Optimization of ultrasonic-assisted extraction of flavor compounds from shrimp by-products and characterization of flavor profile of the extract

Highlights

• •White shrimp heads (WSHs) are utilized for the extraction of flavor compounds.
• •Ultrasonic-assisted extraction increased the content of umami compounds.
• •Ultrasonic extract of WSHs is used to develop shrimp flavor concentrate (SFC).

Abstract

This study aimed to optimize the ultrasonic-assisted extraction conditions of flavor compounds from white shrimp heads (WSHs). The effects of sonication amplitude, sonication time, and solvent-to-solid ratio on the extraction yield (EY) of flavor compounds and the degree of protein modification (DPM) were evaluated by Box-Behnken design and response surface methodology (RSM). The optimum EY (40.87 %) and DPM (26.28 %) were obtained at amplitude, time, and solvent-to-solid ratios of 63.2 %, 20.5 min, and 20.8 mL/g, respectively. The optimum DPM value indicates that sonication markedly influenced the protein denaturation, as evidenced by the higher TCA soluble protein content. Further, we also investigated the taste active composition of ultrasonic extract of shrimp head (USH). Results show that the equivalent umami concentration (EUC) value was significantly (p < 0.05) increased in the ultrasonic extract of shrimp head (USH) (55.44 ± 3.25 g MSG/100 g) compared to the control shrimp head extract (CSH) (8.32 ± 1.07 g MSG/100 g). This study also deals with the development of shrimp flavor concentrate (SFC) using USH by the conventional heating process. Sensory evaluation of SFC revealed that the shrimp-like aroma and umami taste characteristics were predominant. Thus, USH's improved umami taste composition demonstrates its potential utilization for producing SFC with higher EUC.

1. Introduction

Crustacean aquaculture is the largest seafood sector in terms of production all over the world, as they are a rich source of proteins. Shrimps among crustaceans are famous all over the world because of their delicacy and rich nutritional attributes [1]. Among the shrimp species, white shrimp, scientifically known as Litopenaeus vannamei, is the most widely cultivated species worldwide due to its nutritional profile, delicious taste, high market value, and disease resistance [2] and is India's largest shrimp species in terms of its production (815745 tons in 2021) [3]. Generally, shrimps are frozen to store and transport with or without a shell. Hence, the processing of shrimp generates a large amount of solid waste in the form of head, shell, and viscera, constituting approximately 48–56 % of the total weight of shrimps [4]. These by-products cause environmental pollution and disposal problems. Although little quantity of shrimp waste is used as animal feed, large amounts of these by-products are being wasted despite being a valuable source of protein, lipids, minerals, chitin, carotenoids, vitamins, and enzymes.

Some studies reported that the amino acid composition of Penaeus vannamei head contributes to a rich taste and nutritional value [5]. Thus, utilizing white shrimp heads (WSHs) is a promising approach for recovering flavor compounds. Several methods are used to extract proteins and other flavor compounds from shrimp heads, such as the pH-shift method, solvent extraction method, heat treatment, and enzyme/acid hydrolysis. However, the amino acid profile of extracted proteins and the content of flavor compounds varies with the extraction technique applied, and it should be cost-effective and environmentally safe. Hence, there is a need for greener technologies to efficiently extract flavor compounds and proteins from shrimp heads. Ultrasonication is an emerging, non-thermal, and environmentally friendly technology with various advantages, such as energy efficiency, economic feasibility, higher mass transfer rate and maintaining the natural state of foods [6]. During the ultrasonic process, the production of bubbles and cavitation effect leads to the shearing and heating of the material, which results in a high mass transfer rate, and improved diffusion [7], [8]. Moreover, ultrasonication can promote several chemical reactions, such as oxidation, Maillard reaction, proteolysis, and esterification, thus enhancing the aroma profile of food material [9]. It is also reported that the denaturation of proteins is necessary to improve food flavor [10], and ultrasonication could denature the protein by modifying its conformational properties.

In this context, some recent studies have shown that ultrasonication prior to the thermal Maillard reaction has increased the content of volatile and colored compounds [11], [12], [13]. There is an increase of 55 % in the protein yield from ultrasound-assisted alkaline extraction of the chicken liver compared to conventional alkaline extraction [14]. However, there is a lack of potential utilization of WSHs as a source of flavor compounds, and the green extraction techniques of flavor compounds from WSHs need to be explored. Therefore, the present study aimed to (i) optimize the ultrasonic-assisted extraction (UAE) conditions of flavor compounds from WSHs, (ii) characterize the taste-active profile of ultrasonic extract of shrimp head (USH) and iii) the development of shrimp flavor concentrate (SFC) from USH. This study also determined the characteristic flavor and the sensory profiles of SFC.

2. Materials and methods

2.1. Materials

Fresh WSHs were procured from the local seafood market in Mysuru, India and transported in chilled condition to the laboratory. WSHs were washed thoroughly with distilled water and ground to make a uniform mix.

Divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fiber (1 cm length, 50/30 μm thickness), n-alkanes standard (C₇-C₂₀), 1,2-dichlorobenzene, standards of 5′-nucleotides and amino acids, and D-Glucose were purchased from Sigma-Aldrich (St. Louis, MO, USA). Organic acid standards and o-phthaldialdehyde (OPA) were purchased from SRL Pvt. Ltd. (Mumbai, India).

2.2. Experimental design

To optimize the parameters such as the amplitude of sonication (X₁, %), time of sonication (X₂, min), and solvent-to-solid ratio (X₃, mL/g) for maximizing the ultrasonic-assisted extraction of flavor compounds from WSHs, the Design Expert software (trial version 12.0.12, Stat-Ease Inc, Minneapolis, MN, USA) was used to create Box-Behnken design. Extraction yield (EY) and degree of protein modification (DPM) were determined as the response variables. The design consisted of a total of 15 runs, including three replicates at the central point and was conducted in a random order to reduce unpredicted variations. The data was investigated using a quadratic polynomial regression model, as shown in the following equation:

$$Y={\beta }_0+\sum {\beta }_i{x}_i+\sum {\beta }_{\mathit{ii}}{x_i}^2+\sum {\beta }_{\mathit{ij}}{x}_i{x}_j$$ (1)

where Y is the response variable, x_i and x_j are the factors, β₀ is constant, β_i, β_ii, and β_ij are the linear, quadratic and interactive coefficient terms, respectively.

2.3. Ultrasonic-assisted extraction of flavor compounds from WSHs

Minced WSHs were suspended in distilled water at a solvent-to-solid ratio as specified for each trial in the experimental design shown in Table 1. Then, ultrasonication was carried out at 20 kHz using an ultrasound system (Sonics VCX-750, Vibra-Cell Ultrasonic, Newtown, USA) according to the conditions as shown in the Box-Behnken design. The slurry in a conical flask was kept in a beaker containing ice to minimize the heating during sonication. Afterwards, the slurry was subjected to centrifugation (8000xg, 15 min, 20 °C), and the supernatant was collected and lyophilized using a freeze dryer and referred to as ultrasonic extract of shrimp head (USH).

Table 1.

Box-Behnken experimental design with experimental and predicted responses for EY (%) and DPM (%).

Independent Variables				Response Variables
Run order	X1	X2	X3	Y1			Y2
				Experiment	Predict	RD (%)	Experiment	Predict	RD (%)
16	20	10	20	29.37	29.29	0.27	8.97	8.18	8.81
11	100	10	20	30.57	30.37	0.65	10.99	10.64	3.18
14	20	30	20	30.66	30.86	−0.65	9.25	9.60	−3.78
17	100	30	20	30.94	31.01	−0.23	12.37	13.16	−6.39
7	20	20	10	22.76	22.34	1.85	8.15	8.32	−2.09
8	100	20	10	26.00	25.70	1.15	18.49	18.22	1.46
13	20	20	30	34.01	34.31	−0.88	15.00	15.27	−1.80
4	100	20	30	31.76	32.18	−1.32	11.55	11.38	1.47
9	60	10	10	23.01	23.51	−2.17	6.67	7.29	−9.30
2	60	30	10	26.22	26.44	−0.84	9.72	9.20	5.35
5	60	10	30	34.79	34.57	0.63	6.77	7.29	−7.68
12	60	30	30	34.33	33.83	1.46	9.92	9.30	6.25
15	60	20	20	40.70	40.52	0.44	25.85	26.01	−0.62
1	60	20	20	40.39	40.52	−0.32	27.32	26.01	4.80
3	60	20	20	40.11	40.52	−1.02	25.57	26.01	−1.72
10	60	20	20	40.54	40.52	0.05	26.47	26.01	1.74
6	60	20	20	40.87	40.52	0.86	24.83	26.01	−4.75

*Values are expressed as mean (n = 3).

X₁ = Amplitude (%); X₂ = Sonication time (min); X₃ = Solvent-to-solid ratio (mL/g);

Y₁ = EY (%); Y₂ = DPM (%);

RD (relative deviation) = [(Experimental value – predicted value)/ experimental value] × 100.

2.4. Determination of extraction yield (EY) and degree of protein modification (DPM)

The soluble protein content and trichloroacetic acid (TCA) (20 %) soluble protein content of USH was determined by Lowry’s method with bovine serum albumin (BSA) as standard using a UV–VIS spectrophotometer (UV-1800, Shimadzu Corporation, Kyoto, Japan).

The extraction yield of flavor compounds was determined according to Siewe, Kudre, & Narayan [15] using the following equation:

$$Extractionyield\left(EY\%\right)=\frac{\mathit{SolubleproteincontentofUSH}(g)}{\mathit{TotalproteincontentofWSH}(g)}$$ (2)

The degree of protein modification was determined using the following equation:

$$\mathit{Degreeofproteinmodification}(DPM\%)=\frac{\mathit{TCAsolubleproteincontentofUSH}(g)}{\mathit{TotalproteincontentofWSH}(g)}$$ (3)

2.5. Determination of non-volatile taste compounds of USH

Total amino acids (TAAs) of WSHs were extracted according to the method of Siewe, Kudre, & Narayan [15]. The extraction of free amino acids (FAAs) was performed as per the method described in our earlier work [16]. 5′-Nucleotides of USH were extracted according to the method described in our previous work [16]. Organic acids were extracted as per the method for the extraction of 5′-nucleotides. The diode array detector (DAD) at a wavelength of 334 nm, 254 nm, and 215 nm was used for the identification and quantification of amino acids, 5′-nucleotides and organic acids, respectively, by comparing the retention time and peak area of each respective standard.

The content of inorganic ions (Na⁺, K⁺, Mg²⁺, and Ca²⁺) and PO₄^3- in USH and CSH was determined by Microwave Plasma – Atomic Emission Spectrometer (Model no. 4200 MP-AES, Agilent Technologies, Inc., Santa Clara, California, USA) according to the method of Duppeti et al. [16].

2.6. Taste-activity value (TAV)

The TAV is determined by the following equation:

$$\mathit{TAV}=\frac{C_1}{C_2}$$ (4)

C₁ is the concentration of the taste compound (x), and C₂ is the taste threshold concentration of compound (x) obtained from the literature. Compounds with TAV of greater than one are regarded as active taste contributors [17].

2.7. Equivalent umami concentration (EUC)

EUC represents the synergistic relationship between flavor nucleotides and umami amino acids. It is determined using the equation shown below [18]:

$$\mathit{EUC}=\sum a_i{b}_i+1218\left(\sum a_i{b}_i\right)\left(\sum a_j{b}_j\right)$$ (5)

where a_i and b_i denote the concentration and relative umami concentration (RUC) of each umami amino acid (aspartic acid, 0.077 and glutamic acid, 1) respectively, a_j and b_j represent the concentration and relative umami concentration (RUC) of each flavor nucleotide (Adenosine-5′-monophosphate (AMP), 0.18; Guanosine-5′-monophosphate (GMP), 2.3, and Inosine-5′-monophosphate (IMP), 1) respectively. 1218 is a synergistic constant.

2.8. Preparation of shrimp flavor concentrate (SFC)

A mixture of USH (0.5 g), L-cysteine (0.03 g), D-xylose (0.02 g), and 0.01 g glycine was added to 10 mL of ultrapure water and homogenized using a homogenizer (Pro 250, PRO Scientific Inc., Oxford, USA) at 8000 rpm for 1 min. The initial pH of the suspension was measured to be 6.7, and it was transferred into a 20 mL Pyrex screw-sealed vial. After sealing the vial tightly, samples were heated in a thermostatic oil bath at 110 °C for 30, 60, 90, and 120 min and referred to as S30, S60, S90, and S120, respectively. Then, the suspension was cooled immediately to stop the reaction further. Thus, the obtained product was referred to as SFC and stored at − 80 °C until further use.

2.9. Characterization of SFC

2.9.1. Measurement of pH and browning intensity

The pH value of SFC samples was measured at room temperature at a heating time interval of 0, 30, 60, 90, and 120 min using a pH meter (Eutech pH2700, Thermo Fischer Scientific Inc., Massachusetts, USA). Browning intensities of SFC samples were measured at 294 nm and 420 nm using a UV–Vis spectrophotometer (UV-1800, Shimadzu Corporation, Kyoto, Japan).

2.9.2. Analysis of proximate composition and aroma profile of SFC

The proximate composition of SFC was analyzed as per the method described by AOAC [19]. The extraction of volatile compounds from SFC was performed by SPME according to the method reported in our previous work [20]. Briefly, 5 µL of internal standard (1,2-dichlorobenzene, 0.3246 µg/mL in methanol) was added to 5 g of SFC in a 20 mL headspace vial, and the vial was sealed with a PTFE septum. After, the pre-conditioned (270 °C, 30 min) SPME fiber was inserted into the headspace vial and kept in a water bath at 50 °C for 20 min. Then, the fiber was desorbed into the GC injection port for 5 min at 250 °C. The volatile compounds were analyzed by GC (7890B Agilent, USA) in split-less mode, and separation of volatile compounds was carried out on BP-20 (30 m × 250 µm × 0.25 µm; SGE, Australia) wax column and helium as the carrier gas with a flow rate of 1 mL/min. The GC oven program was as follows: Initially, 40 °C for 5 min, then ramp to 115 °C at the rate of 10 °C/ min and then rise to 230 °C at a rate of 30 °C/min for 12 min. The MS was programmed at an ion source temperature of 230 °C, 150 °C of a quadruple temperature, 70 eV of an ionization voltage, and a mass range from 30 to 450 m/z.

Kovat’s retention index (RI) values and a NIST database (version 2.0) were used for the identification of volatile compounds. The RI values were calculated based on the following equation [21]:

$$\mathit{RI}=\left(\frac{Rt\left(x\right)-Rt(n)}{\mathit{Rt}(n+1)-Rt(n)}+n\right)\times 100$$ (6)

where Rt(x) Rt(n), and Rt(n + 1) are the retention times of volatile compound (x), n-alkane eluting before and after the compound (x), respectively, run under similar conditions.

The concentration of volatile compounds was determined by the formula given below:

$$\mathit{Concentration}(\mu g{g}^{-1})=f^{\prime }\times \frac{\mathit{Peakareaofthevolatilecompound}}{\mathit{Peakareaoftheinternalstandard}}\times \frac{\mathit{massofinternalstandard}}{\mathit{massofsample}}$$ (7)

f ′ is a relative correction factor assumed to be 1.

2.9.3. Odor activity value (OAV)

Odor activity value (OAV) was calculated by the following formula:

$$\mathit{OAV}=\frac{C_x}{{\mathit{OT}}_x}$$ (8)

where C_x is the concentration of volatile compound (x), and OT_x is the odor threshold of the compound (x) from the previous literature.

2.9.4. Sensory evaluation of SFC

The sensory characteristics of SFC were evaluated according to the method of Erickson et al. [22] by an experienced sensory panel consisting of three males and seven females (24–32 years old) from the CSIR- Central Food Technological Research Institute, Mysuru, India. Nine sensory attributes (fishy, old shrimp, cooked shrimp, shrimp-like, sour, umami, salty, bitterness, and aftertaste) were selected to evaluate the SFC. The reference taste solutions were prepared as follows: umami, 0.05 % (w/v) monosodium glutamate in ultrapure water; sour, 0.05 % citric acid in water; bitter, 0.05 % caffeine in water; salt, 0.2 % NaCl in water; aftertaste, iodized salt water (0.2 %); shrimp-like, 250 g of L. vannamei meat is boiled in 250 mL ultrapure water for 1 h and filtered. The sensory evaluation took place in a sensory laboratory at room temperature by providing 5 mL of sample (10 % w/v, in ultrapure water) and 5 mL of reference taste solutions to the panelists. The evaluation was implemented using a 0–9 interval scale (0 = none, 9 = extremely strong).

2.10. Statistical analysis

All experiments were carried out in triplicate, and the results were expressed as mean ± standard deviation (SD). The comparison of the means of concentration of taste compounds of USH and CSH was performed using an unpaired t-test using the statistical package for the social sciences (SPSS) software (version 20.0, SPSS Inc., Chicago, IL). Values were significantly different at p ≤ 0.05.

3. Results and discussion

3.1. Model fitting and optimization of EY and DPM

3.1.1. Statistical analysis and model fitting

Response surface methodology (RSM) was used to optimize the ultrasonic-assisted extraction of flavor compounds from WSHs. The effects of amplitude (X₁, %), sonication time (X₂, min), and solvent-to-solid ratio (X₃, mL/g) on the EY (Y₁, %) and DPM (Y₂, %) were observed using a Box-Behnken design. The results of 17 experimental runs and the corresponding responses were shown in Table 1, while the ANOVA results were presented in Table 2. The validation of model coefficients was based on the F-value and p-value, as a low p-value (<0.05) and high F-value indicate a statistical significance of the model terms [15], [23]. As shown in Table 2, both the response models were highly significant (p < 0.0001). The linear (X₂, X₃), the interaction (X₁X₃, X₂X₃), and the quadratic (X₁₂, X₂₂, X₃₂) terms were significant (p ≤ 0.05) for EY, whereas, for DPM, the linear (X₁, X₂), the interaction (X₁X₃) and quadratic (X₁₂, X₂₂, X₃₂) model terms were significant (p ≤ 0.05). It is also revealed that there is a strong interaction between X₂ and X₃ and X₁ and X₃ for EY and DPM, respectively. Fitting of the response surface model for all the responses and independent variables was carried out by the regression analysis. The obtained empirical relationship between the independent and dependent variables as shown:

Y₁=40.52+0.31X₁+0.55X₂+4.61X₃–0.23X₁X₂–1.37X₁X₃–0.92X₂X₃–5.55X₁²–4.59X₂²–6.34X₃² (9)

Y₂=26.08+1.50X₁+0.982X₂+0.026X₃+0.275X₁X₂–3.45X₁X₃+0.025X₂X₃–5.33X₁²–10.35X₂²–7.45X₃² (10)

Table 2.

ANOVA of the regression model for prediction of EY and DPM.

Source	EY				DPM
	SS	df	F-value	p-value	SS	df	F-value	p-value
Model	616.72	9	308.35	< 0.0001	964.03	9	129.08	< 0.0001
X1 (L)	0.7626	1	3.43	0.1064	18.09	1	21.80	0.0023
X1 (Q)	129.51	1	582.77	< 0.0001	119.47	1	143.97	< 0.0001
X2 (L)	2.43	1	10.94	0.0130	7.72	1	9.31	0.0186
X2 (Q)	88.75	1	399.35	< 0.0001	451.41	1	543.96	< 0.0001
X3 (L)	170.20	1	765.88	< 0.0001	0.0055	1	0.0066	0.9373
X3 (Q)	169.43	1	762.42	< 0.0001	233.80	1	281.74	< 0.0001
X1.X2	0.2116	1	0.9522	0.3617	0.3025	1	0.3645	0.5651
X1.X3	7.54	1	33.91	0.0006	47.54	1	57.29	0.0001
X2.X3	3.37	1	15.15	0.0060	0.0025	1	0.0030	0.9578
Lack of fit	1.22	3	4.76	0.0829	2.98	3	1.40	0.3646
Pure Error	0.3403	4			2.83	4
Residual	1.56	7			5.81	7
Corr. Total	618.27	16			969.84	16
R2	0.9975				0.9940
R2Adj	0.9942				0.9863
R2Pred	0.9677				0.9463
CV %	1.44				6.00

L = linear terms; Q = quadratic terms; X₁ = Amplitude (%); X₂ = Time (min); X₃ = Solvent-to-solid ratio (mL/g).

In the regression equation, a positive factor value indicates that the response increases gradually with increasing factor value, whereas a negative value indicates an antagonistic effect. The lack of fit was insignificant (p > 0.05), as indicated by a low F-value and a high p-value for both models. This indicates that the inclusion of statistically parametric values does not improve the model, and the fitted model is good for predicting the responses within the design space. Further validation of the fitness model was determined by the coefficient of determination (R²) and adjusted coefficient of determination (R²_Adj). The high R² of 0.9975 and 0.9940 indicates that the factor terms explain 99.75 % and 99.40 % of the variation in the response model for EY and DPM, respectively. Besides, the R²_Adj values of EY (0.9942) and DPM (0.9863) were found to be close to that of their respective R² values, indicating the high explanatory power of the regression model developed and proved to be reliable. Besides, as shown in Table 2, a lower than 20 % CV value is acceptable [24]. Thus, the established regression models in this study were accurate and reliable for subsequent prediction and optimization steps.

3.1.2. Analysis of the surface plots of interactive effects

The three-dimensional (3D) response surface plots were generated by fixing one variable and varying the other two variables to determine the interactive effects of the independent variables on the dependent variable. The response surface graphs for EY (%) and DPM (%) were displayed in Fig. 1a–c and Fig. 1d–f, respectively. The more curvature of the graph, the more effective the quadratic term on the plot [25]. The response surface plot in Fig. 1a displays the function of amplitude versus sonication time on EY at a constant solvent-to-solid ratio of 20 mL/g. The EY increased when the sonication amplitude increased from 20 to 70 %. The improved EY was due to the accelerated rate of mass transfer under sonication by promoting the hydration and fragmentation reaction and, thus, the extraction [24]. Beyond the sonication amplitude of 70 %, there is a decrease in the EY, which might be due to the induced formation of insoluble protein aggregates [15], [26] and the gentle collapse of the cavitation bubbles by the increased temperature at high amplitudes that reduce the ultrasound efficiency [27]. Fig. 1b shows the effect of sonication time versus solvent-to-solid ratio at a constant amplitude of 60 %.

Fig. 1.

Response surface plots (a, b, c, d, e, and f) showing the interactive effects of amplitude (X₁), time (X₂), and solvent-to-sold ratio (X₃) on EY and DPM.

The sonication time has a significant (p < 0.05) influence on the EY, as shown by the linear increase in EY, as the sonication time rose from 10 to 25 min with further decline beyond 25 min. The response surface 3D plot in Fig. 1c shows the function of amplitude versus solvent-to-solid ratio at a constant sonication time of 20 min. The increase in EY was observed when the solvent-to-solid ratio increased from 10 to 25 mL/g, and then it declined with the further increase in the solvent ratio. This occurrence may be due to the higher ratio caused by a higher diffusion rate and extraction yield explained by the mass transfer principle [24], [26]. The interaction between amplitude and sonication time on DPM in Fig. 1d showed that on increasing the amplitude and time, the DPM starts to increase to a certain point and, with further increase in amplitude and sonication time beyond 70 % and 25 min, respectively, has decreased the DPM. This finding was similar to the results in Fig. 1a, where a longer sonication time and amplitude decreased the EY. It has been indicated that ultrasound waves can destroy the hydrogen bonds and hydrophobic interactions and thus unfold the proteins by changing their conformation [28], [29]. This might cause the higher TCA solubility of proteins and, thus, the DPM. Further increase in sonication amplitude may result in protein degradation as well as protein oxidation caused by free radicals released from the acoustic cavitation [30], [31]. Fig. 1e shows the surface plot as a function of sonication time versus solvent-solid ratio at a fixed amplitude of 60 %. The surface plot revealed that the DPM increased with the sonication time and declined beyond the sonication time of 22 min. The result could be due to excessive exposure to ultrasound treatment at a longer sonication time, resulting in the protein degradation caused by the prolonged sonication or the increase of the enthalpy of denaturation by the aggregation phenomenon [32]. It was also apparent that prolonging the sonication time was economically unfeasible [24], [25], [27]. On the other hand, at a constant sonication time of 20 min, an increase in solvent-solid ratio and amplitude has been found to increase the DPM, as shown in Fig. 1f. However, beyond the sonication amplitude of 80 % and the solvent-to-solid ratio of 25, the DPM is decreased. This finding was associated with the diluting effect of the solvent-solid ratio, which reduced the TCA soluble protein content with a higher diluted solvent [24].

3.1.3. Model validation and optimization

The optimal UAE conditions were predicted by the Design Expert software based on the desirability function. The most desirable solution having a desirability value of 0.976 has been selected in this study for the highest EY (40.87 %) and DPM (26.28 %) and consisted of 63.2 % sonication amplitude with a sonication time of 20.5 min and a solvent-solid ratio of 20.8 mL/g. Experimentally, the maximum EY and DPM were reported to be 41.28 % and 26.25 %, respectively. These values are in conformity with the predicted values generated by the model and, thus, its validation.

3.2. Non-volatile taste composition of USH

3.2.1. Total and free amino acids

The composition of amino acids in WSHs plays a major role in flavor generation when they are used as a raw material for the preparation of flavor concentrate. A total of 17 amino acids were identified in WSHs, and the TAA content of WSHs is displayed in Table 3. Glu (14.06 %) is the major amino acid found in WSHs, followed by Asp (10.85 %) and His (10.46 %). Likewise, Wu et al. [33] also found that Glu is the major amino acid in P. vannamei heads, followed by Arg and Asp. The contents of Gly, Ala, Glu, and Asp are associated with the umami taste of seafood [34], and these are considered delicious amino acids (DAAs). The ratio of DAAs/TAAs of WSHs is found to be 0.40, which indicates that WSHs are substantial ingredients for developing shrimp flavor concentrate. In addition, as shown in Table 3, ultrasound treatment significantly (p < 0.05) increased the TFAA content of WSHs compared to the control by 42.2 %. Tyr is the major FAA found in conventional shrimp head extract (CSH), followed by Gly and Asp, whereas Glu is the major FAA found in USH, followed by Tyr and Gly. The content of DAAs was significantly (p < 0.05) higher in USH (45.75 %) than that in CSH (32.75 %). The ratio of DAAs/TAAs was higher in USH than that of CSH, representing the efficiency of ultrasonication treatment for enhanced umami flavor. Similarly, the content of umami and sweet FAAs was enhanced remarkably (170.77 % and 56.34 %, respectively), and there is a slight reduction in the content of bitter FAAs (1.71 %) upon ultrasonication, as shown in Fig. 2. The TAVs of FAAs were also listed in Table 3. FAAs with TAV > 0.1 contribute to the taste of a given food sample. Glu exhibited the highest TAV in USH, indicating its umami taste potential. Overall, these findings revealed that ultrasonication promoted the increase in the content of DAAs of WSHs, which is a crucial parameter for pleasant flavor generation.

Table 3.

Total amino acid content of WSH, free amino acid content and TAVs of CSH and USH.

S.No.	Amino acid	Total amino acid content of WSH (mg/100 g)	Free amino acid content (mg/100 g)		Taste Threshold (mg/100 mL)	TAV
			CSH	USH		CSH	USH
1	Aspartic acid	474.79 ± 11.25	31.55 ± 5.25b	48.25 ± 3.43a	100	0.316	0.483
2	Glutamic acid	615.11 ± 10.27	22.92 ± 1.22b	99.24 ± 1.57a	30	0.764	3.308
3	Serine	172.36 ± 9.58	8.84 ± 1.90b	9.08 ± 1.10a	150	0.059	0.061
4	Histidine	457.57 ± 12.27	13.54 ± 1.3a2	13.99 ± 2.21a	200	0.068	0.070
5	Glycine	285.73 ± 10.47	43.90 ± 0.42b	54.41 ± 3.21a	130	0.338	0.419
6	Threonine	157.64 ± 9.77	27.83 ± 2.10b	46.00 ± 1.33a	260	0.107	0.177
7	Arginine	299.52 ± 8.58	29.54 ± 0.15b	34.42 ± 0.44a	50	0.591	0.688
8	Alanine	390.55 ± 11.59	11.58 ± 2.17b	34.58 ± 1.59a	60	0.193	0.576
9	Tyrosine	136.34 ± 12.47	70.61 ± 4.46a	65.72 ± 3.73b	90.5	0.780	0.831
10	Cysteine	1.44 ± 0.14	0.48 ± 0.05b	1.15 ± 0.38a	48	0.010	0.024
11	Valine	197.36 ± 7.32	13.45 ± 1.30a	11.51 ± 0.54b	40	0.336	0.288
12	Methionine	84.63 ± 5.48	5.09 ± 0.46b	10.44 ± 0.15a	30	0.170	0.348
13	Tryptophan	184.26 ± 6.87	12.60 ± 3.40b	16.48 ± 0.47a	90.5	0.139	0.293
14	Phenylalanine	200.71 ± 9.46	22.38 ± 0.38a	8.63 ± 0.33b	90	0.249	0.207
15	Isoleucine	143.96 ± 8.47	2.78 ± 0.20b	4.39 ± 0.45a	90	0.031	0.160
16	Leucine	304.86 ± 10.28	10.91 ± 0.35a	6.86 ± 0.13b	190	0.057	0.036
17	Lysine	268.01 ± 12.29	7.74 ± 0.14b	12.28 ± 0.53a	50	0.155	0.246
Total		4374.84 ± 156.56	335.74 ± 25.27b	477.73 ± 21.59a
DAAs/TAAs		0.40	0.33	0.46

*DAAs-Delicious amino acids; TAAs- Total amino acids.

Fig. 2.

FAA content of CSH and USH based on their taste sensation. Bars with significant letters were significantly different (p ≤ 0.05).

3.2.2. Flavor nucleotides

Flavor nucleotides are key umami molecules that can improve the flavor of the food [35]. IMP, AMP, and GMP are the major flavor-enhancing nucleotides that can contribute to umami taste at low concentrations. The content of nucleotides in USH is shown in Table 4. GMP yielded the highest content, followed by IMP, AMP, CMP, and UMP, respectively, in both CSH and USH. Likewise, the TAV of GMP was the highest, followed by IMP and AMP. CMP and UMP did not contribute much to the taste of food, as their taste thresholds were higher. There is a drastic increase in the total nucleotide (TNT) content (673.38 mg/100 g) and flavor nucleotide (FNT) content (610.97 mg/100 g) of USH compared to that of CSH. A similar finding was reported by Siewe et al. [12], who reported an increase in the content of total and flavor nucleotides upon ultrasound-assisted enzymatic extraction than that of conventional enzymatic extraction from Labeo rohita heads. Thus, this outcome suggests that the ultrasound pretreatment produced a higher extraction yield. Further, it is also determined that the EUC (g MSG/100 g) value was increased extremely in USH (55.44 ± 3.25 g MSG/100 g) compared to CSH (8.32 ± 1.07 g MSG/100 g).

Table 4.

The concentrations and TAVs of organic acids and inorganic ions of CSH and USH.

Flavor compounds	Concentration (mg/100 g)		Taste threshold (mg/ 100 mL)	TAV
	CSH	USH		CSH	USH
5′-Nucleotides
AMP	47.14 ± 4.34b	154.69 ± 7.28a	50	0.942	1.093
GMP	93.64 ± 3.41b	294.12 ± 7.62a	12.5	7.491	23.529
IMP	51.75 ± 1.92b	162.16 ± 3.88a	25	2.07	6.486
UMP	6.48 ± 0.45b	13.80 ± 2.86a	625	0.011	0.022
CMP	15.26 ± 0.53b	48.61 ± 2.07a	1292	0.011	0.037
Total	214.27 ± 10.65b	673.38 ± 23.71a
Organic acids
Oxalic acid	182.53 ± 12.11b	269.53 ± 14.44a	50.4	3.621	5.347
Lactic acid	348.28 ± 22.11b	953.72 ± 17.52a	126	2.764	7.569
Citric acid	57.23 ± 3.57b	141.59 ± 8.65a	49.92	1.146	2.836
Succinic acid	447.86 ± 18.65b	772.01 ± 12.31a	10.62	42.17	72.693
Total	1035.9 ± 56.44b	2136.85 ± 52.92a
Inorganic ions
Sodium	4740.13 ± 65.81a	3347.28 ± 65.81b	180	26.334	18.596
Potassium	644.16 ± 16.62a	533.59 ± 17.03b	130	4.955	4.104
Phosphorous	4762.87 ± 63.60a	3131.27 ± 79.79b	130	36.637	24.086
Calcium	6854.01 ± 60.34a	5049.02 ± 85.59b	150	45.693	33.660
Magnesium	229.84 ± 12.47a	185.78 ± 13.16b	96	2.394	1.935
Total	17231.01 ± 218.84a	12246.94 ± 261.38b

Different lowercase letters in the same row indicate a significant difference (p ≤ 0.05).

3.2.3. Organic acids and inorganic ions

As shown in Table 4, succinic acid is the predominant organic acid found in both CSH and USH, followed by lactic acid, while citric acid yielded the lowest concentration. Similar studies reported that succinic acid and lactic acids are the predominant organic acids mainly derived from the abundant glycogen and have contributed to the umami and sourness of many aquatic species [36]. Ultrasonic treatment of WSHs caused an enormous increase in the content of organic acids. Also, the higher TAV of succinic acid in USH (72.693) indicates its significant umami taste contribution. Similarly, the inorganic ions, including Na⁺, K⁺, and PO₄^3- can also affect the sweet and umami tastes of seafood [34]. Table 4 also lists the content, taste thresholds, and TAVs of inorganic ions of WSHs. In contrast to the other non-volatile taste compounds, inorganic ions exhibited an opposite trend in USH by decreasing their concentration. This might be due to the water-insoluble nature of minerals in the residue left after ultrasonic treatment. Based on these results on the variations in the contents of inorganic ions, the salty taste in WSHs was decreased after ultrasonication.

3.3. Characterization of SFC

3.3.1. Changes in pH and browning intensity of SFC during the heating process

The pH value is a critical parameter to monitor the rate of Maillard reaction. The changes in pH values of SFC at different heating periods (S30, S60, S90, and S120) and control are shown in Fig. 3a, and it can be seen that as heating time increased, the pH values declined gradually, which indicates the speed up of the Maillard reaction. Initially, there is a rapid decrease in the pH in first 60 min, thereafter, a slow decrease in the pH is observed until the end of the reaction. The lowest pH value is observed for S90 and S120. This decline in the pH value was due to the generation of organic acids during Maillard reaction and the reduction of the free N-terminal amino groups that led to an increase in the carboxyl group, thereby reducing the pH of SFC [12], [37]. Similar finding was reported by Xiao et al. [37] and Chen et al. [38], who observed a decrease in the pH value of the glucose/rice protein hydrolysate and xylose/bovine casein hydrolysates (BCH) model with the increased Maillard reaction time. Also, the formation of reductones and melanoidins during the Maillard reaction cause a decrease in the pH [39].

Fig. 3.

Changes in a) pH and b) UV absorbance at 294 nm, c) UV absorbance at 420 nm of SFC during the heating process.

The rate of a Maillard reaction is indicated by the browning intensity caused by pigment formation. It is usually monitored by an increase in the absorbance at 294 nm (A294) and 420 nm (A420). A294 used to indicate the content of degradation products of Amadori compounds, a colorless intermediate products of the Maillard reaction and A420 represents the melanoid content of the sample [40]. As shown in Fig. 3b, the A294 value was enhanced with increased reaction time. This increase in UV absorption might be due to the degradation of Heyns-type compounds and the generation of intermediate products [37]. A similar finding was reported by Chen et al. [38], who found that the browning intensity of xylose-BCH Maillard reaction products was significantly increased with a reaction time. Moreover, as shown in Fig. 3c, the absorbance intensity at 420 nm was more pronounced in S90 and S120, indicating a higher content of final brown products than in S0. This rise in the A420 value might be due to the polymerization of reactive cyclic compounds such as hydroxy methyl furfural to yield dark-colored substances that require less acidic conditions [41].

3.3.2. Proximate composition and aroma profile of SFC

The proximate composition of SFC, such as moisture, crude protein, crude fat, and ash contents, was 8.25 ± 0.13 %, 62.59 ± 1.28 %, 9.25 ± 0.53 %, and 7.91 ± 0.45 %, respectively, on a wet weight basis. As summarized in Table 5, 27 volatile compounds are identified and determined in SFC, including 6 ketones, 3 aldehydes, 6 pyrazines, 4 alcohols, 4 acids, and each ester, a heterocyclic compound, aromatic hydrocarbon, and thiazole. Among these, 22 compounds are identified as odor-active compounds as their OAVs were greater than 1. Aldehydes are the important aroma compounds of aquatic species. In this study, benzaldehyde was found to be the major aldehyde with a cherry-like smell. Among the ketones, 2-octanone exhibited the highest OAV, followed by 2-heptanone and 6-methyl-5-hepeten-2-one. 6-Methyl-5-hepeten-2-one contributes to the lemongrass aroma [42]. Furans and pyrazines have nutty and roasted aromas [43]. 2-Pentyl-furan is the only furan identified in SFC, with a beany and grassy aroma. It is also identified in the fish flavoring developed in the study of Siewe et al. [12]. 3-Ethyl-2,5-dimethylpyrazine is the predominant pyrazine with the highest OAV, contributing a nutty and roasted aroma to SFC. It is also the major aroma compound occupying 32.5 % of the total volatile concentration, followed by naphthalene (17.5 %), an aromatic hydrocarbon with a camphor-like odor. However, based on TAV, nonanal is the major odor-contributing compound next to 3-ethyl-2,5-dimethylpyrazine, which smells like fat and citrus [43]. Four short-chain acids were identified in SFC, among which heptanoic acid is the major, followed by acetic acid. But in terms of OAV, octanoic acid displayed the highest value, followed by heptanoic acid. 1-Octanol exhibited the highest OAV among alcohols with a citrus and sweet aroma. 2-Methylbutanoate is the only ester found in SFC that contributes to a fruity aroma of SFC. Similarly, 2-acetylthiazole was the only thiazole identified in SFC with a nutty and popcorn-like aroma. Overall, the SFC prepared from USH exhibited a pleasant aroma with more concentration of volatile compounds, popular for shrimp aroma as determined in previous studies.

Table 5.

Concentration, odor thresholds values and odor activity values of volatiles of SFC.

S.No.	Volatile compound	RT	RI	LRI	Concentration (µg/g)	Odor Threshold (µg/g)	OAV
1	2-Heptanone	8.911	1177.22	1177	50.92 ± 7.05	0.141	361.13
2	Heptanal	8.967	1179.77	1179	27.09 ± 6.62	0.0028	9675
3	2-Pentylfuran	9.879	1225.73	1226	93.29 ± 1.33	0.0058	16084.48
4	6-Methyl-2-heptanone	10.077	1236.63	1236	85.53 ± 5.32	–
5	Methylpyrazine	10.528	1261.48	1261	54.53 ± 7.54	60	0.908
6	2-Octanone	10.848	1279.11	1279	55.30 ± 8.89	0.0502	1101.59
7	Hydroxyacetone	11.168	1296.74	1296	117.11 ± 7.57	–
8	2,5-Dimethylpyrazine	11.478	1316.17	1316	212.44 ± 11.98	0.008	26,555
9	6-Methyl-5-hepten-2-one	11.741	1333.11	1333	21.59 ± 2.29	0.068	317.5
10	2-Ethyl-6-methylpyrazine	12.568	1386.4	1386	165.52 ± 48.18	0.24	689.67
11	Nonanal	12.606	1388.27	1388	58.87 ± 14.08	0.001	58,870
12	2,3,5-Trimethylpyrazine	12.756	1398.51	1398	462.62 ± 95.69	0.297	1557.64
13	3-Ethyl-2,5-dimethylpyrazine	13.452	1444.45	1444	1675.18 ± 91.03	0.001	1,675,180
14	1-Octen-3-ol	13.49	1446.96	1446	74.76 ± 14.16	0.85349	87.59
15	Acetic acid	13.556	1451.32	1451	111.75 ± 22.79	160	0.698
16	3,5-Diethyl-2-methylpyrazine	14.148	1490.42	1491	93.48 ± 18.81	–
17	Benzaldehyde	14.543	1525.56	1525	94.00 ± 37.36	0.0417	2254.19
18	1-Octanol	14.825	1554.39	1554	21.30 ± 6.00	0.001	21,300
19	3,5-Octadien-2-one	14.995	1571.77	1570	43.56 ± 1.28	0.15	290.4
20	2-Octen-1-ol	15.343	1609.81	1609	32.17 ± 10.72	0.02	1608.5
21	2-Acetylthiazole	15.653	1652.76	1652	67.65 ± 34.46	0.01	6765
22	2-Methylbutanoate	15.747	1665.74	1665	227.73 ± 24.90	0.1	2277.3
23	Naphthalene	16.292	1751.83	1751	904.92 ± 43.72	0.06	15,082
24	Hexanoic acid	16.753	1838.54	1838	26.11 ± 7.88	2.52	10.361
25	Benzyl alcohol	16.95	1879.58	1879	132.22 ± 1.38	40.9	3.232
26	Heptanoic acid	17.232	1943.73	1943	170.19 ± 14.77	13.8	12.33
27	Octanoic acid	17.665	2023.17	2023	74.56 ± 1.84	2.7	27.61

Values are expressed as means ± standard deviation (n = 3). *RT = retention time; RI^X = linear retention indices calculated on a BP-20 wax column (30 m × 0.25 mm × 0.25 μm) with a homologous series of n-alkanes (C7–C20); RI^Y = linear retention indices from the literature. “-” indicates “not available”.

3.3.3. Sensory evaluation of SFC

Sensory characteristics of SFC were evaluated by descriptive analysis, and the results are shown in Fig. 4. SFC showed significantly (p < 0.05) higher umami and shrimp-like taste intensity scores compared to other sensory attributes. This higher umami score of SFC produced from USH is in agreement with the higher content of DAAs and flavor nucleotides of USH, as discussed in the previous section. However, there is not much difference between old shrimp and cooked shrimp aroma attributes. The sour and bitter attributes of SFC were almost imperceptible, which was consistent with the study of Siewe, Kudre, & Narayan [15], confirming the reduction in the sour and bitter attributes is a desirable property. This result is in accordance with the enhanced organic acid content of USH, especially succinic acid, which majorly contributes to the umami flavor of SFC.

Fig. 4.

Sensory evaluation of SFC.

4. Conclusion

The demand for natural foods formulated with bio ingredients has been increasing. Hence, industries must develop efficient extraction techniques, especially for flavor compounds and produce natural foods to meet customer demand. This study optimized the ultrasonic-assisted extraction conditions for flavor compounds from WSHs and developing shrimp flavor concentrate (SFC). The experimental data was successfully fitted to the theoretical models to determine the optimum conditions. Ultrasonic-assisted extraction increased the content of taste-active compounds from WSHs. Therefore, it is proven that the ultrasonic-assisted extraction combined with the Maillard reaction could be a promising method for developing SFC, which may have potential uses as a flavor enhancer in various food products. This study also suggests that using WSHs to recover natural umami and other flavor compounds is both sustainable and economical for food industries.

CRediT authorship contribution statement

Haritha Duppeti: Conceptualization, Investigation, Methodology, Validation, Formal analysis, Data curation, Funding acquisition, Writing – original draft, Writing – review & editing. Sachindra Nakkarike Manjabhatta: Conceptualization, Supervision, Validation, Writing – review & editing. Bettadaiah Bheemanakere Kempaiah: Supervision, Validation, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.