Karl J. Siebert

Modeling the inhibitory effects of organic acids on bacteria.

The inhibitory effect of acids on microbial growth has long been used to preserve foods from spoilage. While much of the effect can be accounted for by pH, it is well known that different organic acids vary considerably in their inhibitory effects. Because organic acids are not members of a homologous series, but vary in the numbers of carboxy groups, hydroxy groups and carbon-carbon double bonds in the molecule, it has typically not been possible to predict the magnitude, or in some cases even the direction, of the change in inhibitory effect upon substituting one acid for another or to predict the net result in food systems containing more than one acid. The objective of this investigation was to attempt to construct a mathematical model that would enable such prediction as a function of the physical and chemical properties of organic acids. Principal Components Analysis (PCA) was applied to 11 properties for each of 17 acids commonly found in food systems; this resulted in four significant principal components (PCs), presumably representing fundamental properties of the acids and indicating each acids location along each of these four scales. These properties correspond to polar groups, the number of double bonds, molecular size, and solubility in non-polar solvents. Minimum inhibitory concentrations (MICs) for each of eight acids for six test microorganisms were determined at pH 5.25. The MICs for each organism were modeled as a function of the four PCs using partial least squares (PLS) regression. This produced models with high correlations for five of the bacteria (R2 = 0.856, 0.941, 0.968, 0.968 and 0.970) and one with a slightly lower value (R2 = 0.785). Acid susceptible organisms (Bacillus cereus, Bacillus subtilis, and Alicyclobacillus) exhibited a similar response pattern. There appeared to be two separate response patterns for acid resistant organisms; one was exhibited by the two lactobacilli studied and the other by E. coli. Predicting the inhibitory effects of the organic acids as a function of their chemical and physical properties is clearly possible.

Validation of bacterial growth inhibition models based on molecular properties of organic acids

Organic acids occur naturally in foods and have been used in many food products as preservatives because they inhibit the growth of most microorganisms. The acids commonly found in foods differ greatly in both their structure and inhibitory effects for different bacteria. A way to represent relationships between different acids was previously described in which principal components analysis (PCA) was applied to 11 physical and chemical properties of 17 organic acids, to arrive at principal properties. These were used for development of regression models that related the minimum inhibitory concentrations (MICs) of organic acids to their principal properties. Separate MIC models were constructed for six different bacteria. The objective of the present study was to test the predictive capabilities of the organism models using different organic acids from the ones used to construct the original models. MIC predictions were made for three acids for each of the six bacteria for which models were previously constructed. MIC determinations for these acids were then carried out and compared with the predictions; these were in good agreement, thus validating the models. The new data were combined with that obtained previously to produce similar, but slightly stronger models. These had R(2) values between 0.861 and 0.992.

Modeling the flavor thresholds of organic acids in beer as a function of their molecular properties

Abstract A mathematical model of the flavor thresholds of organic acids in beer as a function of their physical and chemical properties was constructed. Principal components analysis (PCA) was applied to 11 properties for 17 acids; this resulted in four significant principal components (PCs), which accounted for 93% of the variance in the property data. The PCs were dominated by the number of polar groups, the number of double bonds, molecular size, and solubility in non-polar solvents, respectively. PC scores indicating each acids location along each of four scales were obtained. Literature data for acid flavor thresholds in beer were modeled as a function of the principal components using multiple linear regression and partial least squares (PLS) regression. The best-fit and best prediction equations were identical, with an R 2 of 0·905. ©

An alternate mechanism for the astringent sensation of acids

Abstract Tannic acid was added to saliva, the pH was adjusted to 12 different levels in the range 2.5–7.2 and light scattering was observed. The maximum light scattering occurred at pH 4.4, with considerably less scattering at both higher and lower pH. Saliva without added tannic acid had a light scattering maximum at the same pH. The total polyphenol (Folin–Ciocalteu) content of saliva was determined with results between 53 and 93 mg/l. It is proposed that at least part of the mechanism for the astringent sensation of acids in water is caused by an intensification of the interaction between salivary protein and salivary polyphenol. When acid is not present the interaction is less intense and the perception is weak or non-existent. The increased precipitation of salivary protein when acid is added to polyphenol also accounts for the observed increase in astringency.

Modeling Bovine Serum Albumin Binding of Flavor Compounds (Alcohols, Aldehydes, Esters, and Ketones) as a Function of Molecular Properties

Interactions between bovine serum albumin (BSA) and 4 classes of flavor compounds (alcohols, esters, aldehydes, and ketones) in aqueous solution were investigated using solid phase microextraction (SPME) of sample headspace and gas chromatography. Alcohols did not bind significantly to BSA. The binding of fixed amounts of individual flavor compounds with each functional group to increasing amounts of BSA was modeled as a function of several descriptors using partial least squares regression (PLSR). Ester binding was modeled as a function of the number of carbon atoms in a molecule and its boiling point (R(2)= 0.954). Aldehyde binding was modeled as a function of the number of hydrogen atoms and boiling point (R(2)= 0.922). Ketone binding was modeled as a function of the numbers of carbon atoms, length of the longer hydrocarbon chain, and degree of branching (R(2)= 0.961).

Organic acid inhibition models for Listeria innocua, Listeria ivanovii, Pseudomonas aeruginosa and Oenococcus oeni

Abstract Separate mathematical models for each of six bacteria that relate organic acid minimum inhibitory concentrations (MICs) to acid molecular properties were previously constructed and validated. MICs were determined for each of eight acids selected to have maximum modeling power for each of four additional bacteria (Listeria innocua, L. ivanovii, Pseudomonas aeruginosa and Oenococcus oeni). MIC models as a function of acid molecular properties were constructed for each organism using partial least-squares regression; these were very strong, with best prediction R2 values ranging from 0.905 to 0.947. The patterns of the MIC model coefficients were compared with those of the six bacteria models constructed previously. Together these represent one pattern of acid susceptibility and four different patterns of acid resistance (Lactobacillus-type, E. coli-type, Listeria-type and Pseudomonas-type).

Chapter 2 Haze in Beverages

Beverages such as beer, wine, clear fruit juices, teas, and formulated products with similar ingredients are generally expected by consumers to be clear (free of turbidity) and to remain so during the normal shelf life of the product. Hazy products are often regarded as defective and perhaps even potentially harmful. Since consumers are usually more certain of what they perceive visually than of what they taste or smell, the development of haze in a clear product can reduce the likelihood of repeat purchasing of a product and can have serious economic consequences to a producer. Hazes are caused by suspended insoluble particles of colloidal or larger size that can be perceived visually or by instruments. Hazes in clear beverages can arise from a number of causes, but are most often due to protein-polyphenol interaction. The nature of protein-polyphenol interaction and its effect on haze particles, analysis of haze constituents, and stabilization of beverages against haze formation are reviewed.

Human visual perception of haze and relationships with instrumental measurements of turbidity. Thresholds, magnitude estimation and sensory descriptive analysis of haze in model systems

Abstract Spherical polymer beads (0.769, 2.600 and 10.300 μm diameter) were suspended in clear, yellow and red liquids. The samples were measured by turbidimetry and assessed by panelists. Thresholds were determined by the Ascending Method of Limits and ranged from 0.384 to 0.815 NTU. The results were influenced by both particle size and solution color. Visual intensity (assessed by Magnitude Estimation) rose linearly with particle concentration until it reached a plateau. A regression model was developed that expressed visual haze intensity as a function of particle concentration and size, and liquid color ( R 2 =0.949). A relationship between visual and instrumental responses was also developed ( R 2 =0.870); when particle size was included, this improved to R 2 =0.978. Turbidimeter response could be predicted from particle concentration and size ( R 2 =0.986). Principal Components Analysis was applied to Descriptive Analysis results and showed that two factors accounted for 99% of the observed variation. Suspensions of large particles at intermediate concentrations appeared non-homogeneous.

Recent Discoveries in Beer Foam

Combinations of ovalbumin, iso-alpha acid, and ethanol were prepared in buffers of different pH and foamed. A response surface model of foam height was constructed (R2 = 0.876). This indicated that intermediate ethanol levels lead to the best foam, with poorer foam at higher and lower ethanol contents, and that increasing pH leads to poorer foam. When ethanol was added to non-alcoholic beer, the effect on foam was somewhat similar to the model system. When a lager was adjusted in pH, the foam increased with increasing pH, opposite to the model system. Dimethyl formamide, dioxane, and NaCl solution were each added to the model system and beer. Salt greatly reduced foam in the model system, indicating mainly ionic interaction. DMF caused the largest reduction in commercial beer foam, indicating a hydrogen bonding mechanism. Barley lipid transfer protein 1 (LTP1) and proteins Z4 and Z7 have been associated with beer foam. Ovalbumin has considerable similarity to proteins Z4 and Z7, but is quite different from LTP1. A crude extract of barley albumin behaved much like beer in the buffer model system. The results suggest greater involvement of LTP1 than the other two proteins in beer foam.

Modeling physicochemical properties and activity of aspartyl proteinases based on amino acid composition.

A data set containing physicochemical properties and enzymatic activity measurements of aspartyl proteinases was employed for quantitative structure-property relationship (QSPR) and quantitative structure-activity relationship (QSAR) modeling based on either three or five amino acid principal property sums. All but one of the models based on five principal properties were stronger than those based on three properties. Models of zeta potential (R2 = 0.846), circular dichroism (R2 = 0.638), Bigelow average hydrophobicity (R2 = 0.692), accessible surface area (R2 = 0.897), and two dye-based assessments of hydrophobicity (R2 = 0.581 and 0.595) were constructed. Model quality was evaluated by cross-validation and permutation. The amino acids most influential for each modeled property were identified. It is clearly possible to model physicochemical properties of proteins as a function of amino acid principal property sums. Surprisingly, it was also possible to model an enzyme activity ratio (milk clotting/proteolytic activity) in the same manner (R2 = 0.699).