Uncovering cis regulatory codes using synthetic promoter shuffling

Abstract

Revealing the spectrum of combinatorial regulation of transcription at individual promoters is essential for understanding the complex structure of biological networks. However, the computations represented by the integration of various molecular signals at complex promoters are difficult to decipher in the absence of simple cis regulatory codes. Here we synthetically shuffle the regulatory architecture--operator sequences binding activators and repressors--of a canonical bacterial promoter. The resulting library of complex promoters allows for rapid exploration of promoter encoded logic regulation. Among all possible logic functions, NOR and ANDN promoter encoded logics predominate. A simple transcriptional cis regulatory code determines both logics, establishing a straightforward map between promoter structure and logic phenotype. The regulatory code is determined solely by the type of transcriptional regulation combinations: two repressors generate a NOR: NOT (a OR b) whereas a repressor and an activator generate an ANDN: a AND NOT b. Three-input versions of both logics, having an additional repressor as an input, are also present in the library. The resulting complex promoters cover a wide dynamic range of transcriptional strengths. Synthetic promoter shuffling represents a fast and efficient method for exploring the spectrum of complex regulatory functions that can be encoded by complex promoters. From an engineering point of view, synthetic promoter shuffling enables the experimental testing of the functional properties of complex promoters that cannot necessarily be inferred ab initio from the known properties of the individual genetic components. Synthetic promoter shuffling may provide a useful experimental tool for studying naturally occurring promoter shuffling.

Figure 1. Description of synthetic promoter shuffling scheme.

(A) Examples of promoters binding activator ‘A’ (left), repressor ‘R’ (middle) or both, activator and repressor (right). ‘Pol’ labels the RNAP. Activators nearly universally bind upstream of the −35 region in order to make direct, non-interfering contact with RNAP to help it bind. Repressors, on the other hand, can bind anywhere in the immediate region to successfully block RNAP binding. (B) Promoter shuffling scheme. We dissected the bacterial promoter into three regions encoding operator binding sequences (Upstream, Core, Downstream). The −35 and −10 RNAP binding sites and ribosomal binding site (‘RBS’) are indicated. The library consists of double-stranded DNA fragments with region-specific three-nucleotide overhangs allowing for ordered ligation to each other and, collectively, to a backbone vector. The complex promoter controls expression of a yfp gene. (C) Chromosomal insert of transcriptional regulators. araC, lacI, and tetR are transcribed from constitutive promoters, while λcI is regulated by an arabinose inducible promoter P_BAD. Also indicated are the transcriptional terminators t₀, rrnB T_{1 2}, and T1; the gene for spectinomycin resistance spec^r; and the λ phage attachment site, attB.

Figure 2. Promoter fragments used to construct combinatorial library.

The first six rows correspond to the original promoters on which the library is based, the two activators: P_A+ and P_λ+, and the four repressors: P_λ-, P_L1, P_L2, and P_T. Promoter fragments ‘Upstream’, ‘Core’, and ‘Downstream’ of the −35 (blue) and −10 (red) regions are displayed. Each fragment has unique three-nucleotide overhangs, allowing properly-ordered assembly upon ligation to each other and the plasmid backbone. Binding regions of specific regulators are underscored and labeled. “Additional Binding Sites” refers to additional promoter fragments that were created to expand the library. The lone nucleotide in green upstream of the −35 site in P_L2 indicates the accidental insertion of a ‘T’ when we designed this promoter fragment; it has negligible effect on the strength of repression by LacI.

Figure 3. Promoter architectures and transcriptional logic phenotypes of the forward-designed promoter library.

Rows represent different promoter architectures (F1–F29). The two columns labeled ‘Activators’ and ‘Repressors’ indicate the number of distinct activators and repressors capable of binding each promoter. Columns labeled ‘Up’, ‘Core’, and ‘Down’ indicate the three specific DNA fragments coding for various operators (see Fig. 2 for fragment sequences). Parentheses indicate a sequence lacking TR binding sites. For each promoter architecture, gene expression levels are represented by fluorescence measured for individual clones grown in eight wells corresponding to all eight different conditions of binding/nonbinding (+/−) of LacI (L), TetR (T), and AraC/λcI (A/λ). For clarity, we show only the relevant growth conditions for each promoter (expression levels were dependent only on the presence/absence of regulator specific inducers). Fluorescence was determined at an optical density (600 nm) of 0.3. Each row is normalized to the minimum (‘0’, red) and maximum (‘1’, green) fluorescence values for that particular promoter, with the actual minimum and maximum values given in the accompanying histogram to the right (all minimum values were very low and consistent with control cells lacking the yfp gene).

Figure 4. Simple thermodynamic models of NOR and ANDN logic phenotypes.

(A) Two repressors generate a NOR logic. (B) One repressor and one activator generate an ANDN logic. The model assumes non-interactive regulators that compete with (repressors) or help (activators) RNA polymerase bind to the promoters (see Materials and Methods for details). As the concentration of the active repressors or activators (x and y axis) is varied, the transcription of the gene they control changes (z axis).

Figure 5. Bar plots of specific promoters shown in Fig. 3.

(A) Two-input NOR promoters. (B) Two-input ANDN promoters. (C) Three-input NOR promoter. (D) Three-input ANDN promoter.

Figure 6. Randomly assembled promoter library.

Color chart of 29 promoters created through random assembly (same plotting conventions as in Fig. 3). There are three sets of identical sequences (M14/M15, M3/M29, and M20/M26), leaving 26 unique sequences all different from the forward-designed sequences shown in Fig. 3. Promoters are grouped according to effective behavior of components, ordered from the cleanest implementation of the logic to the fuzziest.

Figure 7. Two-input Boolean logic functions at the single promoter level in bacteria.

Boolean logic functions are listed in the first column with their corresponding formal definitions given in the second column. The next two columns indicate the number of distinct activators and repressors required to generate the logic phenotype. The second to last column displays various molecular schemes mostly proposed by Buchler et al. (marked by an asterisk) for implementing specific Boolean logic function at complex promoters. Many of these logic functions require intricate molecular schemes involving either regulator cooperativity or the presence of alternative promoters . Hermsen et al have recently proposed molecular implementations based on cooperative/competitive binding of several TRs for all of these logics. The last column displays Boolean output as a function of the binding states of two transcriptional regulator, TR1 and TR2, inputs at the promoter: ‘+’ (bound) and ‘−’ (not bound). Collectively, these functions represent a complete set of all two-input Boolean functions having outputs that depend on the state of both inputs.