five ␣-helices connected by loops and constituting the closing lid domain. Many hydrolases and  lipases with the ␣/␤-hydrolase fold have been described as making use of a "catalytic triad" of Ser-His-Asp residues to accomplish catalysis. According to this mechanism, the hydroxyl group of the Ser residue is deprotonated and carries out a nucleophilic attack to the carbon atom in the ester bond, while the His and Asp residues stabilize the deprotonated state of the Ser side chain. As depicted in Fig. 8B, Ser 102 , Asp 221 , and His 248 are located in close proximity, in one side of the core domain, facing the lid and near a presumed hinge between domains. Therefore, the model supports the possibility that these residues might conform to the active site of PhaZ.
For docking the substrate on the PhaZ model, a dimer of HO ester was generated in silico and docked in the proximity of the catalytic triad. The presence of a shallow surface across the core domain followed by an elongated deep cavity facilitates the accommodation of mcl-hydrocarbon chains of the substrate (Fig. 8, B and C). Special care was taken to avoid clashes with the lid domain, although it is possible that such lid may move away by a lipase-like interfacial activation mechanism. Fig. 8C depicts a more detailed view of the active site with the substrate bound. In this figure, atoms are represented as van der Waals spheres, and, for the sake of clarity, the lid has been removed. The binding surface of the core domain is highly hydrophobic, and the only polar residues surrounding the substrate are the catalytic triad (Fig. 8D) and Asn 35 , which may take part in substrate recognition as described for the analogous Ser 34 in CumD (46). Finally, it should be pointed out that the side chain of the only cysteine residue in PhaZ (Cys 119 ) is buried and facing off solvent, so that the possibility of disulfide bridge formation is reduced, a result that is in agreement with the lack of effect of reducing reagents on the enzymatic activity of PhaZ (Table 2).

DISCUSSION
The development of sustainable biotechnological processes that can replace conventional chemical processes to generate added value products is one of the main current tendencies in white biotechnology and biocatalysis (54). In this sense, the interest in the study of the PHAs hydrolysis lies not only in their potential use as bioplastics or biomaterials but also in the production of intermediates (chiral (R) hydroxyalkanoic acids (RHAs)), based on the enzymatic hydrolysis of the bacterial biopolymers. RHAs are scaffolds that are used as chiral starting materials in fine chemical, pharmaceutical, and medical industries (8,50). The metabolism of mcl-PHA in different P. putida strains allows the production of a high number of very diverse monomeric RHAs that can be incorporated into the polymer (5). This diversity depends on the side chain that can be saturated, unsaturated, branched methyl and benzyl esters, phenyl, phenoxy, cyanophenoxy, nitrophenoxy, epoxy, cyano, halogenated, among others. Therefore, the transformation of raw material in PHAs or in PHA derivatives with added value can be considered an important input in terms of eco-effective applicability and environmental biotechnology.
In contrast to the efforts made for understanding the enzymatic hydrolysis of scl-PHAs (6), the biodegradation of mcl-PHAs has been rarely studied (15,16,28,55,56). The role of phaZ gene of the pha cluster of Pseudomonads was deduced only from its behavior on in vivo experiments; however, the activity of the PhaZ protein has not been demonstrated in vitro due to the technical difficulties associated with its isolation and biochemical characterization, because it is produced in low amounts, and is an in vivo immobilized enzyme working on an insoluble hydrophobic substrate. Therefore, to experimentally approach these problems we needed to develop highly sensitive methods to determine its activity in vitro, which could allow us to attempt its purification. In this sense, a very simple and sensitive method has been applied to measure the scl-PHA depolymerase activity in vitro based in the quantification of the monomer released on an enzymatic coupled assay with the NAD ϩ -dependent 3-hydroxybutyrate dehydrogenase (19,57). Other less sensitive methods have been also applied to quantify PHA hydrolysis like titration of the released acids, gravimetry, turbidimetry, and others. The intracellular mcl-PHA depolymerase activity was measured for the first time as an autohydrolysis of mcl-PHA granules from a strain of P. oleovorans by using a method consisting in a combination of gravimetry and gas chromatographic analysis of the solubilized products (28). The inhibition caused by the detergents and the effect of other inhibitors suggested that a depolymerase was associated with the polymer boundary of the granule with a serine residue forming part of the active center (15). This procedure also made it possible to determine the mcl-PHA depolymerase activity in a protein mixture obtained from the solubilization of GAPs. Because the activity was ascribed to a fraction containing ϳ32-kDa protein species, which correlated with the calculated molecular mass for the phaZ gene product, it was suggested that the autohydrolysis of mcl-PHA granules was carried out by the product of the phaZ gene and that this enzyme was indeed associated to the granule (15,51,58). Obviously, all these hypotheses based on indirect proofs required a further direct confirmation. Our first attempts to detect the enzyme in E. coli extracts by using conventional methods were unfruitful, and it was not until we developed the radioactive PHA depolymerase assay that we succeed in our approach to purify an active PhaZ mcl-PHA depolymerase. This new procedure allowed us to detect very low levels of PhaZ activity, i.e. as low as 80 -90 units*/mg of protein in crude extracts, being 10-fold more sensitive than turbidimetric or gravimetric/gas chromatographic methods. Another fundamental advantage of this radioactive assay is the possibility of developing long term enzymatic assays, because turbidimetric changes produced by bacterial contamination, substrate aggregation, or other phenomena, which circumstantially could affect turbidimetric measurements, do not affect the radioactive counting of the samples. Moreover, this method facilitates the determination of depolymerase activity on PHA films, and preliminary data suggest that PhaZ is active on these substrates (data not shown).
The comparison between the reported GAP-depolymerase from P. oleovorans and PhaZ of P. putida KT2442 revealed some physiological and biochemical similarities but also clear differences, e.g. the activity of PhaZ described in this work was not affected by the presence of MgCl 2 and reducing agents, in contrast with the inhibitory effect of mercaptoethanol or dithiothreitol, and the enhancing effect of MgCl 2 on the P. oleovorans enzyme (15). In fact, due to the complete loss of activity of the GAP-depolymerase in the presence of reducing agents, the involvement of disulfide linkages affecting the correct folding of the enzyme has been proposed (15). It is worth noting that we did not detect native granule autohydrolysis during our assays with PhaZ, even when the incubation was continued for 60 min (data not shown). There is no evidence for the inducibility of the depolymerase of P. oleovorans, because the PHA granules were always isolated from non-nitrogen-limited growing cells (15), but we cannot discard that the native granules used in our assay contained very low amounts of depolymerase in contrast to that of P. oleovorans. At present, the mechanism and regulation of the intracellular degradation of mcl-PHA are poorly understood (25), but we have determined that the synthesis of PhaZ in P. putida KT2442 is promoted when the cells are growing under non-nitrogen-limited conditions. This is in agreement with previous findings reported for P. putida CA-3, where it has been suggested as a regulatory system for the transcription of phaZ, with carbon limitation being the factor that directs its synthesis (59).
The product intermediates generated by PhaZ activity were determined using P(HO-co-HX) as substrate. In contrast to P(HO) depolymerase GK13 (14,55), which mainly produces dimer esters of HO, the monomer of HO was identified as the main product of PhaZ hydrolysis. In this sense, PhaZ behaves as the extracellular mcl-PHA depolymerase of P. alcaligenes LB19 (56). These findings allowed us to classify PhaZ as an endo/exomcl-PHA depolymerase.
The ability of different strains of P. putida to hydrolyze aliphatic and aromatic mcl-PHAs has been described (16,30,60). Furthermore, it has been shown that overexpression of phaZ in P. putida U avoids the accumulation of PHAs as storage granules (30). An in vivo method for the production of chiral hydroxyalkanoic acid monomers from mcl-PHA has been recently reported by using P. putida strains (50). In addition, a genetically engineered strain of P. putida U (⌬fadBA-phaZ) (CECT 7008) was designed to efficiently transform different n-phenylalkanoic acids into their 3-hydroxyderivatives that are excreted to the culture broth (30). Overexpression of phaZ in this strain avoids the accumulation of PHA or PHPhA as storage granules. Therefore, the ability of P. putida strains to degrade aliphatic and aromatic mcl-PHAs can be now ascribed to PhaZ.
Taking advantage of the PhaZ similarity with other crystallized hydrolases, we have modeled the three-dimensional structure of PhaZ complexed with a dimer of HO (Fig. 8). Using this model, we found that the enzyme appears to be built up from a core, ␣/␤ hydrolase-type domain capped with a lid structure, with an active site containing a catalytic triad (Ser 102 -Asp 221 -His 248 ) buried near the connection between domains. The presence of such a lid suggests a lipase-like mechanism of interfacial activation (61), according to which, the contact with the hydrophobic granule would induce the movement of the lid, exposing the active site to the polymer. The binding site is wide enough to accommodate several mcl-PHA substrates, in concordance with the substrate specificity range of the enzyme ( Table 3). The prototype of extracellular mcl-PHA depolymerase is that of P. fluorescens GK13 (14,55). This enzyme, referred to as P(HO) depolymerase GK13 , is also a member of the serine hydrolase family of enzymes, and like PhaZ, it hydrolyzes mcl-PHA but not scl-PHA or other characteristic substrates for lipases, such as triolein (14). However, in contrast to PhaZ, it is active for soluble esters such as p-nitrophenylacyl esters with six or more carbon atoms in the fatty acid moiety. In this sense, the P(HO) depolymerase GK13 shows a higher sequence similarity with a different family of hydrolases such as the dienelactone hydrolase from P. putida (63) (data not shown). In these enzymes, the lid domain is small or does not exist at all, and the active site is fully exposed to the solvent, i.e. the enzyme cannot undergo any interfacial activation and is often active toward water-soluble esters (e.g. lipase from Bacillus subtilis) (64). Probably, the main difference between PhaZ and P(HO) depolymerase GK13 is the presence of the lid domain in the former. This could explain why the substrate specificity range of the PhaZ depolymerase of P. putida KT2442 is narrower than that of the P(HO) depolymerase GK13 , because the lid might act as a substrate discriminator. Mutagenesis experiments validating this hypothesis are in progress.
Because the amino acid sequence of the first intracellular scl-PHA depolymerase of the paradigmatic PHB producer R. eutropha H16 was reported (18), several isoenzymes have been found in this microorganism, suggesting that a more complex system could be responsible of PHB mobilization in bacteria (17)(18)(19)(20)(21)(22)(23). In this sense, at least 14 similar proteins or PhaZ paralogs encoded in the genome of P. putida KT2440 can be identified by sequence comparison (data not shown), but their corresponding functions have not been determined yet. Interestingly, we have observed the presence of unidentified proteins reactive to anti-PhaZ antibodies in the supernatant fraction of crude extracts from P. putida KT2442 (Fig. 7) that might correspond to PhaZ paralogs, but whether they could be involved in mcl-PHA hydrolysis in this microorganism, taking part of a more complex mcl-PHA hydrolytic system, is under investigation.
These results provide the first formal biochemical evidence that the phaZ gene certainly encodes an intracellular endo/exomcl-PHA depolymerase that could be considered as the paradigmatic representative of this enzyme family. The possibility of assaying and manipulating in vitro this enzyme opens new avenues for develop of new biotechnological processes to industrially exploit the chirality of PHAs as a source of lead compounds.