Collaboration of transcription factors (TFs) and their recognition motifs in DNA is the result of coevolution and forms the basis of gene regulation. However, the way how these short genomic sequences contribute to setting the level of gene products is not understood in sufficient detail. The biological problem to be solved by the cell is complex, because each gene requires a unique regulatory network in each cellular condition using the same genome. Thus far, only some components of these networks have been uncovered. In this review, we compiled the features and principles of the motif grammar, which dictates the characteristics and thus the likelihood of the interactions of the binding TFs and their coregulators. We present how sequence features provide specificity using, as examples, two major TF superfamilies, the bZIP proteins and nuclear receptors. We also discuss the phenomenon of “weak” (low affinity) binding sites, which appear to be components of several important genomic regulatory regions, but paradoxically are barely detectable by the currently used approaches. Assembling the complete set of regulatory regions composed of both weak and strong binding sites will allow one to get more comprehensive lists of factors playing roles in gene regulation, thus making possible the deeper understanding of regulatory networks.

转录因子（TFs）及其在DNA中的识别基序的协作是协同进化的结果，并构成基因调控的基础。然而，这些短基因组序列如何有助于设定基因产物水平的方式尚未得到足够详细的了解。细胞要解决的生物学问题很复杂，因为每个基因在使用相同基因组的每种细胞条件下都需要一个独特的调节网络。到目前为止，仅发现了这些网络的某些组件。在这篇综述中，我们汇编了基序语法的特征和原理，该特征和原理决定了特征性，从而决定了结合的TF及其共调节子相互作用的可能性。我们以两个主要的TF超家族为例，介绍序列特征如何提供特异性，bZIP蛋白和核受体。我们还讨论了“弱”（低亲和力）结合位点的现象，这似乎是几个重要的基因组调控区域的组成部分，但自相矛盾的是，当前使用的方法几乎无法检测到。组装由弱结合位点和强结合位点组成的完整调控区域，将使人们可以获得在基因调控中发挥作用的因子的更全面列表，从而使人们对调控网络有更深入的了解。