How well mRNA transcript levels represent protein abundances has been a controversial issue. Particularly across different environments, correlations between mRNA and protein exhibit remarkable variability from gene to gene. Translational regulation is likely to be one of the key factors contributing to mismatches between mRNA level and protein abundance in bacteria. Here, we quantified genome-wide transcriptome and relative translation efficiency (RTE) under 12 different conditions in Escherichia coli. By quantifying the mRNARTE correlation both across genes and across conditions, we uncovered a diversity of gene-specific translational regulations, cooperating with transcriptional regulations, in response to carbon (C), nitrogen (N), and phosphate (P) limitations. Intriguingly, we found that many genes regulating translation are themselves subject to translational regulation, suggesting possible feedbacks. Furthermore, a random forest model suggests that codon usage partially predicts a gene's cross-condition variability in translation efficiency; such cross-condition variability tends to be an inherent quality of a gene, independent of the specific nutrient limitations. These findings broaden the understanding of translational regulation under different environments and provide novel strategies for the control of translation in synthetic biology. In addition, our data offers a resource for future multi-omics studies. Author summary The central dogma connects DNA, RNA, and protein through transcription and translation. However, with the development of transcriptome and proteomics technology, it has been widely reported that mRNA abundance is not a comprehensive indicator of protein abundance. Translational regulation is critical in resolving this type of mismatch. It has been reported that bacteria respond to heat stress, oxidative stress, and other stressful environments through translational regulation. Nutrient limitations are also fundamental challenges for bacteria, with many unknowns in their adaptation strategies. Using transcriptome and translatome quantification, we uncovered a diversity of gene-specific translational regulations, cooperating with transcriptional regulations, in response to carbon (C), nitrogen (N), and phosphate (P) limitations. Furthermore, we found that codon bias contributes substantially to gene-specific translational regulation. Our findings broaden the understanding of translational regulation under environmental changes and may assist in the design of effective translation strategies in synthetic biology.