A

rF(grv(ds©

K

1

1lbg[itfitmn

0d

Enzyme and Microbial Technology 40 (2007) 1427–1434

Shrimp biowaste fermentation with Pediococcus acidolactici CFR2182:Optimization of fermentation conditions by response surface

methodology and effect of optimized conditions ondeproteination/demineralization and carotenoid recovery

N. Bhaskar ∗, P.V. Suresh, P.Z. Sakhare, N.M. SachindraDepartment of Meat, Fish and Poultry Technology, Central Food Technological Research Institute (CFTRI), Mysore 570 020, Karnataka, India

Received 31 July 2006; received in revised form 21 September 2006; accepted 13 October 2006

bstract

Fermentation of shrimp biowaste was conducted using different lactic acid bacteria (LAB) to select the efficient starter culture based on pHeduction and acid production. Pediococcus acidolactici CFR2182 was found to be the efficient (P ≤ 0.05) among the five starter cultures tested.ermentation conditions viz., inoculum level (X1), sugar level (X2) and incubation time (X3) were optimized using response surface methodologyRSM) to obtain the desirable pH of 4.3 ± 0.1. The optimized conditions were found to be 5% (v/w) inoculum (with 8.28 log cfu ml−1), 15% (w/w)lucose and 72 h of incubation time at 37 ± 1 ◦C to attain a pH of 4.30. The usefulness of the predicted model was further validated by consideringandom combinations of the independent factors. The high correlation (with regression coefficient close to 1.0) between the predicted and observed

alues during validation indicated the validity of the model. The effect of fermentation, by P. acidolactici CFR2182, on the production of chitinas indicated by deproteination and demineralization efficiency) and recovery of carotenoids was also studied. Deproteination of 97.9 ± 0.3% andemineralization of 72.5 ± 1.5% was achieved by fermentation of shrimp biowaste with P. acidolactici. The carotenoid recovery in fermentedhrimp biowaste, as compared to the wet waste, varied between 72.4 and 78.5% during fermentation.

2006 Elsevier Inc. All rights reserved.

on; D

dfobibewufap

eywords: Shrimp biowaste; Pediococcus; Fermentation; RSM; Deproteinizati

. Introduction

Indian shrimp processing industry produces more than00,000 tonnes of industrial wastes that makes it the singleargest industrial fish waste in the country [1]. Shrimp biowasteeing alkaline (pH 7.5–8.0) in its characteristic supports therowth of undesirable putrefying microflora resulting in spoilage2,3]. Although shrimp biowaste is usually dried on the beachest encourages not only environmental pollution but also reduceshe recoverable components in these biowastes. For instance,resh shrimp biowaste allows better pigment retention [4] whilets dried counterpart loses majority of its pigmentation due to

he strong affinity of the constituent pigments to oxygen [5] thus

aking drying unsuitable for a complete recovery of the compo-ents present in such biowastes. Any process that eliminates the

∗ Corresponding author. Tel.: +91 821 2514840; fax: +91 821 2517233.E-mail address: bhasg3@yahoo.co.in (N. Bhaskar).

shaapcd

141-0229/$ – see front matter © 2006 Elsevier Inc. All rights reserved.oi:10.1016/j.enzmictec.2006.10.019

emineralization; Carotenoid

rying stage prevents degradation of some proteins, allowing theull potential of protein recovery to be realized [5]. Conversionf shrimp waste into ensilage advantageously upgrades theiowaste with this approach being eco-friendly, safe, technolog-cally flexible and economically viable [1]. Ensilation of shrimpiowaste can be accomplished either by direct addition of min-ral or organic acids (acid silage) or biologically by fermentationith lactic acid bacteria (biological or fermented silage). Thesefulness of fermentation preservation is mainly due to its eco-riendly nature compared to the more ecologically aggressivend economically unviable preservation methods like acid/alkalireservation or drying [6]. Traditionally chitin preparation fromhrimp waste involves the use of alkalis (usually 4% sodiumydroxide) for deproteination and acids (e.g. 4% hydrochloriccid) for demineralization, making this process ecologically

ggressive and a source of pollution [7]. While the chemicalrocess isolates chitin efficiently, the protein and carotenoidomponents are rendered useless during protein removal andemineralization stages [7,8]. Fermentation of shrimp biowaste

1 robia

u[hrebpwuopa

r1iroviabismar

asa(cftfi(ctwf

2

2

ltacLmoawg

2

af1icTraM

wgflco(pmhtaDEd

2

cae

(dte2eEwidlirbf

2

w1eftowa

428 N. Bhaskar et al. / Enzyme and Mic

sing lactic acid bacteria for the preservation of shrimp waste8–11], production of chitin [5,7,12,13], carotenoids [14], etc.ave been reported. Also, fermentation of shrimp biowaste toecover chitin considerably replaces the expensive and non-nvironmentally friendly chemical process [7,8,15,16]. Thisiotechnological process results in a liquor fraction rich inroteins, minerals and carotenids (especially astaxanthin) alongith a solid fraction rich in chitin. The liquor fraction can betilized as a protein-mineral supplement for human consumptionr as an animal feed [7]. Apart from protein recovery for feedurposes, partial purification of chitin during fermentation islso feasible [16].

Researchers working on shrimp waste fermentation haveeported varying levels of fermentation conditions [1,5,7,8,2,13]. This includes sugar levels varying from 5 to 20%,noculum levels varying from 5 to 20% and incubation timeanging between 48 and 96 h. Most of the researchers workingn shrimp biowaste fermentation standardize conditions byarying one variable at a time [1]. Although a wealth ofnformation is available on biological ensilation using lacticcid bacteria (LAB) cultures, reports on optimized conditionsy considering the significant variables (viz., type of LAB,noculum level, fermentation time and sugar level) in total arecarce. Response surface methodology (RSM) is a statisticalethod that uses quantitative data to simultaneously determine

nd solve multivariable equations, graphically represented asesponse surfaces [17].

The first objective of this work was ascertain the best lacticcid bacteria (LAB) culture, which significantly acidified thehrimp biowaste on fermentation (as determined by pH reductionnd total titrable acidity or TTA), from among the five culturesviz., Lactobacillus plantarum, L. acidophilus, Lactococcusase, Pediococcus acidolactici and a combination of all theseour cultures). Secondly, the LAB that was efficient in acidifyinghe shrimp biowaste was further taken for optimizing theermentation conditions (i.e., incubation time, glucose level andnoculum level) by employing response surface methodologyRSM). Finally, the effect of lactic acid fermentation onarotenoid recovery, de-proteination and de-mineralization ofhe shrimp biowaste for chitin preparation also evaluated. Thisas achieved by assessing several batches of shrimp biowaste

ermented under optimized conditions.

. Materials and methods

.1. Materials

Shrimp waste (Penaeus monodon) was procured from processing plantsocated at Tadri and Mangalore (west coast of India). It was transported tohe lab under frozen condition and kept at −20 ◦C till further use. The lacticcid bacteria (LAB) cultures were obtained from the Institute (CFTRI, Mysore)ulture collection. The bacterial species evaluated included L. plantarum B4496,. acidophilus B4495, Lactococcus lactis B634, P. acidolactici CFR2182 and

ixture of all four cultures at 1:1:1:1 (v/v/v/v). All the cultures were maintained

n Mann–Rogosa–Sharpe (MRS) agar (Hi-Media, India) slants, stored at 4 ◦Cnd sub-cultured periodically [18]. In all the fermentation experiments d-glucoseas used as the sugar source. Chemicals used for the study were of analyticalrade, unless otherwise mentioned.

fwwmc

l Technology 40 (2007) 1427–1434

.2. Screening of lactic acid cultures for starter selection

The frozen shrimp waste was thawed overnight in a refrigerator (4 ◦C)nd minced in a Waring blender (Stephen mill, UM5 Universal, Hong Kong)or 10 min to homogenize the mass. The LAB cultures were grown in00 ml of MRS-Broth (Hi-Media, India) for 24 h at 37 ± 1 ◦C in a shakingncubator (Technico Ltd., India) set at 100 rpm. The cells were harvested byentrifuging (C31 Cooling centrifuge, Remi-India, India) at 3000 × g for 10 min.he harvested cells were washed twice with sterile physiological saline and

esuspended in physiological saline (100 ml). The biomass in the inoculum,fter serial dilution, was assayed by counting colony forming units (cfu) onRS agar (Hi-Media, India) plates.

For the initial screening experiments, shrimp biowaste (100 g) was mixedith distilled water (1:1, w/v) followed by 10% inoculum (v/w), 15% (w/w)lucose and 2% (w/w) salt in a 250 ml conical flask. The flasks were thenushed with nitrogen and sealed with a layer of parafilm (Hi-Media, India) toreate anaerobic conditions. The mixture was allowed to ferment for 72 h in anrbital shaking incubator at 100 rpm and 37 ± 1 ◦C. pH and total titrable acidityTTA) were recorded at 0, 24, 48 and 72 h. pH was determined using a digitalH meter (Cyberscan 1001, Eutech, Singapore). TTA was estimated as per theethod described in Sachindra et al. [14] by determining the ml of 0.1N sodium

ydroxide (NaOH) required for increasing the pH of one gram of fermented masso 8.0. Possible differences among treatments were determined by statisticallynalyzing the data obtained by applying analysis of variance (ANOVA) anduncan’s multiple range test using STATISTICA software (Version 5.5, 1999dition; Statsoft Inc., Tulsa, OK, USA), to decide on the best culture that gaveesirable pH of 4.3.

.3. Optimization of fermentation conditions using P. acidolactici

P. acidolactici CFR2182 was employed, for optimizing the fermentationonditions, based on the screening experiments. The inoculum was prepareds mentioned in the previous section. The inoculum used in the optimizationxperiments had a cell density of 8.28 log cfu ml−1.

The independent factors considered for optimization included inoculum levelX1; % v/w), glucose (X2; % w/w) and incubation time (X3; h). Box–Behnekenesign with 3 factors in one block encompassing 15 runs was employed forhe study. The design consisted of three factors (X1, X2 and X3) at threequidistant levels (5, 10 and 15% for X1; 7.5, 12.5 and 17.5% for X2; and4, 48 and 72 h for X3). In the optimization experiments, unlike screeningxperiments, only pH (Y) was recorded as the response (dependent) variable.ach run comprised 100 g shrimp biowaste mixed with distilled water (1:1,/v) and 2% (w/w) salt along with respective levels of glucose (w/w) and

noculum (v/w). The different ingredients were mixed according to the pre-etermined levels in 250 ml conical flasks, flushed with nitrogen, sealed with aayer of parafilm to create anaerobic condition and incubated at 37 ± 1 ◦C in anncubator for specified time. The designed model was further validated usingandom combinations of the independent variables. All the results were analysedy employing STATISTICA software, to determine the optimum conditions forermentation. pH was determined as described earlier.

.4. Effect of fermentation on recovery of chitin and carotenoids

Homogenized shrimp biowaste (100 g) was mixed with 1:1 (w/v) distilledater and fermented under optimized fermentation conditions (i.e. 5% inoculum,5% sugar, 72 h fermentation with 2% salt) in 250 ml conical flasks. Thexperiment was carried out in four different batches with each batch comprisingour numbers (one each for 0, 24, 48 and 72 h) of 250 ml conical flask containinghe fermentation mass. The inoculum used in these experiments had a cell densityf 8.36 log cfu ml−1. Samples (250 ml flask representing 100 g shrimp biowaste)ere drawn every 24 h up to 72 h (at which time the fermented samples recordedpH of 4.15 ± 0.15). Each sample was filtered using cheese cloth to collect the

ermentation liquor and the residue washed three times with distilled water (1:10,/v). The surface moisture on the washed residue was drained completely andeighed. Part of the wet residue was analysed for carotenoid and the remainingass dried overnight at 55 ± 1 ◦C. The dried sample was referred to as crude

hitin.

N. Bhaskar et al. / Enzyme and Microbial Technology 40 (2007) 1427–1434 1429

Table 1Inoculum level in the fermentation mixture along with changes in pH and total titratable acidity (TTA) of shrimp (Penaeus monodon) biowaste fermented withdifferent lactic acid bacteria (LAB)

Lactic acid bacteria LAB loada pH TTA

24 h 48 h 72 h 24 h 48 h 72 h

Lactobacillus plantarum 8.09 ± 0.01 a 4.91 ± 0.03 a 4.71 ± 0.05 a 4.41 ± 0.06 a 0.9118 ± 0.03 a 1.3045 ± 0.22 a 1.9313 ± 0.10 aL. lactis 8.31 ± 0.01 b 4.90 ± 0.04 a 4.69 ± 0.04 a 4.48 ± 0.05 a 0.9238 ± 0.10 a 1.1866 ± 0.02 a 1.9289 ± 0.10 aL. acidophilus 8.69 ± 0.01 c 4.89 ± 0.08 a 4.63 ± 0.06 a 4.45 ± 0.06 a 0.9202 ± 0.11 a 1.2706 ± 0.10 a 1.9979 ± 0.11 aPediococcus acidolactici 7.49 ± 0.02 d 4.87 ± 0.05 a 4.47 ± 0.03 b 4.34 ± 0.06 b 0.9231 ± 0.10 a 1.4035 ± 0.10 b 2.1263 ± 0.03 bMixed inoculumb 7.69 ± 0.01 e 4.88 ± 0.09 a 4.66 ± 0.04 a 4.37 ± 0.04 b 0.9343 ± 0.10 a 1.2527 ± 0.10 a 2.1005 ± 0.10 b

pH and TTA of the original shrimp biowaste were 8.10 ± 0.1 and 0.0000, respectively. Values with similar letters in the same column are not significantly different(

the coe shri

wpaKofdu

d

%

wwdPr

cctr

2

uO

3

3

otr8c

3

ci

baao(puaPcpaa

3P

had a LAB load of 3.19 ± 0.09 log cfu g .The influence ofinoculum level (X1; % v/w), glucose level (X2; % w/w) andincubation time (X3; h) on the fermentation of shrimp biowastewas determined using Box–Behneken design as mentioned in the

Table 2Response variable (pH; Y) observed during optimization experiments withvarying combinations of independent factors

Runnumber

Inoculum (X1;% v/w)

Glucose (X2;% w/w)

Time(X3; h)

pH (Y)

1 5.00 7.50 48.00 4.922 20.00 7.50 48.00 4.993 5.00 17.50 48.00 4.624 20.00 17.50 48.00 4.455 5.00 12.50 24.00 4.826 20.00 12.50 24.00 4.627 5.00 12.50 72.00 4.278 20.00 12.50 72.00 4.329 12.50 7.50 24.00 5.02

10 12.50 17.50 24.00 4.6811 12.50 7.50 72.00 5.08

P ≥ 0.05).a Total lactic acid bacterial load in the shrimp waste; cfu g−1 calculated fromb All the cultures were mixed at 1:1:1:1 (v/v) and then 10% (v/w) added to th

Moisture, protein and ash content in the fresh waste without fermentationere considered as the basis for computing de-mineralization (DM) and de-roteination (DP) efficiency. Moisture, fat and ash content were determineds per AOAC methods [19]. Crude protein in the fresh waste was estimated byjeldahl’s method as per AOAC [19] and chitin was estimated as per the methodf Spinelli et al. [20]. Protein content in the samples was estimated by a modifiedorm of Biuret method as described by Gornall et al. [21], after digesting 100 mgried material with 10 ml 0.5 M NaOH for 4 h [22] at 40 ◦C and the data wassed for computing DP efficiency.

DM and DP were expressed as percentages and computed by Eq. (1) asescribed in Rao et al. [7].

DP = [(PO × O) − (PR × R)] × 100

PO × O(1)

here PO and PR are protein concentrations (%) before and after fermentation;hile, O and R represent the mass (g) of original sample and fermented residue inry weight basis. DM (%) was calculated using the same equation but replacing

O and PR with AO and AR which represent ash in the original and fermentedesidue, respectively.

The fermentation liquor and the wet residue were analysed for their totalarotenoid content as outlined in Sachindra et al. [23]. The sum of totalarotenoids extracted in both fermentation liquor and residue was comparedo the total carotenoid content in the original shrimp biowaste to express theecovery percentage.

.5. Statistical analyses

All statistical analyses were carried out by employing the specified methodssing STATISTICA software (Version 5.5, 1999 Edition; Statsoft Inc., Tulsa,K, USA).

. Results

.1. General

The shrimp waste used in the study had a moisture contentf 81.4 ± 1.8%. The protein, fat, ash and chitin content ofhe material were 28.4 ± 2.1%, 17.9 ± 1.83% and 19.2 ± 1.7%,espectively, on dry weight basis. The material had a pH.2 ± 0.15, chitin content of 26.3 ± 1.1% (dry weight basis) andarotenoid content of 59.81 ± 0.8 �g g−1 of wet waste.

.2. Selection of starter culture

The original shrimp biowaste was found to harbor a LABount of 3.35 ± 0.5 log cfu g−1. The results of screening exper-ments and information on LAB load per gram of shrimp

1111

unts in the inoculum.mp biowaste.

iowaste after addition of inoculum and during fermentationre presented in Table 1. As can be seen from the table, P.cidolactici CFR2182 was found to be the best culture forbtaining a pH of around 4.30 and was significantly differentP ≤ 0.05) from all the other cultures in terms of cell masser gram of shrimp biowaste. Also, the pH obtained after 72 hsing P. acidolactici was significantly different (P ≤ 0.05) fromll other cultures except the mixed culture (P ≥ 0.05). Hence. acidolactici was chosen for standardization of fermentationonditions and further studies. No difference was observed inH reduction with or without agitation (data not shown); hence,ll subsequent experiments were done under static (withoutgitation) conditions.

.3. Optimization of fermentation conditions by RSM using

. acidolactici

The shrimp biowaste used in the optimization experiments−1

2 12.50 17.50 72.00 4.443 12.50 12.50 48.00 4.474 12.50 12.50 48.00 4.435 12.50 12.50 48.00 4.45

1430 N. Bhaskar et al. / Enzyme and Microbial Technology 40 (2007) 1427–1434

Table 3ANOVA table of pH as affected by inoculum level, sugar level and incubation time during fermentation with Pediococcus acidolactici

SS d.f. MS F P*

Independent variables1. Inoculum (L) 0.0078 1 0.007813 19.531250 0.047576

Inoculum (Q) 0.0000 1 0.000006 0.014423 0.915384

2. Glucose (L) 0.4141 1 0.414050 1035.125000 0.000965Glucose (Q) 0.3241 1 0.324052 810.129808 0.001232

3. Time (L) 0.1326 1 0.132613 331.531250 0.003003Time (Q) 0.0127 1 0.012744 31.860577 0.029982

Interactions1 × 2 0.0144 1 0.014400 36.000000 0.0266711 × 3 0.0156 1 0.015625 39.062500 0.0246572 × 3 0.0225 1 0.022500 56.250000 0.017317

Lack of fit 0.0589 3 0.019625 49.062500 0.020041Pure error 0.0008 2 0.000400

T

p(pfaabs

lpaicwio

Fi

otal SS 0.9982 14

* Level of significance.

revious section. The observed values for the response variablepH) at different combinations of the independent variables areresented in Table 2. As can be seen from the ANOVA tableor pH (Table 3), among the independent variables, glucosend fermentation time had higher significant effect (P < 0.01)s compared to the inoculum level (P < 0.05). The interactionsetween different factors also influenced the pH of fermentedample significantly (P ≤ 0.05).

The response surface graph for pH, as a function of glucoseevel and incubation time at an inoculum level of 12.5% (v/w), isresented in Fig. 1. The figure clearly indicates that pH decreasess glucose increases up to 15% (w/w) beyond which level itncreases. This could probably be due to the product inhibition

aused by excess amount of substrate. Similarly, pH decreasesith time to about 4.3 around 72 h before leveling off. The

nfluence of changes in incubation time and inoculum leveln pH at a glucose level of 12.5% (w/w) is shown in Fig. 2.

ig. 1. Response surface graph for pH as a function of level of glucose andncubation time (inoculum at center of the level—12.5%, v/w).

Fi

Tiiai

X14rdbdlpfo

ig. 2. Response surface graph for pH as a function of level of inoculum andncubation time (glucose at center of the level—12.5%, w/w).

he figure clearly shows that pH decreases with an increase innoculum level. However, the decrease is not considerable atnoculum levels above 5% (v/w). Similarly, effect of inoculumnd glucose levels on pH of fermented shrimp biowaste at anncubation time of 48 h is presented in Fig. 3.

To determine the optimum levels of variables (X1, X2 and3), the response variable (pH) was assigned a desirability of.0 for the lowest observed value for response variable (pH of.27) and a desirability of 0 for the highest observed value of theesponse variable (pH of 5.08), to get the overall desirability. Theesirability function to get optimum carotenoid yield was fittedy the least square method. The level of variable giving highestesirability (1.0) was selected as optimum level. The optimized

evels of variables (X1, X2 and X3) were determined using therofiles (Fig. 4) for predicted values for pH and desirabilityunction. The optimized factors for obtaining the lowest pHf 4.3 were 15% (w/w) of glucose, 5% inoculum (containing

N. Bhaskar et al. / Enzyme and Microbial Technology 40 (2007) 1427–1434 1431

Fg

8i

b(p

p

Table 4pH values observed during validation experiments along with correspondingpredicted values under different combinations of independent variables

Runnumber

Inoculum(X1; % v/w)

Glucose(X2; % w/w)

Time(X3; h)

pH (Y) pH-preda

1 7.00 8.00 24.00 4.94 5.042 6.00 7.00 48.00 5.05 5.023 10.00 15.00 69.00 4.28 4.314 15.00 15.00 72.00 4.29 4.305 13.00 12.00 50.00 4.39 4.466 17.50 10.00 24.00 4.64 4.747 5.00 7.50 24.00 4.99 5.128 20.00 17.50 72.00 4.25 4.349 12.50 12.50 48.00 4.43 4.45

10 8.00 5.00 48.00 5.42 5.4211 18.00 20.00 54.00 4.51 4.6412 14.00 15.00 44.00 4.44 4.4213 5.00 7.50 24.00 5.24 5.1214 5.00 7.50 72.00 4.82 4.8911

oft

t

ig. 3. Response surface graph for pH as a function of levels of inoculum andlucose (incubation time at center of the level—48 h).

.28 log cfu ml−1 of P. acidolactici, i.e. >108 cfu ml−1) andncubation time of 72 h.

The regression equation for pH of the fermented shrimpiowaste, as a function of the three independent variablesX1, X2 and X3) and their linear and quadratic interactions isresented in the following equation:

H = 6.994757 − 0.000278 × X1 − 0.000022 × X12

− 0.291750 × X2 + 0.011850 × X22

− 0.011684 × X3 + 0.000102 × X32

− 0.001600 × X1X2 + 0.000347 × X1X3

− 0.000625 × X2X3 (2)

(eao

Fig. 4. Desirability profiles for pH at optim

5 20.00 12.50 48.00 4.35 4.426 20.00 7.50 24.00 4.99 5.05

a pH predicted based on the model as per Eq. (2).

The predictability of the model was assessed by comparingbserved and predicted values of pH. The regression coefficientor pH observed in the experimental runs was 0.9364, indicatinghe usefulness of the model for prediction.

The model was validated using random combinations (otherhan used for optimization) of the independent variables

Table 4). The regression coefficient and slope of regressionquation for observed versus predicted values of pH was 0.9651nd 1.0091, respectively. These high values confirm the validityf the prediction model.

ized levels of independent factors.

1432 N. Bhaskar et al. / Enzyme and Microbia

Fa

3c

aiobTdbt

e(aodcpma

itoc2b

4

lf5afpiRb

TE

S

0247F

Vf

TE

S

0247F

V

ig. 5. Changes in pH and TTA of shrimp biowaste during fermentation with P.cidolactici.

.4. Effect of fermentation on recovery of chitin andarotenoids

The effect of optimized fermentation conditions on the pHnd TTA of the fermented mass is presented in Fig. 5. It clearlyndicates the significant (P ≤ 0.05) reduction in pH as a resultf acids produced by the lactic culture and the same is indicatedy the increasing TTA (≤0.05) values during fermentation.

he effect of fermentation, under optimized conditions, onemineralization (DM) and deproteination (DP) of shrimpiowaste are presented in Table 5. Fermentation resulted in morehan 72% demineralization and 97.9% of deproteination at the

cice

able 5ffect of fermentation on the de-proteinization (DP) and de-mineralization (DM) of s

ample Weight of residuea (g) Proteinb (g)

h 12.73 ± 0.09 a 0.81 ± 0.01 a4 h 10.68 ± 0.17 b 0.67 ± 0.04 b8 h 08.16 ± 0.04 c 0.20 ± 0.01 c2 h 08.25 ± 0.08 c 0.02 ± 0.00 dresh 18.60 ± 0.12 d 1.11 ± 0.09 e

alues with similar letters in the same column are not significantly different (P ≥ 0.05resh (dried) material.

a All values on dry weight basis.b Mass of the content.

able 6ffect of fermentation on the carotenoid content (�g) of shrimp biowaste and its reco

ample Residue Filtrate

Weight (g) Total carotenoid (�g) (A) Volume (ml)

h 36.31 ± 1.1 a 3052.8 ± 162 106 ± 34 h 41.32 ± 1.9 b 1628.5 ± 45 100 ± 28 h 30.40 ± 0.8 c 1779.7 ± 34 120 ± 32 h 30.41 ± 0.6 c 1767.2 ± 43 125 ± 5resh 5891 ± 77 d �g in 100 g wet waste

alues with similar letters in the same column are not significantly different (P ≥ 0.0

l Technology 40 (2007) 1427–1434

nd of 72 h of fermentation. DM was significantly differentP ≤ 0.05) during the course of fermentation except that DMt 48 h was not significantly different (P ≥ 0.05) from thatbserved at the end of 72 h. Similarly DP was significantlyifferent (P ≤ 0.05) during the course of fermentation. Thislearly indicates the usefulness of fermentation as a tool toroduce chitin biologically instead of the conventional chemicalethods that employ strong mineral acids for demineralization

nd strong alkali for deproteination.The recovery of carotenoids during the course of fermentation

s presented in Table 6. As can be seen from the table,he recovery of carotenoids was more than 72% at the endf fermentation period of 72 h. It can be noticed that totalarotenoids recovered was significantly different between 0,4 and 48 h (P ≤ 0.05); while, it was not significantly differentetween 48 and 72 h of fermentation (P ≥ 0.05).

. Discussion

The effect of different independent variables (i.e., glucoseevel, fermentation time and inoculum level on the pH of theermented shrimp biowaste was investigated. From the results,% of inoculum (containing P. acidolactici at >108 cfu ml−1)long with 15% glucose and a fermentation time of 72 h wereound to be the optimum fermentation conditions to obtain aH reduction from the initial 8.2 ± 0.15 to 4.3 ± 0.05. Similarnoculum level in case of Lactobacillus 541 has been reported byao et al. [7]. However, they adjusted the initial pH of the shrimpiowaste using acetic acid before fermentation with the lactic

ulture. In the present study the reduction in pH was substantialn the first 24 h itself. Low pH as observed in this study is inonformity with many other such studies [1,7,8] although thextent of reduction was different.

hrimp biowaste (original weight of wet shrimp biowaste—100 g) (n = 4)

Ashb (g) DP (%) DM (%)

2.42 ± 0.09 a 26.8 ± 6.6 a 32.3 ± 2.5 a1.93 ± 0.09 b 39.7 ± 4.7 b 45.9 ± 0.9 b0.97 ± 0.09 c 81.5 ± 1.9 c 71.8 ± 1.4 c1.02 ± 0.09 c 97.9 ± 0.3 d 72.5 ± 1.5 c3.57 ± 0.09 d – –

). DP and DM are expressed in relation to mass of the respective content in the

very (original weight of wet shrimp biowaste—100 g) (n = 4)

Total carotenoidcontent (�g) (A + B)

Carotenoid(% of original)

Total carotenoid (�g) (B)

1571.5 ± 45 4264.3 ± 198 a 78.5 ± 2.32748.6 ± 27 4377.1 ± 56 b 74.3 ± 1.12364.1 ± 40 4143.8 ± 61 c 70.3 ± 1.82499.9 ± 117 4267.1 ± 160 c 72.4 ± 2.7

5). Carotenoid content in original taken as 100%.

robial

la[civupcd5tArrat(i

bembloatbaipgpca

crdcrwcpatoattitobb

5

idtecotuso

A

app

R

[

[

[

[

N. Bhaskar et al. / Enzyme and Mic

Deproteination of the shrimp biowaste occurs mainly due toiquefaction of proteins due to the proteolytic activity of lacticcid bacteria culture or by the in situ proteases in the biowaste1]. However, biological ensilation yields solubilization valueslose to 70% [24]. The observed protein solubilization asndicated by the deproteination is within the range of reportedalues and points to the fact that P. acidolactici can besed efficiently in deproteinizing the shrimp waste. Shrimprocessing waste upon fermentation releases nutrients likearotenoids, n − 3 unsaturated fatty acids that are otherwiseifficult to extract [25]. Our study showed that more than0% of the total (recovered) carotenoids were released intohe liquor fraction following fermentation with P. acidolactici.lso, the observed reduction in crude fat content in the

esidue indicates that fermentation ensilation can effectivelyelease this component into the liquor. This could be mainlyttributed to the proteolysis wherein the proteins are hydrolysedhereby releasing the lipid fractions from the lipoproteinmainly carotenoproteins) complexes. This also explains for thencreased carotenoid content in liquor portion.

Higher deproteination (>97%) indicates the usefulness of thisacteria in deproteinating shrimp biowaste through biologicalnsilation. Further it also confirms the fact that proteolysis isore efficient at higher pH [7,26] as the pH was not adjusted

efore the start of fermentation. In addition, the presence ofactic acid bacteria in the protein rich liquor portion is indicativef its use as human or animal food [7]. Further, the strain P.cidolactici CFR2182 showed higher fermentation efficieny inerms of demineralization (72.5 ± 1.5%) also. P. acidolacticieing a homofermentative bacterium [27] produces only lacticcid and this may responsible for the demineralization noticedn the fermented mass. Rao and Stevens [13] working with L.lantarum achieved a demineralization of 81% by the addition oflacial acetic acid before fermentation. They opine that loweringH before fermentation is necessary for bringing down thehances of spoilage. However, in the present study even withoutdjustment of the initial pH there was no spoilage noticed.

Generally, proteolysis results in better recovery ofarotenoids from shrimp waste, especially from the solidesidue. It can be observed that the carotenoids in the residueecreased and the filtrate showed an increase in its carotenoidontent during fermentation. The lower carotenoid content in theesidue could also be due to the fact that the residue undergoesashing before estimating the carotenoid content that allows the

arteonoids to be washed away easily due to the prevailing lowH (4.3 or less) and due to proteolysis by the LAB culture. Thisccounts for the loss of carotenoids as compared to the content inhe original shrimp biowaste. Further release of more than 58%f the available carotenoids in the fermented mass (obviouslys carotenoproteins) into the fermentation liquor, at the end ofhe fermentation period adds to its nutritional quality. Also,he fermentation liquor did not have any perceivable changesn the odor in spite of hydrolysis of proteins in the liquor by

he lactic culture. The absence of any perceivable changes indor due to proteolysis could be explained by the fact thatiological ensilation results in much lower formation of volatilease nitrogen [2].

[

Technology 40 (2007) 1427–1434 1433

. Conclusion

The strain P. acidolactici CFR2182 was found to be efficientn fermenting the shrimp biowaste apart from resulting in highereproteination and considerably high demineralization. Fermen-ation also allowed recovery of carotenoids to a considerablextent (>75%). The protein rich fermentation liquor containingarotenoids could be effectively utilized by concentrating itr co-drying with other feed ingredients. Further studies onhe physico-cheimical quality characteristics of chitin preparedsing biological ensilation methods is needed to ascertain itsuperiority over chitin prepared by conventional chemical meth-ds.

cknowledgements

The authors wish to thank the Ministry of Environmentnd Forests (Govt. of India) for the partial funding of theroject. Thanks to Director, CFTRI for the encouragement andermission to publish this work.

eferences

[1] Mathew P, Nair KGR. Ensilation of shrimp waste by Lactobacillusfermentum. Fish Technol 2006;43:59–64.

[2] Dapkevicius MDLE, Batista I, Nout MJR, Rombouts FM, Houben JH.Lipid and protein changes during ensilage of blue whiting (Microme-sistius poutassou Risso) by acid and biological methods. Food Chem1998;63:97–102.

[3] Rao MS, Stevens WF. Quality parameters of chitosan derived fromfermentation of shrimp biomaterial using a drum reactor. J Chem TechnolBiotechnol 2005;80:1080–7.

[4] Hall GM, Reid CI, Zakaria Z. Fermentation of prawn shell waste by lacticacid bacteria. In: Karnicki Z, editor. Chitin world. NW: Wirtschaftverlag;1994. p. 83–98.

[5] Healy M, Green M, Healy A. Bioprocessing of marine crustacean shellwaste. Acta Biotechnol 2003;23:151–60.

[6] Hall GM, Zakaria Z, Reid CI. Recovery of value added products fromshellfish waste by a lactic acid fermentation process. In: Luton JB,Bonneson T, editors. Seafood from producers to consumers, integratedapproach to quality. London: Elsevier Applied Science; 1997. p. 97–102.

[7] Rao MS, Munoz J, Stevens WF. Critical factors in chitin produc-tion by fermentation of shrimp biowaste. Appl Microbiol Biotechnol2000;54:808–13.

[8] Shirai K, Guerrero I, Huerta S, Saucedo G, Castillo A, Gonzalez RO,et al. Effect of initial glucose concentration and inoculation level oflactic acid bacteria in shrimp waste ensilation. Enz Microbial Technol2001;28:446–52.

[9] Fagberno OA. Preparation, properties and preservation of lactic acidfermented shrimp heads. Food Res Int 1996;29:595–9.

10] Fagbenro OA, Bello-Olusoji OA. Preparation, nutrient composition anddigestability of fermented shrimp head silage. Food Chem 1997;60:489–93.

11] Cira LA, Huerta S, Hall GM, Shirai K. Pilot scale lactic acid fermentationof shrimp wastes for chitin recovery. Process Biochem 2002;37:1359–66.

12] Bautista J, Jover M, Gutierrez JF, Corpas R, Cremades O, Fontiveros E, etal. Preparation of crawfish chitin by in situ lactic acid production. ProcessBiochem 2001;37:229–34.

13] Rao MS, Stevens WF. Fermentation of shrimp biowaste under different salt

concentrations with amylolytic and non-amylolytic Lactobacillus strainsfor chitin production. Food Technol Biotechnol 2006;44:83–7.

14] Sachindra NM, Bhaskar N, Siddegowda GS, Sathisha AD, Suresh PV.Recovery of carotenoids from ensilaged shrimp waste. Biores Technol2007;98:1642–6.

1 robia

[

[

[

[

[

[

[

[

[

[

[

[

1997. p. 56–63.

434 N. Bhaskar et al. / Enzyme and Mic

15] Hall GM, DeSilva S. Lactic acid fermentation of scampi (Penaeusmonodon) waste for chitn recovery. In: Brine CJ, Sandford PA, ZikakisJP, editors. Advances in chitin and chitosan. London: Elsevier AppliedScience; 1992. p. 633–68.

16] Rao MS, Stevens WF. Processing parameters in scale up of Lactobacillusfermentation of shrimp biowaste. In: Domard A, Roberts GAF, Varum K,editors. Advances in chitin sciences, vol. 2. Lyon, France: Jacques Andre;1997. p. 88–93.

17] Giovanni M. Response surface methodology and product optimization.Food Technol (November) 1983:41–5.

18] DeMan JC, Rogosa M, Sharpe ME. A medium for the cultivation oflactobacilli. J Appl Bacteriol 1960;23:130–5.

19] AOAC. Official Methods of Analysis. In: Helirich K, editor. Associationof official analytical chemists. 16th ed. Arlington, VA: AOAC; 1995.

20] Spinelli J, Lehman L, Wieg D. Composition, processing and utilization ofred crab (Pleroncodes planipes) as an aquaculture feed ingredient. J FishRes Bd Can 1974;31:1025–30.

21] Gornall AG, Bardawill CJ, David MM. Determination of serum proteinsby means of the Biuret reaction. J Biol Chem 1949;177:751–66.

[

l Technology 40 (2007) 1427–1434

22] Simpson BK, Haard NF. The use of proteolytic enzymes to extractcarotenoproteins from shrimp wastes. J Appl Biochem 1985;7:212–22.

23] Sachindra NM, Bhaskar N, Mahendrakar NS. Carotenoids in different bodycomponents of Indian shrimps. J Sci Food Agric 2005;85:167–72.

24] Lindgren SE, Pleje M. Silage fermentation of fish or fish waste productwith lactic acid bacteria. J Sci Food Agric 1983;34:1057–67.

25] Guillou A, Khalil M, Adambounou L. Effects of silage preservation onastaxanthin forms and fatty acid profiles of processed shrimp (Pandalusborealis) waste. Aquaculture 1995;130:351–60.

26] Shirai K, Legarreta GI, Rodrigues-Serrano S, Huerta-Ochoa S, Saucedo-Castaneda G, Hall GM. Aspects of protein breakdown during the lacticacid fermentation of prawn waste. In: Domard A, Roberts GAF, Varum K,editors. Advances in chitin sciences, vol. 2. Lyon, France: Jacques Andre;

27] Simpson WJ, Taguchi H. The genus Pediococcus, with a note on the generaTetragenococcus and Aerococcus. In: Wood BJB, Holzapfel WH, editors.The genera of lactic acid bacteria. London: Chapman & Hall; 1995. p.125–72.