Nanoplasmonic SERS reveals previously uncharacterised indole derivative in E. coli metabolism

Nanoplasmonic SERS reveals previously uncharacterised indole derivative in E. coli metabolism | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results Nanoplasmonic SERS reveals previously uncharacterised indole derivative in E. coli metabolism View ORCID Profile Mo Vali , View ORCID Profile Elle W. Wyatt , View ORCID Profile Kieran Abbott , View ORCID Profile Cameron Croft , View ORCID Profile Pietro Lio , View ORCID Profile Thomas F. Krauss , View ORCID Profile Minahil Khan , View ORCID Profile Ashraf Zarkan , View ORCID Profile Jeremy J. Baumberg , View ORCID Profile Diana Fusco doi: https://doi.org/10.1101/2025.11.10.687708 Mo Vali 1 Cavendish Laboratory, Department of Physics, University of Cambridge , Cambridge, CB3 0US, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Mo Vali For correspondence: mv487{at}cam.ac.uk Elle W. Wyatt 1 Cavendish Laboratory, Department of Physics, University of Cambridge , Cambridge, CB3 0US, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Elle W. Wyatt Kieran Abbott 2 Department of Genetics, University of Cambridge , Cambridge, CB2 3EH, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kieran Abbott Cameron Croft 2 Department of Genetics, University of Cambridge , Cambridge, CB2 3EH, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Cameron Croft Pietro Lio 3 Department of Computer Science and Technology, University of Cambridge , Cambridge, CB3 0FD, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Pietro Lio Thomas F. Krauss 4 Department of Physics, Engineering and Technology, University of York , York, YO10 5DD, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Thomas F. Krauss Minahil Khan 4 Department of Physics, Engineering and Technology, University of York , York, YO10 5DD, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Minahil Khan Ashraf Zarkan 2 Department of Genetics, University of Cambridge , Cambridge, CB2 3EH, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Ashraf Zarkan Jeremy J. Baumberg 1 Cavendish Laboratory, Department of Physics, University of Cambridge , Cambridge, CB3 0US, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jeremy J. Baumberg Diana Fusco 1 Cavendish Laboratory, Department of Physics, University of Cambridge , Cambridge, CB3 0US, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Diana Fusco Abstract Full Text Info/History Metrics Supplementary material Preview PDF ABSTRACT Tryptophanase (TnaA) is a promiscuous enzyme that regulates the production of amino acid-derived metabolites essential for bacterial communication. TnaA is known to convert tryptophan into indole, a known signalling molecule, but other routes remain obscure. Here, we show that nanoplasmonic surface-enhanced Raman spectroscopy (SERS) enables label-free detection of indole in Escherichia coli with a limit of detection of 100 nM - over 500-fold more sensitive than previously reported and significantly outperforming existing colorimetric assays. Using this ultrasensitive approach, we perform SERS on wild-type and tnaA knockout E. coli strains and discover a distinct TnaA-dependent metabolic signature in both standard and uropathogenic strains when cultures are supplemented with amino acids other than tryptophan. Quantitative spectral analysis reveals an indole-like metabolite whose Raman signature does not correspond to free indole or any known indole derivative characterised by mass spectrometry. This observation implies previously unidentified bacterial metabolic pathways involving amino acid reallocation which can depend on environmental conditions. Given the central role of TnaA in bacterial signalling and virulence, these findings raise new hypotheses regarding the biological role of indole derivatives and TnaA-mediated byproducts. More broadly, our results establish nanoplasmonic SERS as a powerful technique for probing enzyme activity and bioactive metabolite production in vivo at concentrations well below the detection limits of existing methods. Introduction Bacteria produce a wide chemical portfolio of bioactive extracellular molecules that regulate behavioural phenotypes, cell-to-cell communication and response to environmental cues, as well as biofilm formation 1 – 6 . Many of these metabolites have also been linked to infections, inflammation, neurological disorders and changes in the human microbiome 7 – 11 . These molecules have attracted significant research interest, as their production can be engineered for a wide range of uses, from antibiotics to fertiliser agents 12 , 13 . In vivo measurements of metabolites are thus essential both to better understand their role in bacterial physiology in the wild and to harness their production for practical applications. Indole is an important example of such signalling molecules produced by over 85 Gram-positive and Gram-negative bacterial species 14 – 17 and known to cause a wide range of effects on membrane potential, cytoplasmic pH, transport, biofilm formation, and antibiotic resistance and persistence 16 . Indole is clinically relevant, acting as a biochemical precursor to neurotransmitters such as serotonin, and frequently detected in infections caused by E. coli , particularly in the urinary tract 18 – 21 . Indole is produced by the tetrameric enzyme tryptophanase, or tryptophan indole-lyase, EC 4.1.99.1 (TnaA), which catalyses α, β -elimination and β -substitution reactions of both natural and synthetic amino acids. The tnaA gene encoding the enzyme TnaA has been reported in over 190 species, highlighting the pervasive role of TnaA in metabolism 22 . It is well established that TnaA catalyses the conversion of L-tryptophan (Trp) to indole, pyruvate (via the removal of the carboxyl group), and ammonia (via removal of the amino group) 23 . However, the role of TnaA in other metabolic pathways remains poorly characterised. Seminal work by Newton and Snell in 1964 reported that crystalline preparations of TnaA could catalyse a series of α, β -elimination and β -substitution reactions in amino acids beyond Trp 24 , 25 . Subsequent work has utilised spectrophotometry techniques to identify coenzyme-substrate intermediates using characteristic absorption and circular dichroic spectra 26 – 29 , confirming that other amino acids beyond Trp can act as substrates (and inhibitors) of TnaA in vitro 26 . Few studies have systematically evaluated the role of TnaA, and to our knowledge, there is no study that has carried out a comprehensive comparison of the in vitro enzymatic activity of TnaA versus its in vivo cellular counterpart. Such studies are needed to fully understand the TnaA-regulated metabolic pathways in the living cell, and their role in indole production. Conventional microbial typing and diagnostic approaches for detecting metabolic byproducts or bioactive metabolites typically require extensive sample preparation, often relying on specific reagents or fluorescent markers. Surface-enhanced Raman Spectroscopy (SERS) is becoming increasingly employed as an alternative technology in microbiological contexts, as it has shown high sensitivity for label- and stain-free bacterial detection, identification, and antimicrobial susceptibility testing, both on solid and liquid samples 30 – 33 . In contrast to conventional microbial typing methods, SERS has demonstrated superior sensitivity (10 −9 M to 10 −12 M) and is largely agnostic to the target metabolite, as long as the Raman cross-section is sufficiently large 34 . SERS capabilities have been illustrated for the identification of taxonomically-close bacterial species 35 , 36 , chemical profiling of microbial cells 37 , 38 , single cell analysis 39 and in vivo diagnostics 35 , 40 , 41 . So far, only a limited number of studies have employed SERS for characterising bacterial metabolite profiles, as label-free detection of target analytes in complex biofluids still presents considerable technical challenges in isolating interpretable chemical signatures. Notably, Bodelon et al. developed a method for label-free SERS detection of pyocyanin, a proxy analyte for quorum sensing, in Pseudomonas aeruginosa and Chromobacterium violaceum biofilms. The authors made use of specially designed nanostructured hybrid materials, demonstrating the potential of SERS to track a known target molecule in bacterial aggregates 42 . Nevertheless, to our knowledge, SERS has not been utilised to characterise an enzyme’s metabolic activity in vivo or to identify its primary byproducts without specific extraction. In recent work, we have demonstrated the construction and use of precision nanogaps within sheets of self-assembled gold nanoparticles as reproducible and re-useable SERS devices for sensing bioanalytes such as neurotransmitters 43 – 45 . These ‘MLagg’ films consist of a near-monolayer aggregate of close-packed gold nanoparticles (here 80nm diameter), with 0.9 nm particle spacing precisely defined by the rigid cucurbit[n]uril (CB[n]) spacer molecule employed ( Fig. 1a ). Pre-cleaning these nanogaps to remove all organics inherent in their self-assembly yields precision SERS devices that can be recleaned of all analytes and fouling for reuse. The nanogaps give very large plasmonic SERS enhancements ( > 10 8 ) for small molecules adsorbing within the plasmonic hotspots in the gaps between the nanoparticles, enabling the detection of very low (nM) molecular concentrations. The consistent preparation of these MLaggs and their reproducible response opens up the capability to study many biomolecules in complex environments over long periods of time, affordably and sustainably. Download figure Open in new tab Figure 1. In vitro and in vivo SERS characterisation of TnaA-dependent indole production in E. coli . (a) SERS sensing films, from aggregation of 80nm gold nanoparticles with CB[5] molecules at a liquid-liquid interface before drying into films 43 . (b)i) Enzymatic assays which incubate the purified enzyme TnaA and amino acid in HEPES buffer and PLP cofactor; ii) E. coli grown overnight in media containing M9 media supplemented with amino acid (see Methods). Supernatant is extracted and SERS measured. (c-e) Normalised SERS spectra of (c) enzymatic assay of TnaA combined with Trp (dark green), an identical solution without the enzyme TnaA (light green), and buffer solution only (without Trp, gray); (d) E. coli BW25113 and (e) E. coli 536 of both wild-type (WT) and tnaA gene knockout strains (KO), together with culture media of M9 media supplemented with Trp shown for reference (gray). (f) Subtraction of the two spectra shown in (c-e) and comparison to the spectrum of 10mM pure indole. Significant peaks shown as shaded vertical lines, demonstrating strong alignment between the spectra and presence of supernatant indole in these assays, as confirmed by mass spectrometry (see Fig. 4 ). The goal of our present work is two-fold. First, we demonstrate how SERS with MLagg substrates can be used in combination with appropriately genetically-engineered strains for the quantitative, label-free, in vivo detection of indole within the high chemical complexity of a bacterial suspension, at nanomolar concentrations. Second, we show how to extend this methodology to investigate the broader in vivo TnaA-dependent metabolic activity for each of the 20 canonical L-amino acids, in order to provide mechanistic insights into bacterial physiology. Our results demonstrate that applying SERS to target wild-type and knock-out strains facilitates the detection and quantification of even very low concentrations of byproducts of an enzymatic activity of interest, with minimal sample preparation and strong reproducibility. More broadly, our findings endorse SERS as the method of choice for sensitive, non-invasive, and inexpensive identification and quantification of metabolites in living cell cultures. Results Reconstruction of TnaA-dependent indole signature in E. coli cultures using SERS Bacterial cell cultures are chemically complex, consisting not just of cells and cell debris, but also of all molecules from spent media and byproducts of bacterial growth. The isolation of a SERS spectrum for a target molecule in such complex solutions is challenging, as many compounds contribute with different vibrational modes. To identify a clean SERS signature of indole in this environment, we perform SERS ( Fig. 1 ) on the supernatant of two wild-type (WT) E. coli strains, BW25113 and 536, and their genetically modified counterparts, in which the tnaA gene has been knocked out (KO or Δ tnaA ) ( Fig. 1b ii ), see Methods). Because the supernatants contain the spent media of the cell culture, these assays assess metabolite production from living cells. In parallel, we prepare an in vitro enzymatic assay comprising the purified enzyme TnaA incubated in a buffer containing Trp, and a corresponding control assay without TnaA ( Fig. 1b i ). SERS spectra for WT and KO supernatant assays for both E. coli BW25113 and E. coli 536 show strong peaks ( Fig. 1d,e ), compared to the culture media (shown in gray). Similar peaks are seen for the equivalent in vitro assay with and without TnaA ( Fig. 1c ). A clear difference in spectral profile is observed between WT and KO for supernatant assays, as well as with/without TnaA for enzymatic assays. This can be seen more clearly when looking at the spectral differences ( Fig. 1f ) in which the KO spectrum is subtracted from the WT spectrum to produce a clean TnaA-dependent supernatant spectrum (see Methods). The spectral profile of the difference spectra strongly matches the spectrum for pure indole (10 mM, water, light crimson line in Fig. 1f ), which is the known byproduct of Trp catalysis by TnaA (shared peaks highlighted by shaded vertical lines). The enzymatic assay exhibits an almost identical signature ( Fig. 1f ), demonstrating the ability of SERS to detect an excreted metabolite, such as indole, directly from a bacterial culture with minimal sample manipulation or extraction. The presence of indole in these supernatants is also confirmed independently by mass spectrometry (see SI Fig. 4.1). As expected, indole is produced from Trp by TnaA, and is not present when the enzyme TnaA is removed, as reflected in the SERS spectra. These results are further corroborated by spiking with 10mM of free indole in both strains (SI Fig. 1.1) and in artificial urine, a more complex media (SI Fig. 1.5). To our knowledge, this is the first time that the indole spectrum has been reconstructed using SERS directly in bacterial cultures without specific metabolite extraction 46 , 47 . Importantly, this capability stems from comparing the wild-type strain with the appropriate corresponding genetically modified strain, where the gene ( tnaA ) encoding the enzyme of interest (TnaA) has been knocked out. This provides a control for the non-TnaA dependent metabolite activity. The second crucial factor is the size-selective filtering by the nanogap of only small molecules from the complex mixture, greatly improving the signal-to-noise ratio in the SERS spectra. SERS from MLaggs can detect in vivo indole without specific extraction, at 100 nM concentrations ( Fig. 1g and SI Fig. 1.1), a 500 times increase over prior work 46 . This holds significant promise for detecting indole at very low concentrations in clinical settings, particularly as indole is directly involved in over 20 infections, and the current limits of detection (3 µ M) using colorimetric assays are non-viable for low concentration detection 16 , 48 – 50 . Calibration ( Fig. 1gii )) based on characteristic indole peak areas shows a linear correlation (R 2 = 0.93) between concentration and SERS intensities, from 100 nM to 10 mM. SERS assay profiling across amino acids reveals previously undetected TnaA-dependent activity Having established that SERS can detect indole production directly from the supernatant of bacterial cultures and at very low (nM) metabolite concentrations, we extend this technique to all 20 canonical L-amino acids, to investigate potential metabolic activity of TnaA on other substrates. For enzymatic assays we isolate the effect of TnaA by subtracting the SERS spectrum of solutions of purified TnaA supplemented with the amino acid of interest, from the equivalent SERS spectrum without TnaA ( with TnaA minus without TnaA , Fig. 2a ). Similarly, for the supernatant assays, we compare the difference in SERS spectra from the supernatant of the WT cultures supplemented with the target amino acid, subtracting the SERS spectrum of the corresponding KO strain (WT minus KO, Fig. 2b,c ). We perform this spectral analysis for both E. coli variants, BW25113 and 536. In summary, we utilise three types of assays (enzymatic, cellular E. coli BW25113 and cellular E. coli 536), two types of strains (WT, or with TnaA, and KO, or without TnaA), 20 amino acids, and ten repeats for each spectrum recorded, yielding a total of 1200 spectra (and a vector of size (1200, 4060)), which forms the basis for Fig. 2 . Download figure Open in new tab Figure 2. SERS profiling of enzymatic and supernatant cellular assays. (a) Enzymatic assays: SERS profiling with and without TnaA. (b,c) Supernatant assays: wild-type (WT) minus tnaA gene knockout (KO) assays, inoculated from LB agar into M9 media supplemented with each amino acid, as labelled (see left of figure). In the enzymatic assay, only Trp exhibits significant TnaA-dependent metabolic activity, implying that this is the only amino acid directly targeted by TnaA with a byproduct detectable by SERS. It is known that both Ser and Cys are also TnaA targets, however, the byproducts of these reactions fall outside the fingerprint region considered here (500-1750cm −1 ) 24 . Surprisingly, we also observe substantial TnaA-dependent metabolic activity for several other amino acids in the supernatant assays, suggesting a previously uncharacterised role for TnaA in multiple metabolic pathways within a living bacterial cell. Amino acids L-alanine, L-isoleucine, L-leucine and L-valine show SERS activity in E. coli BW25113 but not in E. coli 536 or enzymatic assays. Conversely, L-arginine, L-asparagine, L-glutamate, L-glycine, L-serine, and L-tyrosine show SERS activity in both strains. L-serine and L-cysteine, which are known substrates of TnaA, show higher activity in E. coli 536 than E. coli BW25113, and no activity in either, respectively. To our knowledge, this is the first time that the role of TnaA in E. coli cultures supplemented with each of the twenty canonical L-amino acids has been characterised. Dimensionality reduction isolates a shared indole-like metabolic signature across amino acids To perform a quantitative comparison between the activity of TnaA on different amino acids, we apply Principal Component Analysis (PCA) on the SERS spectra shown in Fig. 2 (see Methods and SI Fig. 3.1-3.4). Projection of the data on PC1 and PC2 ( Fig. 3 , enzymatic, supernatant E. coli BW25113 and supernatant E. coli 536) shows a surprising trend. By superimposing the projections of pure indole (indigo star) and pure Trp (orange star) spectra on this two-dimensional space, we see that PC1 mostly captures the enrichment/depletion of indole, while PC2 captures the enrichment/depletion of Trp in the solution, as confirmed by the PC1 and PC2 weightings in Fig. 3e . The position of the Trp enzymatic assay (orange square datapoint in Fig. 3a ) confirms its enrichment of indole and depletion of Trp, consistent with the previously characterised role of TnaA in transforming Trp into indole. As suggested by the SERS spectra in Fig. 2 , no other amino acid shows spectral activity in enzymatic assays, confirming that Trp is the only direct substrate of TnaA detectable by SERS in this context. Download figure Open in new tab Figure 3. PCA of supernatant WT - KO spectra. (a) PC1-PC2 projections of enzymatic assay spectra obtained by subtracting assay spectra without TnaA from corresponding spectra with TnaA, and (b,c) of cellular wild-type (WT) spectra minus corresponding tnaA gene knockout (KO) assay spectra for (b) E. coli BW25113 and (c) E. coli 536, coloured by amino acid as in Fig. 2 . Spectrum for 10mM pure indole and 1mM pure Trp are projected onto PC1-PC2 for reference. (d) Rotated PC1-PC2 plane showing the distance of each assay to pure indole. (e) PC1 and PC2 weightings (solid), compared to pure indole and pure Trp spectra (dashed). (f) Ranking of each cellular amino acid assay’s PC1* value as obtained from (d). First two PCA components explain 65% of the variation (see Methods and SI Fig. 3.1-3.4). The PCA for the two supernatant assays ( Fig. 3b and Fig. 3c ) shows a very different picture. First, Trp is similarly positioned across the three assays, confirming that the direct enzymatic reaction between TnaA and Trp is the main metabolic pathway when bacterial cells are exposed to Trp. However, all the other amino acids are found distributed along a line on the PC1-PC2 plane (see SI Fig. 3.1). The differences between enzymatic and supernatant assays clearly indicate that within the bacterial cell, there are indirect interactions between TnaA and amino acids beyond Trp. In addition, the observed linear trend raises the intriguing hypothesis that a common metabolite(s) might be excreted by the cell at different concentrations depending on the amino acid initially present in the media. Given that the SERS spectra of WT minus KO assays for different amino acids predominantly extends along the pure indole projection, we hypothesise the presence of a previously undetected indole-like metabolite, I*, in the supernatant of E. coli assays. To support this hypothesis, we project the SERS spectra of pure indole at different concentrations onto the PC1–PC2 plane (indigo stars, Fig. 3a–c ), which intriguingly fall along the same direction outlined by the amino acids. These results show that this distinctive indole-like signature is not unique to a particular E. coli strain or amino acid. To quantify how similar the WT - KO E. coli assay spectra are to the pure indole spectrum, we rotate the PC1-PC2 projection so that the x -axis runs along the pure indole direction (defined as PC1*, dashed line in Fig. 3d ). We then determine the distance of each amino acid spectrum from the origin, here defined as the centroid of assays without an I* spectral signature (null region, Fig. 3d and Methods). Fig. 3f reports the PC1* values or horizontal distance from the null region in the rotated PCA. As expected, Trp gives the strongest signal, followed by Ser for the E. coli 536 strain, and Tyr and the BCAAs for the E. coli BW25113strain. The other amino acids follow, although no obvious biochemical pattern can be determined in their ranking. Alternative measures of analysing similarities across the spectra have also been explored (see SI Fig 2.1-2.5). LC-MS of supernatant assays and the detection of an indole-like metabolite across several amino acids To investigate the potential identity of I*, the indole-like signature detected by SERS, we perform liquid chromatography–mass spectrometry (LC-MS) on selected supernatant assays for both WT and KO strains (see Methods). Fig. 4a shows the change between WT and KO as ΔMS = log 2 (WT/KO) of normalised peak area intensities for indole and its derivatives found in reference libraries. Only metabolites above the limit of detection (>3 σ ) are included. Unsurprisingly, for indole (top left chart, Fig. 4a ), both E. coli BW25113and E. coli 536 assays in Trp show significant ΔMS changes, while no meaningful differences are detected in the other WT vs. KO assays. The indole derivative 3-(2-hydroxyethyl)indole shows elevated levels in some assays that exhibit the indole-like signature (BW-Tyr, 536-Leu, 536-Gly), but not in others (BW-Leu), and also appears elevated in assays with a very weak indole-like signature (BW-Cys). Based on LC-MS, we find no single molecule (indole-related or otherwise) or combination of molecules displaying a common profile across the amino acids identified by SERS, or meaningful correlation with the I* spectrum observed ( Fig. 2 and SI Fig. 2.3). Overall, 914 metabolites are detected by LC-MS (see SI Fig. 4.1), with considerable variability, both across strains for the same amino acid (e.g. Trp), and across amino acids for the same strain (e.g. Arg and Ser). Further analysis of the metabolite set in cultures supplemented with amino acids reveal potentially interesting shared metabolites. For example, the aromatic amino acids (Trp, Tyr, Phe) share trimethylamine N-oxide (TMAO), a clinically important metabolite found to play a role in cholesterol metabolism, and glycylproline, a metabolite released during the breakdown of collagen (see SI Fig. 4.1d) 51 , 52 . Next, we perform SERS on key indole derivatives 5-hydroxyindole (5-HI), 3-(2-hydroxyethyl)indole (3-2), indole-3-acetic acid (IAA), and 8-hydroxyquinoline (8HQ) identified by LC-MS in some assays. Fig. 4b compares the SERS spectra of E. coli BW25113WT - KO assays in Trp, Tyr and Cys, chosen as reference because they show indole, I* and absence of I*, respectively, with the SERS spectra for indole at different concentrations and multiple indole derivatives. The spectra of these indole derivatives exhibit poor alignment with the I* SERS spectrum. In fact, we find I* aligns more closely with the pure indole spectrum. Relative to the spectrum for pure indole, however, some peaks in I* are wavenumber shifted by ∼1%, while some show significantly reduced intensities (which we explore further below, see Fig. 5 ). These results support our hypothesis that the previously unreported TnaA-mediated metabolic signature shared across several cultures supplemented with amino acids other than Trp is likely an indole derivative. Download figure Open in new tab Figure 4. LC-MS and SERS results of indole derivatives. (a) LC-MS results (ΔMS = log 2 (WT/KO)) for the detection of indole and indole derivatives in selected supernatant WT minus KO cellular assays ( E. coli 536 strain assays shown dashed). (b) Comparison of SERS spectra for indole derivatives versus SERS of WT minus KO Trp-based (indole signature), Tyr-based (I* activity), and Cys-based (no I* activity) assays for E. coli BW25113 strain. Strongest spectral alignment for I* (middle panel) observed with pure indole (see Fig. 5 and discussion). Peak assignments and potential production mechanisms of I* To further explore the molecular differences between indole and I* using SERS, we analyse previously determined band assignments of indole, relative to I*, using the E. coli BW25113Tyr WT - KO spectrum as best representative of the I* signature. As Fig. 5a shows, the indole benzene-ring C–C mode at 1455cm −1 is absent from I*, whilst bands at 1505cm −1 , 1578cm −1 and 1613cm −1 are significantly red-shifted by an average 19cm −1 from I to I*. These bands, attributed to C-C aromatic stretching modes, are sensitive to perturbations in π -electron delocalisation typically caused by oxidation in the indole π system, which redistributes the electron density and weakens or rehybridises C-C bands 53 , 54 . Despite the absence of direct studies probing oxidation effects on free indole’s benzene-ring C–C modes, Chen et al . found an average redshift of 20cm −1 for the same aromatic C–C and C–N bonds associated with tryptophan residues upon oxidation of a chlorophyll dimer within a pigment–protein complex 55 . Other work has found a redshift for C-C and C-N stretches connected to the indole system, upon oxidation to a stable tryptophan radical in an azurin mutant, using resonance Raman techniques 56 . Download figure Open in new tab Figure 5. Detailed comparison of band assignments of free indole versus I*. (a) SERS of free indole (indigo) and I* (crimson, taken as E. coli BW25113Tyr WT - KO spectrum). Gray lines show band assignments, orange lines show key spectral differences between I* and pure indole. (b) Suggested bond vibrations associated with these differences. (c) DFT of pure indole showing alignment with experimentally obtained SERS for indole. Fig. 5b shows these redshifted peaks resulting from the vibrational changes associated with the C2/C3 bonds of the benzene ring. The DFT-derived spectrum for indole ( Fig. 5c , scaling factor of 0.975) shows peaks align well with the obtained experimental indole spectrum (gray lines, Fig. 5a ), and corroborates the observed redshift in I* in the 1400-1600cm −1 region. Beyond the DFT for indole, we perform DFT on various mono-oxidised and protonated forms of indole (see SI Fig. 1.1). We note however that experimental spectra arise from molecules within the MLagg substrate trapped in nanogaps at Au facets, which is known to shift analyte peaks and may additionally depend on environmental factors (pH, solvation) that cannot be easily accommodated in DFT calculations. Based on these results, we posit that I* is likely an oxidised indole derivative. We now consider potential production mechanisms for I*, that is, arising from the effect of TnaA either on a derivatised form of Trp, or on Trp to produce indole which is subsequently converted to I* ( Fig. 6 , and below). Given the tight regulation of intracellular Trp levels, we speculate that supplementation with certain amino acids may perturb this balance, leading cells to reallocate metabolic flux toward utilisation of the externally supplemented amino acid, and thus increasing relative intracellular Trp concentrations 57 . Fig. 6 illustrates a mechanistic example with leucine (which generates the second largest I* signal after Tyr), based on known biosynthesis pathways 58 . The leucine-responsive regulatory protein (Lrp) is a key regulator of amino acid metabolism in E. coli . Lrp binding with leucine changes its conformation, binding differently to the promoters of Lrp-regulated operons 58 . This indirectly or directly downregulates sdaA, sdaB, tdh and other genes involved in the metabolism of other intracellular amino acids such as serine and threonine, resulting in a lower incentive for the cell to harvest aromatic amino acids such as Trp for nitrogen supply, thus increasing relative intracellular Trp concentrations 59 . More broadly, we speculate that an increase in branched chain amino acids supplementation may shift the metabolic balance away from Trp breakdown for nitrogen, as the cell reoptimises use of leucine and other BCAAs to produce glutamic acid as its main nitrogen source. The schematic in Fig. 6 also shows the likely concentration of I* in the outer membrane rather than in solution. Free indole has been shown to have a 90-fold higher affinity for lipids than water, due to hydrophobicity of the aromatic ring 60 . Thus we posit that most of the I* metabolites are likely absorbed by cells (discarded in our assay), leaving low (nanomolar) residual concentrations of indole derivatives in the supernatant, which SERS can detect but existing methods cannot (see Discussion). Download figure Open in new tab Figure 6. Schematic illustrating potential production mechanism for I* and role of amino acids. External supplementation of amino acids perturbs metabolic flux, leading cells to reallocate intracellular levels of amino acids, including Trp (see pathway labelled 1). Increased relative derivatised Trp concentrations is mediated by TnaA, which leads to the production of I*. Higher affinity of indole derivatives for lipids than water, leaves majority of I* in cell membrane (discarded in assay) and very low residual concentrations in the supernatant (numbered 3). Discussion Our results support the finding that a previously unreported TnaA-mediated metabolic signature (I*) shared across several amino acids is an oxidised indole derivative which is produced either from indole or from a derivatised form of Trp. We speculate that its presence has so far been undetected due to 1) the current limits of detection of both LC-MS and colorimetric biochemical assays relative to nanoplasmonic SERS, and 2) the 90-fold higher affinity of indole and indole derivatives for lipids in the cell membrane versus solution, resulting in low extracellular concentrations (see Fig. 6 ). In this work, we have demonstrated a SERS-based methodology for the in vivo , ultrasensitive detection of indole, and discovered a potential indole derivative in multiple assay types, from enzymatic to cellular assays, using SERS when extending this methodology to all 20 canonical L-amino acids. These results contribute to efforts in mapping previously overlooked bacterial signalling pathways which can have importance for clinical diagnostics 48 , 49 . Recent work by Linares-Otoya et al ., for example, characterised the N-acyl-cyclolysine (ACL) system in bacteria using a combination of techniques including LC-MS, pointing to potentially important insights into microbiome-host interactions 61 . I*, the indole-like signature that we have discovered using SERS, has a peak profile which points strongly to an oxidised indole derivative (see Fig. 5 ). We posit this may be due to: 1) an artefact caused by SERS matrix effects, where free indole is oxidised by hot carriers generated from the excitation of surface plasmons, 2) oxidation of indole produced by TnaA, or 3) the degradation of an intracellular Trp derivative into I* 62 – 64 . The absence of free indole in the supernatant of both E. coli BW25113 and E. coli 536 strains across amino acids (with the exception of Trp), confirmed by both LC-MS and colorimetric assays, however, strongly points to the latter two hypotheses. The presence of I* was further validated by performing replicate experiments with additional cell-washing steps (washed three times in PBS prior to inoculation). This series of experiments was performed to exclude the possibility that any indole transferred from cells from LB agar plates on which cells were initially streaked and grown was responsible for the observed signal. We find minimal spectral differences between washed and unwashed assays (see SI Fig. 2.8). Additionally, the abundance of I* is not consistent across cellular supernatant assays supplemented with amino acids, which reinforces our conclusion that specific amino acids may modulate the formation, degradation, or utilisation of indole or indole-like derivatives. Furthermore, to verify that I* is not simply indole at very low concentrations such that the limit of detection of the colorimetric indole assay (3 µ M) prevents its detection, we concentrate hydrophobic metabolites in the cellular supernatant by >50x using C18 solid phase extraction cartridges, thus pushing the limit of detection to ∼0.1 µ M (see Methods). No indole was detected in these colorimetric assays using a modified version of Kovács reagent (QuantiChrom™ Indole Assay Kit, BioAssay Systems). Kovács reagent contains p-dimethylaminobenzaldehyde (DMAB), which binds at the 3-position of the pyrrole ring, suggesting any indole derivative present is modified in this region 65 . Furthermore, the LC-MS results demonstrate minimal presence of previously annotated indole derivatives, with the exception of 3-(2-hydroxyethyl)indole, a hydroxylated (and therefore oxidised) derivative of indole. 3-(2-hydroxyethyl)indole shows significant activity in WT in comparison to KO assays as measured by LC-MS, but deviates from the I* signature found using SERS ( Fig. 3f ). Acquisition of SERS spectra for 3-(2-hydroxyethyl)indole and other derivatives identified by LC-MS confirms that none of these derivatives matches the I* spectrum ( Fig. 4b ). Some work has shown that Trp derivatives can be converted into oxidised indoles 66 , 67 . Such oxidative products of tryptophan metabolism have important biological effects on bacterial signalling and behaviour, including the activation of signalling pathways that regulate immune cell differentiation and function via transcription factors such as the aryl hydrocarbon receptor (AhR), or protective roles in cardiovascular and metabolic diseases 64 , 66 . The results from our enzymatic assays demonstrate that within the spectral range detectable by SERS, only Trp is identified as a direct substrate of TnaA (although Ser and Cys are also known substrates). Our cellular assays, however, report TnaA-mediated production of an indole-like derivative when the bacterial culture is supplemented with amino acids beyond Trp. Because the enzymatic assays rule out a direct interaction between these amino acids and TnaA, we conclude that the effect of these amino acids on the activity of TnaA must be indirect. Given the tight control on the internal Trp concentration, we speculate that the external supplementation of particular amino acids either alters the internal pool of Trp, creating an excess of Trp and Trp derivatives that TnaA converts to indole, or activates TnaA’s expression triggering Trp conversion even at low Trp concentrations. The second option seems unlikely given that it has been demonstrated that constitutively expressing high levels of tnaA does not result in significant degradation of internal pools of Trp 68 . In contrast, amino acid syntheses share multiple pathways and precursors, facilitating indirect interactions between the presence of one amino acid on the synthesis or availability of another 57 . For instance, Tyr, which exhibits among the strongest I* signals in our results, can relieve repression of tryptophan synthesis by acting as an alternative substrate, indirectly promoting tryptophan production 69 . Similarly, Ser is a precursor of Trp, thus its availability can increase Trp synthesis 70 . The supply of leucine, which displays a strong I* signal, may shift the metabolic balance away from energy-dense Trp degradation to produce glutamic acid for nitrogen (see Fig. 6 ). Interestingly, we find that the positions of the amino acids within the biosynthesis pathways that result in indole derivative production reveal no clear or consistent pattern 57 . This suggests distinct regulatory mechanisms may underlie indole derivative production for different amino acids. For example, isoleucine exhibits a strong I* signal, despite being synthesised downstream of threonine, which shows minimal I* signal in SERS. Likewise, tyrosine displays a robust I* signal, whereas phenylalanine does not. Further investigation of these amino acid subgroups may provide valuable insights into the cellular strategies governing amino acid allocation. It is important to note that the concentration of I* released in the supernatant in these conditions is low, as expected given the 90-fold higher affinity of indole with lipids in the cell membrane versus solution, underscoring the key role that ultra-sensitive detection methods like SERS can have. For instance, our work highlights how these methods can be used to investigate how the cell regulates and adapts its internal amino acid pool depending on the external conditions 60 . The differences in indole production between the two E. coli strains for the same supplemented amino acid also show that the optimal amino acid allocation that the cell maintains strongly depends on the genetic background, even for bacteria coming from the same species. Finally, our findings on I* production in these unexpected conditions raise important questions on the role of indole and its derivatives as a signalling molecule among bacteria: how much of the I* production that we detect is simply a byproduct of optimising Trp cellular concentration versus being functionally employed for bacterial communication? Can other bacteria detect the level of I* produced in these conditions and respond to it? Can we associate or relate some of these conditions with the scenarios in which indole production is known to be important, such as biofilm formation or infections? Follow up studies in this direction have the potential to shed new light on the interplay between bacterial metabolism and ecology. Comparison of assays Previous characterisation of TnaA substrates (Trp, Ser, and Cys) has primarily been achieved through the use of enzymatic assays 24 . These assays rely on known reaction products as a readout of activity, which limits their utility in screening for interactions that result in unknown metabolic byproducts. In contrast, LC-MS represents the gold standard for metabolomic studies investigating biochemical reactions by identifying their byproducts 71 . Despite its broad utility and high-throughput, its ability to identify unknown metabolites is restricted by existing annotated libraries 72 . In addition, LC-MS requires invasive sample manipulation and significant processing costs which often make it impractical for clinical applications 73 . Our work shows that combining SERS with clean gene knockouts bridges the gap between these two sets of techniques. On one hand, we show that SERS enables the ultrasensitive detection of previously uncharacterised molecular interactions, with minimal sample preparation. In particular, SERS is found to be equally capable of detecting direct enzymatic activity in vitro , and enzyme-mediated activity in vivo with far superior sensitivity to other existing techniques. We achieve nanomolar-level sensitivity without specifically optimising for the detection limit, demonstrating the potential of current-generation SERS as a readily accessible approach for detecting indole in over 20 associated infections 48 . On the other hand, SERS provides a vibrational ‘fingerprint’ (unlike readouts from conventional in vitro assays) for metabolic byproducts of enzymatic interactions, which can be compared to existing libraries or tested against potential candidates to identify the possible molecule of interest, as is the case here. Importantly, even if band assignments are unclear and reaction byproducts cannot be identified, the precision of SERS can be used for comparative studies across different substrates (i.e. amino acids) or strains (i.e. WT vs. KO) to more accurately assess whether similar or different reactions may be at play across assays. Crucially, we find the comparison of WT and KO SERS spectra ( Fig. 2 ), where the gene encoding the enzyme of interest (in this case TnaA) is knocked out, is key to isolate a clean SERS spectrum for the enzymatic activity of a protein of interest within the complex chemical environment of a bacterial culture. Constructing gene knock-outs in most laboratory strains is relatively straightforward with several such libraries available, making this approach generalisable to many different enzymes or proteins of interest across multiple bacterial species. This work also opens new research possibilities and quantitative studies on the interplay between environment, bacterial physiology and specific biochemical reactions within the cell, orthogonal to current metabolomics efforts. For example, further applications to other gene-related modifications (e.g. knock-downs, integration of other genes) and/or to bacteria grown in different physiological conditions (e.g. nutrient concentrations, temperature, antibiotic susceptibility), combined with the ultrasensitivity of SERS has the potential to yield novel microbiological and clinically relevant insights. In particular, the SERS methodology outlined in this work may be particularly reliable and accessible for low density bacterial samples (e.g. environmental samples or animal/human microbiomes), where the metabolite(s) of interest might be present at very low concentrations. Additionally, SERS may help to shed light on more general real-time aspects of bacterial biochemistry. In a clinical context, the use of precision SERS may hold promise for the non-invasive, in vivo metabolite characterisation at the point-of-care, in a precise, reproducible and low-cost manner, positioning it as a potentially revolutionary tool in microbial diagnostics and public health. Methods Bacterial culture E. coli BW25113 and E. coli 536 strains were used for all experiments. BW25113 Δ tnaA was obtained from the Keio collection 74 . 536 Δ tnaA was generated by λ -red recombination 75 . Bacterial freezer stock held at −80 °C was used to streak an LB agar-based plate, incubated overnight at 37 °C. Single isogenic colonies were collected with sterile plastic loops and used to inoculate 3 mL culture in M9 supplemented with 140 of 1 mM amino acid (filter sterilised). This was incubated overnight at 37 °C. 500 of the solution was extracted and centrifuged for 10 min at 7500 rpm. The supernatant was extracted and SERS performed. Four biological replicates were prepared for all cellular assays. Of these, two biological replicates were derived from two single colonies taken from two different streak plates. Culture medium preparation The M9 medium was prepared using autoclaved DifcoTM M9 Minimal Salts 5x, filter sterilised 20% glucose solution, 1M MgSO 4 and 1M CaCl 2 . 14.1g of powder was dissolved into 250mL of sterile purified water (Milli-Q). The solution was autoclaved at 121 °C for 15 minutes. The autoclaved 5x solution was then diluted to 1x by mixing 50 mL into a 200mL volume of Milli-Q. 5 mL of 20% glucose solution, 0.5mL of sterile 1.0 M MgSO4 solution and 0.025 mL of sterile 1.0 M CaCl 2 solution was added. 50mM of stock concentration of amino acids L-Cysteine (121.16g/mol), L-Arginine (174.2g/mol), L-Alanine (89.09g/mol), L-Asparagine (132.12g/mol), L-Glutamic Acid (147.13g/mol), L-Glycine (75.07g/mol), L-Isoleucine (131.17g/mol), L-Methionine (149.21g/mol), L-Tyrosine (181.19g/mol), L-Tryptophan (204.23g/mol) and L-Serine (105.09g/mol) were prepared. For 3 mL of M9 medium, 140 µ L of each amino acid was added to give a 1mM final concentration. Where specified, cells were cultured in Luria-Bertani (LB) broth (0.5 g/L NaCl), artificial urine, or grown on LB agar plates either containing no antibiotic or 50 µ g/mL kanamycin. Artificial urine media was prepared as described in supplementary methods. λ -Red recombination λ -Red recombination was carried out to engineer the tnaA knock-out of 536 using the pSLTS helper plasmid 76 . Electrocompetent cells with arabinose-induced pSLTS were prepared as described in 76 . The kanamycin resistance cassette within E. coli BW25113 Δ tnaA : kanR was amplified by PCR, using primers designed to incorporate 91 bp and 96 bp homology arms of the sequences upstream and downstream of tnaA in E. coli 536 77 . This PCR product was purified using QIAquick® PCR Purification kit, then 100 ng was added to 50 µ L of the prepared electrocompetent cells, which were electroporated, recovered with 450 of LB, then incubated overnight at 30 °C with 150 rpm shaking in a 15 mL tube. The following day, the cultures were centrifuged at 13,000 rpm for 1 minute, resuspended in 100 µ L of fresh LB, plated on LB agar containing 50 µ g/mL kanamycin, and incubated overnight at 37 °C without shaking. The following day, single colonies from the transformation plate were screened for correct knockouts of tnaA by inoculating the single colonies in LB, incubating overnight at 37 °C with shaking, then using Kovac’s reagent to confirm the absence of indole production 78 . Following the confirmation of the absence of indole production, the strain was sent for whole genome sequencing. Enzymatic assay Enzymatic assays were performed to assess direct TnaA-mediated degradation of individual amino acids. The highly active stock of TnaA (see supplementary methods for protein purification and preparation) in 0.1M HEPES (pH 7.8), 10% glycerol was concentrated using VivaSpin 500 centrifugal concentrators (10 kDa MWCO), then added to 500 µ L of 0.1 M HEPES (pH 7.8) containing PLP and 1 mM final concentration of a single amino acid, and incubated for 2 hours at 37 °C. For L-Trp, 80 µ g of TnaA and 0.5 µ M PLP were included. For L-cys, 400 µ g of TnaA and 20 µ M PLP were included. For L-ser and other L-amino acids, 400 µ g of TnaA and 5 µ M PLP were included. These amounts of TnaA and PLP were identified as sufficient for complete degradation of 1 mM L-trp, L-cys, and L-ser, within 1.5 hours at 37 °C using a lactate dehydrogenase-coupled enzymatic assay 79 . For controls, parallel reactions were set up as above, but adding the same volume of concentrator flow-through (0.1 M HEPES (pH 7.8), 10% glycerol) instead of TnaA. Enzymatic reactions were not filtered prior to SERS measurements, as tests comparing filtered reactions to unfiltered reactions demonstrated that purified TnaA provided negligible SERS signal relative to the buffer solution (see SI Fig. 1.1). Indole concentration and quantification Indole in bacterial supernatant samples was concentrated and quantified using a modified version of the approach described by Zarkan et al 80 . Briefly, C18 solid phase extraction cartridges (Agilent Bond Elut C18, 1 g bed mass, 6 mL) were equilibrated by flowing through 12 mL of methanol, followed by 12 mL of Milli-Q water. 50 mL of supernatant from a centrifuged bacterial culture was loaded into the cartridge, washed with 12 mL of Milli-Q water, and residual water was removed by a 30-second air pull. Indole was eluted from the cartridge with 3 mL of methanol, collected in 500 µ L fractions. The majority of indole was present in the first fractions, resulting in around 50-100x concentration. Indole in the eluted fractions was quantified using QuantiChrom™ Indole Assay Kit (BioAssay Systems). 100 µ L of each fraction was added to individual wells of a 96-well plate, alongside a standard curve of 0 to 100 µ M indole in methanol. 100 µ L of QuantiChrom™ Indole Assay reagent was added to each well, shaken briefly, and absorbance read at 565 nm using a SpectraMax iD3 microplate reader (Molecular Devices). Absorbance values at 565 nm were also measured prior to the addition of QuantiChrom™ Indole Assay reagent, to identify increases in absorbance unrelated to indole concentration, and appropriately subtracted from the final value. The concentration of indole in each fraction was inferred from the absorbance value relative to the absorbance values of the standard curve. The limit of detection of this protocol was approximately 0.1 µ M indole in the original bacterial supernatant. SERS substrate (MLagg) preparation Monolayer aggregates were prepared for the SERS signal enhancement. 500 µ L each of chloroform (CHCl 3 ) and commercial (BBI Solutions) citrate-capped 80 nm gold nanoparticles (AuNPs) were added to an Eppendorf tube. 100 µ L of 1 mM cucurbit[5]uril (CB[5]) solution was then added and shaken for 1 min to initiate aggregation. The mixture was left to settle for the immiscible CHCl 3 and aqueous phases to separate and the aggregated AuNPs to move to the phase interfaces (chloroform-aqueous and aqueous-air). The aqueous phase was washed with three 300 µ L aliquots of DI water to dilute the citrate salts and other supernatants, then concentrated by careful removal of the aqueous phase to form a 5 µ L aggregate droplet floating on the CHCl 3 . The droplet was deposited onto a pre-cleaned borosilicate glass coverslip (0.09 mm thick). Once dried, the resulting AuNP multilayer aggregate (MLagg) was rinsed with DI water and dried with N 2 . The MLaggs were oxygen plasma cleaned for 45 minutes (oxygen mass flow of 30 sccm, 90% RF power) using a plasma etcher (Diener electronic GmbH & Co. KG) to remove CB[5], citrate and other supernatants from the AuNP surfaces (verified using SERS). To re-introduce a scaffolding ligand, the MLaggs were immersed in 1 mM CB[5] solution prepared in 0.5 M HCl for 10 minutes, then rinsed with DI water and dried with N 2 . All chemicals were purchased from Sigma-Aldrich and aqueous solutions were prepared in deionized water (>18.2 MΩcm −1 , Purelab Ultra Scientific system). SERS measurements and pre-processing SERS measurements were collected on a Renishaw InViva Raman confocal microscope, using a 20x objective (NA = 0.4) and 1200 lines mm −1 grating. A 785 nm excitation laser was used in extended scan mode, with 1 s integration time and 0.5% laser power (2.2 mW at the sample). All measurements were taken at room temperature, unless otherwise specified, and the spectra were calibrated with respect to Si. SERS measurements were taken for each MLagg substrate used for all assays, prior to immersing the MLagg in the supernatant solution. 400 µ L of the supernatant solution was pipetted into a polystyrene 96 well plate and the MLagg placed face down in the solution. It was ensured the MLagg was in contact with the solution in the well. The sample was left for 1 hour for molecules to diffuse to the hotspots and the sample to dry, then SERS measurements were taken at 10 spots on the MLagg and averaged (ten technical replicates). All SERS spectra presented here were pre-processed using background correction and substrate normalisation, to ensure comparability across assays. Background correction was performed by iteratively fitting a polynomial to the base of the peaks. Substrate normalisation was performed relative to the cucurbit[5]uril (CB[5]) scaffold used in each of the SERS measurements, to account for potential variations in laser intensity and substrate enhancement factors at the time of measurement. All SERS spectra were normalised to the characteristic CB[5] peak at 829 cm −1 , corresponding to the breathing mode of the macrocycle. This peak was assigned a value of 1, and all other intensities were scaled relative to this peak. Spectra presented in this form are referred to as ‘CB-normalised’ in Results. In addition, the reference CB[5] spectrum, acquired immediately prior to each SERS assay, was subtracted from the sample spectrum(‘CB-subtracted’). This subtraction was performed to show the contributions of the analytes from those intrinsic to the CB[5] scaffold, enabling direct comparison of spectral features arising from the biological samples alone. Semipolar metabolite analysis using LC-MS Semi-polar metabolite profiling was performed by MS-Omics (Vedbak, Denmark). The analysis was carried out using a Vanquish LC (Thermo Fisher Scientific) coupled to a Orbitrap Exploris 240 MS (Thermo Fisher Scientific). The UHPLC used an adapted method described by Doneanu et al. (UPLC/MS Monitoring of Water-Soluble Vitamin Bs in Cell Culture Media in Minutes, Water Application note 2011, 720004042en). An electrospray ionization interface was used as ionization source. Analysis was performed in positive and negative ionization mode under polarity switching. Untargeted data processing Metabolomics processing was performed untargeted using Compound Discoverer 3.3 (Thermo Fisher Scientific) and Skyline (24.1, MacCoss Lab Software) for peak picking and feature grouping, followed by a in-house annotation and curation pipeline written in MatLab (2022b, MathWorks). Identification of compounds were performed at four levels; Level 1: identification by retention times (compared against in-house authentic standards), accurate mass (with an accepted deviation of 3ppm), and MS/MS spectra, Level 2a: identification by retention times (compared against in-house authentic standards), accurate mass (with an accepted deviation of 3ppm). Level 2b: identification by accurate mass (with an accepted deviation of 3ppm), and MS/MS spectra, Level 3: identification by accurate mass alone (with an accepted deviation of 3ppm). Annotations on level 2b are based on accurate mass and MS/MS spectra measured with high resolution Orbitrap ESI-MS in mzCloud (Thermo Fisher Scientific), MassBank of North America (UC Davis) and the European MassBank (Helmholtz Centre for Environmental Research Leipzig). The annotations on level 3 are based on searches in the following libraries: E. coli metabolome database. Metabolomic profiling yielded a total of 914 metabolites (144 annotated / known, 770 unannotated / unknown) across all assays. To quantify differences in metabolite abundance attributable to TnaA, the log 2 fold-change (ΔMS) between WT and KO assays was computed. For each metabolite, the normalised peak area intensity in WT assays was divided by that in the corresponding KO assay, and the resulting ratio expressed as log 2 (WT/KO). Positive ΔMS values indicate enrichment in WT samples (i.e. metabolites associated with TnaA activity), whereas negative values indicate depletion relative to KO. Only metabolites exceeding the detection threshold (>3 σ above background) were included in the analysis. This approach enabled systematic comparison of the metabolic profiles of WT and KO strains and identification of TnaA-dependent metabolites. Analytical techniques The PCA uses an initial vector of size (1200, 4060) utilising three types of assays (enzymatic, cellular E. coli BW25113 and cellular E. coli 536), two types of strains (WT, or with TnaA, and KO, or without TnaA), 20 amino acids, and ten repeats for each spectrum recorded. This is reduced to (60, 1281) for Fig. 3 and is based on deducting KO spectra from WT spectra for each amino acid and using the 550cm −1 to 1700cm −1 wavenumber region. To facilitate a more relevant interpretation of the PCA, we rotated the coordinate system such that the ‘null’ region (defined as spectra with minimal variation between WT and KO samples) was centered at the origin and aligned with the horizontal axis. The centroid of this region was translated to the origin, and subsequently the datapoints rotated to align the principal axis of the null region with the x-axis. This transformation yielded rotated components (PC1*, PC2*), preserving the relative geometry of all points. In order to try and fully identify metabolite signatures in the spectral data, we utilised various dimensionality reduction techniques. We find PCA alone provided a more relevant and interpretable representation than t-SNE or UMAP methods for SERS data ( Fig. 2 ), particularly given the need to quantify the loadings (wavenumbers of interest) and PCA’s deterministic approach. We also utilised a two-step analytical procedure, performing low-rank PCA and then embedding these components with UMAP. This provided cleaner clusters than PCA alone, albeit with less intuitive interpretability. SI Fig. 3.6 shows results for the UMAP technique applied on WT minus KO SERS data, as shown in Fig. 2 . Author contributions statement M.V., A.Z., K.A., D.F. and J.J.B. designed the experiments. M.V. prepared all bacteria-related assays and provided the bacterial samples for SERS. A.Z. supplied both bacteria strains including the genetically modified strains. C.C. generated the 536 Δ tnaA strain. Enzymatic assays were prepared by K.A. E.W.W. fabricated MLagg substrates and carried out SERS measurements, with advice from M.K. and T.F.K. Data analysis was performed by M.V., E.W.W, D.F. and J.J.B. Manuscript drafting was led by M.V. and all authors contributed to data interpretation, figures and manuscript editing. Data availability The data that support the findings of this study are available upon request. Competing interests The authors J.J.B., and E.W.W declare the following competing interests: filed patent, Surface-enhanced spectroscopy substrates, UK 2304765.7, 30/3/2023. All other authors declare no competing interests. Acknowledgements The authors would like to thank Dr. Isabel Askenasy for their assistance with protein purification and for their helpful comments and discussions. A.Z. is a recipient of a Transition To Independence (TTI) fellowship from the School of Biological Sciences at the University of Cambridge and was supported by funding from the Rosetrees Trust (JS16/TTI2021/1), the Isaac Newton Trust (21.22(a)iii) and the School of Biological Sciences at the University of Cambridge. J.J.B is funded by the European Research Council (ERC) under Horizon 2020 research and innovation programme PICOFORCE (Grant Agreement No. 883703), POSEIDON (Grant Agreement No. 861950) and EPSRC (Cambridge NanoDTC EP/L015978/1, EP/L027151/1, EP/X037770/1). Funder Information Declared Rosetrees Trust , JS16/TTI2021/1 Isaac Newton Trust , 21.22(a)iii European Research Council, https://ror.org/0472cxd90 , 883703 , 861950 Engineering and Physical Sciences Research Council , EP/L015978/1 , EP/L027151/1 , EP/X037770/1 Footnotes ↵ * df390{at}cam.ac.uk , jjb12{at}cam.ac.uk , maa77{at}cam.ac.uk References 1. ↵ Phelan , V. V. , Liu , W. T. , Pogliano , K. & Dorrestein , P. C. Microbial Metabolic Exchange: The Chemotype-to-Phenotype Link . Nat Chem Biol 15 ( 2011 ). 2. Ratcliff , W. C. & Denison , R. F. Alternative Actions for Antibiotics . Science 332 ( 2011 ). 3. Davies , B. W. , Bogard , R. W. , Young , T. S. & Mekalanos , J. J. Coordinated Regulation of Accessory Genetic Elements Produces Cyclic Di-Nucleotides for V. cholerae Virulence . Cell 149 ( 2012 ). 4. Stubbendieck , R. M. & Straight , P. D. Multifaceted Interfaces of Bacterial Competition . J. Bacteriol . 198 , 2145 – 2155 ( 2016 ). OpenUrl Abstract / FREE Full Text 5. O’Brien , J. & Wright , G. D. An Ecological Perspective of Microbial Secondary Metabolism . Curr Opin Biotechnol 22 ( 2011 ). 6. ↵ van der Meij , A. , Worsley , S. F. , Hutchings , M. I. & van Wezel , G. P. Chemical Ecology of Antibiotic Production by Actinomycetes . FEMS Microbiol. Rev . 41 ( 2017 ). 7. ↵ Cox , L. M. & Weiner , H. L. Microbiota Signaling Pathways that Influence Neurologic Disease . Neurotherapeutics 15 , 135 – 145 ( 2018 ). OpenUrl CrossRef PubMed 8. Cho , I. & Blaser , M. The Human Microbiome: At the Interface of Health and Disease . Nat Rev Genet . 13 , 260 – 270 ( 2012 ). OpenUrl CrossRef PubMed 9. Wassenaar , T. M. & Panigrahi , P. Is a Foetus Developing in a Sterile Environment? Lett. Appl. Microbiol . 59 , 572 – 579 ( 2014 ). OpenUrl CrossRef PubMed 10. Tremaroli , V. & Bäckhed , F. Functional Interactions Between the Gut Microbiota and Host Metabolism . Nature 489 , 242 – 249 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 11. ↵ Garg , N. et al. Three-dimensional Microbiome and Metabolome Cartography of a Diseased Human Lung . Cell 22 , 705 – 716 ( 2017 ). OpenUrl 12. ↵ Bansal , T. et al. Differential Effects of Epinephrine, Norepinephrine, and Indole on Escherichia coli O157:H7 Chemotaxis, Colonization, and Gene Expression . Infect Immun 75 ( 2007 ). 13. ↵ Harvey , A. et al. The Re-Emergence of Natural Products for Drug Discovery in the Genomics Era . Nature 14 , 111 – 129 ( 2015 ). OpenUrl 14. ↵ Nikaido , E. et al. Effects of Indole on Drug Resistance and Virulence of Salmonella enterica Serovar Typhimurium Revealed by Genome-Wide Analyses . Gut Pathog ( 2012 ). 15. Bommarius , B. et al. A Family of Indoles Regulate Virulence and Shiga Toxin Production in Pathogenic E. coli . PLoS One 8 ( 2013 ). 16. ↵ Zarkan , A. , Liu , J. , Matuszewska , M. , Gaimster , H. & Summers , D. K. Local and universal action: The Paradoxes of Indole Signalling in Bacteria . Trends Microbiol 28 , 566 – 577 ( 2020 ). OpenUrl CrossRef PubMed 17. ↵ Chu , W. et al. Indole Production Promotes Escherichia coli Mixed-Culture Growth with Pseudomonas aeruginosa by Inhibiting Quorum Signaling . Appl Environ Microbiol 78 , 411 – 419 ( 2012 ). OpenUrl Abstract / FREE Full Text 18. ↵ Armstrong , M. D. , Shaw , K. N. F. , Gortatowski , M. J. & Singer , H. The Indole Acids of Human Urine: Paper Chromatography of Indole Acids . J. Biol. Chem . 232 , 17 – 30 , DOI: 10.1016/S0021-9258(18)70370-0 ( 1958 ). OpenUrl FREE Full Text 19. Said , A. , El-Gamal , M. , Abu-Elghait , M. & Salem , S. S. Isolation, Identification, and Antibiotic Susceptibility Pattern of Urinary Tract Infection Bacterial Isolates . Lett. Appl. NanoBioScience 10 , 2820 – 2830 , DOI: 10.33263/LIANBS104.28202830 ( 2021 ). OpenUrl CrossRef 20. Kema , I. P. , de Vries , E. G. E. & Muskiet , F. A. J. Clinical Chemistry of Serotonin and Metabolites . J. Chromatogr. B: Biomed. Sci. Appl . 747 , 33 – 48 , DOI: 10.1016/S0378-4347(00)00341-8 ( 2000 ). OpenUrl CrossRef PubMed 21. ↵ Hu , W. , Garrison , C. , Prasad , R. , Boulton , M. E. & Grant , M. B. Indole Metabolism and Its Role in Diabetic Macrovascular and Microvascular Complications . Am. Hear. J. Plus: Cardiol. Res. Pract . 53 , 100532 , DOI: 10.1016/j.ahjo.2025.100532 ( 2025 ). OpenUrl CrossRef 22. ↵ Boya , B. R. , Kumar , P. , Lee , J. H. & Lee , J. Diversity of the Tryptophanase Gene and Its Evolutionary Implications in Living Organisms . Microorganisms 15 , DOI: 10.3390/microorganisms9102156 ( 2021 ). OpenUrl CrossRef 23. ↵ Wood , W. A. , Gunsalus , I. C. & Umbreit , W. W. Function of Pyridoxal Phosphate: Resolution and Purification of the Tryptophanase Enzyme of Escherichia coli . J. Biol. Chem . 170 , 313 – 321 , DOI: 10.1016/S0021-9258(17)34959-1 ( 1947 ). OpenUrl FREE Full Text 24. ↵ Newton , W. A. , Morino , Y. & Snell , E. E. Properties of Crystalline Tryptophanase . J. Biol. Chem . 240 , 1211 – 1218 , DOI: 10.1016/S0021-9258(18)97562-9 ( 1965 ). OpenUrl FREE Full Text 25. ↵ Snell , E. E. Tryptophanase: Structure, Catalytic Activities, and Mechanism of Action . Adv Enzym. Relat Areas Mol Biol 42 , 287 – 333 , DOI: 10.1002/9780470122877 ( 1975 ). OpenUrl CrossRef PubMed Web of Science 26. ↵ Watanabe , T. & Snell , E. E. The Interaction of Escherichia coli Tryptophanase with Various Amino Acids and Their Analogs: Active Site Mapping . The J. Biochem . 82 , 733 – 745 , DOI: 10.1093/oxfordjournals.jbchem.a131750 ( 1977 ). https://academic.oup.com/jb/article-pdf/82/3/733/2535577/82-3-733.pdf . OpenUrl CrossRef PubMed 27. Zakomirdina , L. N. , Sakharova , I. S. & Torchinsky , Y. M. Linear Dichroism Studies of Tryptophanase and Its Quasisubstrate Complexes Oriented in Polyacrylamide Gel . Eur. J. Biochem . 193 , 243 – 247 , DOI: 10.1111/j.1432-1033.1990.tb19329.x ( 1990 ). https://febs.onlinelibrary.wiley.com/doi/pdf/10.1111/j.1432-1033.1990.tb19329.x . OpenUrl CrossRef PubMed 28. Metzler , C. M. et al. Equilibria and Absorption Spectra of Tryptophanase . J. Biol. Chem . 266 , 9374 – 9381 ( 1991 ). OpenUrl Abstract / FREE Full Text 29. ↵ Phillips , R. S. Reaction of Indole and Analogues with Amino Acid Complexes of Escherichia coli Tryptophan Indole-Lyase: Detection of a New Reaction Intermediate by Rapid-Scanning Stopped-Flow Spectrophotometry . Biochemistry 30 , 5927 – 5934 ( 1991 ). OpenUrl PubMed 30. ↵ Vang , D. et al. Surface-enhanced raman spectroscopy and multivariate analysis for elucidating mechanisms of action in antibacterial agents . ACS Sensors 10 , 2689 – 2698 ( 2025 ). OpenUrl PubMed 31. Premasiri , W. R. , Chen , Y. , Williamson , P. M. et al. Rapid Urinary Tract Infection Diagnostics by Surface-Enhanced Raman Spectroscopy (SERS): Identification and Antibiotic Susceptibilities . Anal Bioanal Chem 409 , 3043 – 3054 , DOI: 10.1007/s00216-017-0244-7 ( 2017 ). OpenUrl CrossRef PubMed 32. Ho , C. S. et al. Rapid Identification of Pathogenic Bacteria Using Raman Spectroscopy and Deep Learning . Nat. Commun . 10 ( 2019 ). 33. ↵ Tahir , M. A. , Dina , N. E. , Cheng , H. , Valev , V. K. & Zhang , L. Surface-enhanced Raman Spectroscopy for Bioanalysis and Diagnosis . Nanoscale 13 , 11593 – 11634 , DOI: 10.1039/D1NR00708D ( 2021 ). OpenUrl CrossRef PubMed 34. ↵ Zhai , Y. et al. Synergistic Enhancement Effect for Boosting Raman Detection Sensitivity of Antibiotics . ACS Sensors 4 , 2958 – 2965 , DOI: 10.1021/acssensors.9b01436 ( 2019 ). OpenUrl CrossRef PubMed 35. ↵ Krafft , C. & Popp , J. The Many Facets of Raman Spectroscopy for Biomedical Analysis . Anal. Bioanal. Chem . 407 , 699 – 717 , DOI: 10.1007/s00216-014-8311-9 ( 2015 ). OpenUrl CrossRef PubMed 36. ↵ Rebrošová , K. et al. Rapid Identification of Staphylococci by Raman Spectroscopy . Sci Rep 7 ( 2017 ). 37. ↵ Liu , C. H. , Das , B. , Glassman , W. et al. Raman, Fluorescence, and Time-Resolved Light Scattering as Optical Diagnostic Techniques to Separate Diseased and Normal Biomedical Media . J. Photochem. Photobiol. B: Biol . 16 , 187 – 209 ( 1992 ). OpenUrl CrossRef PubMed Web of Science 38. ↵ Lorenz , B. , Wichmann , C. , Stöckel , S. , Rösch , P. & Popp , J. Cultivation-free Raman Spectroscopic Investigations of Bacteria . Trends Microbiol . 25 , 413 – 424 ( 2017 ). OpenUrl CrossRef PubMed 39. ↵ Kuku , A. et al. Surface-enhanced Raman Scattering for Label-Free Living Single Cell Analysis . Anal. Chem . 89 , 11160 – 11166 ( 2017 ). OpenUrl 40. ↵ Henry , A. et al. Surface-enhanced Raman Spectroscopy Biosensing: In Vivo Diagnostics and Multimodal Imaging . Anal. Chem . 88 , 6638 – 6647 ( 2016 ). OpenUrl 41. ↵ Cialla-May , D. , Zheng , X.-S. , Weber , K. & Popp , J. Recent Progress in Surface-Enhanced Raman Spectroscopy for Biological and Biomedical Applications: From Cells to Clinics . Chem Soc Rev 46 , 3945 – 3961 ( 2017 ). OpenUrl PubMed 42. ↵ Bodelón , G. , Montes-García , V. , López-Puente , V. et al. Detection and Imaging of Quorum Sensing in Pseudomonas aeruginosa Biofilm Communities by Surface-Enhanced Resonance Raman Scattering . Nat. Mater . 15 , 1203 – 1211 ( 2016 ). OpenUrl CrossRef PubMed 43. ↵ Sibug-Torres , S. , Grys , D. B. , Kang , G. et al. In Situ Electrochemical Regeneration of Nanogap Hotspots for Continuously Reusable Ultrathin SERS Sensors . Nat. Commun . 15 , DOI: 10.1038/s41467-024-46097-y ( 2024 ). OpenUrl CrossRef 44. Niihori , M. et al. Sers Sensing of Dopamine with Fe(III)-Sensitized Nanogaps in Recleanable AuNP Monolayer Films . Small 19 , 2302531 , DOI: 10.1002/smll.202302531 ( 2023 ). https://onlinelibrary.wiley.com/doi/pdf/10.1002/smll.202302531 . OpenUrl CrossRef 45. ↵ Chen , W.-H. et al. Plasmonic Sensing Assay for Long-Term Monitoring (PSALM) of Neurotransmitters in Urine . ACS Nanosci. Au 3 , 161 – 171 , DOI: 10.1021/acsnanoscienceau.2c00048 ( 2023 ). OpenUrl CrossRef PubMed 46. ↵ Jayan , H. et al. Detection of Bioactive Metabolites in Escherichia coli Cultures Using Surface-Enhanced Raman Spectroscopy . Appl. Spectrosc . 76 , 812 – 822 ( 2022 ). OpenUrl PubMed 47. ↵ Dieng , S. D. & Schelvis , J. P. M. Analysis of Measured and Calculated Raman Spectra of Indole, 3-Methylindole, and Tryptophan on the Basis of Observed and Predicted Isotope Shifts . The J. Phys. Chem. A 114 , 10897 – 10905 , DOI: 10.1021/jp107295p ( 2010 ). OpenUrl CrossRef PubMed 48. ↵ Meo , S. R. , Van de Wiele , T. & Defoirdt , T. Indole Signaling in Escherichia coli: A Target for Antivirulence Therapy? Gut Microbes 17 , 2499573 , DOI: 10.1080/19490976.2025.2499573 ( 2025 ). OpenUrl CrossRef PubMed 49. ↵ Lu , Z. et al. The Kynurenine Pathway and Indole Pathway in Tryptophan Metabolism Influence Tumor Progression . Cancer Med 14 , DOI: 10.1002/cam4.70703 ( 2025 ). OpenUrl CrossRef 50. ↵ Darkoh , C. , Chappell , C. , Gonzales , C. & Okhuysen , P. A Rapid and Specific Method for the Detection of Indole in Complex Biological Samples . Appl Environ Microbiol 81 , 8093 – 8097 , DOI: 10.1128/AEM.02787-15 ( 2015 ). OpenUrl Abstract / FREE Full Text 51. ↵ Wang , Z. , Klipfell , E. , Bennett , B. et al. Gut Flora Metabolism of Phosphatidylcholine Promotes Cardiovascular Disease . Nature 472 , 57 – 63 , DOI: 10.1038/nature09922 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 52. ↵ Skov , K. , Oxfeldt , M. , Thøgersen , R. , Hansen , M. & Bertram , H. C. Enzymatic Hydrolysis of a Collagen Hydrolysate Enhances Postprandial Absorption Rate—A Randomized Controlled Trial . Nutrients 13 , DOI: 10.3390/nu11051064 ( 2019 ). OpenUrl CrossRef 53. ↵ Wood , S. et al. Raman Spectroscopy as an Advanced Structural Nanoprobe for Conjugated Molecular Semiconductors . J. Phys. D: Appl. Phys . 50 , DOI: 10.1088/1361-6463/50/7/073001 ? ( 2017 ). OpenUrl CrossRef 54. ↵ Blacha-Grzechnik , A. , Karon , K. & Data , P. Raman and IR Spectroelectrochemical Methods as Tools to Analyze Conjugated Organic Compounds . J Vis Exp 12 , DOI: 10.3791/56653 ( 2018 ). OpenUrl CrossRef 55. ↵ Chen , J. , Bender , S. L. , Keough , J. M. & Barry , B. A. Tryptophan as a Probe of Photosystem I Electron Transfer Reactions: A UV Resonance Raman Study . The J. Phys. Chem. B 113 , 11367 – 11370 , DOI: 10.1021/jp906491r ( 2009 ). OpenUrl CrossRef PubMed 56. ↵ Shafaat , H. S. , Leigh , B. S. , Tauber , M. J. & Kim , J. E. Resonance Raman Characterization of a Stable Tryptophan Radical in an Azurin Mutant . The J. Phys. Chem. B 113 , 382 – 388 , DOI: 10.1021/jp809329a ( 2009 ). OpenUrl CrossRef PubMed 57. ↵ Ren , X. et al. A Comprehensive Review and Comparison of L-Tryptophan Biosynthesis in Saccharomyces cerevisiae and Escherichia coli . Front. Bioeng. Biotechnol . 11 , DOI: 10.3389/fbioe.2023.1261832 ( 2023 ). OpenUrl CrossRef PubMed 58. ↵ Rhee , K. Y. , Parekh , B. S. & Hatfield , G. W. Leucine-Responsive Regulatory Protein–DNA Interactions in the Leader Region of the ilvGMEDA Operon of Escherichia coli . J. Biol. Chem . 271 , 26499 – 26507 , DOI: 10.1074/jbc.271.43.26499 ( 1996 ). OpenUrl Abstract / FREE Full Text 59. ↵ Calvo , J. M. & Matthews , R. G. The Leucine-Responsive Regulatory Protein, a Global Regulator of Metabolism in Escherichia coli . Microbiol Rev 58 , DOI: 10.1128/mr.58.3.466-490.1994 ( 1994 ). OpenUrl CrossRef 60. ↵ Gaimster , H. , Cama , J. , Hernández-Ainsa , S. , Keyser , U. F. & Summers , D. K. The Indole Pulse: A New Perspective on Indole Signalling in Escherichia coli . PLoS ONE 9 , DOI: 10.1371/journal.pone.0093168 ( 2014 ). OpenUrl CrossRef PubMed 61. ↵ Linares-Otoya , L. , Shirkey , J. D. , Chhetri , B. K. et al. Discovery of a Widespread Chemical Signalling Pathway in the Bacteroidota . Nature 646 , 423 – 432 , DOI: 10.1038/s41586-025-09418-9 ( 2025 ). OpenUrl CrossRef PubMed 62. ↵ Qi , Y. et al. Plasmon-Driven Chemistry in Ferri-/Ferrocyanide Gold Nanoparticle Oligomers: A SERS Study . J. Am. Chem. Soc . 142 , 13120 – 13129 , DOI: 10.1021/jacs.0c05031 ( 2020 ). OpenUrl CrossRef PubMed 63. Xiao , S. et al. Expanding the Application of Tryptophan: Industrial Biomanufacturing of Tryptophan Derivatives . Front. Microbiol . 14 , DOI: 10.3389/fmicb.2023.1099098 ( 2023 ). OpenUrl CrossRef PubMed 64. ↵ Melander , R. J. , Minvielle , M. J. & Melander , C. Controlling Bacterial Behavior with Indole-Containing Natural Products and Derivatives . Tetrahedron 70 , 6363 – 6372 , DOI: 10.1016/j.tet.2014.05.089 ( 2014 ). OpenUrl CrossRef PubMed 65. ↵ Lamb , A. , Federico-Perez , R. & Zl., X. Product in indole detection by Ehrlich’s reagent . Anal Biochem . 484 , DOI: 10.1016/j.ab.2015.04.033 ( 2015 ). OpenUrl CrossRef PubMed 66. ↵ Su , X. , Gao , Y. & Yang , R. Gut Microbiota-Derived Tryptophan Metabolites Maintain Gut and Systemic Homeostasis . Cells 11 , DOI: 10.3390/cells11152296 ( 2022 ). OpenUrl CrossRef 67. ↵ Haduch , A. , Bromek , E. , Kuban , W. & Daniel , W. A. The Engagement of Cytochrome P450 Enzymes in Tryptophan Metabolism . Metabolites 13 , DOI: 10.3390/metabo13050629 ( 2023 ). OpenUrl CrossRef PubMed 68. ↵ Li , G. & Young , K. D. Indole Production by the Tryptophanase TnaA in Escherichia coli Is Determined by the Amount of Exogenous Tryptophan . Microbiology 159 , 402 – 410 , DOI: 10.1099/mic.0.064139-0 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 69. ↵ Pittard , J. & Yang , J. Biosynthesis of the Aromatic Amino Acids . EcoSal Plus 3 , DOI: 10.1128/ecosalplus.3.6.1.8 ( 2008 ). OpenUrl CrossRef PubMed 70. ↵ Lane , A. N. & Kirschner , K. The Mechanism of Binding of L-Serine to Tryptophan Synthase from Escherichia coli . Eur. J. Biochem . 129 , 561 – 570 , DOI: 10.1111/j.1432-1033.1983.tb07086.x ( 1983 ). OpenUrl CrossRef PubMed 71. ↵ Fang , Z. Z. & Gonzalez , F. J. LC-MS-based Metabolomics: An Update . Arch Toxicol 88 ( 2014 ). 72. ↵ Schrimpe-Rutledge , A. C. , Codreanu , S. G. , Sherrod , S. D. & McLean , J. A. Untargeted Metabolomics Strategies—Challenges and Emerging Directions . J. Am. Soc. for Mass Spectrom . 27 , 1897 – 1905 , DOI: 10.1007/s13361-016-1469-y ( 2016 ). OpenUrl CrossRef PubMed 73. ↵ Zhou , B. , Xiao , J. F. , Tuli , L. & Ressom , H. W. LC-MS-based Metabolomics . Mol. BioSyst . 8 , 470 – 481 , DOI: 10.1039/C1MB05350G ( 2012 ). OpenUrl CrossRef PubMed 74. ↵ Baba , T. et al. Construction of escherichia coli k-12 in-frame, single-gene knockout mutants: the keio collection . Mol. Syst. Biol . 2 , 2006 .0008 , DOI: 10.1038/msb4100050 ( 2006 ). OpenUrl CrossRef PubMed 75. ↵ Datsenko , K. A. & Wanner , B. L. One-step inactivation of chromosomal genes in escherichia coli k-12 using pcr products . Proc. Natl. Acad. Sci . 97 , 6640 – 6645 , DOI: 10.1073/pnas.120163297 ( 2000 ). OpenUrl Abstract / FREE Full Text 76. ↵ Kim , J. , Webb , A. M. , Kershner , J. P. , Blaskowski , S. & Copley , S. D. A versatile and highly efficient method for scarless genome editing in escherichia coli and salmonella enterica . BMC Biotechnol . 14 , 84 , DOI: 10.1186/1472-6750-14-84 ( 2014 ). OpenUrl CrossRef PubMed 77. ↵ Baba , T. et al. Construction of escherichia coli k-12 in-frame, single-gene knockout mutants: the keio collection . Mol. Syst. Biol . 2 , 2006 .0008 , DOI: 10.1038/msb4100050 ( 2006 ). OpenUrl CrossRef PubMed 78. ↵ Gaimster , H. & Summers , D. Regulation of indole signalling during the transition of e. coli from exponential to stationary phase . PLOS ONE 10 , 1 – 11 , DOI: 10.1371/journal.pone.0136691 ( 2015 ). OpenUrl CrossRef PubMed 79. ↵ Chant , E. L. & Summers , D. K. Indole signalling contributes to the stable maintenance of escherichia coli multicopy plasmids . Mol. Microbiol . 63 , 35 – 43 , DOI: 10.1111/j.1365-2958.2006.05481.x ( 2007 ). OpenUrl CrossRef PubMed Web of Science 80. ↵ Zarkan , A. , Caño-Muñiz , S. , Zhu , J. et al. Indole Pulse Signalling Regulates the Cytoplasmic pH of E. coli in a Memory-Like Manner . Sci Rep 9 , DOI: 10.1038/s41598-019-40560-3 ( 2019 ). OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted November 11, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Nanoplasmonic SERS reveals previously uncharacterised indole derivative in E. coli metabolism Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Nanoplasmonic SERS reveals previously uncharacterised indole derivative in E. coli metabolism Mo Vali , Elle W. Wyatt , Kieran Abbott , Cameron Croft , Pietro Lio , Thomas F. Krauss , Minahil Khan , Ashraf Zarkan , Jeremy J. Baumberg , Diana Fusco bioRxiv 2025.11.10.687708; doi: https://doi.org/10.1101/2025.11.10.687708 Share This Article: Copy Citation Tools Nanoplasmonic SERS reveals previously uncharacterised indole derivative in E. coli metabolism Mo Vali , Elle W. Wyatt , Kieran Abbott , Cameron Croft , Pietro Lio , Thomas F. Krauss , Minahil Khan , Ashraf Zarkan , Jeremy J. Baumberg , Diana Fusco bioRxiv 2025.11.10.687708; doi: https://doi.org/10.1101/2025.11.10.687708 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Microbiology Subject Areas All Articles Animal Behavior and Cognition (7636) Biochemistry (17704) Bioengineering (13898) Bioinformatics (41967) Biophysics (21460) Cancer Biology (18599) Cell Biology (25525) Clinical Trials (138) Developmental Biology (13384) Ecology (19909) Epidemiology (2067) Evolutionary Biology (24326) Genetics (15613) Genomics (22512) Immunology (17740) Microbiology (40423) Molecular Biology (17191) Neuroscience (88645) Paleontology (667) Pathology (2835) Pharmacology and Toxicology (4825) Physiology (7646) Plant Biology (15158) Scientific Communication and Education (2046) Synthetic Biology (4302) Systems Biology (9825) Zoology (2271)