In tropical regions, protein supplementation is a common practice in dairy and beef farming. However, the effect of highly degradable protein in ruminal fermentation and microbial community composition has not yet been investigated in a systematic manner. In this work, we aimed to investigate the impact of casein supplementation on volatile fatty acids (VFA) production, specific activity of deamination (SAD), ammonia concentration and bacterial and archaeal community composition. The experimental design was a 4×4 Latin square balanced for residual effects, with four animals (average initial weight of 280±10 kg) and four experimental periods, each with duration of 29 days. The diet comprised Tifton 85 (Cynodon sp.) hay with an average CP content of 9.8%, on a dry matter basis. Animals received basal forage (control) or infusions of pure casein (230 g) administered direct into the rumen, abomasum or divided (50 : 50 ratio) in the rumen/abomasum. There was no differences (P>0.05) in ruminal pH and microbial protein concentration between supplemented v. non-supplemented animals. However, in steers receiving ruminal infusion of casein the SAD and ruminal ammonia concentration increased 33% and 76%, respectively, compared with the control. The total concentration of VFA increased (P<0.05) in steers receiving rumen infusion of casein. SAD and the microbial protein concentration did not vary significantly among treatments during the feeding cycle, but mean SAD values were greater in steers supplemented in the rumen and rumen/abomasum. Ruminal ammonia concentration was positively correlated with SAD in animals receiving ruminal infusion of casein. Polymerase chain reaction–denaturing gradient gel electrophoresis (PCR-DGGE) analysis revealed low similarity between treatments, animals and time of sample collection. Richness analysis and determination of the Shannon–Wiener index indicated no differences (P>0.05) in species richness and diversity of γ-proteobacteria, firmicutes and archaea between non-supplemented Nellore steers and steers receiving casein supplementation in the rumen. However, species richness and the Shannon–Wiener index were lower (P<0.05) for the phylum bacteroidetes in steers supplemented with casein in the rumen compared with non-supplemented animals. Venn diagrams indicated that the number of unique bands varied considerably among individual animals and was usually higher in number for non-supplemented steers compared with supplemented animals. These results add new knowledge about the effects of ruminal and postruminal protein supplementation on metabolic activities of rumen microbes and the composition of bacterial and archaeal communities in the rumen of steers.