Fundamentals and Exceptions of the LysR-type Transcriptional Regulators

Abstract

LysR-type transcriptional regulators (LTTRs) are emerging as a promising group of macromolecules for the field of biosensors. As the largest family of bacterial transcription factors, the LTTRs represent a vast and mostly untapped repertoire of sensor proteins. To fully harness these regulators for transcription factor-based biosensor development, it is crucial to understand their underlying mechanisms and functionalities. In the first part, this Review discusses the established model and features of LTTRs. As dual-function regulators, these inducible transcription factors exude precise control over their regulatory targets. In the second part of this Review, an overview is given of the exceptions to the “classic” LTTR model. While a general regulatory mechanism has helped elucidate the intricate regulation performed by LTTRs, it is essential to recognize the variations within the family. By combining this knowledge, characterization of new regulators can be done more efficiently and accurately, accelerating the expansion of transcriptional sensors for biosensor development. Unlocking the pool of LTTRs would significantly expand the currently limited range of detectable molecules and regulatory functions available for the implementation of novel synthetic genetic circuitry.

Introduction

The first synthetic genetic circuits was created at the turn of the millennium^1,2 and brought forth the field of synthetic biology.³ The rapid progress of this field is largely attributed to integration of knowledge from diverse disciplines, including microbiology, genetics, genomics, chemistry, physics, information technology, along with general biology and engineering.³ As the field rapidly transitions from its nascent stages to a more advanced scientific discipline, an array of new applications and tools has come forth.⁴⁻⁸ Among these synthetic biology achievements, biosensors have received a great deal of attention over the last years. By leveraging the extensive sensory apparatuses from microorganisms, biosensors enable monitoring of intracellular processes and environmental signals, as well as regulating gene expression in a precise and tunable way.

Such biosensors are composed of input, processing and output parts, as shown in Figure 1A. The input can be a wide range of different sensor molecules, e.g., riboswitches, ribozymes, two-component systems and transcription factors. This review focuses on transcriptional biosensors, more specifically on allosteric transcription factors (TFs). Although two-component systems have demonstrated their utility in biosensing extracellular signals,⁹⁻¹¹ one-component TFs have proven to be highly useful and efficient in a diverse set of synthetic genetic circuits.¹²⁻¹⁵ In addition, the bacterial TF landscape is vast and diverse, with a higher abundance of regulators in comparison to two-component systems,¹⁶ where they exhibit important physiological functions via an array of regulation mechanisms. This large variety makes TFs an ideal source of input sensor parts.

Figure 1.

(A) Schematic overview of transcription factor-based biosensors and their composition. For simplicity, the illustration is limited to inducible activators. The input is comprised of a transcription factor that can sense a ligand of interest. Signal processing occurs during DNA–protein interaction, where the transcription factor alters the expression of a target promoter depending on the input ligand concentration. This target promoter is responsible for the expression of an output gene that is chosen based on the envisioned biosensor application. (B, C) The two defining characteristics of these transcription factor-based biosensors are the response curve and ligand specificity. The response curve describes how biosensors link sensed ligand concentration to the output of choice. The ligand specificity determines how this response curve varies for different ligands. The response curves are given for a range of ligands, wherein the transcription factor exhibits low specificity (gray), as opposed to the high specificity observed for the ligand of interest (orange). Genetic circuit parts are given according to SBOL conventions.^41,42 TF = transcription factor, TFBS = transcription factor binding site, FP = fluorescent protein, LOQ = limit of quantification, ULOQ = upper limit of quantification.

For TF-based biosensors, processing takes place at the protein–DNA interface, i.e., the transcription factor binding sites (TFBS), as shown in the middle panel of Figure 1A. Here, the regulator alters transcription of an output of choice, based on the presence of an input signal in the form of a ligand. By changing the output signal, biosensors can be used for different applications, which have been extensively reviewed elsewhere, including high-throughput screening of intracellular metabolites,¹⁷⁻²¹ biomedical diagnostics,^22,23 environmental pollutant monitoring,²²⁻²⁵ adaptive laboratory evolution²⁶⁻³¹ and dynamic pathway control.^17,32−40

The two main characteristics that define biosensor circuits are the response curve (see Figure 1B) and ligand specificity (see Figure 1C), and are highly dependent on the combination of input, processing and output parts. The response curve of a biosensor describes the relation between a sensed input and the resulting output.^20,43−45 Within a limited range of ligand concentrations, known as the operational range (see highlighted x-axis range in Figure 1B), a biosensor’s response undergoes significant changes, constituting its dynamic range (see highlighted y-axis signal range in Figure 1B).⁴⁴ The lowest limit of quantification (LOQ) is preceded by the “OFF”-state of the biosensor, where any observed output is solely caused by leaky expression of the target promoter, rather than ligand recognition. The “ON” state is reached at the upper limit of concentration quantification (ULOQ). Higher ligand concentrations will not result in a further increase of the biosensor output due to saturation of the system.^20,44 The rate at which the output varies with increasing inducer concentrations characterizes the biosensor’s sensitivity, as shown as the slope in Figure 1B.^46,47 It is strongly influenced by the effect of multiple ligand molecules binding to the transcriptional regulator, which is referred to as cooperativity.⁴⁵ Finally, specificity of a TF can be defined as the affinity of this regulator for a ligand compared to other small molecules, resulting in a more pronounced change in output signal for this ligand.^46,47 This is shown in Figure 1C, where only one ligand results in a significant change in output within the relevant concentration range.

Despite the rich diversity in transcriptional regulators available in nature, and the advancements made on the customization of their characteristics toward applications within the biosensor context, a true repertoire of input sensors is still missing, both in terms of ligand specificity and regulatory mechanism.^17,48,49 To address this challenge, the mostly untapped potential of LysR-type transcriptional regulators (LTTRs), the largest family of bacterial TFs, is being explored. While they show interesting features for biosensor development, including a broad range of ligands and tight gene regulation, research and application of LTTRs has been relatively limited. This is partially ascribed to the complexity of their regulatory systems and the scattered available information. In the following sections, an overview of LTTR family characteristics is provided, shedding light on their intricate mechanisms to illustrate their utility in synthetic genetic circuits, particularly in the context of biosensors.

The LysR-type Transcriptional Regulators as Key Regulators for Bacterial Life

The one-component TFs are further classified in protein families, of which the LTTRs constitutes the largest collection of bacterial transcriptional regulators known to date.⁵⁰⁻⁵² The family was established in 1988, when nine transcriptional regulators where grouped due to similarities in their DNA-binding domain (DBD),^53,54 and named after the most well studied TF of the group at the time, i.e., lysine biosynthesis regulator LysR.^55,56 Since then, the group has grown tremendously, with over 852,000 predicted LTTR sequences reported on the InterPro database (IPR000847 ‘Transcription regulator HTH, LysR’). Similarly to other TF families, newly discovered regulators are assigned as LTTRs due to sequence similarities, mainly in the highly conserved DBD, where all LTTRs show a winged helix-turn-helix (wHTH) motif. The family is ubiquitous in both Gram-positive and Gram-negative bacteria,^57,58 and functional orthologues are found in archaea and algal chloroplasts.^57,59−63 In addition to their widespread occurrence, LTTRs can represent up to 20% of the regulators within one microorganism.^51,64 As an example, Pseudomonas aeruginosa PA01 is reported to contain over 100 different LTTRs.⁶⁵ The group most likely emerged from a common ancestor at an early evolutionary stage, given the widespread distribution of the regulators over extensive genetic distances. The evolutionary complexity of the group is further underscored by its ability to recognize a wide array of molecules, which requires a significant number of evolutionary steps to be achieved.^66,67

The diversity and abundance of this TF family resembles their involvement in all aspects of the microbial life.⁶⁸⁻⁷⁰ LTTRs play an important role in the regulation of key processes often revolving around the cellular response to environmental changes. This also becomes apparent from Table 1, which gives an overview of the currently characterized LTTRs and their biological relevance. It is hypothesized that the evolution of LTTRs was driven by the necessity to respond to these environmental stimuli, hence their ubiquitous presence, especially in free-living microorganisms. Due to continuous exposure to fluctuating environments, such microorganisms require constant adaptations, fueling the evolution of regulatory molecules such as LTTRs.⁶⁷ Examples of key pathways regulated by members of the LTTR family include stress response,^57,71 motility and attachment,^{56,57,66,71,72} quorum sensing and metabolic signaling,^57,71 antibiotic resistance,^58,73 fixation of both CO₂ and N₂,⁷⁴ biosynthesis of amino acids,^75,76 catabolism of aromatic compounds,^{56,57,66,71,72} transport and secretion,^{56,57,66,71,72} virulence factor expression,⁷⁷ initiation of nodulation⁷⁸ and cell division.⁷⁹ Taken together, the family’s substantial size and involvement in a wide diversity of pathways indicates a large potential for biotechnological applications, as it encompasses sensors responsive to an extensive range of compounds. This is further substantiated by the large amount of LTTRs in Actinobacteria, Proteobacteria, and Firmicutes, three phyla that are known for the synthesis of industrially relevant bioproducts.⁸⁰ Examples of such interesting molecules sensed by members of the LTTR family are given in Table 1.

Table 1. Overview of the Currently Characterized LysR-type Transcriptional Regulators and the Different Elements Required for Converting Them into Transcription Factor-Based Biosensors^a.

identifier	origin	biological relevance	target ligand	regulatory target	reference
AalR	Acinetobacter baylyi ADP1	aspartate metabolism	l-aspartate	PaspA, PaspT, PaspY, PracD	(81)
AceR	Acinetobacter baumannii	multidrug efflux pump regulation	chlorhexidine	PaceI	(82, 83)
AdmX	Serratia plymuthica A153	antibiotic synthesis regulation	indole-3-acetic acid	PadmV (operon)	(84−86)
AllS	Escherichia coli	anaerobic allantoin metabolism	allantoin	PallD	(87−89)
AlsR	Bacillus subtilis	acetoin synthesis	acetate	PalsS	(73, 90−92)
AmpR	Citrobacter freundii	β-lactamase synthesis	1,6-anhydro-MurNAc-peptides	PampC	(73, 93−98)
AphB	Vibrio cholerae	virulence	unknown	PtcpPH	(99−103)
AphT	Comamonas testosteroni TA441	aromatic compound degradation	2-hydroxymuconic semialdehyde	PaphCEFGHJI	(104)
ArgP	Escherichia coli	arginine transport	arginine	PargK	(105−107)
AtzR	Pseudomonas spp.	cyanuric acid catabolism	cyanuric acid	PatzDEF	(25, 108, 109)
BauR	Pseudomonas aeruginosa	polyamine utilization	β-alanine	PbauABCD	(110)
BenM	Acinetobacter spp.	aromatic compound degradation	cis,cis-muconate or benzoate	PbenABCDE, PbenPK	(111−118)
BlaA	Streptomyces spp.	β-lactamase synthesis	β-lactam compounds	PblaL	(119)
CatM	Acinetobacter calcolaceticus	catechol catabolism	cis,cis-muconate	PcatBCIJFD, PcatA, PbenPK	(71, 112, 120, 121)
CatR	Pseudomonas putida	catechol catabolism	cis,cis-muconate	PcatBCA, PpheBA	(71, 122−126)
CbbR	Xanthobacter flavus	carbon dioxide fixing	ribulose 1,5-bisphosphate	PcbbLSXFPTAE	(74, 127, 128)
Cbl	Escherichia coli	sulfur acquisition and reduction	thiosulfate	PcysP, PcysT, PtauABCD, PssuEADCB	(129−131)
CbnR	Ralstonia spp.	chlorocatechol catabolism	2-chloro-cis,cis-muconate	PcbnABCD	(17, 71, 132−134)
CcpC	Bacillus subtilis	carbon catabolite control	citrate	PcitB, PcitZ	(135, 136)
CcpE	Staphylococcus aureus	carbon catabolite control	unknown	PcitE	(30, 137)
CfxR	Alcaligenes eutrophus	carbon dioxide fixing	unknown	PcfxLSXYEFP	(138)
ChiR	Serratia marcescens	chitin degradation	unknown	PCBP21	(139)
CidR	Staphylococcus aureus	murein hydrolase regulation	acetic acid	PcidABC	(140)
CitR	Bacillus subtilis	citrate metabolism	unknown	PcitA	(141)
ClcR	Pseudomonas putida	chlorocatechol catabolism	2-chloro-cis,cis-Muconate	PclcABD	(71, 123, 142, 143)
CmpR	Synechococcus elongatus	CO2 fixing	ribulose 1,5-bisphosphate	PcmpABCD	(144−146)
CrgA	Neisseria meningitidis	pili/capsule synthesis and oxidative stress response	α-methylene-γ-butyrolactone	PmdaB	(147−149)
CynR	Escherichia coli	cyanate detoxification	cyanate	PcynTSX	(150)
CysB	Salmonella enterica serovar Typhimurium	cysteine biosynthesis	N-Acetylserine	PcysI	(130, 151−155)
CysL	Bacillus subtilis	sulphite reductase	sulfate, sulfite, thiosulfate	PcysJI	(156)
DarR	Acinetobacter baylyi ADP1	aspartate metabolism	d-aspartate	PaspA, PaspT, PaspY, PracD	(157)
DbdR	Thauera aromatica AR-1	aromatic compound degradation	3,5-dihydroxybenzoate	PdbhL	(158)
DntR	Burkholeria spp.	aromatic compound degradation	2,4-dinitrotoluene, salicylate	PdntA	(50, 159−161)
FdeR	Herbaspirillum seropedicae	aromatic compound degradation	naringenin	PfdeA	(162, 163)
FinR	Pseudomonas putida	ferredoxin-NADP+ reductase regulation	oxidative and osmotic stresses	Pfpr	(164)
GcvA	Escherichia coli	glycine metabolism	glycine and adenine	PgcvTHP, PgcvB	(165, 166)
GigC	Acinetobacter baumannii	cysteine metabolism and virulence	l-cysteine	PcysI	(167)
GltC	Bacillus subtilis	glutamate synthase	α-ketoglutarate	PgltA	(168−170)
HexA	Photorhabdus luminescens	phenotypic heterogeneity, pathogen–symbiont transition	unknown	PpcfABCDEF	(171)
HsdR	Comamonas testosteroni	aromatic compound degradation	testosterone	PhsdA	(172, 173)
HupR	Vibrio vulnificus	haem uptake	haemin	PhupA	(174)
HvrB	Rhodobacter capsulatus	S-adenosyl-l-homocysteine hydrolase expression	light sensitivity	PahcY	(175)
HypT	Salmonella enterica serovar Typhimurium	HOCl response	HOCl	PfhuA	(176)
IlvR	Caulobacter crescentus	isoleucine/valine biosynthesis	unknown	PilvD	(177)
IlvY	Escherichia coli	isoleucine/valine biosynthesis	α-acetolactate, α-acetohydroxy-butyrate	PilvC	(178)
IrgB	Vibrio cholerae	iron-regulated virulence factor	unknown	PirgA	(179)
KaeR	Lactobacillus brevis	flavonoid response	kaempferol	PLVIS1988	(180)
LeuO	Salmonella enterica serovar Typhimurium	bacterial stringent response	unknown	PompS1, PompS2, PassT	(70, 181−183)
LinR	Sphingomonas spp.	aromatic compound degradation	2,5-dichlorohydroquinone, 2,6-dichlorohydroquinone, chlorohydroquinone	PlinE	(71, 184, 185)
LrhA	Escherichia coli	flagella, motility and chemotaxis	unknown	PflhDC	(186)
LttR	Lactobacillus plantarum	conjugated linoleic acid production	linoleic acids	Pcla-dh (operon)	(187)
LysG	Corynebacterium glutamicum	amino acid export	l-lysine and other amino acids	PlysE	(188, 189)
LysR	Escherichia coli	lysine biosynthesis	diaminopimelate	PlysA	(55)
MdcR	Klebsiella pneumoniae	malonate catabolism	malonate	PmdcABCDEGHLM	(190, 191)
MetR	Streptococcus spp.	methionine and cysteine transport/biosynthesis	homocysteine	PmetE	(192−198)
MleR	Streptococcus mutans	malolactic fermentation	l-malate	PmleSP	(199)
MvfR (= PqsR)	Pseudomonas aeruginosa	pathogenicity regulator	4-hydroxy-2-heptylquinone	PpqsA	(200−205)
NagR	Ralstonia eutropha	naphthalene catabolism	salicylate	PnagA	(206, 207)
NahR	Pseudomonas putida	naphthalene/salicylate catabolism	salicylate	PnahABCFDE, PsalGHINL	(208−210)
NdhR (= CcmR)	Synechocystis spp.	CO2-concentrating mechanism	2-phosphoglycolate, 2-oxoglutarate	Pndh3 (operon)	(72, 211)
NhaR	Escherichia coli	Na+/H+ antiporter regulation	Na+	PnhaA	(212)
NocR	Agrobacterium tumefaciens	nopaline catabolism	octopine	PnocPTQM	(213)
NodD	Rhizobium spp.	nitrogen fixation/symbiosis	flavonoids	PnodABCIJ, PnodFEL, PnodMNT, PnodO	(78, 214−217)
NtdR	Acidovorax spp.	aromatic compound degradation	2,4-dinitrotoluene, 2,6-dinitrotoluene, salicylate, anthranilate	PntdA	(71)
OccR	Agrobacterium tumefaciens	octopine catabolism	octopine	PoccQMPT	(218−222)
OdcR	Pigmentiphaga spp.	3,5-dibromo-4-hydroxybenozate catabolism	3-bromo-4-hydroxybenzoate, 3,5-dibromo-4-hydroxybenozate	PodcA, PodcB, PodcC	(223)
OxyR	Escherichia coli	oxidative stress response	redox changes	PkatG, PaphCF, PoxyS, PgorA, PgrxA	(224−231)
PA2206	Pseudomonas aeruginosa	oxidative stress response	redox changes	PpvdS	(232)
PcaQ	Agrobacterium spp.	protocatechuate catabolism	β-carboxy-cis,cis-muconate, γ-carboxymuconate	PpcaDCHGB	(71)
PecT	Dickeya dadantii	virulence	temperature changes	PpelB, PpelC, PpelD, PpelE	(233, 234)
PhcA	Pseudomonas solanacearum	virulence	unknown	PT3SS	(235, 236)
PltR	Pseudomonas protegens	pyoluteorin synthesis	pyoluteorin	PpltLABCDEFG	(237, 238)
QuiR	Listeria monocytogenes	protocatechuate biosynthesis	shikimate	Plmo0489 (operon)	(239)
RbcR	Chromatium vinosum	carbon dioxide fixing	unknown	PrbcAB	(240)
RipR	Salmonella enterica	itaconic acid resistance	isocitrate	PripCBA	(241,242)
RovM	Yersinia pseudotuberculosis	invasion/motility, virulence	unknown	ProvA	(243, 244)
S cmR	Burkholderia thailandensis	secundary metabolism regulation	unknown	Phmq (operon)	(68, 69)
SpvR	Salmonella typhimurium	virulence	unknown	PspvABCD	(245, 246)
SyrM	Rhizobium meliloti	exopolysaccharide synthesis	unknown	PsyrA	(247)
TcbR	Pseudomonas spp.	chlorocatechol metabolism	2-chloro-cis,cis-muconate	PtcbCDEF	(71, 248)
TfdR	Ralstonia spp.	chlorophenol catabolism	2-chloro-cis,cis-muconate	PtfdCDEFB, PtfdA	(71)
TfdT	Burkholderia spp.	aromatic compound degradation	3-chlorocatechol, 4-chlorocatechol, 2-chlorobenzoate, 3-chlorobenzoate	PtfdC	(71)
ThnR	Sphingopyxis granuli	aromatic compound degradation	tetralin	PthnB, PthnC	(249, 250)
ToxR	Burkholderia glumae	toxoflavin biosynthesis	toxoflavin	PtoxABCDE, PtoxFGHI	(251)
TsaR	Comamonas testosteroni	aromatic compound degradation	p-tolunesulfonate	PtsaMBCD	(252−254)
TtuA	Agrobacterium vitis	tartrate metabolism	l-tartrate	PttuB	(255)
VirR	Rhodococcus equi	virulence	unknown	Porf5 (virR operon)	(190)
VqsA	Vibrio alginolyticus	virulence	unknown	PaphA, PluxR	(256)
VtlR (= LsrB)	Agrobacterium tumefaciens	regulation of motility and biofilm formation	unknown	PabcR2	(257−259)
YbeF	Escherichia coli	flagella biosynthesis, putative citrate utilization related 2	unknown	unknown	(260)
YbdO	Escherichia coli	citrate utilization related, flagella biosynthesis	unknown	PybdO	(260)
YbhD	Escherichia coli	l-malate utilization related	unknown	PybhH, PybhI	(260)
YcaN	Escherichia coli	unknown	unknown	PycaC, PycaD	(260)
YeiE	Cronobacter sakazakii	sulfite resistance	sulfite	PcysJI	(261)
YgfI	Escherichia coli	dihydroxyacetone, glycerol or Thr utilization	unknown	PdhaK PyjiT, PpflB, PadhE, PhycBCDEF, PnarZ	(260)
YiaU	Escherichia coli	membrane modification and LPS biosynthesis	unknown	PwaaPSBOJYZU PyjiT, PadeP, PyiaT, PgltD	(260)
YneJ	Escherichia coli	putrescine utilization related	unknown	Psad,PfnrS	(260)
YofA	Bacillus subtilis	cell division	unknown	PftsW	(79)
YtxR	Yersinia entercolitica	ADP-ribosyl-transferase toxin	unknown	PytxAB	(262)

^aFor each regulator, the organism of origin is given, as well as the biological context in which it is active. When available, a ligand recognized by the regulator is provided, as well as target promoters. As these transcription factors often function as global regulators, the “regulatory target” column does not represent all target promoters.

The large variety in bacterial sensor molecules within the LTTR family, shown in Table 1, is highly advantageous for biosensor development. It results in a broad range of input molecules that are readily sensed by transcriptional regulators. Moreover, it makes the LTTRs an interesting starting point for engineering new specificities, as this is done more efficiently when starting from similar compounds. For instance, this approach was utilized in developing biosensors for various flavonoids,¹⁶³ or in engineering the specificity of BenM toward adipic acid.²⁶³ Beyond their diversity, LTTRs are known for stringent regulation over their target genes by binding promoters in both induced and uninduced states.^57,264−266 Next to their role as transcriptional activators upon ligand interaction, continuous association with target promoters causes repression in the uninduced state due to steric hindrance on key promoter elements. Additionally, this causes negative autoregulation, given that most LTTRs are colocalized with a member of their regulon.⁵⁷ This mechanism of transcriptional regulation results in low leaky expression and a fast response to inducing ligands,²⁶⁶ making them ideal inputs for biosensor development. The optimization of naringenin production in E. coli showed the importance if this precise gene control. The concentration of precursor malonyl-COA required careful balancing to prevent toxicity,²⁶⁷ which was achieved by embedding FdeR, a naringenin responsive LTTR, into a multilayer dynamic control system.⁴⁰

Despite the progress made in development of LTTR-based biosensors, it is apparent that the amount of characterized LTTRs and their applications is disproportional to the overall size of this TF family. This is largely due to the limited and scattered knowledge on these transcriptional regulators. Without a clear understanding of their functionalities and properties, applying them in specific contexts is challenging. Moreover, most applications necessitate optimizing the response curve to transform natural TF systems into effective sensors for the intended purpose. Therefore, comprehending the mechanisms underlying transcriptional regulation by LTTRs is essential for unlocking the full potential of this TF family.

A general mechanism for LTTR regulation has been proposed, but many exceptions to this rule exist, underscoring its limitations. Given the extensive size of the LTTR family, it is plausible that the currently ascribed mechanism is only applicable to a subset of the regulators. Therefore, it becomes necessary to reclassify the LTTR family into distinct subfamilies, which will facilitate a more precise allocation of structure and function. The subsequent sections will explore the ’classic’ LTTR regulatory mechanism, starting with an overview of the characteristic ’sliding dimer’ mechanisms. This will be followed by a more detailed discussion on the architecture of regulatory targets of LTTRs, as well as the structural properties of these proteins. Finally, an overview of the currently identified exceptions to this ’classic’ model will be presented, underscoring the need to re-evaluate this extensive transcription factor family.

Overview of the “Sliding Dimer” Mechanism

The regulatory mechanism of the LTTRs is referred to as “sliding dimer” and is visualized in Figure 2. When expressed, LTTR proteins undergo oligomerization, forming dimers that serve as the functional unit for DNA interaction. The targeted DNA sequence for this protein–DNA interaction comprises three binding sites, as shown in Figure 2A. In the absence of the inducing ligand, LTTR dimers interact with the regulatory site (RS) and activation site 1 (AS1). RS serves as an anchor due to highly stable protein–DNA interactions. In contrast, dimers bound to AS1 sites form stable but loose DNA interactions, which is important for the regulatory mechanism.^57,253 A dimer can only bind to AS1 while simultaneously dimerizing with the RS-bound dimer, a process known as cooperative binding.^{108,127,253,268} The tetrameric protein complex covers key promoter regions, as AS1 is situated near the −10 box of the target gene promoter. In addition, the TF-DNA interaction induces DNA bending between 50° and 100° due to the distance between the AS1 site and the RS site, the latter situated around −65 nucleotides from the transcription start site. Both the binding of key sequences and DNA bending cause inaccessibility of the target promoter, preventing RNA polymerase (RNAP) from interacting with it,^116,127,265 as illustrated in Figure 2B.

Figure 2.

The regulatory mechanism of the LysR-type transcriptional regulators is referred to as the “sliding dimer” mechanism, owing to the movement of transcription factor dimers upon ligand interaction. (A) As most LysR-type transcriptional regulators (blue) are expressed from a bidirectional promoter region with one of their target genes (gray), this genetic architecture is given in more detail. (B) In the uninduced state, the tetrameric protein complex binds regulatory site (RS) and activation site 1 (AS1), causing the DNA to bend and preventing the RNA polymerase (RNAP) complex to recognize the promoter sequence. In addition, the location of the transcription factor binding sites overlap with key promoter regions (−10 and −35 boxes) of the target gene, further preventing RNAP complex–DNA interactions via steric hindrance. Due to overlap in the expression systems of the regulator and target gene, the transcription factor also causes negative autoregulation. Upon recognition of the appropriate ligand, the protein complex undergoes conformational changes that forces the dimer bound at AS1 to move to activation site 2 (AS2). These conformational changes are illustrated in a frontal and top view, showing the movement of the monomers within the tetrameric structure upon ligand interaction. As a result, the DNA bending is reduced and RNAP interaction sites become exposed, enabling activation of the target promoter transcription. On the surface of the ligand-bound regulator, RNAP recruitment sites appear after the conformational changes, further facilitating and guiding the interaction between RNAP and the target promoter. Genetic circuit parts are given according to SBOL conventions.^41,42 TF = transcription factor, RS = regulatory site, AS = activation site, RNAP complex = RNA polymerase with appropriate sigma factor, LTTR = LysR-type transcriptional regulator.

Upon ligand interaction, the tetrameric structure tightens due to interactions among residues within the ligand-binding domain (LBD) and the ligand. These conformational modifications within the protein structure are transduced to the DBD,¹¹⁶ and forces the dimer bound to AS1, to transition to the AS2 site.^50,116,127 The latter is located upstream, proximate to the −35 box, but beyond the crucial promoter regions. By moving to AS2, the protein complex exposes the promoter for RNAP, and reduces the DNA bending to a range of 9–50°. Furthermore, the induced tetramer actively recruits RNAP. The conformational changes cause a migration of internal protein sequences to the tetramer surface, which interact with the α-subunits of RNAP.^{116,127,253,268,269}

Transcription Factor Binding Sites and Gene Architecture

The regulatory mechanism of LTTRs involves three different sites for TF interactions, namely the high affinity regulatory binding site RS and lower affinity activating binding sites AS1 and AS2, with the latter two usually found together and often overlapping in sequence.²⁴⁹ This is visualized in Figure 3, which shows the genetic architecture of target promoter regions of LTTRs. The differences in affinity between binding sites arise from variations in the nucleotide sequence,^116,223 as discussed below and shown in Figure 3. The RS site, usually situated between −80 to −50 bp from the target gene transcription initiation site, is seen as an anchor for the TF complex.^{66,127,253,268} Its high affinity allows the interacting dimer to form a stable interaction, subsequently enabling a second dimer to bind to the lower affinity AS1 sequence. The latter is situated downstream of the RS site and overlaps key transcription control elements of the targeted promoter, such as the −35 and −10 boxes. While binding of an LTTR dimer onto the AS1 site can occur, it is believed to be unstable until the dimer oligomerizes with the RS bound dimer.⁵⁶ This loose interaction between the regulator complex and AS1 enables the “sliding dimer” mechanism upon ligand induction, as shown in Figure 2. In addition to the differences in sequence within one promoter, TFBS sequences can also vary over different target promoters. This allows divergent regulation of LTTRs over their regulon, adding an extra layer to the transcriptional regulation.

Figure 3.

Overview on the transcription factor binding sites of LysR-type transcriptional regulators. The target promoters of these regulators show three different binding sites, namely the regulatory site (RS) and two activation sites (AS1 and AS2). These regulatory sites follow the characteristic T-N₁₁-A sequence, with symmetry in the nucleotides surrounding the thymine and adenine of this sequence. During DNA binding, the RS site is first bound by a dimer of regulator molecules. The high degree of symmetry in this site, as shown in bold for regulators BenM, DntR, and OccR, allows for a stable and strong interaction between the dimer and the DNA. A second dimer binds AS1, which shows a much lower degree of dyad symmetry. Hence, binding at this site is only possible while simultaneously forming the full tetrameric structure together with the RS-bound dimer. Due to the distance between RS and AS1, the tetramerization and DNA binding is accompanied by DNA bending. Similarly to the variation in AS1 sequence, AS2 also shows imperfect symmetry. The AS1 and AS2 sites cover important promoter regions, such as the −10 and −35 box. The location of these boxes, as well as the transcription start site, are given per promoter for both the target promoter system (gray) and the transcription factor expression system (blue). Genetic circuit parts are given according to SBOL conventions.^41,42 TF = transcription factor, RS = regulatory site, AS = activation site.

In general, TFBS sequences of LTTRs have a characteristic T-N₁₁-A conserved sequence that shows imperfect dyad symmetry.^66,265,270 The degree of symmetry within the recognition region determines the strength of TF binding,^{66,127,253,268,271} with a certain degree of asymmetry required for effective regulation.²⁷² The T-N₁₁-A motive is important for guiding the TF to the correct DNA sequence, but true recognition and specificity of TF-DNA interactions is caused by the surrounding nucleotides.⁶⁶ For example, the recognition logo for BenM was expanded to ATAC-N₇-GTAT due to the importance of these neighboring nucleotides,¹¹⁶ which is depicted in Figure 3. To enable the sliding of a dimer from AS1 to AS2, the interaction at these sites should be stable but weaker than the RS-dimer interaction. To facilitate this, AS sequences tend to show less dyad symmetry than the RS sequence, as seen for BenM in Figure 3, where both AS sites show one nucleotide difference with the RS site. These variations can be more drastically different, as is seen for regulator DntR. The AS sites of P_dntA, which are given in Figure 3, could only be found via experimental methods, as they do not resemble the classic LTTR binding site sequence. In contrast, the RS shows a high dyad symmetry T-N₁₁-A sequence.⁵⁰ Similarly, the discovery of AS sites of OccR was impaired due to overlap of these sites.^213,218,221 Nevertheless, this information is important for optimization of LTTR-based biosensors, as different parameters of the response curve, discussed in Figure 1B, can be significantly altered by changing the expression level of either the TF or target promoter, or altering the TFBS sequences.^20,162,273

Initially, due to their co-occurrence with target genes, as illustrated in Figure 3, LTTRs were described as cis-acting regulators. Further research into the regulons of various LTTR members has unveiled that a significant number of them function as global regulators.^158,265 Much of their transcriptional regulation is thus done as trans-acting regulators, but their colocalization with a specific target gene remains a characterizing feature of the family.⁵⁷ This genomic architecture results in the TF divergently transcribed from a target gene, surrounding a bidirectional promoter region with overlapping operator sequences, as shown in Figure 3. This is a common architecture in bacteria, which promotes efficient regulation of multiple genes due to the TFBSs spanning both promoters.^178,274 Hence, upon binding the target promoter, the LTTR simultaneously represses its own expression, also referred to as negative autoregulation (see Figure 2B). The mechanism behind the autoregulation is based on steric hindrance when the TF is bound to the bidirectional promoter,^{66,172,178,222} as the RS of the target gene overlaps with key regulatory elements of the TF promoter.¹¹⁶ This is postulated to prevent overproduction of TFs, which could put an unnecessary burden upon the cell.^{60,127,219,221,225,253,274} It has also been shown that negative autoregulation in combination with weak expression systems controlling TF synthesis, results in an increased sensitivity, while simultaneously reducing response time and lowering leaky expression.²⁶⁶ Both parameters are essential for a swift reaction upon ligand sensing, which is of great importance for the vital biological functions regulated by LTTRs.

In summary, the ’sliding dimer’ mechanism of LTTRs involves interactions at high-affinity regulatory binding sites and lower-affinity activating sites. These sites typically feature a T-N₁₁-A motif with varying dyad symmetry affecting binding strength and regulatory effectiveness. Next to the DNA–protein interactions, transcriptional regulation by LTTRs also depends on ligand-protein and protein–protein interactions, as discussed below.

Protein Structure and Oligomerization of LysR-type Transcriptional Regulators

At the TFBS, multiple monomers oligomerize into a tetrameric protein structure that forms the interface between the input ligands and output gene expression. This is achieved by a complex interplay between the different protein domains that make up LTTRs. For example, recent work on the connecting region between the DBD and LBD, referred to as the linker helix, showed that this connector has big influences on both DNA- and ligand recognition and binding.²⁷⁵ Furthermore, there are several regions on the LTTRs that are essential for their regulatory mechanism, such as the regions for oligomerization^{115,117,172,265} and RNAP-recruitment.^{123,165,193,208,276}

While this complexity hinders rational engineering of LTTRs,²⁶³ insights into the structure of both monomeric and multimeric forms facilitate the engineering and understanding of mutant regulators. For example, altering the specificity of LTTR regulators (see Figure 1C) is accelerated by knowledge on the ligand-binding pocket and the way these residues interact with the ligand of interest.^277,278 More recently, LTTRs have been subjected to domain swapping to combine DNA- and ligand binding properties of interest.¹⁶³ Here, structural information is vital to accurately define domains, design domain swapping chimera and to circumvent hurdles related to structural incompatibility.

Determining the protein structure of LTTRs has seen many hurdles.^115,117 Due to the structural flexibility of DBDs, initial attempts at obtaining crystal structures were limited to LBDs.^151,224,244 While providing valuable insights, the isolated domains lack information regarding overall protein structure and domain interactions.²⁵³ In addition to protein flexibility, oligomerization of these regulators further complicates structural studies due to the formation of aggregates during crystallization.¹¹⁶ Because oligomerization influences the structure of the monomers,²⁵³ higher order oligomers are needed to obtain valuable information on protein structure and quaternary interactions. However, tetramerization only occurs in the presence of DNA, and crystallization of LTTRs in interaction with DNA has proven difficult. The main problem is the high salt concentrations needed for stabilizing LTTR protein structures, which is a difficult environment for the crystallization of DNA complexes.¹³² As a result, only 16 full structures have been elucidated of this TF family. The available structures are given in Table 2.

Table 2. Overview of the LysR-type Transcriptional Regulators for Which the Full Structure Is Experimentally Determined^a279.

regulator	identifier	origin	ligand	reference
AphB	3SZP	Vibrio cholerae	pH, O2	(102)
ArgP	3ISP	Mycobacterium tuberculosis	l-arginine, l-lysine	(268)
BenM	3K1N	Acinetobacter baylyi	cis,cis-muconate, benzoate	(115)
CbnR	7D98	Cupriavidus necator	cis,cis-muconate	(132, 133)
CrgA	3HHG	Neisseria meningitidis	unknown	(149)
DarR	7DWO	Vibrio fischeri	d-aspartate	(280)
DntR	5AE5	Burkholderia cepacia	salicylate	(50)
Hink	6M5F	Pseudomonas aeruginosa	unknown	unpublished
HypT	5YDW	Salmonella enterica serovar Typhimurium	HOCl	(176)
LysG	6XTU	Corynebacterium glutamicum	l-arginine, l-lysine, l-histidine	(281)
NdhR/CcmR	5Y2V	Synechocystis sp.	2-phosphoglycolate, 2-oxoglutarate	(72)
OxyR	4X6G	Pseudomonas aeruginosa	H2O2	(228)
TsaR	3FXQ	Comamonas testosterone	p-toluenesulfonate	(253)
VV2_1132	5Y9S	Vibrio vulnificus	unknown	(282)
PA01	3FZV	Pseudomonas aeruginosa	unknown	unpublished
PA0477	2ESN	Pseudomonas aeruginosa	unknown	unpublished

^aInformation gathered from the RCSB PDB database (https://www.rcsb.org/).

The first full LTTR that has been crystallized was the cis,cis-muconate-responsive regulator CbnR from Cupriavidus necator.¹³³ Recently, further research on the structure of CbnR resulted in tetrameric protein complexes cocrystallized with target DNA, both in the presence and absence of the inducing ligand, for the first time in this family.¹³² This work has confirmed some of the hypotheses surrounding LTTR regulation such as the “sliding dimer” mechanism and the occurrence of conformational changes upon ligand binding. Due to the importance of the structure for functionality, it is further discussed below, starting from the monomeric protein up to the full tetramer protein complex. As the most extensively studied LTTR in terms of protein structure, the CbnR protein will be central in this discussion.

Monomer Domain Composition and Structure

Upon translation of the regulator gene, a monomeric TF unit is produced. None of the currently discovered LTTRs regulate gene expression as monomers, but its structure has important consequences for the higher order protein complexes. The encoded LTTR protein is roughly 300 amino acids long and is composed of several domains, namely the DNA-binding domain (DBD), the ligand-binding domain (LBD) and connecting the two is a linker helix (LH) and hinge (H)^66,283 as shown in Figure 4. While many of the available structural studies annotate the proteins on their different domains, the exact boundaries are not well-defined. Establishing such boundaries facilitates the development of chimeric transcription factors via engineering techniques such as domain swapping.^{278,284−286} Within the LTTR family, this has led to novel transcriptional regulators with custom DNA- and ligand binding specificities.¹⁶³

Figure 4.

(A) Annotation of the LysR-type transcriptional regulator (LTTR) protein. Using sequence and structural alignments, an amino acid consensus sequence was created, and specific boundaries of the protein domains were set. The sequence alignment was based on 134 unique and curated sequences from the Uniprot database (https://www.uniprot.org/),²⁸⁷ which can be found in Supplementary Table 1. The protein models used for the structural alignment were extracted from either the RCSB database (https://www.rcsb.org/)²⁷⁹ or the AlphaFold database (https://alphafold.com/),²⁸⁸ and are given in Supplementary Figure 1. (B) The structure of regulator CbnR is given, as well as a schematic illustration of the secondary structure of LTTRs, both annotated according to the domain definitions. The α-helices are given as squares and numbered from N-terminus to C-terminus, while the β-sheets are similarly given as arrows and letters A–F. Figure adapted from ref (265). DBD = DNA-binding domain, LH = linker helix, H = hinge, LBD = ligand-binding domain, RD = subdomain of the LBD, h = hydrophobic amino acids, X = any amino acid.

To further understand the domain organization of LTTRs, a sequence alignment was performed, of which the results are shown in Figure 4A. From the Uniprot database,²⁸⁷ 134 different LTTR sequences were extracted, which are given in Supplementary Table 1. The sequences were selected among the “Reviewed” protein entries within the ‘LysR transcriptional regulatory family’ section of the database. This results in 433 proteins, of which many are homologues with highly similar sequences. Hence, to prevent a bias toward such clusters during sequence alignment, only one sequence was chosen per regulator, resulting in a total of 134 proteins. The boundaries of the domains were subsequently found by combining the alignment information with available protein models and crystal structures, the latter given in Supplementary Figure 1. An overview of the LTTR protein structure with boundaries for the DBD, LH, H and LBD is presented in Figure 4A.

The DNA-Binding Domain

The DBD structure is visualized in Figure 4B. It is located at the N-terminus, spanning approximately 65 amino acids that fold into three α-helices and two β-strands. Within this domain, α₂, α₃ and both β-strands make up the LTTR-specific wHTH structure,^116,253,289 which is responsible for protein–DNA interactions. While there are important residues on α₂ for stabilizing the DNA–protein complex via interactions with the phosphate backbone, most of the selective residues are located in α₃.^133,149 The 13 base pair long TFBS spans two DNA major grooves that each hold conserved nucleotides, including the thymine and adenine of the T-N₁₁-A sequence logo. By entering deeply into the major grooves of the DNA, the α₃-helices interact with these two nucleotides, which are important for sequence recognition. Subsequently, the selective interactions enabling LTTRs to target their specific TFBSs, are made with the surrounding nucleotides.¹³⁴ The winged β-strands interact with adjacent DNA minor grooves, where they contact the sugar–phosphate moiety of the DNA backbone to provide stability for protein–DNA interaction.^{116,133,134,149} In addition, the wing assists in positioning the DBD’s α-helices to allow proper DNA recognition and binding.¹¹⁶ In total, a dimer of LTTRs interacts with roughly 25 nucleotides spread over both DNA strands.¹³⁴ Outside of the wHTH, α₁ is made up of a conserved hydrophobic core and less conserved polar residues at the surface facing the DNA sequences. The latter are believed to aid in the interaction with target DNA, although their exact function has yet to be determined.^56,133

The Ligand-Binding Domain

The 200 C-terminal residues of the regulator gene encode for the ligand recognition domain. The LBD is split into two subdomains, RD1 and RD2, connected by two short antiparallel β- strands, as shown in Figure 4B. RD1 is composed of two different parts of the polypeptide chain, as the protein loops back onto itself.¹⁵¹ The subdomains are structured in α/β Rossmann like folds, consisting of nine α-helices and ten β-strands.^121,290 At the interface of the subdomains, between the crossover strands, a cavity is formed that holds the ligand-binding cleft.^61,133,151 Due to the nature of inducing molecules varying greatly between members of the LTTR family, the architecture of the ligand-binding cleft can be vastly different.²⁶⁵ Strongly polar ligands form multiple hydrogen bonds with amino acids lining the cleft interior,^{111,137,159,239,253,261} while other ligands are engulfed in a hydrophobic cavity.^200,201 LTTRs can bind their target DNA in the absence of the inducing ligand, but this molecule is required for transcriptional activation of a target promoter. Upon recognition and association with the corresponding ligand, LTTRs alter the transcription of their target gene between 6- and 200-fold. Conformational changes occur when the inducing molecule enters the RD hinge region, changing the overall protein complex structure.¹³²

In contrast to the DBD, there is a much lower conservation noted in the LBD, both in sequence composition and in length. Therefore, determining those residues essential for structure or the key amino acids for ligand recognition has been proven difficult.¹¹¹ Different mutational studies highlighted the importance of both proximal and distal residues from the ligand-binding cleft for ligand recognition. This is likely due to the subsequent signal transduction, which is equally as important as ligand recognition for transcriptional activation of LTTRs.^{149,210,217,253,264,291} Furthermore, the C-terminus plays an important role in oligomerization and nonspecific DNA interactions,^66,71,253 and additional protein sites are important for transcriptional activation because they allow protein–protein interactions with RNAP. These sites are hidden in the uninduced protein complex and are exposed by the conformational changes following ligand interactions.²⁹² Next to the location of the mutation, the nature of the replacing residue in the LBD shows equal importance. As an example, comparing mutations of CysB, the cysteine biosynthesis regulator from Salmonella typhimurium, at the same site revealed that T149 M results in an activation upon cysteine binding comparable to the wild type, whereas T149P resulted in only 10% of this activity. The introduction of a proline reduces overall flexibility which potentially interferes with conformational changes required for the ligand inducing.¹⁵³

The Linker Helix and Hinge Region

Coinciding with the transition from DBD to LH domains, a highly conserved TxxG sequence is noted in the amino acid consensus sequence, shown in Figure 4A. The LH is a single α-helix domain that is important for correct homodimerization,⁶⁵ as well as positioning the DBDs for optimal contact with the TFBS. In contrast, the boundary between the hinge and LH is more difficult to establish. The hinge is defined here as the loop between the α-helix structure of the LH, and the first β-sheet of the LBD. Upon ligand interaction in the LBD, conformational changes in the LBD cause the DBDs of one dimer in the protein complex to move to a new TFBS. This occurs around the hinge, which transduces the signal of ligand interaction toward the DBDs via the LH. In addition, the hinge is believed to hold important residues for determining the degree of oligomerization. When analyzing the octameric regulator CrgA from Neisseria meningitidis, responsible for oxidative stress response, the hinge showed a more rigid structure due to the presence of a proline residue in comparison to the more mobile structure of the tetramer CbnR.¹⁴⁹ This difference in structural flexibility could influence the higher order oligomerization pattern.¹⁴⁹ The full extent of the influence of LH and hinge in the overall functionality of these regulators is not yet established. Small changes in the amino acid sequence of these domains have shown to cause changes in specificity¹⁶³ and altered strength of transcriptional regulation upon ligand induction,²⁷⁵ emphasizing the importance of gaining deeper insights into this underexplored part of the LTTR protein structure.

Understanding the structure and function of the different domains constituting LTTR monomers is essential for accurate application and further optimization of these regulators. The DBD, located at the N-terminus, is crucial for specific DNA interactions, while the LBD, at the C-terminus, recognizes and binds ligands, inducing conformational changes necessary for transcriptional activation. The LH and hinge regions facilitate dimerization and signal transduction from the LBD to the DBD. These structural elements set the stage for understanding how these monomers oligomerize into multimeric structures, which form the functional units for transcriptional regulation.

LTTR Dimers As the Functional Unit

From early on in the discovery of LTTRs, it was clear that these TFs oligomerize into higher order protein complexes due to the size, dyad symmetry, number and distance between DNA sites that are bound simultaneously.^66,122 Later it was postulated that the LTTRs form dimers in the absence of DNA, and further dimerize when they bind their target DNA, resulting in what is now called the “dimer of dimers” conformation. This was confirmed with the first successful crystallization of a full LTTR protein complex, namely CbnR.^132,133 Dimerization occurs around the linker helix, with antiparallel coiled coil interactions between the helices,^{57,71,133,221} as shown in Figure 5B. Their hydrophobic residues interact resulting in a hydrophobic patch. At the helix ends, several hydrogen bonds occur, resulting in a strong interaction that further influences the protein structure of both monomers in the dimer.¹³³ Within this structure, further stabilization is achieved via interactions of the C-terminal regions of each monomer. Due to the antiparallel linker helix interaction, the DBDs are mirrored toward each other and show a pseudo-2-fold axis, which matches the dyad symmetry palindromic sequence of the TFBSs.²⁵³ In addition, the DBDs are separated roughly 30 Å, resembling the distance between major grooves in the DNA helix, where the binding sites are located.²⁵³

Figure 5.

Oligomerization of CbnR as a model transcription factor for the LysR-type transcriptional regulators. (A) The monomers form two distinct conformations upon dimerization, i.e., the extended (left) and compact (right) form. These structural differences are stabilized by interacting residues between the linker helix and ligand-binding domain. A close-up of this region is given above the structures. (B) Dimerization occurs primarily around the linker helix in an antiparallel manner. As a result, the DNA-binding domains (DBDs) are positioned to enter two adjacent major grooves of the DNA, as shown in the box containing a close-up of the DBDs of CbnR in association with DNA. The wings of the winged helix-turn-helix structures interact with the nascent minor grooves. (C) Two dimers interact, forming the “dimer of dimers” conformations. The tetramerization interface is mainly located in the ligand-binding domains.

While the LTTRs form homodimers, the monomers show two different conformations, namely compact and extended subunits,^133,253,268 as shown in Figure 5A. The difference between these conformations is found at the hinge domain, where different angles between the linker helix and LBD are noted. For CbnR, it was found that the compact monomer showed an angle of 50°.¹³³ A set of three direct hydrogen bonds and several water-mediated hydrogen bonds were found between the two domains. In contrast, no interactions were noted for the extended subunit, where an angle of 130° is noted. This organization of different conformations was also noted in the structures of TsaR and ArgP.^253,268 The underlying mechanism for the formation of these two conformations between monomers with the same amino acid composition has not been elucidated. It is possible that in the extended form, the residues responsible for hydrogen bond formation are occupied. This would happen randomly to some monomers, while others form the hydrogen bonds and become compact subunits upon expression. Alternatively, it is possible that the determination of conformations is proximity based. During dimerization, the monomers would then be pushed into either extended or compact forms due to protein–protein interactions, causing the interacting residues to be brought into proximity for the formation of hydrogen bonds.^253,268

Formation of the Tetrameric Protein Complex

As dimers, the LTTRs are able to recognize the DNA-binding sites, but simultaneous tetramerization and DNA binding is required for the creation of a stable interaction at the AS1 site. The tetramerization interface is mainly situated in the LBDs, where several interactions occur to form the eventual protein complex.¹³² The full tetramer structure of regulator CbnR is given in Figure 5C. The monomers collectively constitute a parallelogram configuration with the compact monomer from the dimers interacting not only with its extended monomer counterpart, but also with both units from the other dimer.^71,133 In this context, the RD1 and RD2 domains of the extended and compact monomers interact through the formation of salt bridges^{65,71,133,253} and hydrophobic interactions.²²¹ The C-terminal region has also shown importance for tetramerization, yet the exact mechanism has not been found.²⁶⁸

Tetramerization can only occur by bending the DNA to bring the dimers within proximity. However, DNA bending already starts at the dimer state. When comparing regulators, the cumulative bending angles within a dimer interacting with DNA got up to roughly 46° for CatM, and 45° for BenM. Most of this is achieved between the recognition half sites, i.e., between the thymine and adenine of the T-N₁₁-A LTTR sequence.¹¹⁶ Both the bending by the individual dimers, as the full DNA bending by the tetramer are caused by the presence of extended and compact monomers, which results in an asymmetric overall structure and creates a V-shape of DBDs all situated in the same plane.^{116,127,133,293} Consequently, an uninduced tetramer binds the target promoter in key RNAP-binding regions and causes the DNA to bend up to 100°. The combined effect of steric hindrance and DNA bending impedes the polymerase’s ability to attach to the promoter and initiate transcription.^57,265

The functional mechanism of the LTTRs is referred to as the “sliding dimer” mechanism and is visualized in Figure 2. The dimer interacting with binding site AS1 moves to a site closer to the RS-bound dimer when the LTTR complex interacts with its ligand. This became evident in DNase I footprinting pattern studies, which depict variations in DNA binding when comparing induced and uninduced states.^{124,142,143,268} These studies also highlight hypersensitivity of DNA to DNase I, which is linked to the DNA-bending mechanism, and they depict a shift in this hypersensitivity upon ligand addition.^71,123,142 To move the DBDs of the AS1-bound dimer, conformational changes within the protein complex are required in order to form a more compact protein structure and to bring the DBD pairs closer. This has been demonstrated through multiple crystal structure studies, revealing varying quaternary structures depending upon the cocrystallization of the inducing ligand.^50,132,253 When the structures of CbnR¹³³ and TsaR²⁵³ were first compared, large differences were noted in their tetrameric structure. The former was crystallized in the absence of any inducing ligand and showed interactions between helices of nearby RD2 domains. No such interactions were found for the structure of TsaR, which was cocrystallized p-toluenesulfonate. This led to a mechanism describing the conformational changes following ligand binding including a break in such helix interactions, leading to the transition from a closed to an open form,^159,253 as also illustrated in Figure 2B. This has been confirmed by several LTTRs structures, including ArgP,²⁶⁸ DntR,¹⁵⁹ BenM^111,118 and AphB.¹⁰² However, there are large differences in these conformational changes depending on the regulator, indicating that there is more research needed on the subsequent events following ligand interaction.

In addition to the reduced DNA bending and release of key promoter sequences, ligand-induced LTTR tetramers actively recruit RNAP to the bound promoter system. During analysis of this mechanism in CatR, it became clear that C-terminal mutations in the α-subunit of RNAP resulted in a lack of transcription activation upon CatR induction.¹²³ These findings were complemented with research on NahR where direct interactions with the RNAP complex were found.²⁰⁸ The exact mechanism of RNAP recruitment by LTTRs is not yet fully understood, but it was hypothesized that because of conformational changes following ligand induction, RNAP recruitment patches migrate from internal protein positions to the tetramer surface of the LTTR complex. Here, they would be able to interact with the α-subunit of RNAP^{122,125,208,215,226} to guide the polymerase to the promoter’s UP region and increase the stability of RNAP-DNA interactions.^71,294,295 Additionally, residues in the DBD could interact with the α-subunit of RNAP, as seen for BenM.¹²⁰

While further research is necessary, current insights into the structural composition of LTTRs have led to the development of a standard regulatory model centered on the ’sliding dimer’ mechanism. The organization of monomers within one dimer, and their subsequent dimerization into the ’dimer of dimers’, supports this regulatory mechanism. Moreover, conformational changes within the tetrameric structure facilitate the sliding of the complex upon interaction with the target ligand, as well as recruitment of RNAP. However, the immense diversity of regulators within the LTTR family, along with challenges in obtaining structural information, remains a significant hurdle for further refining this ’classic’ model.

Exceptions Among the LysR-type Transcriptional Regulators

The “classic” model for LTTR regulation, based on the “sliding dimer” principle, applies to a large part of the currently characterized LTTRs. Nevertheless, since the founding of this protein family, many exceptions to these rules have emerged. What were initially considered outliers have now become increasingly common. The CysB-like regulator, Cbl, is a prime example of the extent to which LTTRs can deviate from what is considered the LTTR regulatory mechanism.⁶¹ Cbl is an important regulator in Escherichia coli where it regulates many genes related to sulfur acquisition and virulence.^{61,131,296,297} The promoters of these genes show only a single Cbl interaction site,²⁹⁷ which led to the hypothesis that it is functional as a dimer.¹³⁰ The regulator activates target promoters independently, without requiring any ligand interactions. Moreover, ligand interactions have a suppressive effect on the transcriptional activation. For instance, adenosine phosphosulfate leads to the repression of Cbl-bound promoters,²⁹⁸ while thiosulfate impedes DNA interactions.¹³¹ There was also no autoregulation noted for this LTTR.¹²⁹ As a result, converting this regulator into an LTTR-based biosensor requires significantly different consideration in comparison to the ’classic’ model.

As it is the largest bacterial transcriptional regulator family with a wide range of important biological functions, it is unsurprising that a single model is inadequate to describe all LTTRs. The introduction of a subfamily system based on functional mechanisms, oligomerization patterns, or other distinguishing features, could greatly enhance our comprehension of this family. Such a system would also accelerate the characterization of newly discovered LTTRs by more accurately linking them to known functional mechanisms. It would also allow the development of biosensor construction strategies and designs for each subfamily. Figure 6A summarizes the currently known exceptions for each of the key features of transcriptional regulation by LTTRs, together with the impact that these variations can have on the regulatory mechanism (visualized in Figure 6B). In what follows, a selection of different characteristics, showing a significant number of exceptions to the “classic” LTTR mechanism, is given as an argument for the need for a subdivision of the LTTR family.

Figure 6.

Overview of the currently identified alterations in LysR-type transcriptional regulators. (A) The components of the “classic” LysR-type transcriptional regulator model, given with “*”; their exceptions are summarized in five categories. Red arrows indicate repression, green arrows indicate activation, and gray arrows indicate no effect on transcription. (B) The possible effects of the exceptions described in A on the response curves of these regulators are visualized per parameter (as described in Figure 1B). Genetic circuit parts are given according to SBOL conventions.^41,42T = Temperature, TF = transcription factor, TFBS = transcription factor binding site.

Transcription Factor Binding Site Organization and Regulation of Target Genes

The “sliding dimer” mechanism requires three TFBSs, yet an increasing number of LTTRs deviate from this, as exemplified in Figure 6A. It is expected that, due to the diversity of LTTRs, a multitude of regulatory mechanisms are embedded in the family, each with specific TFBS architectures.²⁶⁵ Similar to the previously mentioned Cbl, aspartate transport and metabolism regulator AalR from Acinetobacter baylyi ADP1 shows single binding sites in their target promoters.⁸¹ The occurrence of two binding sites, which were also longer in sequence than the classic T-N₁₁-A LTTR box, was noted for regulators AlsR and CrgA. Given that these regulators form octameric structures, their regulation mechanism is likely to vary, corresponding to their mode of oligomerization.^149,299

As global regulators, LTTRs are known to regulate their target genes in different ways, which is mostly achieved by the variation in TFBS sequences that result in altered affinity of the TF.¹⁵¹ For CysB, it was noted that next to variation in the sequences, the amount of TFBSs also differs over its target promoters.^130,131,300 This is taken even further for ArgP from E. coli, which interacts with seemingly random DNA sequences. While LTTR binding sites have been found for ArgP, it was noted that it is able to regulate promoters by nonspecific binding of AT-rich regions of DNA.¹⁰⁶ However, one of the most challenging parts of characterizing LTTRs is the discovery of their binding sites due to this natural variation in TFBS sequences. Therefore, whether these variations in the number and nature of TFBS sequences are a result of undiscovered sites or altered regulatory mechanisms, remains unclear.

Due to colocalization of LTTRs with a target gene, binding to these TFBSs also enables the TF to regulate its own expression. This is an important mechanism, as it determines the TF concentration, which greatly impacts the overall response of the regulator. However, it was established early on that a variety of autoregulatory mechanisms are present within the family,⁶⁶ as shown in Figure 6A. First, negative autoregulation has been reported for LTTRs that are not found around bidirectional promoters. This has been shown for CysB, which is expressed seemingly isolated from any of its regulatory targets but is negatively autoregulated. Another example is VirR, which is the first protein expressed in the virulence operon of Rhodococcus equi. VirR represses transcription of the operon, resulting in negative autoregulation, but is able to regulate the other genes within the operon at an additional internal promoter, past the VirR sequence.¹⁹⁰ Second, different LTTRs have been proven to lack autoregulation, examples including the previously mentioned ArgP,¹⁰⁵ as well as PA2206 from Pseudomonas aeruginosa.²³² Third, despite negative autoregulation being regarded as the standard for LTTRs, there is a seemingly large set of regulators that show positive autoregulation, including SpvR from Salmonella typhimurium(245) and YtxR from Yersinia enterocolitica.²⁶² Several other regulators, e.g., KaeR from Lactobacillus brevis,¹⁸⁰ LttR from Lactobacillus plantarum,¹⁸⁷ ThnR from Sphingomonas macrogolitabida strain TFA³⁰¹ or DbdR from Thauera aromatica AR-1,¹⁵⁸ show positive autoregulation upon ligand induction. Positive autoregulation could give an advantage to systems in environments where the inducing ligand shows limited availability, due to the increased sensitivity for signal changes.³⁰² Lastly, there is a group of transcriptional repressors which show positive autoregulation, including LrhA from E. coli,¹⁸⁶ homologues PecT from Dickeya dadantii(234) and HexA from Photorhabdus luminescens,¹⁷¹ and RovM from Yersinia pseudotuberculosis.²⁴³ For these repressors, positive autoregulation can aid in repression mechanisms by ensuring the availability of adequate concentrations of the regulator.³⁰¹

In accordance with the variation in autoregulation, an increasing number of regulators are found to repress their target promoter upon ligand induction. Two regulators, GltC and NdhR, have the ability to perform both activation and repression depending on the interacting ligand. In the presence of α-ketoglutarate, GltC of Bacillus subtilis performs transcriptional activation over its regulon, but in the presence of glutamate, the regulator represses target promoters.¹⁶⁸ Similarly, NdhR is important for coordinating carbon and nitrogen metabolisms in cyanobacteria and activates appropriate genes in the presence of 2-phosphoglycolate. In times of nitrogen starvation, 2-oxoglutarate accumulates and is able to interact with NdhR, causing the regulator to repress its targets.⁷² However, it is not clear if this repression is different from the uninduced repression noted for all regulators. While the presence of corepressors makes it likely that these are actual repressors, it is possible that the molecules are blocking the activator from interacting and thus prevent activation, rather than cause repression. Competitive interaction has been noted for ClcR from Pseudomonas putida, where fumarate competes with 2-chloro-cis,cis-muconate for interaction at the ligand-binding cleft.¹⁴² As the difference between competitive binding and corepressor is small, further investigation into the regulatory mechanism of these regulators and potential conformational shifts upon repressing molecule binding, could provide a clear definition.

Ligand Recognition

Several regulators seem to lack an inducing ligand, and their regulation is altered based on different signals, as illustrated in Figure 6A. The LTTR NAC, or nitrogen assimilation control protein, has been shown to possess a rare functional mechanism that does not require any inducing ligands.³⁰³ This TF of Klebsiella pneumoniae is a key regulator in nitrogen metabolism, and regulates many pathways within this context, both as a repressor and activator, depending on the target promoter. It does so in often nonspecific ways, similar to ArgP, where it seemingly binds all available AT-rich DNA regions.³⁰³ Instead of ligand induction, the activity of this regulator is mainly dependent on its intracellular concentration. A two-component system that monitors nitrogen availability in the environment, regulates the transcription of NAC. This transcriptional activation is balanced out by a unique form of negative autoregulation where NAC prevents DNA bending of its own promoter region, counteracting the functional mechanism of the activator.^304,305

A different set of regulators are influenced in their activity by external factors or environmental conditions. The virulence genes of Dickeya dadantii are controlled by PecT, which can only bind and repress the target genes at higher temperatures (roughly 35–38°). In these conditions, the regulator undergoes conformational changes resulting in the formation of active complexes. In addition, the relaxation of DNA supercoiling at these temperatures helps the active PecT with binding its target DNA.²³⁴ OxyR from Pseudomonas aeruginosa is a well characterized LTTR that aids in the prevention of oxidation damage by sensing hydrogen peroxide (H₂O₂) and regulating antioxidant genes.³⁰⁶ While technically H₂O₂ functions as a ligand for this regulator, it is not bound by the TF, but causes conformational changes by oxidizing the protein resulting in the formation of disulfide bonds.^228,229 In doing so, the regulator reduces the H₂O₂ molecule, which counters oxidative stress in addition to the subsequent activation of antioxidant genes.²³⁰ Similarly, HypT from Salmonella enterica serovar typhimurium helps the microorganism withstand hypochlorous acid (HOCl) by regulating methionine and cysteine biosynthesis pathways as well as reducing the toxic molecule upon interaction.¹⁷⁶ Structural changes were noted upon reduction of HOCl, which also triggered the activation of target genes, similar to the mechanism of OxyR.¹⁷⁶ AphB from Vibrio cholerae is responsible for the activation of virulence genes under oxygen-limited situations.^100,101 Due to the decrease in oxygen, a cysteine residue is reduced, causing the regulator to undergo structural changes and form an active complex, as well as increase dimerization stability.¹⁰¹

Other exceptions to the classic LTTR ligand recognition include binding of multiple ligands. This was discussed above for regulators GltC and NdhR, which change transcriptional regulation patterns based on the available ligand. For these TFs, the ligands both bind at the ligand-binding cleft situated between RD subdomains of the LBD. An alternative mechanism of ligand interactions based on the availability of multiple ligand-binding sites has been seen for BenM and DntR.¹¹⁸ This enables synergistic actions of both inducing molecules upon the regulator, resulting in unique regulatory control over the target gene by integrating multiple inputs.^159,293,307 For BenM, a benzoate degradation regulator in Acinetobacter sp. Strain ADP1, both cis,cis-muconate and benzoate are bound by the regulator on different sites.^118,293,307 In Burkholderia sp. strain DNT, DntR is known for its role in the oxidative degradation of 2,4-dinitrotoluene (2,4-DNT) by binding this molecule and activating the appropriate genes in response.^308,309 It was further shown that the regulator recognizes salicylate at a different ligand-binding site.¹⁵⁹ For both BenM and DntR, it is known that conformational changes within the protein structures are only fully developed with the binding of both molecules. Similarly, sulfate can interact with the ligand-binding site of CysB, causing the protein to act as a repressor. By binding in a different location, O-acetylserine excludes sulfate from the regulator due to conformational changes, transitioning CysB to an activator. Exclusion of sulfate enables N-acetylserine to bind the regulator, causing further stabilization of the activated form.¹⁵⁴

Oligomerization

Determining the degree of oligomerization of LTTRs has been challenging because of the issues related to crystallizing higher order protein structures in the absence of DNA, as discussed above. Hence, many of the initially characterized LTTRs were classified as dimers. However, none of the current oligomerization interfaces facilitate a functional dimeric LTTR.²⁶⁵ It was later established that most regulators from this family form tetramers. Even within the tetrameric structures, differences in organization of monomers have been noted, which has recently been reviewed by Baugh et al.²⁶⁵ Higher order oligomerization creates intriguing structures of octameric protein complexes, which can further dimerize to hexadecamers in association with DNA.¹⁴⁹ CrgA of Neisseria meningitidis is the most well-known example of octameric regulators.¹⁴⁹ When compared to the process of tetramerization, the main differences are found in the organization of monomeric structures. While the tetrameric regulator CbnR shows extended and compact subunits, as shown in Figure 5A, all CrgA monomers show exactly the same conformation. There are no hydrogen bonds formed between the hinge and LBD, and a much more rigid hinge sequence is found due to the presence of a PxG motive instead of the more mobile GxxG motif of CbnR.¹⁴⁹ Several other regulators have been hypothesized to form octamers, e.g., AlsR from Bacillus subtilis(90) or HsdR from Comamonas testosteroni,¹⁷² but CrgA remains the only LTTR for which this was experimentally determined.

Subclassification of the LysR-type Transcriptional Regulators

The diversity embedded within the LTTRs has become extensive, and it is likely that more exceptions will arise during continuous expansion of this TF family. As illustrated in Figure 6, these variations significantly impact the regulator’s effect on the target promoter, resulting in distinct response curves when used as biosensors. While this variability can be advantageous for developing biosensors for a wide range of genetic circuit applications, it also complicates their design due to the difficulty of predicting their effects in vivo. To accelerate and simplify LTTR-based biosensor development, it would be beneficial to create a subclassification system that groups LTTRs with similar variations on the ’classic’ regulatory mechanism.

In the past, several attempts have been made to organize this diversity in subfamilies, yet this has been hampered by the large number of regulator features that vary. These attempts were mainly focused on the classification of LTTRs based on sequence similarities. However, homology seemingly lacks many of the intricate differences that are embedded in the LTTR amino acid sequence. In the hinge region, the presence of a single proline residue is potentially enough to prevent the formation of extended and compact monomer structures, resulting in octameric regulators. As an alternative to homology-based clustering, some suggestions can be made from the discussion presented here. Oligomerization is seemingly linked to many of the other features which are varied in the exceptions presented above. It is directly related to the overall structure of the protein complex and the DNA-interaction, making it important for the precise method of transcriptional regulation. The main problem is determining the degree of oligomerization, especially as higher order complexes are usually only formed in the presence of target DNA sequences. Several techniques have been used in the past, including SDS-PAGE, size-exclusion chromatograph, microscopy-based techniques and X-ray crystallography or NMR spectroscopy, yet the requirement of DNA–protein complexes, the difficulties with extracting LTTRs, and the low throughput of these methods, inhibit the use of oligomerization as an efficient grouping parameter. Alternatively, the architecture of TFBSs has important implications on the mechanism behind transcriptional regulation. Similar to oligomerization, accurate and reliable TFBS prediction is difficult due to the extent of variation in these sequences.

To effectively group LTTRs into subfamilies, comprehensive information on a wide range of parameters is essential. By integrating this data with amino acid sequence analysis, it is possible to identify patterns that can serve as the basis for establishing a more precise classification system. This process can be significantly expedited through the application of machine learning techniques, which excel at intricate pattern analysis when provided with sufficient data. Therefore, by establishing high-throughput experimental methods for characterizing LTTRs and integrating computational biology, significant progress can be achieved in understanding and organizing the LTTR family.

Conclusions

Despite advances in modifying natural transcriptional regulators for biosensor use, there is a need for a broader repertoire of sensors as genetic circuit inputs in order to fully exploit the technology. To expand on the currently limited set of used transcriptional regulators, the untapped pool of TFs from the LTTR family can be exploited. LTTRs are the largest group of bacterial transcriptional regulators and play pivotal roles in various biological processes. They have evolved into highly efficient dual-regulatory molecules that provide precise control over their regulon via the “sliding-dimers” mechanism. The tetrameric protein complexes bind their target promoters, repressing transcription by interacting with key RNAP interaction sites, and making them further inaccessible by bending the DNA. After interaction with their respective ligand, the LTTRs undergo conformational changes, causing the protein complex to bind an upstream binding site resulting in RNAP recruitment via direct and indirect mechanisms.

However, with the size of the family comes a highly diverse set of regulators that show many exceptions to this “classic” LTTR model. Due to the increasing quantity of exceptions, it is becoming clear that one model is not sufficient to describe a TF family of this size. Reorganization of the family, potentially into subfamilies of regulators with similar mechanisms of regulation, will aid in the understanding both known and newly discovered LTTRs. Despite the added complexity, the array of different mechanisms opens options to new applications and novel biosensors. Expanding our understanding of their structure and regulatory mechanisms is essential for harnessing their capabilities and expanding the toolbox of sensors in synthetic biology.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acssynbio.4c00219.

• Protein structures used to validate the LTTR domain definition and list of proteins used to establish the domain definition (PDF)

Author Contributions

All authors were involved in the conceptualization and design of the manuscript. W.D. drafted the manuscript, which was followed by critical revision by all authors.

The authors declare no competing financial interest.