Determination of the protein quality of almonds (Prunus dulcis L.) as assessed by in vitro and in vivo methodologies

Abstract Almonds (Prunus dulcis), such as all nuts, are positioned within the protein foods grouping within the current U.S. Dietary Guidelines. The ability to make claims related to the protein content of almonds, within the United States, requires substantiation via the use of the Protein Digestibility‐Corrected Amino Acid Score (PDCAAS). The present study was designed to provide current estimates of PDCAAS, using both in vivo and in vitro assays, of key almond varietals from the 2017 California harvest. Additionally, historical protein and amino acid composition data on 73 separate analyses, performed from 2000 to 2014, were analyzed. Amino acid analysis confirmed lysine as the first‐limiting amino acid, generating amino acid scores of 0.53, 0.52, 0.49, and 0.56 for Butte, Independence, Monterey, and Nonpareil varietals, respectively. True fecal protein digestibility coefficients ranged from 85.7% to 89.9% yielding PDCAAS values of 44.3–47.8, being highest for Nonpareil. Similar, albeit lower, results were obtained from the in vitro assessment protocol. Analysis of the historical data again positioned lysine as the limiting amino acid and yielded information on the natural variability present within the protein and amino acid profiles of almonds. Comparison of the 2017 AA profile, averaged across almond varietals, to the historical data provided strong evidence of persistence of amino acid composition and indices of protein quality over time.

While the current dietary guidelines recognize the importance of nuts, including almonds, to dietary protein intakes, communicating messages regarding the value of specific foods for the provision of protein necessitates meeting key regulatory requirements related to protein quality. In the United States, protein content claims on labels are based on the amount of quality-corrected protein within a representative amount customarily consumed (RACC), where protein quality is assessed using the Protein Digestibility-Corrected Amino Acid Score (PDCAAS) (US Food & Drug Administration, 2018). The latter is determined as the product of the Amino Acid Score (AAS) and the percent true fecal protein digestibility of the food in question. The AAS reflects the most limiting amino acid supplied by the food in comparison with an established reference requirement pattern and thus reflects a chemical assay. True protein digestibility, as stipulated by the FDA, requires the use of biological rodent assay to determine the amount of fecal nitrogen excreted per unit of dietary nitrogen consumed. The final PDCAAS values, which range from 0% to 100% (due to truncation) (FAO/WHO, 1991), are multiplied against the amount of crude protein in a RACC. In order to make protein content claims on foods, they must contain between 10% and 19% of the Daily Value (50 g) of PDCAAS-corrected protein within a RACC for a "Good Source" claim, or 20% or greater for an "Excellent Source" claim. In the case of almonds (RACC = 30 g), the PDCAAS would need to be approximately 80% in order to qualify for a "Good Source of Protein" claim on labels. Attaining this value is a challenge for all nuts, including almonds, due primarily to limitations in their content of the indispensable AA lysine. As such, the requirement to attain the aforementioned quality benchmarks, as stipulated by the FDA, is somewhat incongruous to the message within the MyPlate dietary patterns regarding the importance of this food group in contributing protein to the human diet, a challenge described previously (Marinangeli & House, 2017).
Despite the above challenges, it is important to understand factors influencing the protein quality of almonds, as the sector seeks to position foods to satisfy the increasing demands by consumers for plant proteins. Published data by Ahrens et al. (2005), in 2005, provided an estimate for the AAS of almonds as being 0.26, with the sulfur AAs methionine and cysteine being limiting. Based on the AA content of almonds published in the USDA Nutrient Database, the AAS for almonds is calculated to be 0.46, with lysine being the limiting AA (USDA, 2018). Given these discrepancies, together with the interest in positioning plant-based proteins to consumers, new estimates of the PDCAAS of almond varieties are warranted. The objectives of the current study were as follows: (a) to determine the PDCAAS (using both in vivo and in vitro estimates of protein digestibility) of 4 almond varietals and (b) to summarize historical protein and AA composition data on almonds gathered over 15 years, as generated by commercial laboratories.

| Materials
Composite samples (5,000 g) of shelled, raw Butte, Independence, Monterey, and Nonpareil almond varietals were provided by the Almond Board of California, Modesto, CA. All chemicals and reagents, including the NIST Standard Reference Material 3234 soy flour, were purchased from Sigma. Dietary ingredients for the in vivo true fecal protein digestibility study were procured from Dyets Inc..

| Sample preparation and analysis
Prior to the in vivo and in vitro analysis of the 4 almond varieties plus a casein (high nitrogen) control, all test articles were ground, using a Wiley Mill (Thomas Scientific) to pass through a 2-mm screen.
Subsamples for proximate and AA analysis were further ground, using a handheld electric mill, to pass through a 1-mm screen.
Samples were stored in airtight containers at −20°C prior to analysis. For quality control, the NIST soy flour Standard Reference Material 3234 was used for all AA analyses. For amino acid analyses, the hydrated molecular weights of amino acids were used for quantitation.

| Protein digestibility
In order to calculate PDCAAS for the test articles, the percent True Fecal Protein Digestibility (%TFPD) was measured according to AOAC Official Method 991.29 (Association of Official Analytical Chemists, 1995), with minor modifications to account for advances in rodent nutrition since the date of first action of the published method (1991). Principally, the AIN-93G vitamin and mineral premixes were employed instead of the AIN-76 formulations (Reeves, Nielsen, & Fahey, 1993), as the authors have described previously (House, Neufeld, & Leson, 2010;Nosworthy, Medina, et al., 2017). As an additional measure, the percent in vitro protein digestibility (%IVPD) was determined via the pH drop method, in duplicate, as previously described (Tinus, Damour, Riel, & Sopade, 2012). In addition to the PDCAAS calculation, the in vitro PDCAAS was also determined as the product of the AAS and %IVPD ). Tryptophan and the sulfur AA were not measured for all samples.

| Historical analytical data procurement
The analytical laboratories are USDA-certified and thus follow approved methodology for respective nutrient analyses. The AAS values were determined as described above. When tryptophan or the sulfur AA was absent, lysine was assumed to be the first-limiting AA for AAS calculations.

| Statistical analyses
Data for composite samples are presented as the mean of duplicate analyses. Data for %TFPD were subjected to one-way ANOVA, with p-value <.05 taken to indicate significance and post hoc analyses conducted by Tukey's HSD method, using Prism 7 (GraphPad Software). Historical nutrient composition data were examined for outliers, using the nonlinear regression ROUT method (GraphPad Software), with a conservative false discovery rate of 0.1% and measures of central tendency and variation computed.

| Nutritional quality of almond protein
The determination of the AA profile is needed to establish the AAS ( for reasons cited above. This AAS places almonds in a similar range (calculated from USDA, 2018) of AAS as walnuts (0.51) and pecans (0.54), along with other foods that are similarly limiting in lysine, including cereal grains and certain seeds, including hemp (House et al., 2010). Other plant proteins may have different limiting AAs, such as pulses, which tend to be limiting in either sulfur AA or tryptophan.
For example, field peas present with an AAS of 0.8 (Nosworthy, Medina, et al., 2017).   (Hall et al., 1958). Further efforts should be placed to understand the impact of almond processing on measures of protein and AA digestibility. The latter may be more important as efforts have been made to position a new method for assessing protein quality, based on the Digestible Indispensable Amino Acid Score (DIAAS) system (FAO/WHO, 2013). This method requires a more invasive approach to measuring ileal AA digestibility and may prove challenging to implement for a diverse and complex human foodscape. Given the potential for in vitro approaches to approximate in vivo digestibility coefficients, efforts should be placed on validating and approving in vitro approaches for assessing protein quality.
In an effort to further explore variation in the protein and AA content of almonds, data from 73 individual sample analysis submissions made to commercial laboratories were evaluated ( value of .99. The latter result reflects general agreement between the analytical methods used in the current study to those used in commercial laboratories and also confirms the resiliency of the protein and AA content of the major almond varieties over time.

| CON CLUS ION
In conclusion, the current data support a PDCAAS value for raw almonds of between 44.3 and 47.8, for the varieties tested. Concurrent assessment of in vitro digestibility and growth provided additional evidence of the quality of almond proteins. The data can be used to guide varietal selection for amino acid content; however, the natural variability between varieties, particularly in lysine content, may not be sufficient to make substantial improvements. Of the varieties test, Nonpareil presented with the most consistent pattern of highest protein quality. Given the importance of Nonpareil to the California almond industry, the current results support continued attention to this variety.

ACK N OWLED G M ENTS
We are grateful to Judy Scott-McKay, Almond Board of California, for coordinating the shipment of almonds and transmission of the historical composition data.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest with regard to the described research, the publication of results, or financial issues.