Employing aromatic tuning to modulate output from two-component signaling circuits
- 2.3k Downloads
Two-component signaling circuits (TCSs) govern the majority of environmental, pathogenic and industrial processes undertaken by bacteria. Therefore, controlling signal output from these circuits in a stimulus-independent manner is of central importance to synthetic microbiologists. Aromatic tuning, or repositioning the aromatic residues commonly found at the cytoplasmic end of the final TM helix has been shown to modulate signal output from the aspartate chemoreceptor (Tar) and the major osmosensor (EnvZ) of Escherichia coli. Aromatic residues are found in a similar location within other bacterial membrane-spanning receptors, suggesting that aromatic tuning could be harnessed for a wide-range of applications. Here, a brief synopsis of the data underpinning aromatic tuning, the initial successes with the method and the inherent advantages over those previously employed for modulating TCS signal output are presented.
KeywordsSynthetic microbiology Aromatic tuning Receptor engineering Signal modulation Synthetic circuits
Two-component signaling circuit
Open reading frame
Sensor histidine kinase
Trp-flanked peptides with a poly-Ala-Leu core
Tyr-flanked peptides with a poly-Ala-Leu core
Circular dichroism spectroscopy
Nuclear magnetic resonance spectroscopy
Minimum glycosylation distance
Second transmembrane helix
Course-grained molecular dynamics simulations
Domain responsible for catalytic ATPase activity
Domain responsible for dimerization and histidylphosphotransfer
Two-component signaling circuits (TCSs) are a ubiquitous mechanism by which bacteria sense, respond and adapt to external stimuli. TCSs facilitate responses to a wide range of environmental parameters, such as ambient temperature, availability of nutrients or external osmolarity . Medically relevant multiorganism phenomena including quorum-sensing  and host-pathogen interaction are also governed by TCSs . Furthermore, these systems control essential agricultural and environmental processes such as chloroplast synthesis  and nitrogen fixation , which is involved in root nodule formation. More than 400,000 open reading frames (ORFs) believed to encode TCSs have been sequenced and annotated suggesting that these systems control an almost unlimited number of processes that could be harnessed by synthetic microbiologists [6, 7]. Based on this premise, engineering TCSs with novel functionality is a very active field , especially because it has been recently shown that these modified systems can be systematically transferred to mammalian cells . However, rationally designing TCSs, either by moving intact components from one microorganism to another or by forming chimeric components from the domains of two different proteins, can result in aberrant signal output or loss of stimulus-perception, or both . Here, an aromatic tuning methodology that has been established to circumvent these issues and is generally applicable to both native and rationally designed TCSs is described.
Two-component signaling circuits
Biophysical and biochemical underpinnings of aromatic tuning
From a biochemical perspective, another methodology, known as glycosylation mapping, also demonstrated the contribution of aromatic residues to positioning α-helices within a lipid bilayer . Glycosylation mapping utilizes the lumenally positioned endoplasmic reticulum enzyme oligosaccharyltransferase (OST) to add a glycan to the Asn residue in Asn-Xaa-(Ser/Thr) glycosylation acceptor sites. In this manner, OST acts as a molecular ruler because each acceptor site will be glycosylated to an extent that correlates with the distance between the active site of OST and the acceptor site. In Fig. 2b, the red acceptor sites are not far enough from the lumenal membrane to become glycosylated, whereas the green sites are distal enough to become glycosylated. From this information, minimum glycosylation distance (MGD), or the distance required to attain half-maximal glycosylation, can be determined as schematically shown in Fig. 2c. These MGDs can then be compared between segments to determine whether any helix repositioning occurred. Upon comparison, a decrease in MGD suggests that lumenal boundary of TM2 has been displaced into the lumen, while an increase in MGD demonstrates that the TM2 boundary has been repositioned into the membrane (Fig. 2c). A series of TM segments that are identical except for the position of the flanking aromatic residues at one end can be assessed for changes in MGD that can be assigned to the repositioning ability of those aromatic residues.
Changes in MGD have been reported when Trp residues are moved throughout the C-terminal half of a poly-Leu TM segment , thereby demonstrating that bidirectional repositioning of the TM segment is due to the affinity of Trp residues for the polar/hydrophobic interfaces (Fig. 2b). In a slightly modified version of this assay that considers all possible positions within the TM helix, Tyr residues where shown to possess character similar to Trp residues . Thus, before aromatic tuning was conceived, previously existing biophysical and biochemical evidence suggested that Trp and Tyr residues possess affinity for the polar/hydrophobic regions located within membranes and suggested that moving them within an intact membrane-spanning protein could reposition individual TM helices.
Aromatic residues at the ends of Tar TM2 govern its position in the membrane
Armed with these outputs to monitor changes in Tar signaling, initial arginyl- and cysteinyl-scanning mutagenesis of Tar TM2 from Salmonella enteritica serovar Typhimurium was performed and resulted in several mutations that highlighted the importance of aromatic residues in governing helix position within the cytoplasmic membrane . Large changes in signal output consistent with a displacement of TM2 toward the periplasm, similar to those observed in the apo conformation of Tar, namely increased CheA kinase activity and decreased methylation, were observed when either Phe-189 or Trp-192 was replaced with an Arg residue at the periplasmic end of TM2. This was also seen when Phe-189 was replaced with a Cys residue. Conversely, reduced kinase activity and increased levels of covalent methylation, both consistent with the aspartate-bound confirmation of Tar, were observed when Trp-209 was replaced with Arg . Independently, the importance of Trp-192 and Trp-209 in maintaining baseline signal output from E. coli Tar was demonstrated when they were substituted for alanyl residues . In the case of the Tar W192A variant, the receptor was undermethylated and exhibited higher signal output, consistent with apo state and displacement of TM2 toward the periplasm. In contrast, the W209A mutant was overmethylated and possessed reduced signal output, suggesting a displacement of TM2 toward the cytoplasm as proposed to occur within the aspartate-bound confirmation. In summary, these results demonstrate that aromatic residues within TM2 of Tar are critical for maintaining normal signal output and suggest that they govern the baseline position of the helix within the membrane to an extent that could be employed to manipulate Tar signal output.
Incremental tuning of Tar signal output
Previous results show that the intracellular concentration of CheY-P governs the clockwise flagellar (CW) rotational bias (Fig. 4b). Therefore, Tar signal output could be estimated from the aromatically tuned variants by monitoring flagellar rotation. To accomplish this, individual flagella from single E. coli cells were monitored for 20 s and classified into one of five categories that represent increasing levels of intracellular CheY-P: those rotating exclusively CCW, flagella rotating mostly CCW with occasional switching to CW rotation, flagella rapidly switching between CCW and CW with no inherent bias, flagella rotating mostly CW with occasional switching to CCW rotation, and finally those that rotate exclusively CW . When flagella from cells expressing the aromatically tuned variants described in Fig. 4a were analyzed, a clear shift from CCW to CW rotational bias was observed as the aromatic residues were moved from WY-3 toward WY + 3 (Fig. 4c). This indicates that the signal output from the minus-series of tuned Tar variants (i.e., WY-3 to WY-1) was so far biased toward reduced signal output that the compensatory action of covalent methylation failed to completely overcome this signaling bias. Conversely, the WY + 2 and WY + 3 variants were biased toward increased signal output to the extent that methylation could also not compensate. As expected, in the absence of receptor (− ctrl), only CCW rotation was observed (Fig. 4c). It is also important to note that, as predicted, overmethylation of Tar was observed for the minus-series (WY-3 to WY-1) of receptors, while undermethylation was seen for the plus-series (WY + 1 to WY + 3) of aromatically tuned variants . When taken together, these results demonstrate that signal output from Tar is modulated in an incremental manner when the aromatic residues were moved at the cytoplasmic end of TM2.
While these results served as an interesting foray into aromatic tuning, they do not empirically demonstrate that moving the aromatic residues at the cytoplasmic end of TM2 results in a physical repositioning of the helix. Therefore, two subsequent studies, one computational and one biochemical, were performed to assess repositioning of Tar TM2 upon moving the Trp-Tyr tandem. Coarse-grained molecular dynamics (CG-MD) simulations that were employed to examine the ability of aromatic tuning to displace Tar TM2 in the presence of an explicit membrane and solvent demonstrated that moving the Trp-Tyr residue was sufficient to induce small TM2 displacements of up to 1.5 Å . In addition, the glycosylation-mapping assay described above (Fig. 2b and c) was performed in order to determine whether moving the Trp-Tyr pair at the end of TM2 would yield analogous results in a biological membrane. Assuming that the region in Lep that contains the glycosylation-accepting site is in an extended conformation, a shift in MGD of 0.5 residues as seen for the Tar constructs corresponds to a shift in the positioning of the TM2 helix of 1.6-1.7 Å , in close agreement with the CG-MD results . Furthermore, within both the CG-MD simulations and MGD analysis, similar patterns of displacement were observed. A grouping of the minus-series of receptors with similar displacements toward the cytoplasm (WY-3 through WY-1), a baseline position for the wild-type (WY 0), two receptors that are slightly displaced toward the periplasm (WY + 1 and WY + 2) and a larger shift toward the periplasm for the WY + 3 variant was observed with both techniques. Therefore, this combination of in vivo, in vitro and in silico results demonstrate that repositioning the Trp-Tyr tandem at the cytoplasmic end is sufficient to generate a physical displacement of TM2 and the resulting modulation of Tar signal output in an incremental manner.
Non-incremental tuning of EnvZ signal output
In order to determine whether aromatic tuning would work within a canonical SHK, its effectiveness was examined within the major E. coli osmosensor (EnvZ), where a rotation of TM2 was initially proposed as the mechanism of transmembrane communication [34, 35, 36, 37, 38]. More recently, a regulated unfolding  model has been suggested, in which SHKs are modular proteins composed of individually folding domains that each contribute a distinct functionality. Regulated unfolding suggests that the effector domain within an SHK is maintained in an inactive conformation by a rigid connection between the stimulus perception and effector domains. Upon perception of stimulus, this connection disengages allowing the effector domain to adopt an active conformation . This region connecting the TM to the cytoplasmic domain in bacterial receptors is colloquially referred to as a “control cable” because its residue composition governs coupling of signal transduction between adjacent domains [26, 27, 40, 41, 42, 43, 44, 45, 46, 47].
To analyze steady-state signal output from EnvZ/OmpR osmosensing circuits containing aromatically tuned receptors, a two-color fluorescent reporter strain was employed. This E. coli strain possesses transcriptional fusions of cfp to ompC and of yfp to ompF within its chromosome (Fig. 5b) . Quantifying the ratio of CFP to YFP fluorescence provides a rapid and sensitive measure of the ratio of ompC to ompF transcription, which estimates the intracellular level of phosphoylated OmpR. Therefore, the ratio of CFP to YFP can be employed to estimate signal output from the aromatically tuned EnvZ variants. To confirm this, we previously demonstrated that the transcription of CFP and YFP behaved as expected upon addition of sucrose to the growth medium (Fig. 5c) .
Due to the proposal of non-piston signaling mechanisms, it was difficult to predict what pattern of signal outputs would be observed upon aromatic tuning of an SHK. In the case of EnvZ, a Trp-Leu-Phe triplet was moved from five residues into the membrane (WLF-5) through two residues out of the membrane (WLF + 2) in single residue steps (Fig. 5d). In this case, the surface of TM2 that the aromatic residues reside upon was found to be the major determinant of EnvZ signal output rather than their absolute vertical position as was the case with Tar (Fig. 5E) . These results clearly demonstrated that even though a different mechanism of TM communication has been proposed, aromatic tuning remained successful in modulating signal output. Similar results were obtained with a Trp-Tyr-Ala triplet that is more similar to the Trp-Tyr tandem employed during tuning of Tar suggesting that in the case of EnvZ, the composition of the residues that are moved is not a major determinant in the pattern of signal output . However, additional experimentation is required to support the concept that disturbance of a particular helical surface is the cause of changes in EnvZ signal output. It should also be noted that these different patterns of signal output modulation are not surprising given that the concept of various signaling mechanisms being employed by different subclasses of bacterial receptors has been recently put forward [56, 57, 58]. Therefore, regardless of whichever signaling mechanism is employed by a receptor, aromatic tuning may be successful in modulating signal output.
Glycosylation mapping experiments with TM2 of EnvZ demonstrated that moving the Trp-Leu-Phe triplet resulted in almost no change in MGD. An exception was observed for WLF-3, which is most likely a specific case where the positive charge within the side-chain of Arg-182 would not be able to interact with the headgroups of the membrane any longer due to it being displaced very far out of the membrane. Thus, in the case of most aromatically tuned EnvZ segments, changes in MGD are very small and not steadily increasing in contrast to what was observed with the aromatically tuned Tar TM2 segments . Initial CG-MD analysis with these aromatically tuned EnvZ segments also does not support incremental displacement (B. Hall, personal communication).
Employment of aromatic tuning within a wide variety of membrane-spanning receptors
Aromatic tuning was successful in modulating signal output from both Tar and EnvZ, however, a difference in the pattern of signal outputs was observed [7, 26]. This pattern of signal outputs shows that even though aromatic tuning did not displace the TM2 helix of EnvZ , it was still effective in modulating signal output from the full-length receptor in vivo. Aromatic tuning has also been successfully employed in two different Tar-EnvZ chimeric receptors (Lehning and Draheim, unpublished observations), when a Trp-Tyr tandem was used, and within a NarX-Tar chimera (Reinhard and Draheim, unpublished observations), when a single Trp residue was moved.
Published sequence alignments demonstrate that aromatic residues are often found at the cytoplasmic end of the final transmembrane helix within bacterial membrane-spanning receptors [27, 59] suggesting that aromatic tuning could be useful for research groups working with other two-component circuits. It is important to note that the majority of aromatically tuned Tar, EnvZ and chimeric receptor variants retain the ability to respond to stimulus suggesting that their signal output is biased but not locked in either a stimulus-deprived or a stimulus-saturated conformation. In this regard, aromatic tuning is advantageous compared to deletion of entire SHKs  or substitution of the conserved His residue involved in autophosphorylation and phosphotransfer because these methods may result in complete loss of kinase or phosphatase activity, which has been shown to result in non-physiological cross-talk between various two-component signaling pathways within a cell [61, 62]. Thus, by apparently biasing, rather than abolishing activity and stimulus-perception, aromatic tuning represents a significant improvement over previous attempts at stimulus-independent modulation of signal output. Even though these initial attempts have been successful, we have only observed aromatic tuning in membrane-spanning receptors that possess two TM helices. However, we feel strongly that in the case of SHKs containing only two TM helices, that aromatic tuning will facilitate stimulus-independent modulation in the majority of cases. Furthermore, we are currently expanding our range of targets to include SHKs that possess more than two TMs such as AgrC from Staphylococcus aureus that has recently been shown to possess seven TMs .
In addition to rectifying issues with signal output or stimulus-perception, we believe that aromatic tuning has other potential uses. For example, the input stimulus has been identified for only a small fraction of SHKs, thereby making it challenging to investigate downstream signaling pathways without extensive mutagenesis and subsequent screening for changes in phenotypic output. Aromatic tuning could be employed to circumvent these limitations and allow manipulation of signal output from a specifically targeted SHK. One potential use would be the rapid assignment of downstream physiological and developmental processes to particular SHKs. Each SHK within an organism could be independently subjected to aromatic tuning and subsequent monitoring for the phenotype of interest. If the appearance of the phenotype correlated with aromatic tuning of a particular SHK, this would suggest that the desired phenotype is governed by the aromatically tuned receptor. In addition to assignment of phenotypes to specific SHKs, aromatic tuning could be employed to facilitate induction of medically relevant bacterial phenotypes in the absence of complex host-pathogen interactions, thus reducing the burden of complex interkingdom laboratory model systems. Finally, aromatic tuning of individual SHKs could be coupled with pre-existing transcriptional reporter libraries  rather than observable phenotypes, to rapidly unravel the signaling pathways governed by each SHK. This would facilitate rapid and cost-effective organism-level signal pathway mapping.
The aromatic residues described here are found at the cytoplasmic boundary of the final TM helix within many bacterial membrane-spanning receptors, suggesting that aromatic tuning may useful in a wide-range of applications involving synthetic microbiology. Here, we provide a brief synopsis of the data supporting aromatic tuning, its initial successes and the advantages of this method over those previously employed for modulating TCS signal output. We have shown that aromatic tuning works within two well-characterized systems and hope that it is presented in sufficient detail to provide other research groups with the opportunity to employ the methodology within their TCSs of interest.
RY was generously supported by the Indonesian Endowment Fund for Education (S-1999/LPDP; 3/2014).
- 14.Killian JA, Salemink I, de Planque MR, Lindblom G, Koeppe 2nd RE, Greathouse DV. Induction of nonbilayer structures in diacylphosphatidylcholine model membranes by transmembrane alpha-helical peptides: importance of hydrophobic mismatch and proposed role of tryptophans. Biochemistry. 1996;35:1037–45.CrossRefGoogle Scholar
- 15.de Planque MR, Boots JW, Rijkers DT, Liskamp RM, Greathouse DV, Killian JA. The effects of hydrophobic mismatch between phosphatidylcholine bilayers and transmembrane alpha-helical peptides depend on the nature of interfacially exposed aromatic and charged residues. Biochemistry. 2002;41:8396–404.CrossRefGoogle Scholar
- 17.de Planque MR, Greathouse DV, Koeppe 2nd RE, Schafer H, Marsh D, Killian JA. Influence of lipid/peptide hydrophobic mismatch on the thickness of diacylphosphatidylcholine bilayers. A 2H NMR and ESR study using designed transmembrane alpha-helical peptides and gramicidin A. Biochemistry. 1998;37:93–45.CrossRefGoogle Scholar
- 23.Draheim RR. The role of protein-membrane interactions in modulation of signaling by bacterial chemoreceptors. Texas A&M University: Ph.D. dissertation; 2007.Google Scholar
- 50.Hoch JA, Silhavy TJ. Two-Component Signal Transduction. Washington, DC: Am. Soc. Microbiol. Press; 1995.Google Scholar
- 52.Forst S, Delgado J, Rampersaud A, Inouye M. In vivo phosphorylation of OmpR, the transcription activator of the ompF and ompC genes in Escherichia coli. J Bacteriol. 1990;172:3473–7.Google Scholar
- 53.Lan CY, Igo MM. Differential expression of the OmpF and OmpC porin proteins in Escherichia coli K-12 depends upon the level of active OmpR. J Bacteriol. 1998;180:171–4.Google Scholar
- 62.Siryaporn A, Goulian M. Characterizing cross-talk in vivo avoiding pitfalls and overinterpretation. Methods Enzymol. 2010;471:1–16.Google Scholar
- 68.Rampersaud A, Inouye M. Procaine, a local anesthetic, signals through the EnvZ receptor to change the DNA binding affinity of the transcriptional activator protein OmpR. J Bacteriol. 1991;173:6882–8.Google Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.