The cellular translation process should obey the principle of maximizing efficiency and minimizing resource and energy costs. Here, we validated this principle by focusing on the basic translation components of tRNAs and amino acids. To most efficiently utilize these components, we reasoned that the quantities of the 20 tRNAs and their corresponding amino acids would be consistent in an organism. The two values should match at both the organismal and protein scales. For the former, they co-vary to meet the need to translate more proteins in fast-growing or larger cells. For the latter, they are consistent to different extents for various proteins in an organism to comply with different needs of translation speed. In this work, 310 out of 410 genomes in three domains had significant co-adaptions between the tRNA gene copy number and amino acid composition, and thus validating the principle at the organism scale. Furthermore, fast-growing bacteria co-adapt better than slow-growing ones. Highly expressed proteins and those connected to acute responses have better co-adaption, illustrating the principle at the individual protein scale. Experimentally, manipulating the tRNA gene copy number to optimize co-adaption between enhanced green fluorescent protein (EGFP) and tRNA gene set of Escherichia coli indeed lifted the translation rate (speed). Our results also contribute to revealing a translation rate-associated factor with universal and global effects. From a practical perspective, our findings suggest a strategy to increase the expression of target proteins and have implications for designing chassis cells in the field of synthetic biology field.