Plant original Massilia isolates producing polyhydroxybutyrate, including one exhibiting high yields from glycerol


Gabriella Molinari, Environmental Microbiology Laboratory, Helmholtz Centre for Infection Research (HZI), 38124 Braunschweig, Germany. E-mail:


Aims:  The purpose of this study was to isolate new and potentially better polyhydroxyalkanoate (PHA)-producing bacteria, with a view to obtaining high yields from inexpensive substrates like glycerol, a major by-product of the biodiesel process.

Methods and Results:  Eleven new plant original isolates of the genus Massilia, a poorly studied lineage within the Betaproteobacteria, were isolated and characterized. Two isolates, 2C4 and 4D3c, could not be assigned to a validated Massilia species and probably represent new species. Six isolates were found to produce poly-3-hydroxybutyrate (P3HB) when cultured with glucose or glycerol as carbon source. Isolate 4D6 accumulated up to 50 wt% of cell mass as polyhydroxybutyrate (PHB) when grown on glycerol.

Conclusions:  The phyllosphere may be a good source of bacteria unrelated or weakly related to human/animal pathogens for screening for new PHA producers for industrial application. Isolate 4D6 was capable of accumulating particularly high levels of PHB from glycerol.

Significance and Impact of the Study:  With the increase in biodiesel production, which generates increasing amounts of glycerol as a by-product, there is a major interest in exploiting this compound as feedstock for the synthesis of interesting products, like biopolymers, such as PHA. The new Massilia sp. 4D6 isolate described in this study may be a useful candidate as a cell factory for the industrial production of PHA from glycerol.


Polyhydroxyalkanoates are optically active polyesters produced as intracellular granules by a large number of organisms when experiencing an excess of carbon and/or depletion of some other essential nutrients, such as nitrogen, phosphorus or oxygen (Anderson and Dawes 1990). Polyhydroxyalkanoate (PHA) composition and physicochemical properties are influenced by diverse factors, such as the nature of substrates provided (alkanes, alkenes, alcohols, fatty acids, carbohydrates, etc.) and the specific active metabolic pathways present. Thus far, more than 150 different monomers have been described (Steinbüchel and Valentin 1995) and an array of taxonomically different micro-organisms have been shown to produce PHA.

Different PHAs possess distinct thermoplastic and elastomeric properties that endow them with potential for diverse industrial, medical and consumer applications. PHAs are potential alternatives to conventional petrochemical plastics because they and their chemically modified variants exhibit similar physical properties. Moreover, because of their enantiomeric purity (they are mainly composed of [R]-3-hydroxycarboxylic acid), PHA monomers obtained either by natural PHA depolymerization or by chemical hydrolysis of the polymer are considered to be useful as chiral synthons (Zinn et al. 2001). PHA monomers have also been described to have potential antimicrobial and/or antiviral activities (Ruth et al. 2006; Chen 2009).

A key barrier to the large-scale exploitation of PHAs is their production cost. Therefore, much effort is currently concentrated in developing less expensive processes. This includes the search for new and better PHA producers, particularly environmental isolates not affiliated with known pathogens, with higher production yields, and/or having more efficient or more specific PHA synthases that might allow the use of cheaper or functionalized substrates that extend or enhance the characteristics of the polymers. For example, new bacteria producing PHA from inexpensive and abundant substrates like glycerol, a major by-product of the biodiesel production process (∼10% of the final weight of biodiesel), are being actively sought. Here, we describe new plant original isolates of the genus Massilia, a poorly studied lineage within the Betaproteobacteria, and show that they produce PHA. One isolate was found to accumulate polyhydroxybutyrate (PHB) up to 50 wt% of cell mass when grown on glucose or glycerol.

Materials and Methods

Isolation of plant original bacteria

Plants were sampled in May 2008 on the campus of the University of Balearic Islands (Mallorca Island, Spain, 39°38′11″N; 2°38′50″E). The samples (leaves, branches and flowers) were immediately placed on R2A agar plates (Reasoner and Geldreich 1985), which were incubated at room temperature for up to 2 weeks. Purified bacterial clones initially developing on the plates were identified by 16S rDNA sequence analysis. All isolates were routinely grown in R2A at room temperature, except strain 5F6, which was cultured in double-strength R2A (R2A2) or OM medium.

OM medium contained 1·0 g starch, 1·0 g glucose, 1·0 g peptone, 1·5 g yeast extract, 10 ml of solution A, 10 ml of solution B, 1 ml of solution C and 1 ml solution D per litre. Solution A contained (in g l−1) KH2PO4 (5·0) and K2HPO4 (5·0). Solution B contained (in g l−1) MgSO4 (17·0), NaCl (1·0), MnSO4 (0·7) and CuSO4 (0·06). Solution C contained (in g/100 ml): FeSO4·7H2O (0·1), sodium citrate dihydrate (2·2), ammonium acetate (2·0), sodium thioglycolate (0·75) and sodium succinate hexahydrate (3·3); Solution D contained (in mg/100 ml): biotin (10·0), nicotinic acid amide (35·0), thiamine hydrochloride (30·0), p-aminobenzoic acid (20·0), pyridoxal hydrochloride (10·0), calcium pantothenate (10·0) and vitamin B12 (5·0). All reagents were purchased by Difco (Heidelberg, Germany) and Sigma-Aldrich (Seelze, Germany).

Strain Massilia plicata 76T (DSM17505) obtained from the German Type Culture Collection (DSMZ) was grown in R2A and OM media.

Screening of isolates for accumulation of hydrophobic substances

Isolates were streaked on Nile Red plates, prepared by adding one ml of Nile Red solution (Sigma-Aldrich, St Louis, MD, USA; 0·25 mg dissolved in 1 ml dimethylsulfoxide, DMSO) to 500 ml of R2A or OM agar, as a prescreen for those producing lipophilic intracellular inclusions. As Nile red has an excitation wavelength of 542 nm and an emission wavelength of 598 nm, the plates were viewed under ultraviolet light using a Bio-Rad ChemiDoc XRS system (Hercules, CA, USA), after different incubation periods (Spiekermann et al. 1999), and scored for fluorescence.

Positive colonies were then further examined by fluorescence microcopy as follows. One ml of liquid culture was mixed with two drops of Nile red solution (see above) in an Eppendorf tube, centrifuged at 16 000 g at 4°C for 5 min (Centrifuge 5417R; Eppendorf, Hamburg, Germany), and the supernatant fluid was discarded. Pellets were washed twice with 2 ml MgCl2 (10 mmol l−1), resuspended in 500 μl of MgCl2 (10 mmol l−1) and examined by fluorescence microscopy for the presence of PHA granules (Bassas 2010), using a Zeiss Axio Imager A1 epifluorescence microscope equipped with the AxioVision rel 4·6·3 software (Zeiss Imaging solutions GmbH, Oberkochen, Germany).

The Nile red plate screen gave a high proportion of positive colonies, but fluorescence microscopy of Nile red-stained samples of liquid cultures proved to be a more reliable method for the identification of isolates accumulating lipidic inclusions.

Identification of isolates by 16S rDNA sequence analysis

Bacterial DNA was extracted from 2 ml of overnight cultures, using the Promega Wizard SV 96 Genomic DNA purification kit, according to the manufacturer’s protocol. Nearly, complete 16S rRNA gene was PCR amplified using universal 16S rDNA primers f27 and r1492 (Lane 1991). The amplicons were sequenced, using the BigDye terminator cycle sequencing kit v3.1 (Applied Biosystems, Foster City, CA, USA), according to the manufacturer’s instructions and a 3130 Genetic Analyzer (Applied Biosystems, Darmstadt, Germany). 16S rDNA sequences were compared with reference sequences in the EBI Nucleotide Sequence Database (Kulikova et al. 2006), using the FASTA program (Pearson and Lipman 1988), and subsequently aligned with reference sequences from the SILVA SSU102 release, using the SINA web aligner (Pruesse et al. 2007). All of the described type species of genus Massilia and related species were included. Aligned sequences were imported into the ARB package (Ludwig et al. 2004) for the calculation of sequence similarities and evolutionary distances and for the construction of phylogenetic trees (distance, parsimony and maximum likelihood methods). The tree shown in the manuscript was calculated using the neighbour-joining algorithm (Saitou and Nei 1987), using a distance matrix calculated with the Jukes and Cantor correction (Jukes and Cantor 1969). Bootstrap values over 50% (1000 replicates) for those branches including the novel plant original isolates are shown.

Cultivation conditions for PHA production

Isolates were grown on OM and R2A medium supplemented with glucose (5 or 10 g l−1) at 30°C, obtaining the best results in OM media (data not shown). 500-ml Erlenmeyer flasks containing 200 ml OM medium were inoculated with 0·5 ml of an overnight preculture grown in the same medium, to an approximate density of 8 × 106 CFU ml−1, and incubated at 30°C in rotary shaker at 120 rpm for 48–72 h. PHA production was also assessed in cells grown on other carbon sources: arabinose, fructose, galactose, glycerol, mannose, rhamnose, pyruvic acid, sucrose and xylose (5 g l−1 each), linoleic acid (0·5%) and octanoic acid (15 mmol l−1). For these cultures, 100-ml flasks with 40 ml of OM medium and incubation at 30°C were used.

Cell growth and biomass quantitation

Growth over time was monitored by taking culture samples and measuring their OD550 and by determination of their dry weight. In the latter case, the biomass in 10-ml aliquots was harvested by centrifugation for 15 min at 6500 g at 4°C (Allegra 25R; Beckman Coulter, Fullerton, CA, USA), washed twice with distilled water and dried at 80°C for 24 h in a hot air oven to a constant weight. Biomass production was defined as cellular dry weight (CDW) per litre of culture broth.

PHA content

PHA content of cultures was determined gravimetrically. 10 ml culture samples were centrifuged at 6500 g for 15 min at 4°C, frozen at −20°C, lyophilized in a lyophilizer Alpha 1-4 LSC (Christ, Osterode, Germany) at −59°C and 0·140 mbar and subsequently extracted with 10 ml of chloroform for 3 h at 80°C. The chloroform solution was then filtered to remove any cell debris and concentrated by rotary evaporation. The PHA content (wt%) was defined as the percentage of the CDW represented by the poly-(3-hydroxybutyrate).

PHA purification

Total lyophilized biomass was extracted with chloroform at a ratio of 1 : 15 (w/v) for 4 h at 80°C using a Soxhlet system. The chloroform solution was filtered to remove any cell debris and concentrated by rotary evaporation. PHA was purified by precipitation of the chloroform solution [1% (w/v)] through dropwise addition to a cold methanol. The methanol–chloroform mixture was decanted and the pure polymer was washed with fresh iced methanol. The purified polymer was stored under nitrogen at −20°C.

Transmission electron microscopy

Bacteria were fixed by chilling the cultures to 4°C and addition of glutaraldehyde (2%) and formaldehyde (5%). They were then washed with cacodylate buffer (0·1 mol l−1 cacodylate, 0·01 mol l−1 CaCl2, 0·01 mol l−1 MgCl2·6H2O, 0·09 mol l−1 sucrose, pH 6·9) and stained with 1% aqueous osmium for 1 h at room temperature. Samples were then dehydrated with a graded series of acetone (10, 30, 50, 70, 90 and 100%) with incubation for 30 min at each concentration, except for the 70% acetone, which contained 2% uranyl acetate and was performed overnight. Samples were infiltrated with an epoxy resin, according to the Spurr formula for hard resin (Spurr 1969), for several days with pure resin. Ultrathin sections were cut with a diamond knife, counterstained with uranyl acetate and lead citrate and examined in a TEM910 transmission electron microscope (Carl Zeiss, Oberkochen, Germany) at an acceleration voltage of 80 kV. Images were taken at calibrated magnifications using a line replica. Images were recorded digitally with a Slow-Scan CCD-Camera (ProScan, 1024 × 1024, Scheuring, Germany) with ITEM-Software (Olympus Soft Imaging Solutions, Münster, Germany).

Field emission scanning electron microscopy

Samples were fixed as above, washed with cacodylate buffer and then washed with TE-buffer (20 mmol l−1 TRIS, 1 mmol l−1 EDTA, pH 6.9). 50 μl of washed bacteria were applied to poly-l-lysine precoated cover slips (12 mm in diameter), which were left for 5 min, washed in TE-buffer, incubated with 2% glutaraldehyde in TE-buffer for 15 min and washed again with TE-buffer. Dehydration was carried out with a graded series of acetone (10, 30, 50, 70, 90, 100%) on ice for 15 min for each step, followed by 100% acetone at room temperature and critical-point drying with liquid CO2 (CPD 30; Bal-Tec, Balzers, Liechtenstein). Samples were then gold shadowed by sputter coating (SCD 500; Bal-Tec) and examined with a field emission scanning electron microscope Zeiss DSM 982 Gemini (Carl Zeiss, Oberkochen, Germany), using the Everhart Thornley SE detector and the inlens detector in a 50 : 50 ratio at an acceleration voltage of 5 kV. Images were recorded onto a MO-disc. Contrast and brightness were adjusted with Adobe Photoshop CS3.

NMR analysis

For 1H-NMR analysis, 5–10 mg of polymer was dissolved in 0·7 ml of deuterated chloroform (CDCl3, whereas 15–20 mg of polymer was used for recording the 13C spectra. 1H and 13C NMR spectra were recorded at 300 K on a Bruker DPX-300 NMR Spectrometer (Brucker, Billerica, MA, USA) locked to the deuterium resonance of the solvent (CDCl3). Chemical shifts are given in ppm relative to the signal of the solvent (1H: 7·26, 13C: 77·3) and coupling constants in Hz. Standard Bruker pulse programs were used throughout.

Gel permeation chromatography

Average molecular weights were determined by Gel permeation chromatography (GPC) (Waters, Milford, MA, USA) consisting in a HPLC system with a combination of three styragel columns series (Styragel HR3, HR4 and HR5) and equipped with a 2410 differential refractive index detector and 2487 UV/Vis detector (Waters). Chloroform was used as eluent at 35°C and a flow rate of 1·0 ml min−1 (isocratic). Sample concentration and injection volume were 10–20 mg ml−1 and 20 μl, respectively. The calibration curve was obtained using polystyrene standards (Polymer Laboratories, Church Stretton, UK) in the Mw range of 580–1930 000 g mol−l.

Thermal properties

Thermal properties of the microbial polyesters were determined by differential scanning calorimetry (DSC) on 10–20 mg of the purified polymer in each analysis. DSC analyses were performed with a DSC-30 Mettler Toledo Instruments, New York, NY, USA) and the samples were placed on an aluminium pan and heated from −100°C to 400°C at 10°C/min under nitrogen (80 ml min−1).

Nucleotide sequence accession numbers

The accession numbers of the 16S rDNA nucleotide sequences of the isolates are FR865952FR865962.


Phylogenetic assignment of plant original, PHA-producing Massilia isolates

After phylogenetic assignment of PHA-producing isolates by 16S rRNA gene sequence analysis, eleven strains from nine different samples (leaves, branches or flowers; Table S1) that grouped within the genus Massilia, of the Betaproteobacteria, were selected for further characterization.

Different treeing methods were employed to determine the phylogenetic relationships of isolates and, in all cases, the affiliations obtained were the same (Fig. 1). Seven affiliated with good bootstrap support with the type strain of M. aurea (99·6–99·9% 16S rDNA sequence similarity), one isolate (4D6) with M. plicata (99·9% sequence similarity) and another (5F6) with M. niabensis (99·3% sequence similarity). Two isolates (2C4 and 4D3c) affiliated weakly with M. niastensis and M. aerilata, showing slightly higher sequence similarities with the former (97·5–97·9%).

Figure 1.

 Phylogenetic tree based on 16S rRNA gene sequences showing the relationships of the plant original strains isolated in this study, the Massilia species and other related bacteria within the family Oxalobacteraceae of the class Betaproteobacteria (GenBank/EMBL accession numbers are given in brackets). The pylogenetic tree was constructed using the ARB software package (Ludwig et al. 2004), after alignment of data with the ARB alignment tool and the SILVA SSU102 database (Pruesse et al. 2007). The tree was calculated using the neighbour-joining algorithm (Saitou and Nei 1987) using a distance matrix calculated by the Jukes–Cantor method (Jukes and Cantor 1969). The bar indicates sequence divergence. Bootstrap values over 50% (1000 replicates) are indicated at the nodes.

Growth and Morphology

The shape and colour of colonies was examined on R2A plates. Colonies were pale-translucent yellow, in agreement with the description of the validated Massilia species (Gallego et al. 2006; Zul et al. 2008). However, when strains were plated on OM, most of the colonies showed stronger coloration (yellow-orange). Strain 4D6 (99·9% homology with the M. plicata strain; Zhang et al. 2006) also formed translucent to pale white-yellow colonies on R2A and OM agar plates, though older colonies tended to develop a characteristic wrinkled morphotype (Fig S1).

Of the 11 Massilia strains, six were selected for further study on the basis of the phylogenetic distance and the ability to accumulate lipidic inclusions. Fluorescence microscopy of Nile red-stained cells from cultures in modified OM media with 10 g l−1 glucose during 48 h revealed different morphologies and degrees of intracellular granules (Fig. 2). Strains 5F6 and 4A1 grew as short rods with some Nile red positive granules within the cells (Fig. 2a,b), whereas isolate 5F8 grew as long rods with several bright granules (Fig. 2c). Strain 4D3c showed a diffuse staining which delineated the long and large rods with inclusions (Fig. 2d). Massilia 2C4 cells (Fig. 2e) elongated soon after 24 h of incubation, acquiring the shape of long filaments with large round granules. Filamentation was also observed with the M. brevitalea byr-23-80T type strain when grown in rich medium, but not in 1 : 10 diluted media (Zul et al. 2008). The cell shape of strain 4D6 varied from very short rods to round cells, with numerous brightly fluorescent granules which almost filled the entire cytoplasm (Fig. 2f).

Figure 2.

 Fluorescence microscopy observations of Massilia isolates stained with Nile red after grown in OM medium with 10 g l−1 glucose during 48 h. Nile red stains intracellular lipid inclusions. A representative image from the different strains is shown in the panels as follow: (a) 5F6, (b) 4A1, (c) 5F8, (d), 4D3c, (e) 2C4, (f) 4D6. The inset shows larger magnification of the area indicated by the square in (f). Scale bars correspond to 2 μm.

The structural changes occurring in strains 2C4 and 4D6 were further characterized by scanning (FESEM) and transmission (TEM) electron microscopy. Cells of strain 2C4 were short rods only during the initial period of incubation, whilst after 24 and 48 h only long filaments with few granules, which were visible as refractive inclusion bodies, were observed (Fig. 3a,b,c). Strain 4D6 did not form filaments, either at 48 h (Fig. 3d) or after long incubation periods (5 days) but increased in cell size over time whilst accumulating granules distributed throughout the cytoplasm area (Fig. 3e,f).

Figure 3.

 Field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM) of isolates 2C4 and 4D6. Isolates were cultured in OM with 10 g l−1 glucose. (a and b) FESEM of strain 2C4 after 18 and 48 h culture, respectively; (c) TEM showing part of a long filament from strain 2C4 with PHA granules after 48 h of culture; (d) FESEM of strain 4D6 after 48 h of culture; (e and f) TEM showing 4D6 bacterial cells with several PHA granules distributed in the cytoplasm after 48 h culture. PHB granules are indicated with white arrows.

Poly-(3-hydroxybutyrate) accumulation

To characterize PHA production, the new isolates were cultivated in OM supplemented with 5 or 10 g l−1 glucose, whereby the highest PHA yields were obtained with the highest glucose concentration (Table 1). Although biomass yields were similar for all isolates, ranging from 1·02 to 1·68 g l−1, considerable differences in PHA yields were observed, with Massilia strain 4D6 being the best PHA producer with a biomass yield of 1·68 ± 0·08 g/l and PHA accumulation of 45·7 wt%. As 4D6 is affiliated with M. plicata, we analysed PHA production by M. plicata 76T (DSM17505), the type strain of this species. This strain gave a CDW of 1·47 ± 0·04 g l−1 and a PHA production of 0·69 ± 0·03 g l−1 representing 44·5 wt% of PHA accumulation, after 48 h growth, so M. plicata seems to be a good PHA producer under these conditions.

Table 1.   Biomass and PHB production by Massilia isolates
Strain noCDW (g l−1)*PHB (wt%)†
  1. All isolates were cultured in OM with 10 g l−1 glucose for 48 h at 30°C and 120 rpm.

  2. *CDW, cell dry weight. Values are means of triplicates ± SD.

  3. †PHB (wt%), PHA content relative to CDW. Values are means of triplicates ± SD.

  4. Traces: PHB accumulation < 1(wt%).

2C41·12 ± 0·015·4 ± 2·4
4A11·27 ± 0·0410·2 ± 2·3
4D3c1·26 ± 0·042·9 ± 1·8
4D61·68 ± 0·0845·7 ± 2·4
5F61·20 ± 0·12Traces
5F81·02 ± 0·116·9 ± 1·6

PHA characterization

The polymers produced by the isolates were extracted, purified and characterized by NMR. The analysis revealed that all the strains produced a homopolymer of poly-3-hydroxybutyrate (P3HB) when cultivated with glucose as a carbon source (Fig. 4).

Figure 4.

 NMR spectra of the PHB polymer produced by isolate 4D6. The isolate was cultivated in OM with 10 g l−1 glucose. Identical spectra were obtained for all the PHB polymers produced by the other isolates. (a) 1H-NMR spectrum and (b) 13C-NMR spectrum.

Molecular weight distribution and thermal properties were determined by GPC and DSC, respectively (Table 2). Massilia strains 5F6 and 5F8 produced a P3HB polymer with average molecular mass (Mw) of 5·86 × 105 Da, whereas isolates 2C4 and 4D6 produced polymers with higher molecular weights: 7·62 × 105 Da and 8·25 × 105 Da, respectively. The melting temperature (Tm) values of the P3HB polymers obtained ranged from 175°C to 181°C, and the decomposition temperature (Td) values were 295–298°C. These values are in agreement with the previously described properties of P3HB (Abe et al. 1994; Budde et al. 2011).

Table 2.   Thermal properties and molecular weights of PHB produced by Massilia isolates
Strain noTm (°C)*ΔHm (J g−1)†Td (°C)‡Mw (kDa)§Mn (kDa)¶PI**
  1. All isolates were cultured in OM with 10 g l−1 glucose for 48 h.

  2. *Tm (°C), melting temperature.

  3. †ΔHm (J g−1), enthalpy of fusion.

  4. Td (°C), decomposition temperature.

  5. §Mw (kDa), weight-average molecular weight.

  6. Mn (kDa), number-average molecular weight.

  7. **PI, Polydispersity index (Mw/Mn).


Growth and PHB accumulation with different substrates

All Massilia strains grew with all the different carbohydrates offered as carbon sources, though not all substrates served for the production of PHA (Table 3). Both M. plicata 76T type strain and isolate 4D6 produced the highest levels of intracellular polymer with the different substrates, and large numbers of granules were observed when the strains were grown on fructose, galactose, glucose, mannose, ribose and glycerol. In contrast, fatty acids were not good substrates for PHA production, and only isolate 5F6 showed few inclusions after 72 h when cultivated with linoleic acid. Of the 12 substrates tested, the aldohexoses (galactose, glucose and mannose) proved to be the best substrates for PHA accumulation. M. plicata 76T and isolate 4D6 strains were the best producers of PHA, with high biomass production (optical densities of OD550 nm 12–16; Fig. S2) and accumulation of high amounts of intracellular PHB (40–50 wt%).

Table 3.   Growth and PHB accumulation of the isolates cultured on different substrates
  1. Cells were grown in modified OM medium supplemented with 5 g l−1 of the different sugars and glycerol, pyruvic acid (20 mmol l−1), octanoic acid (15 mmol l−1) and linoleic acid (0·5% v/v). One hundred-ml flasks with 50 ml culture were used. Growth was monitored by OD550 measurements and granule formation was followed by fluorescence microscopy of NR-stained cells. The table shows the results after 72 h of incubation. Growth and presence of stainable granules are indicated by + or −, before and after the forward slash, respectively. (++) indicates that the cell contained at least three granules. The results are means of duplicates.

  2. (nd) not determined.

Pyruvic acid+/−+/−+/−+/++/++/−+/+
Octanoic acid+/−−/−nd−/−−/−+/−−/−
Linoleic acid+/−+/−nd+/−+/−+/++/−

Production of PHB from glycerol by the isolate 4D6

Massilia sp. 4D6 was cultivated in OM medium with glycerol concentrations of 0·5, 1 and 2% (v/v) and found to grow and produce PHB at all concentrations. On 0·5% glycerol, the maximum yield of PHA was obtained after 72 h of cultivation, with biomass reaching 1·98 ± 0·11 g l−1 CDW and PHB accumulating to 36·9 wt%. At higher glycerol concentrations, higher PHB yields but lower biomass levels were attained: 1·66 ± 0·04 g l−1 CDW and 43·2 wt% of PHB at 1% glycerol, and 0·86 ± 0·17 g l−1 CDW and 49·6 wt% of PHB at 2% glycerol.

During cultivation on glycerol, the medium became increasingly viscous, which hindered recovery of bacteria from the cultures, and necessitating additional dilution-washing steps. PHA recovered from glycerol cultures of isolate 4D6 was shown by 1H-NMR and 13C-NMR to be a homopolymer of poly-3-hyroxybutyrate, P(3HB). Differential scanning calorimetry (DSC) gave a Tm value of 174·08°C (ΔHf of 89·03 J g−1).


A considerable number of diverse bacteria have been found to produce PHA. However, few have been exploited for industrial production of PHA polymers because of the high costs of the process. As a result of the increasing cost of petroleum-based chemicals, a major emphasis has been put on green chemistry in recent years. Different approaches, based on nonpolluting processes and products, and the move towards sustainable practices requiring the use of renewable feedstocks are being explored. Within this context, there has been an increasing interest in the exploration and exploitation of new biodiversity in the search for new processes and products, including commercially viable processes for PHA production (Ibrahim and Steinbüchel 2010a; Budde et al. 2011; Sudesh et al. 2011). With the increase in biodiesel production, with glycerol as a by-product, interesting products from glycerol as a feedstock, including PHA, are being actively sought (Papanikolaou et al. 2008; Ibrahim and Steinbüchel 2010b; Zhu et al. 2010).

The genus Massilia belongs to the family Oxalobacteraceae (Class Betaproteobacteria) and was first described in 1998 (La Scola et al. 1998). Since then, a total of 18 species have been validated, 14 of them isolated from different environmental samples (Gallego et al. 2006; Zhang et al. 2006; Weon et al. 2008, 2009, 2010; Zul et al. 2008; Kämpfer et al. 2010, 2011; Wang et al. 2011). Members of the genus are aerobic, gram-negative, motile short or straight rods. Some species were reported to metabolize glucose (Padmanabhan et al. 2003) and to have the ability to degrade phenanthrene (Bodour et al. 2003) or other aromatic compounds (Khammar et al. 2005).

In this study, we report the isolation of 11 new strains of Massilia from the phyllosphere of different plants in Mallorca. According to their 16S rRNA gene sequences, seven isolates were closely related to M. aurea (99·6–99·9%), one – 4D6 – to M. plicata (99·9%) and another – 5F6 – to M. niabensis (99·3%). On the basis of the average levels of sequence identity between species in a genus (Yarza et al. 2008), these nine isolates should be considered as new strains of these three species. Two isolates, 2C4 and 4D3c, could not be assigned to a validated Massilia species and probably represent new species, although this should be confirmed by further taxonomic analyses.

All eleven plant original Massilia isolates produced polyhydroxyalkanoates. Microscopic observation of PHA-producing cultures of all isolates revealed a diversity of cell/PHA granule morphology. Six isolates were subsequently selected for further study. PHA yields varied considerably among these strains, with isolate 4D6 giving the highest amounts (45–50 wt%) of PHB accumulation when cultured with glucose or glycerol as carbon source. As the plant original 4D6 strain showed 99·9% homology with the validated type strain of M. plicata 76T (DSM17505), we compared the two strains and found them to give similar yields of biomass and PHB. Significantly, none of the isolates clustered phylogenetically with the clade containing the human-associated species M. timonae, M. oculi and M. consociata, so the phylosphere may represent a good source of biotechnologically interesting Massilia isolates distantly related to pathogens.

During the preparation of this manuscript, a first report (Cerrone et al. 2011) appeared of PHA production by seven different strains of the genus Massilia (Massilia type species M. plicata, M. dura, M. aurea, M. albidiflava, M. brevitalea, M. lutea and M. aerilata), grown in ISP2 medium with glucose or starch as substrates. The best yields (approximately. 30 wt% of accumulation) were obtained from M. lutea cultured on starch, and some of the poorest from M. plicata on the two substrates tested. In our study, however, the same M. plicata type strain gave some of the highest yields of PHA ∼46 wt% when grown on glucose in OM medium.

The polymers obtained from the new isolates were analysed and found to exhibit thermal properties similar to those of other PHB-polymers described previously (Tm∼175–181°C) (Sudesh et al. 2000; Fiorese et al. 2009; Zhu et al. 2010), although they did exhibit differences in the molecular masses of the polymers: ∼ 6 × 105 for PHAs produced by isolates 5F6 and 5F8, and ∼ 8 × 105 for PHAs from isolates 4D6 and 2C4. These values are within the mass range of PHB polymers produced by other environmental strains, except for that of Ralstonia eutropha H16 (formerly Alcaligenes eutrophus and now renamed as Cupravidus necator), which produces a PHB with a molecular weight of 3·0 × 106 (Khanna and Srivastava 2005; Budde et al. 2010, 2011).

PHA production from glycerol was an important aspect of the strains isolated in this study. Isolate 4D6 was found to achieve PHA production yields of 42–50 wt% when grown in different glycerol concentrations, yields considerably better than those reported for Pseudomonas corrugata (19·7 wt%), P. oleovorans (26·8 wt%) and Eschericia coli (9·8–34 wt%) (De Almeida et al. 2007; Papanikolaou et al. 2008), but lower than those of B. cepacia ATCC 17759 and Zobellella denitrificans strain MW1 [up to 80 wt% (Ibrahim and Steinbüchel 2010b; Zhu et al. 2010)]. However, these types of comparisons are difficult because efforts to optimize yields can change relative performances significantly, and issues like responses to process parameters and affiliation of strains to known pathogens, can have major impacts on the utility of strains for industrial production.

In the case of 4D6, increasing glycerol concentrations in the media resulted in reduced biomass and an increase in the viscositiy of the medium. Cerrone et al. (2011) have also described that when Massilia strains are cultured with starch as carbon source the viscosity of the medium increased (supposedly due to α-amylase activity causing accumulation of glucose monomers in the media). However, to our opinion, the increase in the viscosity may be correlated with exopolysaccharide (EPS) production as it has been described previously in several PHB producers (Zhu et al. 2010). The mucous-wrinkled-colony morphology of the 4D6 strain suggests that high levels of EPS production is a characteristic of this strain and may explain the high viscosity of the medium (Goff et al. 2009; Ibrahim and Steinbüchel 2010b). Similar behaviour was described by Zhu et al. (2010) when Burkholderia cepacia ATCCC 17759 was cultivated using crude glycerol as a carbon source.

In conclusion, we have described the isolation of new plant original species of environmental Massilia. One isolate, M. plicata, gives high yields of PHB from the inexpensive feedstock glycerol. Massilia strain 4D6 may well be suitable for process optimization and exploition for PHB production from glycerol.


We thank Marita Sylla and Susanne Müller for their excellent technical assistance and Ina Schleicher for assistance with the sample preparation for TEM. G. Molinari acknowledges the financial support of the DAAD through the project 50151681 from the Acciones Integradas Program Spain-Germany. B. Nogales acknowledges the financial support of the Spanish MICINN through project CTM2008-02574/MAR (with FEDER co-funding) and the Acciones Integradas program (DE2009-0104). K.N. Timmis gratefully acknowledges generous support from the University of the Balearic Islands as a visiting professor and the Fonds der chemischen Industrie.