Use of milk proteins as biomarkers of changes in the rumen metaproteome of Holstein cows fed low-fiber, high-starch diets

Abstract

Dietary levels of undegraded neutral detergent fiber (uNDF240) and rumen-fermentable starch (RFS) can affect the rumen microbiome and milk composition. The objective of the study is to investigate the use of milk proteins as biomarkers of rumen microbial activity through a comparative evaluation of the rumen microbial and milk protein profiles produced by Holstein cows fed diets with varying contents of physically effective uNDF240 (peuNDF240) and RFS. Eight ruminally cannulated lactating Holstein cows were included in a larger study as part of a 4 × 4 Latin square design with 4 28-d periods to assess 4 diets varying in peuNDF240 and RFS content. For this experiment, cows received one of 2 dietary treatments: (1) low-peuNDF240, high-RFS (LNHR) diet or (2) high-peuNDF240, low-RFS (HNLR) diet. Within each period, rumen fluid samples were collected from each cow on d 26 (1400 h) and d 27 (0600 h and 1000 h), and milk samples were collected from each cow on d 25 (2030 h), d 26 (0430 h, 1230 h, and 2030 h), and d 27 (0430 h and 1230 h). Microbial proteins were isolated from each rumen fluid sample. For milk samples, milk proteins were fractionated, and the whey fraction was subsequently isolated. Isolated proteins within each rumen fluid or milk sample were isobarically labeled and analyzed by liquid chromatography-tandem mass spectrometry. Product ion spectra acquired from rumen fluid samples were searched using SEQUEST against 71 composite databases. In contrast, product ion spectra acquired from milk samples were searched against the Bos taurus database. Data were analyzed using the PROC MIXED procedure in SAS 9.4 to assess the effect of diet and time of sampling. To increase stringency, the false discovery rate-adjusted P-value (P_FDR) was also calculated to account for multiple comparisons. Using the mixed procedure, a total of 129 rumen microbial proteins were quantified across 24 searched microbial species. Of these, the abundance of 14 proteins across 9 microbial species was affected due to diet and diet × time interaction, including 7 proteins associated with energetics pathways. Among the 159 quantified milk proteins, the abundance of 21 proteins was affected due to the diet and diet × time interaction. The abundance of 19 of these milk proteins was affected due to diet × time interactions. Of these, 16 proteins had the disparity across diets at the 0430 h sampling time, including proteins involved in host defense, nutrient synthesis, and transportation, suggesting that biological shifts resulting from diet-induced rumen changes are not diurnally uniform across milkings. The concentration of lipoprotein lipase (LPL) was statistically higher in the milk from the cows fed with the LNHR diet, which was numerically confirmed with an ELISA. Further, as determined by ELISA, the LPL concentration was significantly higher in the milk from the cows fed with the LNHR diet at 0430 h sampling point, suggesting that LPL concentration may indicate dietary carbohydrate-induced ruminal changes. The results of this study suggest that diet-induced rumen changes can be reflected in milk in a diurnal pattern, further highlighting the need to consider sampling time points for using milk proteins as a representative biomarker of rumen microbial activity.

Keywords: milk proteins; physically effective undegraded neutral detergent fiber; rumen microbial proteins; rumen-fermentable starch.