BIOTECHNOLOGICAL PRODUCTS AND PROCESS ENGINEERING

Medium chain length polyhydroxyalkanoate (mcl-PHA)production from volatile fatty acids derived fromthe anaerobic digestion of grass

Federico Cerrone & Santosh K. Choudhari & Reeta Davis & Denise Cysneiros &Vincent O’Flaherty & Gearoid Duane & Eoin Casey & Maciej W. Guzik &

Shane T. Kenny & Ramesh P. Babu & Kevin O’Connor

Received: 26 August 2013 /Revised: 7 October 2013 /Accepted: 8 October 2013 /Published online: 27 October 2013# Springer-Verlag Berlin Heidelberg 2013

Abstract A two step biological process for the conversion ofgrass biomass to the biodegradable polymer medium chainlength polyhydroxyalkanoate (mcl-PHA) was achievedthrough the use of anaerobic and aerobic microbial processes.Anaerobic digestion (mixed culture) of ensiled grass wasachieved with a recirculated leach bed bioreactor resulting inthe production of a leachate, containing 15.3 g/l of volatilefatty acids (VFAs) ranging from acetic to valeric acid with

butyric acid predominating (12.8 g/l). The VFA mixture wasconcentrated to 732.5 g/l with a 93.3 % yield of butyric acid(643.9 g/l). Three individual Pseudomonas putida strains,KT2440, CA-3 and GO16 (single pure cultures), differed intheir ability to grow and accumulate PHA from VFAs. P.putida CA-3 achieved the highest biomass and PHA onaverage with individual fatty acids, exhibited the greatesttolerance to higher concentrations of butyric acid (up to40 mM) compared to the other strains and exhibited amaximum growth rate (μMAX=0.45 h−1). Based on theseobservations P. putida CA-3 was chosen as the test strainwith the concentrated VFA mixture derived from the ADleachate. P. putida CA-3 achieved 1.56 g of biomass/l andaccumulated 39 % of the cell dry weight as PHA (nitrogenlimitation) in shake flasks. The PHA was composedpredominantly of 3-hydroxydecanoic acid (>65 mol%).

Keywords mcl-Polyhydroxyalkanoates . Grass biomass .

Biorefining . Anaerobic digestion . VFAs

Introduction

Grass biomass is one of the most diffuse biomes in the world.It covers 40 % of the terrestrial surface and represents 70 % ofthe agricultural land (Panunzi 2008). In Ireland 61 % of theland is dedicated to agriculture and 55 % of this is composedof grassland (Central Statistics Office report 2007). Includingpasture and rough grazing areas the grassland percentage risesup to 86 % (O'Mara 2008). Grass can grow on land that isunsuitable for the growth of other crops and is a second-generation (lignocelluosic) biomass resource. The need todevelop technologies for the production of polymers fromrenewable lignocellulosic resources is well recognised

Electronic supplementary material The online version of this article(doi:10.1007/s00253-013-5323-x) contains supplementary material,which is available to authorized users.

F. Cerrone : S. K. Choudhari :R. Davis :D. Cysneiros :V. O’Flaherty :R. P. Babu :K. O’ConnorTechnology Centre for Biorefining and Bioenergy, Orbsen Building,National University of Ireland, Galway, Ireland

F. Cerrone :R. Davis :M. W. Guzik : S. T. Kenny :K. O’Connor (*)School of Biomolecular and Biomedical Sciences, UCD ConwayInstitute and Earth Institute, University College Dublin, BelfieldDublin 4, Irelande-mail: kevin.oconnor@ucd.ie

S. K. Choudhari :R. P. BabuCentre for Research on Adaptive Nanostructure and Nanodevices,Trinity College Dublin, Dublin 2, Ireland

D. Cysneiros :V. O’FlahertyDepartment of Microbiology, School of Natural Sciences andEnvironmental Change, National University of Ireland,Galway, Ireland

G. Duane : E. CaseySchool of Biochemical and Bioprocessing Engineering, UniversityCollege Dublin, Belfield Dublin 4, Ireland

R. P. BabuSchool of Physics, Trinity College Dublin, Dublin 2, Ireland

Appl Microbiol Biotechnol (2014) 98:611–620DOI 10.1007/s00253-013-5323-x

(Satyanarayana et al. 2009). The use of such resources that donot compete with food is an important criterion forsustainability. Given the diffuse nature of grass and its lackof competition with food crops it is an ideal starting materialfor biopolymer production.

The microbial breakdown of grass is well reported (Harutaet al. 2002; Witzig et al. 2010). Grass is composed of cellulose(25–40%), hemicellulose (35–50%), lipids (3 %), lignin (10–30 %) and free sugars (10–26 %) (e.g., sucrose; Koch et al.2010; Ellis et al. 2012) and is thus a rich source of fermentablecarbon substrates (Sanchez 2009). Grass has been the subjectof numerous studies for anaerobic digestion and theproduction of biogas (Lehtomaki et al. 2008; Seppala et al.2009). However grass has not been studied as a substrate forpolyhydroxyalkanoate (PHA) production. Anaerobicdigestion is a bioprocess where different microbialpopulat ions, namely acidogenic, acetogenic andmethanogenic bacteria, act in sequence to degrade complexsubstrates to produce methane and carbon dioxide (Nizamiet al. 2009; Cysneiros et al. 2012a). The substrates produced ateach step support the metabolism of the following populationin the chain (Cirne et al. 2007). Key intermediates in theanaerobic digestion process are short chain fatty acids suchas acetic, propionic, butyric, and valeric acid (Du and Yu2002a; Siegert and Banks 2005). Varying the conditions ofthe anaerobic digestion bioreactor, can enrich for particularfatty acids, e.g., acidic conditions of incubation enriched forbutyric acid production, while neutral pH conditions result ina mixture of both acetic and butyric acid production in equalproportion (Cysneiros et al. 2012a). Volatile fatty acids(VFAs) are well-known substrates for bacteria accumulatingthe biodegradable polymer polyhydroxybutyrate (PHB)(Albuquerque et al. 2011; Lemos et al. 2006) but not mediumchain length PHA (mcl-PHA). PHAs are a family ofpolyesters with side chains that vary in length from a simplemethyl moiety (PHB) to alkyl chains up to 11 carbons long.They are broadly divided into two divisions namely, short sidechain length (scl) with up to two carbons in the side chain andmedium chain length (mcl) with three or more carbons in theside chain (Khanna and Srivastava 2005). The properties ofthe polymers are affected by the length of the side chain aswell as the molecular weight of the polymer (Hazer andSteinbüchel 2007). PHB, an sclPHA, is a highly crystallinepolymer which is brittle, while mcl-PHA properties rangefrom rubber to tacky polymers (Steinbüchel 2003). Themajority of studies on the accumulation of medium chainlength PHA from fatty acids have focused on substrates suchas octanoic acid (Jiang et al. 2012; Le Meur et al. 2012) withno studies focusing onVFA tomcl-PHA. This is the first studyto report about the anaerobic digestion of a widely availablerenewable lignocellulosic resource to produce substrates(VFAs) for the production of the biodegradable polymermcl-PHA.

Material and methods

Chemicals

Fatty acids such as acetic acid, propionic acid, butyricacid, and valeric acid, all of 99 % purity, were purchasedfrom Sigma Aldrich (Dublin, Ireland) and used as carbonsources for growth and PHA accumulation by bacteria. AVFA mixture produced from the anaerobic digestion ofgrass was generated within the project. The medium usedfor bacterial cultivation was MSM media (Schlegel et al.1961). All other chemicals were purchased from SigmaAldrich.

Bacterial strains and culture conditions

Three strains of Pseudomonas putida were chosen basedon their ability to grow with the short chain fatty acidbutyric acid. They were P. putida GO16 (Kenny et al.2012) CA-3 (O'Connor et al. 1995) and KT2440(Regenhardt et al. 2002). The strains were maintainedas freeze dried (lyophilised) cultures for long termstorage and on MSM agar plates supplemented withsodium gluconate as the sole carbon and energy sourcefor periods of up to 2 weeks. From the MSM plates, asingle colony was innoculated into 5 ml of MSM mediawith 20 mM sodium butyrate as C source and incubatedfor 18 h at 30 °C shaking. This pre-inocula (1 ml) was used toinoculate 50 ml of MSM media with either 1 g/l NH4Cl (nonlimiting nitrogen concentration) or 0.1 g/l NH4Cl (Nitrogenlimiting growth and thus PHA accumulating conditions). Arange of VFA concentrations (10–60 mM) was tested forgrowth and PHA accumulation. The cultures were incubatedfor 48 h at 30 °C with shaking (200 rpm).

Nutrients analysis

The concentration of nitrogen in the growth media wasmonitored over time using the previously described methodof Scheiner (1976). Phosphate concentration in the growthmedium (PO4

2−) was determined according to the US EPAcolorimetric method (USEPA 1978). Fatty acid consumptionin the growth medium was determined by subjecting samplesto methylation with a 1:1 solution of chloroform/methanolafter NaCl addition with an acidic digestion (H2SO4, 50 %vol/vol). Fatty acid methylesters were analyzed with GC-FIDdetector using an HP-INNOWAX capillary column (25 m×0.25 mm, 0.32-μm film thickness; SGE Analytical Sciences).A temperature program was used in order to separate thedifferent peaks of the alkanoic acid methylesters; theconditions for the capillary column were the following:120 °C for 1 min; temperature ramp for 10 °C/min up to240 °C for a total of 13 min.

612 Appl Microbiol Biotechnol (2014) 98:611–620

PHA extraction

The cells pellets arising from centrifugation of 50 mlculture samples from shaken flasks were frozen at−80 °C and freeze-dried overnight. A method ofmethylation/sulphuric acid extraction was used aspreviously described to generate methylesters of thePHA monomers (Lageveen et al. 1988). The sampleswere analyzed in a Hewlett-Packard 6890 N gaschromatograph equipped with an HP-INNOWAXcapillary column (25 m×0.25 mm, 0.32-μm filmthickness; SGE Analytical Sciences) and a flameionisation detector to detect 3-hydroxyalkanoic methylesters with the temperature program previously described(Lageveen et al. 1988).

Grass substrate

Whole-crop of perennial grass, chopped to achieve anaverage length of ~1.5 cm, was used as substrate foranaerobic digestion. The grass was harvested from thefield, ensiled for approximately 3 months andsubsequently pressed in a pilot scale screw press(Model PP-7RL-S; Ponndorf Maschinenfabrik, GmbH,Kassel, Germany) to have the solid and liquid phasesseparated. The screw was set at a pitch 1:7.5 androtated with 11–14 revolutions/min and sample pressingwas conducted with a conical screw press with a screenperforation of 1.5 mm in diameter. Both liquid and solidphases were frozen at −20 °C, then thawed overnightprior to use in the experiments.

Design of the leach bed reactors

The leach bed reactors were made from a 400-mm-longperspex pipe with a 150-mm diameter which was closed ateach end. This forms a closed vessel with a 4-l workingvolume, from which 3.5 l was the leach bed (leachate plusgrass). This reactor was mounted vertically and the bottomplate was fitted with a drainage tube covered with a stainlesssteel mesh on which the bed of substrate was supported,preventing the drainage tube from clogging. The top platehad a port for the recirculation tube and another one for therelease of gas into a gas collection bag. The leach bed wasoperated in a downflow mode which was achieved bypumping the leachate through a tube inserted in the base ofthe reactor to an inlet above the level of the substrate bed.

Leach bed reactor operational temperature

Each reactor was heated bymeans of a water jacket around thereactor, which was made of another 400 mm piece of aPerspex tube with a 250-mm diameter placed outside the

reac tor. The tempera ture was main ta ined by athermocirculator placed in a water bath and operated tomaintain the reactor temperature at 37 °C.

Leach bed reactor experimental set-up and operation mode

Three leach bed reactors were operated for 14 days each. Tostart the batch, all reactors were filled with ~1.1 kg of a mixcontaining ~280 g of granular sludge derived from a full scalereactor digesting dairy wastewater and ~820 g of ensiledgrass. This gave a volatile solid (VS) content of ~200 g, ofwhich ~170 g was fresh substrate.

After the reactors were filled with the solid mixture, 2 l ofleachate solution was added to the control reactor and thesame leachate containing sodium bicarbonate (NaHCO3)was added to the experimental reactors. The sodiumbicarbonate concentration was calculated to maintain pH atthe desired range (5.0 or 6.0). The reactors were then sealed.The leachate was continuously recirculated from the bottom ofthe reactor to the top by a peristaltic pump.

Two leach beds were used as experimental reactors, one ofwhich was maintained at pH 5.0 and one at pH 6.0, while onereactor was used as a control at pH ~6.5. pH was monitoreddaily and when necessary, additional NaHCO3 wassupplemented to the experimental reactors.

Sampling and analysis

The initial mixture of feedstock and inoculum at the beginningand the digestate at the end of each batch were weighed and asample was taken to determine total solids (TS) and VSanalysis. Leachates were sampled, from a sampling portplaced in the recirculation line, every alternate day and wereanalysed for chemical oxygen demand (COD) and VFAcontent. The pH was measured daily directly in the samplingport. The biogas produced was collected in gas bags and itsvolume and composition was analysed every alternate day.

VFA leachate collection

At the end of each AD leachate bed reactor batch, theleachates were drained, collected and stored at −20 °C. TheVFAs present in the leachate samples were subjected toconcentration and subsequently used as substrates for PHAaccumulation tests.

Concentration of volatile fatty acids from anaerobic digestionleachate

VFAs coming from anaerobic digestion were furtherconcentrated using a combination of salting out and liquidextraction as previously described (Alkaya et al. 2009;Cherkasov and Il'in 2009).

Appl Microbiol Biotechnol (2014) 98:611–620 613

Results

Anaerobic reactor conditions to alter the VFAs composition

The anaerobic digestion of grass resulted in the production aVFA mixture composed of acetic acid, propionic acid, butyricacid and valeric acid in a ratio (mol%) of 10.6:5.5:80.5:3.2,respectively, and a total concentration of 15.5 g/l. Based on thecomposition of the grass AD leachate and while developingthe literature method to concentrate VFA mixtures we tested arange of commercially available VFAs as substrates for thegrowth and mcl-PHA accumulation by three P. putida strains.

Screening bacterial strains for PHA accumulationusing a range of volatile fatty acids

P. putida KT2440, a TOL-plasmid deficient variant ofP. putida mt-2 whose genome is fully sequenced showedsimilar growth with all the four VFAs achievingapproximately 0.5 g CDW/l (Fig. 1a). P. putida CA-3(NCIMB 41162) previously reported to utilise a broad rangeof fatty acids (Ward and O'Connor 2005; Tobin et al. 2007)

achieved maximum biomass (0.73 g/l) with butyric acid as thecarbon source. All other substrates achieved an average celldry weight of 0.6 g/l. (Fig. 1a). VFAs with an even number ofcarbons in the aliphatic chain resulted in higher biomass ofP. putida GO16 (NCIMB 41538; Kenny et al. 2012)compared to VFAs with uneven chain length. This strainwas able to achieve the highest biomass (0.82 g CDW/l) withbutyric acid as the sole carbon source. The other three VFAs(acetic, propionic and valeric acid) resulted in a cell dryweight of 0.43, 0.11 and 0.16 g/l, respectively.

PHA accumulation by P. putida strains when supplied VFAs

Nitrogen limitation in a chemically defined media (MSM) isone of the easier ways to induce PHA accumulation in P.putida (Hoffmann and Rehm 2005). Thus all experimentswere performed using a defined medium with a limitingconcentration of nitrogen (0.1 g/l).

P. putida KT2440 was able to accumulate 26 and 29 % ofthe cell dry weight as PHAwith acetic acid and propionic acidbut produced the highest PHA with butyric acid (44 % ofCDW). Interestingly, valeric acid was a poorer substrate thanbutyric acid for PHA accumulation with only 19 % of theCDW of strain KT2440 composed of the polymer (Fig. 1b).P. putida CA-3 showed better storage ability, accumulatingsimilar levels of PHAwith acetic, propionic, butyric acid (36–44 % of CDW) and accumulated lowest levels of PHA (24 %of CDW) with valeric acid as the sole carbon and energysource in flasks culture (Fig. 1b). P. putida GO16 accumulated28 % of CDW as PHAwith butyric acid as the carbon sourcebut all other substrates resulted in lower PHA accumulation(Fig. 1b).

The monomer composition of the PHA accumulatedby P. putida KT2440 and CA-3 was similar for allsubstrates tested with (R )-3-hydroxydecanoic acid (3-HD)as the predominant monomer and two other monomers(R ) -3-hydroxyoctanoic ac id (3-HO) and (R ) -3-hydroxydodecanoic acid (3-HDD) as minor components(Fig. 2a,b). (R )-3-Hydroxytetradecanoic acid (3-HTD)was observed in trace amounts in PHA accumulated by allthe three strains (Fig. 2). While 3-HD was also thepredominant monomer in PHA accumulated by P. putidaGO16, the polymer also contains higher levels of 3-HDDmonomers compared to PHA accumulated by P. putidaKT2440 and CA-3 (Fig. 2c). The level of 3-HDD in PHAaccumulated by P. putida GO16 appears to be straindependent. Interestingly when uneven carbon numberedfatty acids were used (i.e., propionic and valeric acid) asC source, PHA composed of uneven carbon numbered PHAmonomers (especially (R )-3-hydroxynonanoic and (R )-3-hydroxyundecanoic) was accumulated in all three strains,with the highest value observed for P. putida GO16 (16 %of (R )-3-hydroxyundecanoic acid (3-HUD) (Fig. 2).

Fig. 1 a growth of P. putida strains (KT2440, CA-3 and GO16,respectively) with sclVFAs as C source (A acetic acid, P propionic acid,B butyric acid, V valeric acid). The numbers above each bar refer to thegrowth yield (g CDW/g VFA). b Tot PHA production (% of CDW) by P.putida strains (KT2440, CA3 and GO16, respectively) with sclVFAs as Csource (A acetic acid, P propionic acid, B butyric acid, V valeric acid)

614 Appl Microbiol Biotechnol (2014) 98:611–620

Maximum growth rate determination (μMAX)

From the shake flask experiments, it was evident that all threebacteria showed best growth with butyric acid (Fig. 1a). Thegrowth rate of these strains on butyric acid and the maximumtolerated concentration were determined for the three strains.Fatty acids have been reported to be toxic to the growth of

bacteria (Cheung et al. 2010) and we observed an increase inthe lag period at concentrations above 30 mM (KT2440) andat 40 mM butyric acid (GO16) (3.4 g/l) (Fig. S1). P. PutidaCA-3 exhibited the highest growth rate of all three strains withbutyric acid (Table 1) as the sole carbon and energy source.The growth rate of P. putida CA-3 was the same for all thetested concentrations and was 1.4- and 1.8-fold highercompared to the μMAX of P. putida KT2440 and P. putidaG016, respectively, at 40 mM butyric acid (Table 1). Noincrease of lag period for P. putida CA-3 was detected withan increase in butyric acid concentration (up to 40 mM)(Fig. S1). The specific PHA production yield was very similarfor P. Putida KT2440 and P. putida CA-3 with increasingconcentrations of sodium butyrate, while the specificproduction yield of P. putida GO16 was the same as the othertwo strains up to 20 mM concentrations but lower at higherconcentrations (Table 1).

Based on the observations that P. putida CA-3 had thehighest growth rate and best tolerance to high concentrationsof butyric acid, it was selected for further study with aconcentrated VFAmixture produced from anaerobic digestionof grass. The VFA solution arising from an anaerobic digestoris too dilute to be used in a subsequent bioprocess and needs tobe concentrated. The following section describes theproduction and purification of butyric acid from the VFAmixture and subsequent conversion to PHA.

Concentrating the AD VFAs solution

Two litres of the VFA mixture (15.5 g/l) arising from theanaerobic digester were precipitated as described above(materials and methods) resulting in a 37.13-ml solution.This concentrated solution contained a mixture of acetic acid,propionic acid, butyric acid, and valeric acid in a ratio of3.3:4.7:87.4:5.5 (mol%) The butyric acid was concentrated51-fold, and this amount represents a 93.3 % yield (Table 2).100 % of the valeric acid was recovered while 69 % and 30 %recovery of propionic and acetic acid, respectively, wasobserved. This solution was subsequently used for growthand PHA accumulation by P. putida CA-3. Given that theleachate CODwas 17 g/l and the resulting total VFAs after theanaerobic digestion were 15.3 g/l, it can be estimated a totalrecovery of the 79 % of the total COD due to a total VFAsrecovery of 13.4 g/l (after the extraction process).

Concentrated anaerobic digestate (AD VFAs) as C sourcefor growth and PHA accumulation

The concentrated VFA solution was used as the sole carbonsource for shake flask experiments, where either nitrogen (N)or inorganic phosphate (P) limitation was imposed to inducePHA accumulation. Two concentrations of the total mixture ofVFA were used (1.69 and 3.38 g/l). PHA accumulation was

Fig. 2 PHA monomer composition (mol%) by P. putida strains(a KT2440, b CA-3, c GO16, respectively) using differentsclVFAs. 3 -HO 3-hydroxyoctanoate, 3 -HN 3-hydroxynonanoate,3 -HD 3-hydroxydecanoate, 3 -HUD 3-hydroxyundecanoate,3 -HDD 3-hydroxydodecanoate, 3 -HTrD 3-hydroxytridecanoate,3 -HTD 3-hydroxytetradecanoate

Appl Microbiol Biotechnol (2014) 98:611–620 615

39 % of CDW after 48 h of growth with 3.38 g/l ofconcentrated VFA mixture in limited N growth conditions.The total biomass, under these conditions was 1.56 g ofCDW/l. P limited growth conditions at 3.38 g/l resulted in asimilar (35 %) PHA accumulation, while the total biomass, inP limited conditions was 1.31 g CDW/l (data not shown).Interestingly the predominant PHA monomer was 3-hydroxydecanoic acid (3-HD) (65–66 mol%) under allgrowth conditions with 3-hydroxyoctanoic acid (3-HO)and a 3-HDD as minor monomers. Traces of 3-hydroxytetradecanoic acid (3-HTD) were found in all PHAsamples (Fig. 3).

Discussion

Grass is used primarily as food for grazing animals used toproduce milk and meat. Due to its abundance, high cellulose/hemicellulose content and low lignin content (Balat 2011) it isa good source of fermentable (hexoses/pentoses) sugars forPHA accumulation by bacteria. However, few PHAaccumulating bacteria utilise pentoses compared to hexoseswhich can limit the available sugars for a subsequentfermentation process. Converting the sugars to VFAs throughanaerobic digestion of grass can indirectly make the carbonmore available to PHA accumulating bacteria.

While the metabolic versatility of P. putida is welldescribed with a large number of substrates for growth andPHA production (Rodrigues et al. 2005; Sohn et al. 2010),growth ofPseuodomonas with VFA is less well described andno reports of VFA to mcl-PHA have been described.Furthermore, the conversion of grass to VFAs andsubsequently to mcl-PHA is unreported.

While P. putida KT2440 and CA-3 are genetically verysimilar (O'Leary et al. 2005), they showed differences ingrowth with butyric acid and valeric acid as substrates in thecurrent study. Interestingly, the levels of PHA accumulationand monomer composition were similar for both strains for allsubstrates tested. P. putida GO16 showed only good growthand PHA accumulation with fatty acids with an even numberof carbons. P. putida U has also been shown to grow andT

able1

Growth

ratesof

thethreedifferentb

acteria(P.putidaKT2440,C

A-3

andGO16)with

differentconcentratio

nof

butyricacid

(range

10–40mM)

P.P

utidaKT2440

P.putidaCA-3

P.putidaGO16

Maxim

umGrowth

rate

(μMAX=h−

1)

R2

PHAproductio

nyield(g/g)

Maxim

umGrowth

rate

(μMAX=h−

1)

R2

PHAproductio

nyield(g/g)

Maxim

umGrowth

rate(μ

MAX=h−

1)

R2

PHAproductio

nyield(g/g)

10mM

0.28

0.83

0.006±0.001

0.43

0.72

0.006±0.001

0.21

0.89

0.006±0.003

20mM

0.30

0.85

0.007±0.001

0.44

0.94

0.007±0.001

0.23

0.85

0.005±0.001

30mM

0.34

0.96

0.08

±0.023

0.45

0.88

0.08

±0.011

0.21

0.87

0.007±0.004

40mM

0.33

0.87

0.44

±0.024

0.45

0.89

0.38

±0.057

0.24

0.90

0.29

±0.040

SpecificPH

Aproductio

nyieldisgivenas

gof

PHAforgof

CDW

Table 2 VFAs composition of the anaerobic digestate and concentration(g/l) of each fatty acid before and after extraction process

Aceticacid (g/l)

Propionicacid (g/l)

Butyricacid (g/l)

Valericacid (g/l)

Initial fatty acidconcentration

1.1 0.8 12.8 0.6

Fatty acid concentrationspost purification

18.2 30.1 643.9 40.3

% Recovery of each acids 30.4 68.8 93.3 100

616 Appl Microbiol Biotechnol (2014) 98:611–620

accumulate PHA better from even carbon number mediumchain fatty acids compared to uneven carbon substrates (Ariaset al. 2008).

Previous reports suggested that scl VFAs, such as C5, aremore likely to be used for energy supply and cell materialsynthesis rather than mcl-PHA accumulation (Du and Yu2002b). In the current study we observed accumulation ofPHA mainly composed by C8 and C10 monomers but alsowith uneven monomers when valeric acid used as a substrate(Fig. 2b). A reverse β-ketothiolase mediated chain elongationof PHA precursors has previously been described (Huijbertset al. 1992; Olivera et al. 2001; Steinbüchel and Lotke-Eversloh 2003), and such a reaction could explain thepresence of these uneven chain monomers. We also observemonomers with an uneven carbon chain in PHA isolated fromstrains grown with propionate (Fig. 2b) but to a lesser extentcompared to that observed with valeric acid, suggestingpropionate or its metabolic intermediates are less efficientsubstrates for chain elongation.

The toxicity of VFAs to bacteria is well known (Cheunget al. 2010). The inhibition is due to a change in osmolarity ofthe cell due to the need for bacteria to actively pump outanions to counter the effect of the acidic pH in the growthmedium. Furthermore, the uptake of the organic acid into thecell releases protons contributing to acidification of thecytoplasm and consequent energy consumption to maintainthe gradient inside the cell. While P. putida CA-3 and P.putida KT2440 showed very similar PHA accumulation withincreasing concentrations of sodium butyrate P. putida CA-3was chosen for further study as it exhibited the highest growthrate at 40 mM butyrate (Table 1). The latter is a critical factoras growth rate at higher concentrations is a reality of fed batchcultivations for PHA production in stirred tank reactors (futuredevelopment work) and butyrate is the major fatty acid in the

VFA mix generated from anaerobic digestion of grass. Thegreater tolerance of higher butyric acid concentrations (up to40 mM) by P. putida CA-3 over KT2440 may be explainedby higher polyP storage ability (50 % more than P. putidaKT2440) (Tobin et al. 2007) as this inorganic biopolymer isknown to enhance stress tolerance and provide energy for thecell (Casey et al. 2013).

VFA generated from the anaerobic digestion of food wastehas been used as a substrate for PHB production withCupriavidus necator cells accumulating 87 % of the cell dryweight as PHB (Hafuka et al. 2011). However the conversionof VFAs to medium chain length PHA has not been described.The highest level of PHA accumulated by P. putida CA-3from the concentrated VFA mix was 39 % of CDW. Thiscompares favourably with mcl-PHA accumulation from otherPHA unrelated substrates such as glucose, glycerol (Kennyet al. 2012), and short chain fatty acids (Reddy et al. 2012) butis lower than PHA accumulated from longer chain fatty acids(PHA related substrates) such as octanoic acid (70 % ofCDW) (Ramsay et al. 1992).

In general, AD processes are designed to avoid or minimisethe generation of VFAs as they inhibit methanogenic bacteria(Nizami et al. 2010; Jagadabhi et al. 2010). However, thedesign of a process that maximises VFA production wouldbe beneficial for the production of biopolymers such as PHA.While the maximisation of VFA production is highly desired,it is also necessary to prevent the negative effect of VFAs onthe biomass hydrolytic process (Bhattacharyya et al. 2008).Unsuitable digester and process design can result in highconcentrations of VFAs which negatively impact microbialactivity and result in both inocula replacement and loss ofhydrolytic potential (Nizami et al. 2010). Thus, lower organicloading rates (OLR) are needed to avoid such VFA poisoningof the AD process. The configuration used in the study can

Fig. 3 Monomer (mol%)composition in the extracted PHA(1.7–3.4 g/l of VFAs as C source)with either limited in P or limitedin N media. 3-HH3-hydroxyhexanoate, 3-HO3-hydroxyoctanoate, 3-HD3-hydroxydecanoate, 3-HDD3-hydroxydodecanoate, 3-HTD3-hydroxytetradecanoate

Appl Microbiol Biotechnol (2014) 98:611–620 617

work with higher OLR as the VFA is harvested from therecirculating leachate but the enzymes are kept in the digesterto promote hydrolysis and thus further VFA production. Thisprevents a too high concentration of VFAs thus avoidingsubstrate inhibition, inocula replacement, and allows moresustainable VFA production (Lu et al. 2008; Cysneiros et al.2012b). The combination of anaerobic digestion and aerobicfermentation can convert a complex mixture of biopolymerspresence in grass into a single biodegradable thermoplasticelastomer PHA. P. putida CA-3 was the most robust of threePHA accumulating strains tested, and the data arising fromthis study will form the basis for future development of a fed-batch process for PHA production from grass in a stirred tankbioreactor.

Two major approaches for PHA production have beenproposed in the scientific literature namely monocultures(Steinbüchel and Lotke-Eversloh 2003; Cerrone et al. 2010;Riedel et al. 2012; Muhr et al. 2013; Chuah et al. 2013) oropen mixed cultures (Lemos et al. 2006; Albuquerque et al.2011; Jiang et al. 2011). The open mixed culture approach hasto date reported uniquely on PHB or PHBV accumulatingcultures as mcl-PHA accumulators either do notpredominate/exist in the startingmaterials that act as inoculumfor open cultures or the selective pressures do not allow theirsurvival in open mixed cultures (Jiang et al. 2011). The twoapproaches are not necessarily mutually exclusive and the useof both can be envisaged within a biorefining context whereboth scl-PHA and mcl-PHA can be produced from differentsubstrates/co-substrate streams emerging from the samestarting material. Indeed, a variety of approaches will mostlikely be adopted by industry depending on the substrate thatis available locally and also the type of PHA that is targetedfor applications (Xu et al. 2010) Integration of PHAproduction activities within a biorefining concept is the mostlikely way to achieve maximum economic and environmentalbenefit (Nikodinovic-Runic et al. 2013).

Acknowledgments This work was supported by the TechnologyCenter for Biorefining and Bioenergy (project n° CC20090004) www.tcbb.ie with the financial support of Enterprise Ireland and the IrishIndustrial Development Agency (IDA).

References

Albuquerque MAM, Martino V, Pollet E, Averous L, Reis MAM (2011)Mixed culture polyhydroxyalkanoate (PHA) production fromvolatile fatty acid (VFA)-rich streams: effect of substratecomposition and feeding regime on PHA productivity, compositionand properties. J Biotechnol 151(1):66–76

Alkaya E, Kaptan S, Ozkan L, Uludag-Demirer S, Demirer GN (2009)Recovery of acids from anaerobic acidification broth by liquid–liquid extraction. Chemosphere 77:1137–1142

Arias S, Sandoval A, ArcosM, Caedo LM,Maestro B, Sanz JM, NaharroG, Luengo JM (2008) Poly-3-hydroxyalkanoate synthases from

Pseudomonas putida U: substrate specificity and ultrastructuralstudies. Microb Biotechnol 1(2):170–176

Balat M (2011) Production of bioethanol from lignocellulosic materialsvia the biochemical pathway: a review. Energy Convers Manage 52:858–875

Bhattacharyya JK, Kumar S, Devotta S (2008) Studies on acidification intwo-phase biomethanation process of municipal solid waste. WasteManage 28:164–169

Casey WT, Nikodinovic-Runic J, Fonseca-Garcia P, Guzik MW,McGrath JW, Quinn JP, Cagney G, Prieto MA, O'Connor KE(2013) The effect of polyphosphate kinase (ppk ) deletion onpolyhydroxyalkanoate accumulation and carbon metabolism inPseudomonas putida KT2440 Env Microbiol Rep, in press

Cerrone F, Sánchez-Peinado MDM, Juárez-Jimenez B, González-LópezJ, Pozo C (2010) Biological treatment of two-phase olive millwastewater (TPOMW, alpeorujo): polyhydroxyalkanoates (PHAs)production by Azotobacter strains. J Microbiol Biotechnol 20(3):594–601

Cherkasov DG, Il'in KK (2009) Salting out of butyric acid from aqueoussolutions with potassium chloride. Russ J Appl Chem 82(5):920–924

Cheung HNB, Huang GH, Yu H (2010) Microbial-growth inhibitionduring composting of food waste: effects of organic acids. BioresTechnol 101(15):5925–5934

Chuah JA, Yamada M, Taguchi S, Sudesh K, Doi Y, Numata K (2013)Biosynthesis and characterization of polyhydroxyalkanoatecontaining 5-hydroxyvalerate units: effects of 5HV units onbiodegradability, cytotoxicity, mechanical and thermal properties.Polym Degrad Stabil 98(1):331–338

Cirne DG, Lethomaki A, Bjornsson L, Blackall LL (2007) Hydrolysisand microbial community analyses in two-stage anaerobic digestionof energy crops. J Appl Microb 103:516–527

CSO (Central Statistics Office) (2007) Census 2006. Volume 1(Population classified by area) and Volume 8 (Occupations).Central Statistics Office, Cork

Cysneiros D, Banks CJ, Heaven S, Karatzas Kimon-Andreas G (2012a)The effect of pH control and 'hydraulic flush' on hydrolysis andVolatile Fatty Acids (VFA) production and profile in anaerobic leachbed reactors digesting a high solids content substrate. BioresTechnol 123:263–271

Cysneiros D, Banks CJ, Heaven S, Karatzas Kimon-Andreas G (2012b)The role of phase separation and feed cycle length in leach bedscoupled to methanogenic reactors for digestion of a solid substrate(Part 1): optimisation of reactors' performance. Biores Technol 103(1):56–63

Du G, Yu J (2002a) Green technology for conversion of food scraps tobiodegradable thermoplastic polyhydroxyalkanoates. Env SciTechnol 36:5511–5516

Du G, Yu J (2002b) Metabolic analysis on fatty acid utilization byPseudomonas oleovorans : mcl-poly(3-hydroxyalkanoates)synthesis versus β-oxidation. Process Biochem 38(3):325–332

Ellis JL, Dijkstra J, France J, Parsons AJ, Edwards GR, Rasmussen S,Kebreab E, Bannink A (2012) Effect of high-sugar grasses onmethane emissions simulated using a dynamic model. J Dairy Sci95:272–285

Hafuka A, Sakaida K, Satoh H, Takahashi M, Watanabe Y, Okabe S(2011) Effect of feeding regimens on polyhydroxybutyrateproduction from food wastes by Cupriavidus necator. BioresTechnol 102(3):3551–3553

Haruta S, Cui Z, Huang Z, Li M, Ishii M, Igarashi Y (2002) Constructionof a stable microbial community with high cellulose-degradationability. Appl Microbiol Biotechnol 59(4–5):529–534

Hazer B, Steinbüchel A (2007) Increased diversification ofpolyhydroxyalkanoates by modification reactions for industrialand medical applications. Minireview. Appl Microbiol Biotechnol74(1):1–12

618 Appl Microbiol Biotechnol (2014) 98:611–620

Hoffmann N, Rehm BHA (2005) Nitrogen-dependent regulation ofmedium-chain length polyhydroxyalkanoate biosynthesis genes inpseudomonads . Biotechnol Lett 27:279–282

Huijberts GNM, Eggink G, de Waard P, Huisman GW, Witholt B (1992)Pseudomonas putida KT2442 cultivated on glucose accumulatespoly(3-hydroxyalkanoates) consisting of saturated and unsaturatedmonomers. Appl Environ Microbiol 58:536–544

Jagadabhi GS, Kaparaju P, Rintala J (2010) Effect of micro-aeration andleachate replacement on COD solubilisation and VFA productionduring mono-digestion of grass-silage in one-stage leach-bedreactors. Biores Technol 101(8):2818–2824

Jiang Y, Hebly M, Kleerebezem R, Muyzer G, van Loosdrecht MCM(2011) Metabolic modeling of mixed substrate uptake forpolyhydroxyalkanoate (PHA) production. Water Res 45(3):1309–1321

Jiang X, Sun Z, Marchessault RH, Ramsay JA, Ramsay B (2012)Biosynthesis and propert ies of medium-chain- lengthpolyhydroxyalkanoates with enriched content of the dominantmonomer. Biomacromol 13(9):2926–2932

Kenny TS, Runic JN, Kaminsky W, Woods T, Babu RP, O'Connor KE(2012) Development of a bioprocess to convert PET derivedterephthalic acid and biodiesel derived glycerol to medium chainlength polyhydroxyalkanoate. Appl Microbiol Biotechnol 95:623–633

Khanna S, Srivastava AK (2005) Recent advances in microbialpolyhydroxyalkanoates. Process Biochem 40(2):607–619

Koch K, Lübken M, Gehring T, Wichern M, Horn H (2010) Biogas fromgrass silage – measurements and modeling with ADM1. BioresTechnol 101(21):8158–8165

Lageveen R, Huisman G, Preusting H, Witholt B (1988) Formation ofpolyesters by Pseudomonas oleovorans : effect of substrates onformation and composition of poly-(R)-3-hydroxyalkanoates andpoly-(R)-3-hydroxyalkanoates. Appl Environ Microbiol 54(12):2924–2932

LeMeur S, ZinnM, Egli T, Thöny-Meyer L, Ren Q (2012) Production ofmedium-chain-length polyhydroxyalkanoates by sequential feedingof xylose and octanoic acid in engineered Pseudomonas putidaKT2440. BMC Biotechnol 12:53

Lemos PC, Serafim LS, Reis MAM (2006) Synthesis ofpolyhydroxyalkanoates from different short-chain fatty acids bymixed cultures submitted to aerobic dynamic feeding. J Biotechnol122(2):226–238

Lehtomaki A, Huttunen S, Lehtinen TM, Rintala JA (2008) Anaerobicdigestion of grass silage in batch leach bed processes for methaneproduction. Biores Technol 99(8):3267–3278

Lu F, He PJ, Hao L, Shao LM (2008) Impact of recycled effluent on thehydrolysis during anaerobic digestion of vegetable and flowerwaste. Water Sci Technol 58(8):1637–1643

Muhr A, Rechberger EM, Salerno A, Reiterer A, Malli K, Strohmeier K,Schober S, Mittlebach M, Koller M (2013) Novel description ofmcl-PHA biosynthesis by Pseudomonas chlororaphis from animal-derived waste. J Biotechnol 165(1):45–61

Nikodinovic-Runic J, GuzikM, Kenny ST, Babu R, Werker A, O'ConnorKE (2013) Chapter Four. Carbon-rich wastes as feedstocks forbiodegradable polymer (polyhydroxyalkanoate) production usingbacteria. Adv Appl Microbiol 84:139–200

Nizami AS, Korres NE, Murphy JD (2009) Review of the integratedprocess for the production of grass biomethane. Env Sci Technol 43(22):8496–8508

Nizami AS, Thamisiriroj T, Singh A,Murphy JD (2010) Role of leachingand hydrolysis in a two-phase grass digestion system. Energ Fuel24:4549–4559

O'Connor K, Buckley CM, Hartmans S, Dobson ADW (1995) Possibleregulatory role for nonaromatic carbon sources in styrenedegradation by Pseudomonas putida CA-3. Appl EnvironMicrobiol 61:544–548

O'Leary N, O'Connor KE,Ward P, GoffM, DobsonADW (2005) Geneticcharacterization of accumulation of polyhydroxyalkanoate fromstyrene in Pseudomonas putida CA-3. Appl Env Microbiol 71(8):4380–4387

Olivera ER, Carnicero D, Garcia B, Minambres B, Moreno MA,Cañedo L, Di Russo CC, Naharro G, Luengo JM (2001) Twodifferent pathways are involved in the β-oxidation of n-alkanoicand n-phenylalkanoic acids in Pseudomonas putida U: geneticstudies and biotechnological applications. Mol Microbiol 39(4):863–874

O'Mara F (2008) Country Pasture/Forage Resource Profile. FAO Plantproduction and protection Division report Agriculture Dept. Cropand Grassland Service, FAO, 2008; http://www.fao.org/ag/AGP/AGPC/doc/Counprof/Ireland/Ireland.htm

Panunzi E (2008) Are grasslands under threat? Brief analysis of FAOstatistical data on pasture and fodder crops. http://www.fao.org/ag/agp/agpc/doc/grass_stats/grassstats.htm

Ramsay BA, Saracovan I, Ramsay JA, Marchessault RH (1992) Effect ofnitrogen limitation on long-side-chain poly-beta-hydroxyalkanoatesynthesis by Pseudomonas resinovorans . Appl Env Microb 58(2):744–746

Reddy MV, Nikhil GN, Mohan SV, Swamy YV, Sarma PN (2012)Pseudomonas ot i t idis as a potent ia l biocatalys t forpolyhydroxyalkanoates (PHA) synthesis using synthetic wastewaterand acidogenic effluents. Biores Technol 123:471–479

Regenhardt D, Heuer H, Heim S, Fernandez DU, Strömpl C, MooreERB, Timmis KN (2002) Pedigree and taxonomic credentials ofPseudomonas putida strain KT2440. Environ Microbiol 4:912–915

Riedel SL, Bader J, Brigham CJ, Budde CF, Yusof ZAM, Rha C, SinskeyAJ (2012) Production of poly(3-hydroxybutyrate-co -3-hydroxyhexanoate) by Ralstonia eutropha in high cell density palmoil fermentations. Biotech Bioeng 109(1):74–83

Rodrigues AC, Wuertz S, Brito AG, Melo LF (2005) Fluorene andphenanthrene uptake by Pseudomonas putida ATCC 17514: kineticsand physiological aspects. Biotechnol Bioeng 90(3):281–289

Sanchez C (2009) Lignocellulosic residues: biodegradation andbioconversion by fungi. Biotechnol Adv 27:185–194

Satyanarayana KG, Arizaga GGC, Wypych F (2009) Biodegradablecomposites based on lignocellulosic fibers—an overview. ProgrPolym Sci 34(9):982–1021

Scheiner D (1976) Determination of ammonia and Kjeldahl nitrogen byindophenol method. Water Res 10(1):31–36

Schlegel H, Kaltwasser H, Gottschalk G (1961) Ein Submersverfahren zurKultur wasserstoffoxidierender Bakterien:WachstumsphysiologischeUntersuchungen. Arch Mikrobiol 38:209–222

Seppala M, Pavola T, Lehtomaki A, Rintala J (2009) Biogas productionfrom boreal herbaceous grasses – specific methane yield andmethane yield per hectare. Biores Technol 100(12):2952–2958

Siegert I, Banks C (2005) The effect of volatile fatty acid additions on theanaerobic digestion of cellulose and glucose in batch reactors.Process Biochem 40(11):3412–3418

Sohn SB, Kim TY, Park JM, Lee SY (2010) In silico genome-scalemetabolic analysis of Pseudomonas putida KT2440 forpolyhydroxyalkanoate synthesis, degradation of aromatics andanaerobic survival. Biotechnol J 5(7):739–750

Steinbüchel A (2003) Production of rubber-like polymers bymicroorganisms. Curr Op Microbiol 6(3):261–270

Steinbüchel A, Lotke-Eversloh T (2003) Metabolic engineering andpathway construction for biotechnological production of relevantpolyhydroxyalkanoates in microorganisms. Biochem Eng J 16:81–96

Tobin KM, McGrath JW, Mullan A, Quinn JP, O'Connor KE (2007)Polyphosphate accumulation by Pseudomonas putida CA-3 andother medium-chain-length polyhydroxyalkanoate-accumulatingbacteria under aerobic growth conditions. Appl Environ Micrbiol73(4):1383–1387

Appl Microbiol Biotechnol (2014) 98:611–620 619

USEPA (1978) Method 365.3, Phosphorus, all forms, colorimetric,ascorbic acid, two reagent. Issued 1978. Table IB, Note 1

Xu Y, Wang RH, Koutinas AA, Webb C (2010) Microbial biodegradableplastic production from a wheat-based biorefining strategy. ProcBiochem 45(2):153–163

Ward PG, O'Connor KE (2005) Induction and quantification ofphenylacyl-CoA ligase enzyme activities in Pseudomonas putida

CA-3 grown on aromatic carboxylic acids. FEMS Microbiol Lett251:227–232

Witzig M, Boghun J, Kleinsteuber S, Fetzer I, Rodehustcord M(2010) Effect of the corn silage to grass silage ratio and feedparticle size of diets for ruminants on the ruminalBacteroides–Prevotella community in vitro . Anaer 16:412–419

620 Appl Microbiol Biotechnol (2014) 98:611–620