Mass spectrometry-cleavable photo-crosslinkers for proteome-wide photo-crosslinking mass spectrometry

Abstract

The presently disclosed subject matter provides compositions and methods for the synthesizing and the application of heterobifunctional photo-crosslinkers for the analysis of protein-protein interactions in mass spectrometry, tandem mass spectrometry, and cross-linking mass spectrometry.

Description

[0001] MASS SPECTROMETRY-CLEAVABLE PHOTO-CROSSLINKERS FOR PROTEOME- WIDE PHOTO-CROSSLINKING MASS SPECTROMETRY FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT This invention was made with government support under grant GM140926 awarded by NIH. The government has certain rights in the invention. CROSS-REFERENCE TO RELATED APPLICATIONS This application claims benefit of U.S. Provisional Patent Application No.63 / 583,709 filed on September 19, 2023, which is incorporated herein by reference in its entirety. BACKGROUND Crosslinking mass spectrometry (XL-MS) is an emerging technique situated at the intersection of structural and cellular biology and plays a pivotal role in deciphering the intricacies of biological macromolecules. XL-MS offers residue-level resolution for identifying protein-protein interactions across the proteome by utilizing chemical crosslinkers. These crosslinkers break down fragments during XL-MS analysis, aiding in component identification and structural characterization. While effective for extensive single protein coverage, they pose challenges in complex samples, such as lysates, cells, and tissues. This complexity arises from the so-called N2problem, where the search space for potential crosslinked peptides scales quadratically with the number of linear peptides, thereby complicating spectral identification and increasing the risk of false discovery. Although, new XL-MS studies have emerged with a newer generation of crosslinkers to reduce false discovery for proteome wide XL-MS, none have produced chemically inert photo-reactive functional groups, which upon UV irradiation, generate highly reactive moieties that rapidly and covalently capture nearby proteins to identify crosslinks in proteome-wide experiments. SUMMARY In some aspects, the presently disclosed subject matter provides a compound of formula (I): 1 42282.601_P17605-02 I); wherein n is an integ , and 8. In certain aspects, the compound of formula (I) is selected from: . In other conjugate comprising a first peptide (P1), a second peptide (P2), and a urea-based linker derived from a compound of formula (I): ; wherein m and n are each 1, 2, 3, 4, 5, 6, 7, and 8. In other aspects, the presently disclosed subject matter provides a method for detecting protein-protein interactions in a sample, the method comprising: (a) contacting one or more proteins in the sample with a compound of formula (I) comprising a diazirine functional group and an NHS-carbamate functional group: (I); wherein n is an integer 7, and 8, whereby the NHS-carbamate functional group reacts with one or more lysine or N-terminal amine groups of the one or more proteins to generate a conjugate comprising a urea linked to the diazirine functional group and a first peptide residue, wherein the linkage between the urea and the first peptide residue is a first mass spectrometer (MS)-cleavable linkage; (b) irradiating the conjugate with ultraviolet (UV) radiation whereby the diazirine functional group forms a carbene group that reacts with a second peptide residue and links the second peptide residue to the urea, wherein the linkage between the urea and the second peptide residue is a second mass 2 42282.601_P17605-02 spectrometer (MS)-cleavable linkage; and (c) cleaving the urea at the first MS-cleavable linkage or the second MS-cleavable linkage to generate a first peptide product comprising an amine fragment and a second peptide product comprising an isocyanate fragment, wherein the first peptide product and the second peptide product are on the same protein or different proteins; and (c) analyzing the sample by mass spectrometry to detect protein-protein interactions in the sample, wherein the first peptide product and the second peptide product form characteristic doublets in a MS spectrum. In some aspects, the method further comprises identifying interprotein and / or intraprotein crosslinks. In some aspects, the Cα–Cα distance between the first peptide residue and the second peptide residue is below 30 Å, 25 Å, 20 Å, or 15 Å. In some aspects, the Cα–Cα distance between the first peptide residue and the second peptide residue is below 15 Å. In some aspects, the Cα–Cα distance between the first peptide residue and the second peptide residue is greater than 50 Å. In some aspects, the first peptide residue and / or the second peptide reside is a non-acidic residue. In some aspects, the first peptide residue and / or the second peptide reside is a Ala, Gly, Val, or Tyr. In some aspect, the sample is selected from a lysate, a cell, and a tissue sample. In some aspects, steps (a) and (b) are performed in vivo or in cellulo. In some aspects, the mass spectrometry comprises liquid chromatography (LC)-mass spectrometry / mass spectrometry (MS / MS). In some aspects, cleaving comprises higher-energy collision-induced dissociation (HCD) conditions. In some aspects, the UV radiation wavelength is 365 nm. In some aspects, the UV radiation wavelength is 365 nm. In some aspects, the UV radiation wavelength is in the range of 315-400 nm, 350-380 nm, 355-375nm, 360-370nm, 363-368nm. In some aspects, UV radiation is with a UV LED. In some aspects, the UV radiation is for 1, 2, or 3 seconds. In some aspects, the method further comprises adding a crosslinking catalyst during step (a). In some aspects, the method further comprises adding a quenching reagent prior to step (b). In some aspects, the method further comprises one or more of the following steps prior to the mass spectrometry analysis of step (c): (I) denaturing the sample; (II) reducing the sample by addition of dithiothreitol (DTT); (III) alkylating the sample; (IV) either: (i) digesting the sample with trypsin; or (ii) digesting serially with Lys-C and then trypsin; (V) desalting the digested samples; and / or (VI) fractionating and enriching the crosslinked 3 42282.601_P17605-02 sample by size exclusion chromatography (SEC). In some aspects, the method further comprises performing a spectral search for one or more crosslinked peptides. Certain aspects of the presently disclosed subject matter having been stated hereinabove, which are addressed in whole or in part by the presently disclosed subject matter, other aspects will become evident as the description proceeds when taken in connection with the accompanying Examples and Figures as best described herein below. BRIEF DESCRIPTION OF THE FIGURES The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee. Having thus described the presently disclosed subject matter in general terms, reference will now be made to the accompanying Figures, which are not necessarily drawn to scale, and wherein: FIG.1A, FIG.1B, FIG.1C, and FIG.1D show the structure and properties of MS- cleavable heterobifunctional photo-crosslinkers. (FIG.1A) Structures of photo-crosslinkers known in the art, including DSBU, a urea-based (teal box) cleavable crosslinker; NNP-9, an NHS-carbamate (red box) containing-crosslinker (Nury et al., 2015); SDA, a heterobifunctional diazirine (green box) crosslinker; and SDASO, a heterobifunctional sulfoxide-based (yellow box) cleavable photo-crosslinker. In contrast, shown on the right are the presently disclosed heterobifunctional NHS-carbamate photo-crosslinkers, DizSEC and DizSPC. (FIG.1B) Reaction of the NHS-carbamate functional group in DizSEC or DizSPC with lysine or N-terminal amines generates ureas. Following UV irradiation, the diazirine forms a carbene, which rapidly reacts with proximal proteins to form covalent crosslinks. (FIG.1C) Ureas are doubly-fissile under higher-energy collision-induced dissociation (HCD) (across each N-carbonyl bond) and generate two distinct sets of fragment ions (green (i) and purple (ii) dashed boxes), generating amine and isocyanate fragments on either peptide product (nomenclature used for these fragment ions shown). These crosslinked peptide fragments form characteristic doublets in the MS spectrum. (FIG.1D) A representative crosslinked spectrum match (XSM) generated by the treatment of BSA with DizSEC, along with a structure (PDB: 4F5S) illustrating identified crosslinks. The spectrum 4 42282.601_P17605-02 contains all four possible products from cleavage of the MS-cleavable crosslink (that is, the modified parental peptides); FIG.2A, FIG.2B, FIG.2C, FIG.2D, FIG 2E, and FIG.2F demonstrate proteome- wide identification of photo-crosslinks. (FIG.2A) The experimental workflow to prepare crosslinks for LC-MS / MS identification using DizSEC (or DizSPC) on E. coli extracts. (FIG.2B) Identified crosslinks from MeroX and XiSearch mapped onto the 70S E. coli ribosome (PDB: 7CPJ). (FIG.2C) Summary data showing the number of crosslinks mapped onto the ribosome for both search algorithms. (FIG.2D) Cumulative distribution plot showing the Cα–Cα distances of the crosslinks identified. The majority of crosslinks identified in both search algorithms agree with a cryo-EM structure of the ribosome (<15 Å). (FIG.2E) A structural representation of crosslinks mapped onto the 70S ribosome complexed with EF-G (PDB: 7N2C). The inset shows a detailed view of these crosslinks, which have lengths that agree with the structure. (FIG.2F) A representative interprotein crosslink spectrum match of RplT (αPep) crosslinked with RplU (βPep) with DizSPC); FIG.3A, FIG.3B, FIG.3C, FIG.3D, FIG.3E, and FIG.3F demonstrate assessment of false discovery across algorithms. (FIG.3A) Summary data comparing the number of unique crosslinks identified by MeroX and Xi in crosslinking in cellulo experiments with DizSEC / DizSPC. (FIG.3B) MeroX finds many interprotein crosslinks between proteins with low or unknown copy number; this finding is not recapitulated in the Xi workflow. (FIG.3C) MeroX often finds impossible crosslinks when a double validation is performed. Erroneous yeast crosslinks make up 22% of identifications for this fraction and 81% of interprotein identifications. Conversely, there are far fewer erroneous XiSearch IDs for the DizSEC in-cell fractions. (FIG.3D) Number of crosslinks identified in the Xi workflow for the DizSEC in cellulo experiment, sorted by the identity of the “non-lysine” attachment site. Below, the enrichment for each residue normalized to its frequency in the proteome. (FIG. 3E) Doubly-fissile crosslinkers leave a modification on 3 of the 4 possible cleavage products, unlike scenarios in which simple diazirines (e.g., photo-Met) attach to Asp or Glu, which do not leave modifications on the fragments ions. These unmodified cleavage products are indistinguishable from associated co-isolating peptides. (FIG.3F) An MS2spectrum showing a NusA peptide (at K117) crosslinked with an RpoB peptide (at E899) via DizSEC. Three sets of modified, cleaved, parental peptide doublets are highlighted in 5 42282.601_P17605-02 yellow, resulting from cleavage at the urea moiety. The pink arrow points to a product formed upon cleavage at the ester, which produces the unmodified nonlysine peptide; FIG.4A, FIG.4B, and FIG.4C show crosslink validation by comparison to structures of heterocomplexes. (FIG.4A) Identified crosslinks from XiSearch are mapped onto the EF-Tu / EF-Ts complex (PDB: 4PC3). (FIG.4B) Identified crosslinks from XiSearch are mapped onto the H-NS (AlphaFold: P0ACG) / StpA (AlphaFold: P0ACF8) complex, docked in HADDOCK. The N-termini of the proteins interact with one another, while the C- terminal domain and linker interact with DNA. (FIG.4C) Structural model generated in which Alphafold structures of ManX (AlphaFold: P69797) and ManY (AlphaFold: P69801) are docked based on crosslinks, and then ManY aligned with the cryo-EM structure (PDB: 6K1H) of the whole pore (ManY3ManZ3). The mannose bound to loops in ManY and ManZ is shown in magenta, and catalytic phosphohistidines in ManX are shown in orange. Crosslinks used for docking are highlighted with yellow background; FIG.5A, FIG.5B, and FIG.5C show subnetworks among dynamic complexes reveals structural insights. (FIG.5A) Subnetwork featuring known interactions between RNA Polymerase and NusA (PDB: 6FLQ), and an unreported interaction between NusA and HupA, a histone-like protein in E. coli. A structural model in which HupAB (PDB: 4YEW) is docked and cumulative distribution plot of Cα–Cα distances are provided. The crosslink used for docking is highlighted with yellow background. (FIG.5B) Structural model, network plot, and cumulative distribution plot for the ATP synthase complex (PDB: 6OQR). (FIG.5C) Subnetwork associated with the E. coli degradosome, featuring known interactions between PNPase, RNAse E, and the ribosome. Unreported interaction between enolase and the ribosome suggests a potential role as an adaptor; FIG.6 shows Scheme 3: shows a Xi workflow flow-chart diagram illustrating the computational workflow; FIG.7A, FIG.7B, FIG.7C, FIG.7D, FIG.7E, FIG.7F demonstrate proteome-wide identification of photo-crosslinks. FIG.7A shows the experimental workflow to prepare crosslinks for LC-MS / MS identification using DizSEC (or DizSPC) on E. coli extracts. FIG. 7B shows a bar chart showing crosslink-spectrum match (XSM) type distribution; FIG.7C shows a bar chart showing unique crosslink type distribution. FIG.7D shows a model to explain DizSEC / DizSPC’s selectivity for interprotein crosslink formation due to 6 42282.601_P17605-02 their size and lifetime compared to DSBU / DSSO chemical crosslinking. FIG.7E shows a representative MS2 spectrum from the DizSPC / +Lys-C experiment that matches a crosslinked peptide in which the terminal lysine of one peptide is the attachment site for DizSPC – implying proteolysis at the modified lysine. FIG.7F shows a Euler diagram illustrating the general absence of overlap amongst identified crosslinks across the four experimental conditions tested. Figure generated with BioVenn; FIG.8A, FIG.8B, FIG.8C, and FIG.8D demonstrate residue-level protein-protein interaction network of E. coli lysate. FIG.8A shows a network plot representing 1986 distinct crosslinks by pooling data from all four experimental conditions on E. coli extracts. This network can be interrogated interactively at bit.ly / DizSEC_1 through the web-based XiView tool. FIG.8B shows identified crosslinks associated with ribosomal proteins are in excellent agreement with a cryo-EM structure of the 70S E. coli ribosome (PDB: 4V9D). Five zoom-in views highlight that all Cα–Cα distances are less than 30 Å in Euclidean distance. FIG.8C and FIG.8D show interaction networks of two E. coli proteins (RplL (FIG.8C) and AceF (FIG.8D)) that are hubs for interactions with many other E. coli proteins. As XL-MS provides information on both protein identity and residue, these networks illustrate regions of the hub proteins that are “hot spots” (65-75, 80-90, and 100- 105 on RplL; and 90-105, 190-210, 290-315 on AceF; FIG.9A, FIG.9B, FIG.9C, FIG.9D, and FIG.9E demonstrate in-cell Photo- crosslinking Visualizes E. coli Molecular Sociology at Residue-Level Resolution. FIG.9A shows a workflow for in-cell photolysis. Cells are pumped through a flow cell, where the UV-LED array rapidly crosslinks the samples, which are then collected in a separate vessel. A photo of the setup in operando is shown. FIG.9B shows a network plot representing 2131 distinct crosslinks by pooling data from two in-cell photo-crosslinking experiments. This network can be interrogated interactively at bit.ly / DizSEC_2 through the web-based XiView tool. FIG.9C shows a subnetwork spanning between the proteins RpoA, RpoB, RpoC, NusA, HupA, and HupB. Crosslinks between RpoA / NusA and RpoB / NusA are mapped onto a cryo-EM structure of the RNA polymerase / NusA complex (PDB: 6FLQ), and are in excellent agreement. Crosslinks between HupA / HupB are mapped onto an x-ray structure of the HupA / HupB heterodimer (PDB: 4YEW), and are also in excellent agreement. We additionally discover a crosslink (red dotted-line) between these two structures in vivo, 7 42282.601_P17605-02 putatively involved in transcriptional regulation. FIG.9D shows comparison of the number and type of XSMs in lysate and in-cell experiments. FIG.9E shows comparison of the number and type of unique crosslinks identified in lysate and in-cell experiments; FIG.10A, FIG.10B, and FIG.10C demonstrate photo-crosslinks of the E. coli proteostasis machinery. FIG.10A shows chaperone-client crosslink networks derived from the in-cell crosslinking experiments, demonstrating the ability of XL-MS to capture interactions, including those with a cryptic chaperone, HtpG (highlighted in yellow). FIG. 10B shows the number of unique inter-protein crosslinks detected in in-cell photo-XL-MS experiments, categorized by the abundance of the two proteins connected by the crosslink. Low-abundance proteins are defined as those with ≤ 200 copies / cell. Proteins without annotated abundance in ref.73 are treated as low abundance; green (high-high), yellow (high-low); red (low-low). FIG.10C shows the number of proteins identified in in-cell photo-XL-MS experiments, categorized by their Protein Existence (PE) score. Enrichment factors are calculated relative to search on linear peptides; FIG.11A and FIG.11B show crosslinking of lysozyme with DizSEC and DizSPC. (FIG.11A) Crosslinks of lysozyme with DizSEC and DizSPC from all trial experiments shown in linear chart form. FIG.11A shows circular chart form. FIG.11B shows crosslinks mapped onto a hen-egg lysozyme crystal structure 3LYZ; and FIG.12 shows enrichment of crosslinks using size exclusion chromatography (SEC). Crosslinks were enriched on an Akta Go FPLC system. Crosslinked fractions were collected between 12 mL and 20 mL of elution volume. Between 12 mL and 18 mL of elution volume, up to twelve fractions of 0.5 mL were collected for analysis. Between 18 mL and 25 mL, up to 7 fractions of 1 mL were collected for analysis. DETAILED DESCRIPTION The presently disclosed subject matter now will be described more fully hereinafter with reference to the accompanying Figures, in which some, but not all embodiments of the inventions are shown. Like numbers refer to like elements throughout. The presently disclosed subject matter may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. Indeed, many 8 42282.601_P17605-02 modifications and other embodiments of the presently disclosed subject matter set forth herein will come to mind to one skilled in the art to which the presently disclosed subject matter pertains having the benefit of the teachings presented in the foregoing descriptions and the associated Figures. Therefore, it is to be understood that the presently disclosed subject matter is not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. In some embodiments, the presently disclosed subject matter provides a compound of formula (I): ; wherein n is an integer In certain embodiments, the compound of formula (I) is selected from: . In other a conjugate comprising a first peptide (P1), a second peptide (P2), and a urea-based linker derived from a compound of formula (I): ; wherein m and n are each 1, 2, 3, 4, 5 , 6 , 7, and 8. In yet other embodiments, the presently disclosed subject matter provides a method for detecting protein-protein interactions in a sample, the method comprising: (a) contacting one or more proteins in the sample with a compound of formula (I) comprising a diazirine functional group and an NHS-carbamate functional group: 9 42282.601_P17605-02 (I); wherein n is an integer selec , 7, and 8, whereby the NHS-carbamate function group reacts with one or more lysine or N-terminal amine groups of the one or more proteins to generate a conjugate comprising the diazirine functional group linked to the one or more proteins through a urea linkage; (b) irradiating the conjugate with ultraviolet (UV) radiation whereby the diazirine functional group forms a carbene group, which reacts with one or more proximal proteins to form a covalent crosslink between one or more peptides of the one or more proximal proteins; and (c) analyzing the sample by mass spectrometry to detect protein-protein interactions in the sample. In particular embodiments, the method further comprises performing a spectral search for one or more crosslinked peptides as provided herein. In certain embodiments, the sample is selected from a lysate, a cell, and a tissue sample. In particular embodiments, the method is performed in vivo or in cellulo. In certain embodiments, the mass spectrometry comprises liquid chromatography (LC)-mass spectrometry / mass spectrometry (MS / MS). In particular embodiments, the mass spectrometry is performed under higher-energy collision-induced dissociation (HCD) conditions. In certain embodiments, the mass spectrometry analysis generates amine and / or isocyanate fragments of the one or more crosslinked peptides. In certain embodiments, the amine and / or isocyanate fragments of the one or more crosslinked peptides form doublets in a mass spectrometry spectrum thereof. In particular embodiments, the UV radiation is about 365 nm. In certain embodiments, the method further comprises adding a crosslinking catalyst during step (a). In certain embodiments, the method further comprises adding a quenching reagent prior to step (b). In certain embodiments, the method further comprises one or more of the following steps prior to the mass spectrometry analysis of step (c): 10 42282.601_P17605-02 (a) denaturing the sample; (b) reducing the sample by addition of dithiothreitol (DTT); (c) alkylating the sample; (d) either: (i) digesting the sample with trypsin; or (ii) digesting serially with Lys-C and then trypsin; (e) desalting the digested samples; and / or (f) fractionating and enriching the crosslinked sample by size exclusion chromatography (SEC). Following long-standing patent law convention, the terms “a,” “an,” and “the” refer to “one or more” when used in this application, including the claims. Thus, for example, reference to “a subject” includes a plurality of subjects, unless the context clearly is to the contrary (e.g., a plurality of subjects), and so forth. Throughout this specification and the claims, the terms “comprise,” “comprises,” and “comprising” are used in a non-exclusive sense, except where the context requires otherwise. Likewise, the term “include” and its grammatical variants are intended to be non-limiting, such that recitation of items in a list is not to the exclusion of other like items that can be substituted or added to the listed items. For the purposes of this specification and appended claims, unless otherwise indicated, all numbers expressing amounts, sizes, dimensions, proportions, shapes, formulations, parameters, percentages, quantities, characteristics, and other numerical values used in the specification and claims, are to be understood as being modified in all instances by the term “about” even though the term “about” may not expressly appear with the value, amount or range. Accordingly, unless indicated to the contrary, the numerical parameters set forth in the following specification and attached claims are not and need not be exact, but may be approximate and / or larger or smaller as desired, reflecting tolerances, conversion factors, rounding off, measurement error and the like, and other factors known to those of skill in the art depending on the desired properties sought to be obtained by the presently disclosed subject matter. For example, the term “about,” when referring to a value can be meant to encompass variations of, in some embodiments, ± 20%, in some embodiments ± 15%, in some embodiments ± 10%, in some embodiments ± 5%, in some embodiments 11 42282.601_P17605-02 ±1%, in some embodiments ± 0.5%, and in some embodiments ± 0.1% from the specified amount, as such variations are appropriate to perform the disclosed methods or employ the disclosed compositions. Further, the term “about” when used in connection with one or more numbers or numerical ranges, should be understood to refer to all such numbers, including all numbers in a range and modifies that range by extending the boundaries above and below the numerical values set forth. The recitation of numerical ranges by endpoints includes all numbers, e.g., whole integers, including fractions thereof, subsumed within that range (for example, the recitation of 1 to 5 includes 1, 2, 3, 4, and 5, as well as fractions thereof, e.g., 1.5, 2.25, 3.75, 4.1, and the like) and any range within that range. EXAMPLES The following Examples have been included to provide guidance to one of ordinary skill in the art for practicing representative embodiments of the presently disclosed subject matter. In light of the present disclosure and the general level of skill in the art, those of skill can appreciate that the following Examples are intended to be exemplary only and that numerous changes, modifications, and alterations can be employed without departing from the scope of the presently disclosed subject matter. The descriptions and specific examples that follow are only intended for the purposes of illustration, and are not to be construed as limiting in any manner. EXAMPLE 1 Progress Toward Proteome-Wide Photo-Crosslinking to Enable Residue-Level Visualization of Protein Structures and Networks in vivo 1.1 Overview Crosslinking mass spectrometry (XL-MS) is emerging as a method at the crossroads of structural and cellular biology, uniquely capable of identifying protein-protein interactions with residue-level resolution and on the proteome-wide scale. With the development of crosslinkers that can form linkages inside cells and easily cleave during fragmentation on the mass spectrometer (i.e., MS-cleavable crosslinks), it has become increasingly facile to identify contacts between any two proteins in complex samples, 12 42282.601_P17605-02 including in live cells or tissues. Photo-crosslinkers possess the advantages of high temporal resolution and high reactivity, thereby engaging all residue-types (rather than just lysine); nevertheless, photo-crosslinkers have not enjoyed widespread use, and have yet to be employed for proteome-wide studies, because their products are challenging to identify. The presently disclosed subject matter demonstrates the synthesis and application of two heterobifunctional photo-crosslinkers that feature diazirines and N-hydroxy- succinimidyl carbamate groups, the latter of which unveil doubly-fissile MS-cleavable linkages upon acyl transfer to protein targets. Moreover, these crosslinkers demonstrate high water-solubility and cell-permeability. Using these compounds, the feasibility of proteome- wide photo-crosslinking in cellulo was demonstrated. These studies elucidate a small portion of E. coli’s interaction network, albeit with residue-level resolution. With further optimization, these methods will enable the detection of protein quinary interaction networks in their native environment at residue-level resolution, and it is expected they will prove useful toward the effort to explore the molecular sociology of the cell. 1.2 Background With recent technical and technological advances, crosslinking mass spectrometry (XL-MS) is emerging as a powerful structural proteomics technique – uniquely capable of elucidating protein-protein interaction (PPI) networks at residue-level resolution and on the proteome-wide scale. Iacobucci et al., 2020; Mishra et al., 2020; Lenz et al., 2021; Wheat et al., 2021. More particularly, improvements in high-resolution mass analyzers, chemical crosslinkers, and analysis software have joined forces to move XL-MS from a structural biology method (applied to purified protein complexes) to an “-omics” technique that bridges between cellular and structural biology. For example, in a recent report, O'Reilly and co-workers combined in-cell XL-MS with cryo-electron tomography (cryo-ET) to develop a structural model of the RNA polymerase-ribosome complex performing coupled transcription-translation in situ. O’Reilly et al., 2020. Early crosslinkers applied in XL-MS studies, such as disuccinimidyl suberate (DSS) and bis(sulfosuccinimidyl)suberate (BS3), are homobifunctional compounds consisting of two NHS-esters connected by a hydrocarbon linker. Leitner et al., 2010; Sinz 2003; Trakselis et al., 2005; Bishop et al., 2001; Petrotchenko et al., 2010; Tang et al., 2010. Following chemical crosslinking and trypsinolysis, these compounds generate pairs of 13 42282.601_P17605-02 peptides (called crosslinked peptides) irreversibly adjoined between two nucleophilic residues (typically lysines). These crosslinkers proved useful for obtaining high coverage on single proteins; however, they were challenging to apply to complex samples, such as lysates, cells, and tissues. Without wishing to bound to any one particular theory, it is thought that this challenge arises because the complexity of the search space associated with possible crosslinked peptides scales as the number of linear peptides squared (known as the N2problem), Liu et al., 2015, which complicates spectral identification and raises the chances of false discovery. Proteome-wide XL-MS studies emerged with the onset of a newer generation of crosslinkers, Arora et al., 2017; Matzinger and Mechtler, 2020, such as disuccinimidyl sulfoxide (DSSO) and disuccinimidyl dibutyric urea (DSBU, shown in FIG.1A), which can be cleaved in the gas phase during collision-induced dissociation (CID) or higher-energy collision-induced dissociation (HCD). These crosslinkers feature a functional group (cyan and yellow boxes in FIG.1) which is readily broken, splitting the crosslinked peptide into two modified linear peptides. Since the functional group also is symmetrical, fissure can occur in one of two ways (green (FIG.1C(i)) and purple dotted lines (FIG.1C(ii))), which further facilitates the identification of spectra of crosslinked peptides by generating doublets with a characteristic separation (FIG.1C). These features collectively enable MS-cleavable crosslinkers to mitigate the N2problem by reducing the database size against which spectra are searched and allowing complex samples to be interrogated by XL-MS. While numerous functional groups have been reported to be MS-cleavable, Matzinger and Mechtler, 2020, sulfoxides and ureas are among the most frequently employed. O’Reilly and Rappsilber, 2018; Götze et al., 2019. With these chemical tools, several proteome-wide crosslinking studies have been reported, such as on Bacillus subtilis extracts, De Jong et al., 2017, E. coli extracts, De Jong et al., 2017; Liu et al., 2017, yeast extracts, Fürsch et al., 2020, Drosophila embryonic extracts, Götze et al., 2019, Mycoplasma pneumoniae cells, O’Reilly et al., 2020; Xue et al., 2022, murine mitochondria, Schweppe et al., 2017, murine heart tissues, Chavez et al., 2018, human adenocarcinoma extracts, Klykov et al., 2018, and in HEK-293 cells. Kaake et al., 2014; Jiao et al., 2022. 14 42282.601_P17605-02 Photo-crosslinkers (also called photo-affinity labels (PALs)) possess chemically inert photo-reactive functional groups, which upon UV irradiation, generate highly reactive moieties that rapidly and covalently capture nearby proteins. Moss, 2006; Das, 2011. Though azides, benzophenones and diazirines are the most utilized PALs in biological systems, recent studies have reported aryl-tetrazoles and o-nitrobenzyl alcohols as PALs to capture protein-protein interactions. Deng et al., 2020; Hu et al., 2019. Diazirines are “zero-length” crosslinkers that are stable in aqueous media at room temperature, are synthetically tractable in a broad range of architectures, and offer predictable photo-crosslinking products; hence, they have emerged as a popular “universal” photo-affinity label. Dubinsky, 2012; Zhuo et al., 2022; West et al., 2021. Additionally, diazirines are photo-activated at relatively benign UV wavelengths (approximately 365 nm), avoiding protein damage that can occur at shorter wavelengths (less than 300 nm). Halloran and Lumb, 2019. Upon photoactivation, diazirines expel N2and form reactive carbenes that can insert across a diverse array of chemical bonds, including N-H, C-H, O-H, and S-H. Aliphatic carbenes evolved from diazirines have short lifetimes (ns–µs) and are reactive toward any amino acid in proximity. Hence this combination of properties enables diazirine-based photo-crosslinkers to capture rapid or transient interactions and significantly expands “crosslinkable” chemical space, given that most chemical crosslinkers attach to a restricted profile of residue-types (primarily Lys or Cys). Das et al., 2011. Heterobifunctional photo-crosslinkers, such as succinimidyl-4,4’-azipentanoate (SDA) (FIG.1A) contain an amine-reactive NHS-ester and a diazirine (green box) separated by an alkyl chain. This family of crosslinkers was initially introduced to assist investigations of PPIs by covalently coupling interacting proteins to a target protein of interest. Yang et al., 2015; Yang et al., 2016a; Yang et al., 2016b; Lin et al., 2014. They utilize a “plant-and-cast” strategy whereby the amine-reactive NHS-ester, Kalkhof and Sinz, 2008, plants the compound onto a target protein, casting the photo-crosslinking moiety to react with proximal residues upon subsequent UV irradiation. The crosslinks generated by heterobifunctional photo-crosslinkers also can be sequenced with mass spectrometry; however, identifying photo-crosslinked peptides is technically challenging because: (i) their higher reactivity results frequently in quenching and consequently lower crosslinking yields; 15 42282.601_P17605-02 and (ii) they can react with all residue-types through a variety of linkages. Wang et al., 2008; Blencowe and Hayes, 2005. Hence, photo-crosslinkers have found some application in structural studies to generate interaction maps spanning diverse residue types (yielding orthogonal and complementary information to what is obtained with conventional chemical crosslinkers); Gutierrez et al., 2021; Häupl et al., 2017, however, the challenges associated with photo- crosslinking have resulted in limited applicability to proteome-wide experiments. One study by Iacobucci and co-workers demonstrated that diazirines can generate MS-cleavable ester moieties when their photoproducts crosslink with Asp or Glu residues, facilitating identification on several purified proteins (e.g., calmodulin and GCAP-2). Iacobucci et al., 2018. Moreover, a recent study by Gutierrez et al. reported a series of diazirine-based crosslinkers bearing the MS-cleavable sulfoxide moiety (SDASO, shown in FIG.1A). Gutierrez et al., 2021. These crosslinkers have advanced the study on single purified proteins (or protein complexes), however are not without limitations. Ester linkages are MS-cleavable but do not leave behind modifications upon cleavage. SDASO’s MS- cleavable crosslink does produce mass modifications after cleavage but is singly-fissile, Yu et al., 2020, and hence does not generate signature doublets (e.g., FIG.1C). We conjectured that these spectral features could be important for facilitating accurate identification of crosslinked spectra from complex proteome-wide samples. 1.3 Scope To our knowledge, a photo-crosslinker has not yet been deployed to identify crosslinks in a proteome-wide experiment. It was thought that a heterobifunctional photo- crosslinker that is MS-cleavable in a doubly-fissile manner (thereby producing doublet peaks) could improve confidence in reliably identifying crosslinks from highly complex samples. To this end, we designed two heterobifunctional crosslinkers comprising an NHS- carbamate. Following a nucleophilic attack by lysine, NHS-carbamates yield urea functional groups at the initial attachment site (FIG.1B). This group would possess doubly-fissile MS- cleavability, which could facilitate spectral search (FIG.1C), though this latent capacity of NHS-carbamates has not been previously exploited. 16 42282.601_P17605-02 Accordingly, in some embodiments, the presently disclosed subject matter provides the synthesis and evaluation of DizSEC and DizSPC, photo-crosslinkers featuring NHS- carbamates and aliphatic diazirines with ethyl and propyl spacer arms, respectively, separating the chemical and photo-crosslinking moieties. In the following, we demonstrate that DizSEC and DizSPC could generate readily identifiable crosslinks in pilot studies on purified proteins. We demonstrate that DizSEC and DizSPC are cell-permeable and develop a method to perform rapid in vivo photo-crosslinking on whole E. coli cells. Next, we show that these crosslinkers’ doubly-fissile MS-cleavability, in combination with an experimental and computational workflow developed here, enables confident identification of crosslinks in the E. coli proteome. The data provide a partial interaction map of E. coli with residue- level resolution, capturing some previously unknown protein-protein interactions and providing some structural insights that are discussed (along with potential biological implications) in the following. Therefore, the compounds and methods described in this Example provide a foundation to perform proteome-wide photo-crosslinking mass spectrometry, which we hope will become a valuable tool to elucidate the molecular sociology of the cell with high temporal, spatial, and compositional resolution. Mahamid et al., 2016; Pfeffer and Mahamid, 2018. 1.4 Experimental Methods 1.4.1 Synthetic Methods Heterobifunctional photo-crosslinkers DizSEC and DizSPC were synthesized in a seven-step protocol from commercial reagents with yields of 86.7% (111 mg) and 85.9% (116 mg), respectively. Detailed synthetic procedures and characterization are reported in Example 2 (Scheme 1). 1.4.2 Crosslinking Lysozome and BSA with DizSEC and DizSPC Lysozyme and bovine serum albumin (BSA) were purchased from Sigma-Aldrich, was prepared in reaction buffer (50-mM Na-HEPES pH 8.0, 100-mM NaCl, 10-µM final concentration protein) at a 50-µL scale to which 0.5 µL of freshly prepared photo- crosslinkers DizSEC, DizSPC, or sulfo-SDA (as positive control) were added as stocks in DMSO (100-mM or 1-M stock concentration) to a final concentration of 1 mM or 10 mM. Some samples received 4-dimethylaminopyridine (DMAP) to a final concentration of 0.1 mM or 1 mM. Freer and McKillop, 1996; Höfle et al., 1978. 17 42282.601_P17605-02 Samples were gently mixed by inversion, then incubated at room temperature on an end-over-end rotator for 2 h (10 rpm). Chemical crosslinking reactions were quenched by the addition of Tris-HCl pH 8.0 (100 mM, final concentration) followed by incubation for 15 min at room temperature. Samples were transferred to a 1.0-mm path length quartz cuvette and photolyzed for 5.0 s with a 4.0-cm2Hönle LED Spot W UV (365 nm)-LED array rated with a 14 W / cm2max output (20% or 80% intensity). The distance between the cuvette and the LED head was minimized (ca.1–2 mm) by affixing the cuvette with tape to the array. 1.4.3 Proteome-Wide Crosslinking of E. Coli Extracts E. coli cells (K12 subst. MG1655) were grown overnight in LB (5 mL) at 37 °C, 220 rpm. Overnight cultures were diluted into 50-mL LB with an initial OD600of 0.05 until a final OD600 of 1.0 was reached. The cultures were pelleted in a 5910R swing-bucket benchtop centrifuge (Eppendorf) at 4000 g for 10 mins (4 °C), and the supernatant was discarded. The cell pellet was vigorously resuspended in 1 mL of HEPES reaction buffer and then lysed by sonication on water-ice with an FB120 probe sonicator (Fisher Scientific) at 45% amplitude for 10 min in 10 s / 10 s on-off cycles. E. coli lysates were centrifuged at 14,000 g for 10 minutes on a 5425R benchtop centrifuge (Eppendorf) to clarify the samples. BCA assay (Pierce) was performed following a manufacturer's protocol to determine the protein concentration. Samples were normalized to 1.0-mg / mL protein with reaction buffer. The crosslinking reaction was performed on a 1-mL scale of the normalized extracts. DizSEC or DizSPC (10-mM final concentration) were added from 1-M stocks in DMSO. Some samples also were supplied with 1-mM DMAP (final concentration) as a crosslinking catalyst. Samples were gently mixed by inversion, then incubated at room temperature on an end-over-end rotator for 2 h (10 rpm). Chemical crosslinking reactions were quenched by the addition of Tris pH 8.0 (100 mM, final concentration) followed by incubation for 15 min at room temperature. Samples were transferred to a 1.0-mm path length quartz cuvette (Starna Cells, PN: 21-Q-1) and photolyzed for 5.0 s with a 4.0-cm2Hönle LED Spot W UV (365 nm)-LED array rated with a 14 W / cm2max output (20% or 80% intensity). The UV- LED array was connected to a 200-W LED PowerDrive (Hönle). 1.4.4 In-Cell Photo-Crosslinking of E. coli 18 42282.601_P17605-02 Overnight and day cultures were prepared identically to that described above.50-mL LB cultures (OD6001.0) were pelleted in a 5910R swing-bucket benchtop centrifuge (Eppendorf) at 4000 g for 10 mins (4 °C), and the supernatant was discarded. Cells were resuspended and washed in 20 mL of HEPES reaction buffer four times. After the final centrifugation, cells were resuspended in 1-mL HEPES reaction buffer, and the NHS- carbamate coupling reactions were performed by the addition of DizSEC or DizSPC (10- mM final concentration). Cell suspensions were gently mixed by inversion, then incubated at room temperature on an end-over-end rotator for 2 h (10 rpm). The crosslinking reaction was quenched by the addition of Tris pH 8 (100 mM, final concentration), and incubation for 15 minutes at room temperature. Cell suspensions were diluted 10-fold by the addition of 9 mL HEPES reaction buffer before photolysis to prevent excessive scattering. The photolysis apparatus consisted of the LED Spot W UV-LED array with a 1-mm pathlength quartz flow-cell (Starna Cells, PN: 45-Q-1) affixed directly to the LED array surface with tape. The flow cell was connected to a Masterflex L / S Model 77200-50 peristaltic pump via 1.6 mm ID C-Flex biocompatible tubing (Masterflex). The tubing and flow cell were primed with water at max speed to remove bubbles. The cell suspension was pumped through the flow cell at 50 rpm (approximately 1 s exposure time), collected, and centrifuged at 4000 g for 10 min. The supernatant was discarded and crosslinked cells were resuspended in 1-mL HEPES reaction buffer, then lysed by sonication and normalized to 1.0-mg / mL protein using the BCA assay as described above. 1.4.5 MS Sample Preparation Crosslinked protein mixtures (typically 1 mL of 1.0 mg / mL) were denatured by addition of urea (1.92 g), ammonium bicarbonate (Ambic, 15.8 mg) and diluted to a final volume of 4 mL by addition of Millipore water to generate samples that are 0.25 mg / mL protein, 8 M urea, and 50-mM Ambic (pH 8). These samples were reduced by the addition of dithiothreitol (DTT) to 10-mM final concentration and incubation at 37 ºC with agitation (700 rpm) in a benchtop ThermoMixer (Eppendorf) for 45 minutes. Samples were then alkylated by the addition of iodoacetamide (IAA) to 40-mM final concentration and incubation at room temperature in the dark for 30 min. Samples were either: (i) diluted with 3 volumes of 50-mM Ambic and digested with trypsin (1:50 enzyme:protein w / w ratio, Pierce) overnight (ca.16 h) at 25 ºC, 700 RPM in a ThermoMixer; or (ii) digested serially – 19 42282.601_P17605-02 first with Lys-C (1:100 enzyme:protein w / w ratio, Pierce) for 2 h at 37 ºC, then diluted with 3 volumes of 50-mM Ambic followed by second digestion with trypsin overnight (ca.16 h). Desalting of protein digests was performed using Sep-Pak C18 cartridges (Waters). Lysozyme and BSA samples were desalted with C18 cartridges (Waters Corporation). First, digests were acidified with HPLC-grade trifluoroacetic acid (TFA, Alfa Aesar) to a final concentration of 1% v / v and diluted to 1 mL with 0.5% TFA in Optima LC / MS grade water (Fisher chemical, Buffer A). Cartridges were placed onto a vacuum manifold, conditioned with 2 × 1 mL 0.5% TFA in 80% Optima LC / MS grade acetonitrile (Fisher chemical, Buffer B), and equilibrated with 4 × 1 mL Buffer A. Acidified digests were loaded onto the cartridges under a reduced vacuum (such that it took approximately 5 min for 1 mL to passage through the sorbent) and then washed with 4 × 1 mL Buffer A. The cartridges were placed on top of 15 mL conical tubes, and 1 mL of Buffer B was pipetted onto the resin. The cartridges were gently eluted under centrifugal force in swing-buckets spun at 350 rpm in a 5910R centrifuge (Eppendorf) for 3 min and then reduced to dryness in a Vacufuge centrifugal concentrator (Eppendorf). E. coli extracts (either crosslinked in lysate or in cellulo) were desalted analogously, except using high-capacity C18 cartridges (Waters Corporation). Samples were acidified as described above. High-capacity cartridges were placed onto a vacuum manifold, conditioned with 2 × 10 mL Buffer B, and equilibrated with 4 × 10 mL Buffer A. Acidified digests were loaded onto the cartridges under a reduced vacuum (such that it took approximately 5 min for the ca.12 mL to passage through the sorbent) and then washed with 4 × 10 mL Buffer A. The cartridges were placed on top of 50-mL conical tubes, and 5 mL of Buffer B was pipetted onto the resin. The cartridges were gently eluted under centrifugal force in swing-buckets spun at 350 rpm on a 5910R centrifuge for 5 min and then reduced to dryness using a Vacufuge centrifugal concentrator. Dried peptides were stored at -80˚C until further use. 1.4.6 Enrichment and Fractionation of Crosslinks by Size Exclusion Chromatography Crosslinked peptides derived from E. coli extracts and in-cell crosslinked samples were enriched by peptide size exclusion chromatography (SEC). Dried peptides were reconstituted in 250 µL of 3% acetonitrile, 0.1% TFA with vigorous sonication. Peptides were fractionated on an AKTA Go FPLC (Cytiva) with a Superdex 3010 / 300 GL gel 20 42282.601_P17605-02 filtration column. Chromatography used an isocratic method with Solvent A (3% acetonitrile, 0.1% TFA) and a flow rate of 0.4 mL / min (approximate column pressure 2.5 MPa) with a column temperature of 4 ºC. Prior to fractionation, the column was equilibrated with Solvent A for 2 hours. 250-µL peptides were injected onto the column after being loaded into a 500-µL loop, and 0.2-column volumes (the void volume) were passed to waste without being collected. Elution volumes were set to 0.5 mL for fractions 1–12, with collection beginning 8 mL after injection; elution volumes were set to 1.0 mL for fractions 12–24. Out of 24 fractions, 4–20 were found to be enriched with crosslinked peptides, and these fractions were selected for LC-MS / MS analysis. Fractions were reduced to dryness in a Vacufuge centrifugal concentrator and stored at -80˚C until further use. 1.4.7 LC-MS / MS This study utilized an UltiMate3000 UHPLC system (Thermo Fisher) coupled with a Q-Exactive HF-X Orbitrap mass spectrometer (Thermo Fisher). Peptide fractions were resuspended in 30 µL of 0.1% Optima formic acid (FA) in Optima water, vigorously vortexed, and sonicated for five minutes. The amount of peptide material in each resuspended sample was approximately quantified with a NanoDrop OneCmicrovolume UV-Vis spectrophotometer (Thermo Fisher Scientific), and typically approximately1 µg of the peptide was injected onto the LC-MS system for each run. If a fraction contained insufficient material to inject one µg of peptide, 10 µL of the sample (one-third) was injected instead. The column and trap cartridge temperatures were set to 40 ºC, and the flow rate was set to 300 µL / minute. Mobile phase Solvent A comprised 0.1% FA, 2% Optima acetonitrile (Fisher Scientific), 98% Optima water. Mobile phase Solvent B comprised of 0.1% FA in 98% Optima acetonitrile, 2% Optima water (Fisher Scientific). The injected sample was loaded onto a trap column cartridge (Acclaim PepMap 100, C18, 75 µm × 2 cm, 3 µm, 100-Å column) and washed with Solvent A for 10 minutes. Next, the trap column was switched to be in line with the separating column (Acclaim Pepmap RSLC, C18, 75 µm × 25 cm, 2 µm, 100-Å column). The sample was separated in a linear gradient from 2% B to 35% B over 100 minutes, to 40% B over 25 minutes, and to 90% B over 5 minutes. The end of the LC method employed a saw-tooth gradient to remove sample residue from the resolving column. 21 42282.601_P17605-02 XL-MS data were acquired in positive ion mode with each full MS scan using a mass range of 350–1500 m / z. The resolution of the full MS scans was set to 120k, the automatic gain control (AGC) target was set to 3E6, the maximum injection time (IT) was set to 100 ms, and a stepped HCD fragmentation scheme using a three-tiered stepped normalized collisional energy (NCE) scheme. Various NCEs were tested to optimize these parameters for the different experiment types; however, 22%, 25%, 28% proved acceptable for most of the experiments. After each full MS scan, ten data-dependent MS2scans were collected at a resolution of 15k, an AGC target of 2E5, a minimum AGC target of 8E3, a maximum IT of 250 ms, an isolation window of 1.0 m / z (some lysozyme samples used an isolation window of 1.4 and 2.0 m / z), and a dynamic exclusion time of 60.0 s. Ions with charges less than +3 or greater than +8 and isotopomeric peaks were excluded from data-dependent MS2scans. 1.4.8 MeroX Data Analysis Workflow Raw files were converted to the mzML format as centroid spectra using the MSConvert GUI from the ProteoWizard toolkit, Chambers et al., 2012. and a spectral search for crosslinked peptides was conducted with MeroX (version 2.0.1.4). Götze et al., 2011. For both DizSEC and DizSPC, protease cleavage sites were allowed after Arg and Lys, with Lys cleavage blocked by Pro, and a maximum of three missed cleavages were allowed. The minimum peptide length was set to 3 residues, and the maximum peptide length was set to 30 residues. Static modifications included the carbamidomethylation of cysteine (C to B), and variable modifications included a maximum of one oxidation of methionine (M to m) per peptide (greater than one prohibitively expands the search space). The precursor precision was set to 10.0 ppm, and the fragment ion precision was set to 20.0 ppm. Mass recalibration for each raw file was determined manually by performing a separate search on each raw file for linear peptides in Proteome Discoverer 2.4 (Thermo Fisher Scientific) without spectrum recalibration, and the median Δm / z (typically between 5–8 m / z) was relayed to the MeroX settings file for each individual search. The lower mass limit was set to 350 m / z, the upper mass limit was set to 1500 m / z, the signal-to-noise ratio was set to 2.0, b- and y- ions only were enabled, the minimum number of fragments per peptide was set to 3, the minimum MS1 charge was set to 3, and deisotoping mode was selected. We used a pre-score intensity of 10% and an FDR cutoff of 1%. Decoys were generated by shuffling sequences but keeping protease sites within the supplied FASTA files. Typical scores of 22 42282.601_P17605-02 identified peptides ranged between 20 and 120. Lysozyme and BSA samples used quadratic mode, whilst crosslinks derived from E. coli (both from crosslinking in lysate or in cellulo) utilized proteome-wide mode developed by Götze and co-workers. Götze et al., 2019. For the lysozyme and BSA samples, a FASTA file was prepared that contained the sequence of Hen egg white lysozyme (P00698) or bovine serum albumin (P02769). For the E. coli lysates and in-cell crosslinked samples, the reference FASTA for E. coli K-12 substrain MG1655 (UP000000625) was used. To encode the crosslinkers in MeroX, for the DizSEC-crosslinked samples, the composition mass was set to C5NOH7 (97.05276383 Da). The first specificity site for the NHS-carbamate was Lys and the N-terminus, and the second specificity site for the diazirine was set to all amino acids, both termini, carbamidomethylated cysteine, and oxidized methionine. Modifications at site 1 had composition masses of COH–2 (25.97926456 Da; lysCO) and the unmodified peptide (i.e., 0; lyspep). Modifications at site 2 had composition masses of C4NH9(71.07349927 Da; nonamine), C5NOH7(97.05276383 Da; nonisocyanate), and unmodified peptide (0; discussed more in the following). For the DizSPC-crosslinked samples, the composition mass was set to C6NOH9 (111.06841389 Da). Specificity sites and modifications at site 1 were the same as in the DizSEC settings. Modifications at site 2 had composition masses of C5NH11 (85.08914933 Da; nonamine), C6NOH9 (111.06841389 Da; nonisocyanate), and unmodified peptide (0). For double-validation “entrapment” searches, a FASTA file was prepared which contained the 1500 most abundant E. coli proteins (as based on pooling fractions, performing a standard linear peptide search, and sorting based on peptide-spectrum matches (PSMs)) as well as 1500 randomly chosen yeast proteins (taken from UP000002311). Search and FDR settings were otherwise identical to the production searches. Method files for running these searches in MeroX and an executable script to automate job submission are provided (10.5281 / zenodo.7778901). 1.4.9 XiSearch and XiFDR Data Analysis Workflow Raw files were converted to the .mgf file format with mass calibration conducted by MSConvert (run via a script, also available on the Zenodo deposition). MSConvert was directed to identify the optimal mass shift by comparing spectra against a FASTA file 23 42282.601_P17605-02 containing the 400 most abundant proteins in E. coli (as based on a ranked list of number of PSMs from a separate search of linear peptides conducted in Proteome Discoverer). Recalibrated .mgf files, settings files, and FASTA files were moved to the Rockfish computing cluster at Advanced Research Computing at Hopkins (ARCH) via the file transfer program, WinSCP (Version 5.21.7). Individual recalibrated .mgf files were searched with XiSearch (Version 1.7.6.7) for crosslinked peptides using two search settings (both inclusive and conservative), as described herein. All versions of the settings included the variable modifications of the oxidation of methionine (+15.99491463 Da) and the crosslinker loop-link modification at Lys and the N-terminus (+97.05276383 Da for DizSEC or +111.06841389 for DizSPC) as well as the fixed modification of the carbamidomethylation of Cys (+57.021464). XiSearch was set to search only b- and y- ions, and the digestion enzyme was set to “constrained” (blocked by proline) trypsin. Water loss (-18.01056027 Da) was allowed at Ser, Thr, Asp, Glu, and the C-terminus. Ammonia loss was allowed at Arg, Lys, Gln, Asn, and the N-terminus. DizSEC was encoded as a crosslinker with a mass of 97.05276383 Da with the first linked amino acid as Lys or the N- terminus and all canonical amino acids and the termini allowed as the second linked amino acid. DizSEC cleavage products were encoded as modifications of 25.97926456 Da at Lys or the N-terminus (lysCO), 0 Da at Lys or the N-terminus (lyspep), and 71.07349927 Da (nonamine) or 97.05276383 Da (nonisocyanate) at any amino acid. DizSPC cleavage products were encoded as losses of 25.97926456 Da at Lys or the N-terminus (lysCO), 0 Da at Lys or the N-terminus (lyspep), and 85.08914933 Da (nonamine) or 111.0684189 Da (nonisocyanate) at any amino acid. For the “inclusive” search settings, 3 missed cleavages and 3 modifications per peptide were allowed. The inclusive settings used wider tolerances of 10.0 ppm at the MS1 level and 20.0 ppm at the MS2 level. The inclusive search used the full E. coli FASTA file, consisting of 4448 proteins. These searches were computationally expensive and required high-performance computational clusters with high memory allocations (500 GB) to run to completion. The conservative settings allowed only one missed cleavage and one modification per peptide. These settings used a tighter mass tolerance of 5.0 ppm at the MS1 level and 15.0 ppm at the MS2 level. The conservative search used a smaller FASTA file that 24 42282.601_P17605-02 contained the top 2000 proteins by number of PSMs as identified by Proteome Discoverer. These searches could be run on a conventional lab workstation (ours has 8 cores with 2.934 GHz processors and 64 GB RAM). Searches were run on Rockfish using a multi-submission script (provided on Zenodo deposition) using 500 GB of RAM, which was critical for completion of searches using the inclusive settings. Search output files that corresponded to a single experimental condition (e.g., all fractions that were collected using the DizSEC crosslinker for in cellulo crosslinking) were concatenated using a script (available on Zenodo). The concatenated data file was provided to the XiFDR GUI along with the appropriate settings file and FASTA. Several validation settings were tested; the setting we settled on was XiFDR’s default with two modifications: a prefilter for doublet counts (CCfragmentdoubletcount) to be greater than zero and boosting on “between” (interprotein) crosslinks. FIG.6 shows Scheme 3, a Xi workflow flow-chart diagram illustrating the computational workflow. As provided in Scheme 3, raw MS files are converted to the .mgf file format using MSConvert and a recalibration script. These files are then transferred to Rockfish via WinSCP, where they are searched using XiSearch. Each fraction job was submitted to Rockfish using PuTTY and the Job_Submitter.sh script, which refers back to the 35 fraction job scripts (DizSEC_Cluster.sh or DizSPC_Cluster.sh; each of these scripts is edited with respect to the appropriate .mgf file). The 35 XiSearch .csv outputs are then transferred off of Rockfish via WinSCP and concatenated by the JoinCSV.py script per experiment (DizSEC or DizSPC). FDR analysis was performed on the DizSEC and DizSPC .csv files separately in XiFDR. Results from the FDR analysis were uploaded to XiView and aggregated to create a network plot. All scripts (blue filenames) are provided in the Zenodo deposition. 1.4.10 Network Analysis MeroX data and XiFDR data were exported from their respective programs in the XiView format, including identification and sequence files. These files were uploaded to XiView, Graham et al., 2019, with a Met oxidation modification mass of 15.99491 Da and a Cys carbamidomethylation modification mass of 57.02146 Da. Each set of experimental data were aggregated together on the XiView server to form network diagrams. Protein Data 25 42282.601_P17605-02 Bank (PDB) codes were imported into XiView and Euclidian crosslink distance measurements were exported into PyMol scripts. Structures with mapped crosslinks were generated in PyMOL by importing the PDB file and running the appropriate PyMOL command script. 1.4.11 Protein Docking HADDOCK .tbl restraint files were exported from XiView for each complex to be docked. AlphaFold or PDB files for the appropriated structures were edited with PDBTools to remove unwanted heteroatoms and headers, then submitted to the HADDOCK 2.4 webserver. Honorato et al., 2021; Van Zundert et al., 2016. Only the polypeptide chains involved in the crosslinking interaction were selected for docking. Crosslinks were provided as ambiguous restraints with an upper distance bound of 30 Å, except for the docking of ManZ and ManY, for which the two crosslinks were provided as unambiguous restraints. All other settings used for docking were the default HADDOCK settings. Models for docked structures were either taken from the Protein Data Bank if there was an available E. coli structure (the highest resolution was selected if there were multiple) or from AlphaFold if there was not. No violations were observed in any of the docked structures. Docked structures were then aligned with their original (full) PDB or AlphaFold structure in Pymol. 1.5 Results 1.5.1 Piloting Photo-crosslinkers on Purified Proteins Initially, we combined DizSEC with N-Boc-L-lysine in a HEPES buffer (pH 8.0) to measure the kinetics of the acyl transfer reaction (Scheme 2). DizSEC. This experiment confirmed that the NHS-carbamate was a comparable (though somewhat slower) electrophile to the more standard NHS-ester, with the reaction coming to completion at 60 min. Next, we screened several different experimental conditions with 26 42282.601_P17605-02 DizSEC and DizSPC on a single protein (lysozyme). Parameters screened included concentration of the crosslinker, power of the UV-LED array, presence or absence of a Lys- C pre-digestion step, the addition of DMAP as an acyl-transfer catalyst, and variable collisional energies. Preliminary results confirmed the success of the crosslinking experiment, and crosslinked peptide identification was not very sensitive to the power of the UV lamp, the presence or absence of Lys-C or DMAP, or crosslinker concentration – suggesting these compounds function as robust protein crosslinkers. The identification of crosslinked peptides was explicitly dependent on treatment with a chemical crosslinker. To test that one of these conditions would generalize well to an unrelated protein, we repeated this experiment on BSA. FIG.1D shows an example of one of the assigned spectra from MeroX on one crosslinked peptide from BSA. Altogether, we identified 249 spectra that matched to intraprotein crosslinked peptides (and 1689 intrapeptidal crosslinks), corresponding to 143 unique intraprotein crosslinks. The crosslink-spectrum match (XSM) shown in FIG.1D corresponds to a lysine (K476) crosslinked to a leucine residue (L456) with support from 73 matched fragment ions. In this case, all four of the characteristic fragment ions in which fissure occurred at the crosslink (but nowhere else, releasing modified linear peptides) are detected, resulting in two doublets each separated by 13 m / z (for the 2+ charge state). The fragmentation at the crosslink leaves behind characteristic modifications on the non-lysine crosslinking site (an alkyl amine and alkyl isocyanate, cf. Pepαon FIG.1C). These are frequently found in the b- and y-ion series and facilitate the identification of the attachment site selected by the photo-activated diazirine. As shown in FIG.1D, y-ions of the α-peptide containing the isocyanate modifications are apparent from y13 down to y7, but starting with y6 the isocyanate modification disappears. Hence, the attachment site can be uniquely assigned to L8. This finding confirms that the urea generated at the DizSEC crosslink junction (formed by acyl transfer from lysine to the NHS- carbamate) can fragment more readily than standard peptide bonds, similar to the popular crosslinker DSBU. The number of crosslinks identified with DizSEC on lysozyme were similar to that of the commercially available non-cleavable photo-crosslinker, sulfo-SDA. Hence, for studies on a single protein, MS-cleavability is not a necessity. With DizSEC, however, we typically identify more fragment ions containing the crosslink site, so that it is easier to 27 42282.601_P17605-02 accurately pinpoint the non-lysine attachment site with DizSEC compared to sulfo-SDA. This observation is consistent with the recent finding that MS-cleavable crosslinkers improve backbone fragmentation. Kolbowski et al., 2022. DizSPC had notably fewer crosslinks on lysozyme, but gave comparable results to DizSEC on BSA. The median Cα–Cα crosslink distances were very short for both model proteins (17.3 Å and 13.0 Å for lysozyme and BSA, respectively), confirming the capacity of the diazirine to function as a zero-length crosslinker, and suggesting that our assignments of the identity of the crosslinked residue on the non-lysine-containing peptides are generally accurate. 1.5.2 Developing a Workflow for Proteome-wide Photo-crosslinking in cellulo To investigate if DizSEC and DizSPC were capable of probing protein structure and interaction networks in their native cellular context, we incubated E. coli with the compounds after washing away the primary amine-rich LB media and performed photolysis on live cells. These experiments take advantage of the high-power UV LED array, coupled to a flow cell (FIG.2A, see also Experimental Methods) such that each cell is illuminated for approximately 1 s. With such rapid irradiation, we anticipated the experiment would capture a snapshot of the cell’s “state” during photolysis and trap transient interactions. This photolysis method reduces diazirine crosslinking time substantially from typical irradiation methods (several minutes), and hence reduces the likelihood that results would be influenced by how cells might behave downstream to UV stress. Following rapid photolysis, cells were collected and lysed. The extracts were subject to double in-solution digest (first with Lys-C in 8 M urea, then with trypsin in 2 M urea) and the peptides cleaned up with solid phase extraction. Anticipating high sample complexity, we enriched for crosslinked peptides and fractionated them by size-exclusion chromatography. Sixteen fractions were collected that were deemed likely to contain crosslinks and were subsequently analyzed by LC-MS / MS as described in the experimental section. Because the complexity of these samples is greater than most crosslinking experiments (any lysine in any E. coli protein can crosslink to any other residue in the E. coli proteome), we wanted to test and compare search and validation workflows with two leading software platforms for crosslinking mass spectrometry, MeroX (developed by Sinz 28 42282.601_P17605-02 and colleagues), Iacobucci et al., 2018, and Xi (developed by Rappsilber and colleagues). Mendes et al., 2019; Graham et al., 2019. In MeroX, we searched each of our Raw files separately, filtering to a 1% FDR at the XSM level for each search. Validated peptides were then compiled together. In Xi, we examined several distinct search settings, because of the greater processing and memory resources this software needed for this large search space and our desire to streamline the computational workflow. With Xi, we concatenated the pre-validated search results from all the fractions that corresponded to the same experiment (see Scheme 33 for a flow chart) and performed validation on these in XiFDR with a 5% FDR cutoff at the residue pair level. Once again, several settings were tested and we found that prefiltering for doublet counts (CCfragmentdoubletcount) to be greater than zero and boosting on “between” (interprotein) crosslinks were beneficial. Hence, we applied these XiFDR settings to all of our experiments. In general, the inclusive settings resulted in significantly more crosslinks being identified. To check if these crosslinks were reliable, we decided to focus on one of the most structurally well-characterized complexes in E. coli, namely the ribosome (FIG.2B-FIG. 2D). Both MeroX and Xi performed well by this test. MeroX identified 230 crosslinks involving only ribosomal proteins, whilst Xi identified 511; whereas most of these were intraprotein crosslinks (i.e., between two peptides on the same ribosomal protein), some were interprotein and these were equally structurally plausible (FIG.2B-FIG.2C). For each crosslinked peptide identified in MeroX and Xi, we measured the Cα–Cα distance between the crosslinked residues (as based on one cryo-EM structure of the E. coli ribosome, PDB: 7CPJ) and FIG.2D shows the cumulative frequency distributions for these distances. Remarkably, 75% of the crosslinks have Cα-Cα distances below 15 Å, representing “zero- length” crosslinks (given that diazirines will be approximately 10 Å away from the Cα of the lysine to which it is attached, and most other residues’ side chains extend approximately 5 Å from their Cα). The dataset also was found to be high-quality using the fraction of structure- corroborating identifications (FSI) metric introduced by Yugandhar et al., 2020. The identification of many short-distance crosslinks within and between ribosomal proteins highlights an exciting advantage of photo-crosslinkers to provide high-resolution structural information, and also suggests that accurate identifications from whole proteomes 29 42282.601_P17605-02 are possible, at least for ribosomal proteins. FIG.2F shows an inter-protein crosslink between the N-terminus of RplU and D102 of RplT. The assignment of D102 as the attachment site is supported by 15 y-ions that contain the modification and is moreover consistent with the structure (these residues have a Cα–Cα distances of 10.5 Å). In addition to finding crosslinks between ribosomal proteins, we also identified two crosslinks between ribosomal proteins (RplL and RplB) and elongation factor EF-G, responsible for translocation (FIG.2E). These two crosslinks also were resolved down to the residue-level and cohered closely to a cryo-EM structure (PDB: 7N2C) with the elongation factor bound to the A-site (with Cα–Cα distances of 9.2 and 17.7 Å). 1.5.3 The Xi Workflow Better Controls for False Discovery Because we obtained poorer quality results in our experiments in which crosslinking was conducted in lysate, we focused on the conditions in which crosslinking was conducted in cellulo. The number of unique crosslinks pooled together from the 16 and 15 respective fractions of the DizSEC and DizSPC crosslinking experiment are shown in FIG.3A, using either MeroX or Xi workflows (green or orange respectively). In general, Xi identified more intraprotein crosslinks for both crosslinkers. Intraprotein crosslinks provide structural information, however interprotein crosslinks are generally considered more interesting from the perspective of building protein-protein interaction networks and assisting molecular docking. Graziadei and Rappsilber, 2022; Klykov et al., 2018; Piersimoni et al., 2021. We were initially surprised to see that MeroX identified significantly (approximately 3-fold, FIG.3A) more interprotein crosslinks than Xi, and this result made us wonder if perhaps XiFDR was stricter in its validation criteria for interprotein crosslinks. Manually inspecting MeroX’s identifications, we found it curious that so many of the crosslinks involved low-abundance or uncharacterized proteins, including those that weren’t even identified in standard linear searches. To interrogate this finding more quantitatively, we categorized E. coli proteins as high-abundance (greater than 200 copies / cell), low-abundance (less than 200 copies / cell), or unknown abundance (as based on Ribo-seq data on E. coli, Li et al., 2014, documented in EcoCyc, Keseler et al., 2017; Keseler et al., 2021) and sorted each crosslink into categories based on the abundances of the two proteins it connects. The results are shown in the frequency tables in FIG.3B. Consistent with our suspicion, MeroX identifies many 30 42282.601_P17605-02 crosslinks to low-abundance proteins, but this curiosity is not the case using the Xi workflow with validation by XiFDR. As a further test, we performed an entrapment search on one of our fractions (DizSEC, in cellulo, fraction 13) in MeroX with a FASTA file in which half of the records are impossible (taken from the yeast proteome rather than the E. coli proteome). Consistent with the other observations, incorrect identifications involving yeast peptides were rather common among interprotein crosslinks (FIG.3C). We performed an entrapment search using Xi on the DizSEC in-cell fractions and identified far fewer erroneous yeast crosslinks. On the balance, these further tests lead us to conclude that at present the Xi workflow provides more reliable identifications of inter-protein crosslinks in proteome-wide photo-crosslinking experiments. This finding is broadly consistent with the view that FDR estimation for inter-protein crosslinks is particularly challenging, requiring a careful propagation of error from XSMs to peptide pairs to protein pairs. Lenz et al., 2021. 1.5.4 Diazirines Can Efficiently Engage Non-acidic Residues Iacobucci and coworkers reported the finding that activated diazirines can react with carboxylic acids (as on Asp and Glu residues) to form MS-cleavable ester linkages. Iacobucci et al., 2018. By instructing the search algorithm to look for these cleavage products, crosslinks were more readily identified using the diazirine-containing photo-amino acids (photo-Met and photo-Leu) and it was suggested this mechanism may be the dominant mode (ca.75%) for diazirine crosslink formation. Separate studies in which diazirine photoaffinity labels were combined with individual amino acids have given support to the view that alkyl diazirines preferentially react with acidic residues particularly in the protonated state. West et al., 2021. On the other hand, since MS-cleavable ester crosslinks result in MS2spectra that are more easily identified, the notion that diazirines preferentially react with acidic residues might be biased. Because the doubly-fissile urea linkage formed by DizSEC and DizSPC enable reliable identification of crosslinked peptides regardless of the photoproduct’s insertion site, these compounds could enable an unbiased test of this hypothesis. FIG.3D shows the distribution for the non-lysine attachment site for the DizSEC proteome-wide in cellulo crosslinking experiment, which is highly non-random as based on a P-value of 10-239according to the Chi-square test. Moura et al., 2013. 31 42282.601_P17605-02 Glu is found to be the most frequent crosslink site (22.5%), a 4-fold enrichment above its frequency in the E. coli proteome. This finding is consistent with the theory that diazirines preferentially react with protonated carboxylic acids, since Glu is less acidic than Asp, and Asp has a comparatively modest enrichment factor at 1.8-fold. Between the two acidic residues, they account for 32% of all attachment sites from the photo-crosslink (33% for DizSPC). In contrast with the view that acidic residues are the universal targets, however, we find the second most frequent insertion site is Ala (13.2%), which is 1.3-fold enriched above its frequency in the E. coli proteome (and 1.7-fold enriched once acidic sites have been discounted). Once acidic residues have been discounted, the amino acid types that are most enriched (relative to proteomic frequency) for serving as attachment sites are Ala, Gly, Val, and Tyr (in both the DizSEC and DizSPC in cellulo experiments). Of note, Gly was a surprisingly frequent crosslinking site (8.3%), despite its lack of a side chain. This feature is desirable because Gly is common in disordered regions and flexible loops. It is important to note that in cases where fragmentation patterns are incomplete, XiSearch cannot unambiguously call the placement of the non-lysine attachment site, and so the crosslinking preferences shown in FIG.3D should be viewed with some caution. Nevertheless, such non-random distribution of attachment sites with respect to overall proteomic frequencies with acidic residues discounted (P-value 10-45by Chi-square test) indicates that XiSearch can call attachment sites with some fidelity. Finally, we observed that for 10% of the identified crosslinked peptides, there was no acidic residue present on the photo-crosslinked (non-lysine) side, providing further evidence for participation of non- acidic residues. Even though DizSEC crosslinks frequently with Asp and Glu, there is another subtle reason why the doubly-fissile urea fragmentation is an important improvement for reliable identification of photo-crosslinks proteome-wide. As shown in FIG.3E, the ester linkage between a conventional diazirine (e.g., photo-methionine) and Glu will cleave to form an alkene and unmodified Glu. Since alkenes also are a potential quench (“dead-end”) product from diazirine photolysis, the precursor mass / MS2spectrum from a cleaved ester-derived crosslink would be indistinguishable from a scenario in which a peptide with a quenched diazirine and a peptide with an Asp / Glu form an association complex in the gas phase (FIG. 3E). This ambiguity is not present when the photo-crosslinker is incorporated onto a 32 42282.601_P17605-02 doubly-fissile MS-cleavable system. FIG.3F shows a representative MS2spectrum in which the non-lysine crosslinking site is Glu. Fragmentation at the urea generates the previously-discussed modified linear peptides arranged into doublets (lysCO and lyspep (found in two charge states); nonamineand nonisocyanate). Fragmentation at the ester results in a distinct product, an unmodified linear form of the peptide with the non-lysine attachment site (would be called nonpep in our nomenclature scheme). As shown in FIG.3F, the nonpep ion is present – consistent with the well-known fact that esters are labile functional groups under HCD. If esters were much more labile than ureas, however, then the amine-bearing and isocyanate-bearing modifications on the non-lysine peptide would be rare. This possibility is apparently not the case, as the intensity of the fragmentation-at-urea products are comparable to the fragmentation-at-ester product (FIG.3F). Manual interrogation of many spectra showed us that this feature is common amongst MS2spectra in which DizSEC crosslinked at Asp or Glu. 1.5.5 Structural Validation from Heterocomplexes Intermolecular crosslinks are generally more interesting but also more challenging to reliably identify in the enormous search space encompassed by proteome-wide photo- crosslinking. Lenz et al., 2021. Structurally well-defined heterocomplexes should have restricted crosslink distances if inter-protein crosslinked peptides are identified properly. Therefore, we mapped our crosslinks against reported structures of hetercomplexes to obtain a more visual form of validation, although some reports have suggested that other validation strategies are more stringent. Yugandhar et al., 2020. For our “final” dataset on protein-protein interactions, we performed XiSearch / XiFDR on fractions from DizSEC and DizSPC treated cells, crosslinked in cellulo. The samples in which photo-crosslinking was conducted in lysate generated very few inter- protein crosslinks when searched and validated in the Xi workflow. In addition to the ribosome (generally one of the easier targets for XL-MS studies), we also obtained a dense network of crosslinks interconnecting the heterocomplex between EF-Tu and EF-Ts, a GTPase associated with translation elongation and its associated guanosine exchange factor (FIG.4A). Superposing these crosslinks onto the x-ray structure of the complex we once again detect only structurally reasonable crosslinks with a median Cα–Cα distance of 15.7 Å. Kawashima et al., 1996. 33 42282.601_P17605-02 As another example shown in FIG.4B, we obtained a decent number of crosslinks between Hns (histone-like nucleoid structuring protein) and its paralogous partner, StpA (sometimes also called HnsB) which had previously been shown to interact. Williams et al., 1996. These proteins promote compaction of DNA and transcriptional repression in response to stress. Using our crosslinks alone we were able to obtain a reasonable dimer structure docking these two proteins’ AlphaFold structural models with HADDOCK, which recapitulates the known function of the N-terminal domain in oligomerization. Rimsky, 2004; Dorman et al., 1999. The crosslinks on Hns also show contacts between the long N- terminal helix and an ostensibly-distal C-terminal disordered region, implying an apparent pincer movement about the loop which forms a joint at the bottom of the “V.” This observation is consistent with previous observations on mutants that assigned an important functional role of the linker region in gene repression. One advantage of crosslinking-based structural approaches is that they are amenable to interrogating disordered regions even though these portions of proteins are typically poorly resolved in x-ray or cryo-EM structures. We obtained one crosslink bridging ManX and ManY, two components of the phosphotransferase system (PTS) responsible for active uptake of mannose (FIG.4C). ManX comprises two domains that correspond to EIIA and EIIB; the EIIA domain accepts phosphate from a phosphocarrier protein at His10, and then transfers it to His175 on the EIIB domain. Stolz et al., 1993. The EIIB domain docks with EIIC (corresponding to ManY for the mannose PTS), the transmembrane channel which imports incoming mannose and transfers phosphate to it. Liu et al., 2019. The channel is itself a hexamer comprised of three copies of ManY and three copies of ManZ. Although the structures of the EIIAB protein (ManX) and the channel (ManYZ) have both been solved, Stolz et al., 1993; Liu et al., 2019, the manner in which ManX docks and donates phosphate to mannose has not been seen in structures even though the interaction is supported by genetics and biochemical evidence. Erni et al., 1987. To visualize this interaction, we docked ManX to ManY using our crosslinking data with AlphaFold structures (which contain all disordered regions) and found that EIIB on ManX interacts with the disordered cytoplasmic-facing C-terminus of ManY. Hence, in-cell crosslinking can highlight the importance of disordered regions in promoting a biochemical interaction that could be challenging to resolve. 34 42282.601_P17605-02 1.5.6 Insights into Cellular Biology from Proteome-Wide Photo-crosslinking We pooled data from two in-cell experimental crosslinking conditions (DizSEC and DizSPC, both with Lys-C pre-digestion, and obtained 3202 XSMs and 2362 unique crosslinks (257 interprotein and 2105 intraprotein) representing 177 different protein-protein interactions. The result is a small portion of E. coli’s interaction network, albeit with residue-level resolution. Currently, the coverage from this dataset is modest relative to the total E. coli interactome, and we would like to point out here that other interactomic methods (such as by high-throughput affinity-purification mass-spectrometry (AP-MS), Hu et al., 2009, and yeast two-hybrid screens, Rajagopala et al., 2014), have to date provided deeper coverage than the present one. Nevertheless, probing interactions in their native cellular context and with residue level-resolution are interesting advantages afforded by proteome-wide photo-crosslinking. Hence, we interrogated this network and focused on a few hotspots where there were denser sets of crosslinks available, and in the following describe three examples that highlight how the results from photo-crosslinking can provide new insights at the intersection of cellular and structural biology. One example focuses on a sub-network centering on RNA polymerase (RNAP, FIG. 5A), whose core enzyme in E. coli is a tetramer, [(RpoA)2(RpoB)(RpoC)]. The XL-MS data recapitulate the key interactions that make up the core enzyme, with crosslinks bridging between all three polypeptides. We further identified a linkage between RpoC and RpoZ, the ω-subunit of RNA polymerase, which is an assembly factor that has been shown to facilitate the incorporation of RpoC into the holo-complex. Ghosh et al., 2001. As before, 75% of the crosslinks approach zero-length, highlighting the close-range of these crosslinks. We also detected two crosslinks between RNA polymerase and the protein NusA (consistent with a cryo-EM structure of the complex, shown in left inset in FIG.5A), which has been implicated in stabilizing transcriptional pauses and anti-termination in E. coli. Guo et al., 2018. The presently disclosed crosslinks show that the N-terminal helix of NusA binds against a hairpin between the first two alpha helices of HupA (right inset in FIG.5A), a subunit of the HU histone-like DNA binding complex. To visualize this interaction, we used these data to dock HupAB onto the RNAP:NusA complex. HU is a heterodimer containing one monomer of HupA and HupB and is implicated in nucleoid organization and 35 42282.601_P17605-02 regulation. Kar et al., 2005. Because XL-MS data provide spatial resolution for these linkage sites, this crosslink could be used as a distance restraint for future studies focusing on the structural basis of transcriptional regulation in E. coli. Moreover, these observations show the role of NusA is a multifaceted adaptor protein, involved in bridging RNAP to both termination factors and to nucleoid proteins. A second example concerns ATP synthase, another abundant macromolecular complex in E. coli responsible for converting proton gradients across the inner membrane into ATP synthesis (FIG.5B). In total 44 crosslinks were observed, including some non- obvious intermolecular crosslinks spanning the F1 subunit and the rotor (AtpA–AtpB). While the majority of the crosslinks (75%) are under 15 Å, in this case we observe a “fat tail” to the Cα–Cα distance distribution, with 10% of the crosslinks exceeding 50 Å in the cryo-EM structure of state 1A. All these long-distance crosslinks occurred between the stator (AtpF) and apical portions of the F1 subunit (comprising of AtpA and AtpD). Whilst difficult to reconcile with a single ATP synthase complex (including in its alternative rotational states, Sobti et al., 2020), these crosslinks could make sense in the context of the “rows” of ATP synthases that have been resolved in cryo-electron tomography reconstructions on yeast mitochondria in which F1 subunits come very close to one another. Davies et al., 2012. Although this formation has not been previously shown in E. coli, these results are suggestive that higher order organization of ATP synthases may exist in bacteria, as well. As a final vignette, we were intrigued by a set of crosslinks that interconnect RNase E, polynucleotide phosphorylase (PNPase), enolase, and the large subunit of the ribosome (FIG.5C). The degradosome is a large multi-enzyme complex consisting of RNase E, PNPase, enolase, and an RNA helicase (RhlB – not in our observed subnetwork) responsible for mRNA decay in E. coli. Py et al., 1994. RNase E has been described as the scaffold for the degradosome, particularly its disordered C-terminal half. Indeed, our crosslinks recapitulate a known interaction between the C-terminus of RNase E and PNPase. Kido et al., 1996; Vanzo et al., 1998. A report from Tsai et al. showed that the degradosome associates stably with ribosomes, which might coordinate the degradation of mRNA following its translation. Tsai et al., 2012; Aït-Bara et al., 2015. Tsai and co-workers found that the C-terminal portion of 36 42282.601_P17605-02 RNase E binds to ribosomes in vitro through electrophoretic mobility shift assays through a putative RNA-binding domain termed the AR2 region (residues 796-819). Consistent with this result, we identified a crosslink between RPL6 and a site in RNase E adjacent to the AR2 region. In addition, we also found several crosslinks between enolase and the ribosome, specifically at RPL1, RPL2, and RPL23. These three ribosomal proteins face the E-site of the ribosome, hence enolase could serve as an adaptor that helps orientate RNase E to the outgoing side of the mRNA channel. This positioning would facilitate degradation of the 5’ side of the mRNA as it issues from the trailing end of the polysome, a model initially proposed by Luisi and colleagues. An important caveat to all these discussions is that the crosslink density in our dataset of any one of the three complexes discussed (RNAP, ATP synthase, and degradosome) is not high enough to develop fully-fledged models. Such would require more detailed crosslink maps of the individual complexes, likely achieved through tagging and enrichment by affinity purification. Nevertheless, we hope to convey that proteome- wide photo-crosslinking can be a powerful hypothesis generation tool, revealing structural relationships between proteins in situ to stimulate further study. 1.5.7 Optimization of Crosslinking Parameters on Lysozyme. Initially, we screened several different experimental conditions with DizSEC and DizSPC on a single protein (lysozyme). Parameters screened included concentration of the crosslinker, power of the UV- LED array, presence or absence of a Lys-C pre-digestion step, the addition of DMAP as an acyl-transfer catalyst, and variable collisional energies. The results of these experiments are compiled in Tables 1 and 2, and the crosslinks identified are mapped in FIG.11A, FIG.11B, and FIG.11C. Preliminary results confirmed the success of the crosslinking experiment, and crosslinked peptide identification was not very sensitive to the power of the UV lamp, the presence or absence of Lys-C or DMAP, different collisional energies, or even crosslinker concentration— suggesting these compounds function as robust protein crosslinkers. As expected, the identification of crosslinked peptides was explicitly dependent on treatment with a chemical crosslinker and the UV irradiation step (Tables 1-2). FIG.1D shows an example of an assigned spectrum from MeroX of one crosslinked peptide from lysozyme. The spectrum has 50 peaks that match predicted fragment ions, of which six fragment ions contain the crosslink. Importantly, we detected 37 42282.601_P17605-02 three of the four fragment ions corresponding to the cleaved crosslink but with no fragmentation of the attached peptides (shown schematically in FIG.1C and annotated in FIG.1D), thereby confirming the generation of an MS-cleavable urea group upon nucleophilic attack by Lys on the NHS-carbamate. Crosslinking results obtained from the DizSEC experimental trials were similar to that of the commercially available non-cleavable photo-crosslinker, sulfo-SDA. Hence, for studies on a single protein, DizSEC could function as an alternative: Table 1: Summary of DizSEC crosslinks on lysozyme with varied experimental and crosslinking conditions. sulfo-SDA: Table 2: Summary of DizSPC crosslinks on lysozyme with varied experimental and crosslinking conditions. 38 42282.601_P17605-02

[0002] To perform XL-MS on whole extracts from E. coli, we used a method illustrated schematically in FIG.7A. Cells were lysed, and lysates were treated with DizSEC / DizSPC, photolyzed, and digested. The complexity of such samples is vast, necessitating a degree of fractionation and enrichment to prevent pervasive co-isolation during LC-MS / MS. We adopted the strategy of peptide size- exclusion chromatography (alternatives include affinity pull-down on crosslinkers that include an affinity label or strong cation exchange chromatography). Iacobucci et al., Nat. Protoc.2018, 13, 2864–2889; Vellucci; J Am Soc Mass Spectrom 2010, 21 (8), 1432–1445; Fritzsche et al., Rapid Communications in Mass Spectrometry 2012, 26 (6), 653–658; Schmidt e al., Anal Bioanal Chem 2017, 409 (9), 2393–2400. To optimize parameters for identifying crosslinks from proteome-wide photo- crosslinking experiments, we screened several conditions (presence or absence of DMAP, inclusion or exclusion of Lys-C, and HCD energies) and carried out LC-MS / MS on a trial fraction (typically 13 or 14, which tends to be rich in crosslinks; see Tables 3 and 4 below and FIG.12: Table 3: Summary of DizSEC crosslinks from preliminary screening of enriched E. coli extracts. 39 42282.601_P17605-02

[0003] extracts. collisional energies of 22, 25, 28%, and an isolation window of 1.0 m / z. In the production set of experiments, these settings were typically applied to fractions 4–19, ca.16 fractions were separately injected and searched, and the resulting identified crosslinks were pooled together. We performed this experiment four times (a total of 63 LC-MS / MS runs) to generate four distinct datasets, corresponding to crosslinking with either DizSEC or DizSPC and with or without pre-digestion with Lys-C (FIG.7B-C). As shown in FIG.7B, these crosslinkers generated a satisfactory number of assignable spectra (crosslink-spectrum matches, XSMs). Despite the fact aliphatic diazirine- based crosslinkers are sometimes considered too reactive and primarily quench with water (resulting in quenched, dead-end products), we found in all four cases that dead-end products (FIG.7B, gray) were identified less frequently than spectra in which crosslinks 40 42282.601_P17605-02 formed between amino acids (black and red). In all cases, the majority were intra-peptidal (red, as many as 9152 in the DizSEC / –Lys-C condition), as expected for crosslinkers with relatively short spacer arms. Next, we focused on the inter-peptidal XSMs, grouped together redundant ones, andasked among these unique crosslinks, how many bridged across different proteins (interprotein) or within the same protein (intraprotein) (FIG.7C). Strikingly, the majority were interprotein in all conditions (between 75-97%). These findings suggest that DizSEC and DizSPC could be particularly useful for exploring PPIs. However, we were surprised because typically, one sees more intraprotein crosslinks in most XL-MS studies. Keller et al; J Proteome Res 2019, 18 (8), 3077–3085. These observations led us to speculate why photo-crosslinkers might be better suited for identifying protein interactions. FIG.7D illustrates our model. With homo-bifunctional chemical crosslinkers, the relatively long persistence time of the NHS-ester moiety enables the crosslinker to “wag” around its protein after initial attachment, exploring the surface, until it eventually finds a suitable partner (another nucleophilic residue). On the other hand, in a hetero-bifunctional photo-crosslinker, the activated carbene has a very short lifetime. Thus, the scenario in which the carbene would have to explore around the surface to find a second attachment site is more likely to result in quenching. On the other hand, at a protein-protein interface, the carbene will be buried and more likely to react with the partner protein. Additionally, DizSEC and DizSPC have short spacer arms, and after the initial reaction, they result in relatively small modifications (cf. FIG.1B), akin to post-translational modification on Lys. This modification is less likely to sterically block a PPI compared to most homo-bifunctional chemical crosslinkers, which will still be bulky after initial attachment. Consequently, one can imagine a surface lysine reacting with DizSEC / DizSPC to form a modified lysine, which later gets buried at a PPI and crosslinked during UV irradiation. West et al.; J Am Chem Soc 2021, 143 (17), 6691–6700. Upon searching for crosslinked spectra, we encountered a sizable number in which the crosslinked lysine occupied a terminal position (as high as 1239 in DizSPC / –Lys-C), implying that both trypsin and Lys-C can cleave proteins at carbamylated lysines which form upon acyl transfer with the NHS-carbamate group. FIG.7E shows an example of a fragmentation spectrum of a terminal lysine-linked peptide, highlighting the substantial 41 42282.601_P17605-02 number (18) of fragment ions containing the crosslink attached to the assigned lysine in question. We infer that many other XL-MS studies could increase the number of identified crosslinks if the search criteria were adjusted to allow for such linkages. For all four experimental conditions, 24.1% of unique crosslinks were bonded to acidic residues, consistent with a previous report which found acidic residues to be common attachment sites for diazirines.28 Finally, we observed that the unique crosslinks across these conditions were mostly unique (FIG.7F), implying that different crosslinkers or proteolysis conditions can be combined to increase coverage. It is also likely a result of the stochasticity of data- dependent acquisition of such complex samples. 1.5.9 Protein Interaction Networks Derived from Photo-crosslinks. Pooling together the data obtained on all four experimental crosslinking conditions on E. coli extracts, we identified 1986 unique crosslinks (1636 interprotein and 350 intraprotein), 1882 unique crosslinked proteins, and 1617 PPIs, some of which include putative or uncharacterized proteins. FIG.8A illustrates a high-level view of this protein interaction network. Details are compiled in Tables 5–8. Table 5: Summary of DizSEC crosslinks from enriched E. coli extracts pre-digested by Lys- C. 42 42282.601_P17605-02

[0004] Table 6: Summary of DizSEC crosslinks from enriched E. coli extracts without pre- digestion by Lys-C. 43 42282.601_P17605-02

[0005] C. 44 42282.601_P17605-02

[0006] by Lys-C. 45 42282.601_P17605-02

[0007] The ribosome is among the most well structurally characterized complexes and hence can serve as a fiducial for the reliability of the crosslinks identified (FIG.8B). In total, we observed 21 crosslinks amongst ribosomal proteins, of which 100% had Cα–Cα distances below 30 Å. Fig.8B shows a few regions within the ribosome with a dense set of crosslinks, in which all of the crosslinked residues were within close proximity. In addition to being a stable complex, the ribosome is a hub for numerous interacting partners, which is also captured in our dataset and highlights a unique advantage of XL-MS over other structural methods. For instance, we found 34 distinct interactors with RplL (including its well-known interaction with the GTPase elongation factor EF-G (FusA), Dunkle et al., Science (1979) 2011, 332 (6032), 981–984. FIG.8C). E. coli RplL is one of the more elusive ribosomal proteins, absent in most x-ray and cryo-EM structures; Montesano-Roditis et al., Journal of Biological Chemistry 2001, 276 (17), 14117–14123; our study suggests this may be because it is a dynamic adaptor, introducing a wide array of factors to the ribosome. This includes some anticipated proteins, such as the stringent response initiator RelA (which recognizes deacylated tRNA in the ribosomal A-site; Zhao et 46 42282.601_P17605-02 al., J Bacteriol 1996, 178 (1), 232–239), the co-translational chaperone DnaK (Hsp70), several aminoacyl-tRNA synthetases (ThrS and MetG), and methyltransferases that modify the ribosomal RNA (RlmI and RsmG). On the other hand, RplL’s interactome also consists of several uncharacterized enzymes not previously associated with the ribosome (FIG.8C). Drawing an example from central metabolism, our network demonstrates pyruvatedehydrogenase (AceF) as a hub (FIG.8D). AceF is an important metabolic enzyme that mediates between glycolysis and the citric acid cycle. We see crosslinking between several important metabolic proteins, including phosphoglycerate dehydrogenase (SerA), involved in serine biosynthesis and reduction of α-ketoglutarate to generate 2 hydroxyglutarate. Zhao et al., J Bacteriol 1996, 178 (1), 232–239. Another crosslink was observed with a metabolic enzyme lipoate synthase (LipA), which is well-known to interact directly with E2 subunits and is important for AceF functions. Stepp et al., Biochemistry 1981, 20 (16), 4555–4560. Other than that, AceF also interacted with several uncharacterized proteins, such as a zinc-binding dehydrogenase, YdjJ; a KpLE2 phage-like protein, YjhV; and a carbon-nitrogen hydrolase family protein, YobB. In both sub- networks, the central proteins are partnered with several putative or uncharacterized proteins. Stelmashenko et al., Mol Pharmacol 2012, 82 (4), 760–766. We propose that photo-crosslinking can give insight into their potential roles by capturing their interaction with more abundant and well-understood partners. 1.5.10 In-cell Proteome-wide Photo-crosslinking. To see if DizSEC and DizSPC were capable of probing PPIs in their native cellular context, we incubated cells with the compounds after washing away the primary amine-rich LB media and directly performed photolysis on live cells. These experiments explicitly took advantage of the high-power UV LED array used in this study, coupled to a flow cell (FIG. 9A, see also Experimental Methods) such that each cell is illuminated for approximately 1 s. With such rapid irradiation, we anticipated the results would capture a snapshot of the cell’s “state” during photolysis and trap transient interactions. To our delight, in-cell crosslinking studies proved to be an even richer source of identifiable crosslinks. We identified 1486 and 1122 unique crosslinks in-cell vs.281 and 511 unique crosslinks in the lysate (for DizSEC and DizSPC, respectively, FIG.9D and FIG.9E). 47 42282.601_P17605-02 This success might be attributed to the cytosol-containing proteins at a higher concentration (>100 mg / mL) and possessing water at a lower concentration than lysates. It also shows that DizSEC and DizSPC are quite cell-permeable, even for a Gram-negative bacterium with two membranes. We point out that whilst SDA is cell-permeable, it is poorly soluble in water, which prompts many users to use sulfo-SDA, an analog that is highly water-soluble but not cell-permeable. Our results show that such a trade-off does not exist for DizSEC and DizSPC: they are both reasonably soluble in water (when pre-dissolved in DMSO and diluted) and cell-permeable. 1.5.11 Protein-Protein Interactions in vivo Captured in 3-D. We pooled data from two in-cell experimental crosslinking conditions (DizSEC and DizSPC, both with Lys-C pre-digestion) and obtained 14,941 XSMs and 2131 unique crosslinks spanning across 1243 unique proteins. The result is a partial interaction map of the E. coli cytosol with residue-level resolution (illustrated schematically in FIG.9B). One example is a sub-network centering on RNA polymerase (FIG.9C), whose core enzyme in E. coli is a tetramer, [(RpoA)2(RpoB)(RpoC)]. The XL-MS data recapitulate the key interactions that make up the core enzyme, with crosslinks bridging between all three polypeptides. We further identified a linkage between RpoC and RpoZ, the ω-subunit of RNA polymerase, which is an assembly factor that has been shown to facilitate the incorporation of RpoC into the holo-complex. Ghosh et al., Eur J Biochem 2001, 268 (17), 4621–4627. A previous XL-MS / cryo-ET study in Mycoplasma showed that NusA is an adaptorprotein that connects RNA polymerase with ribosomes to constitute the “expressosome,” the complex responsible for coupled transcription-translation. O’Reilly et al., Science (1979) 2020, 369 (6503), 554–557. Our results show that the C-terminal helix of NusA binds against a hairpin between the first two alpha helices of HupA, a highly abundant subunit of the HU histone-like DNA binding complex. HU is a heterodimer containing one monomer of HupA and HupB, and is implicated in nucleoid organization and regulation. Kar et al., Proceedings of the National Academy of Sciences 2005, 102 (45), 16397–16402. Because XL-MS data provide spatial resolution for these linkage sites, this crosslink could be used as distance restraint for future studies focusing on the structural basis of 48 42282.601_P17605-02 transcriptional regulation in vivo. Moreover, these observations cement the role of NusA as an adaptor between RNA polymerase and its various interactors. To further illustrate the power of our approach with one final example, we mapped out residue-level interaction networks of selected components of the E. coli proteostasis network (FIG.10A-C). It is commonly accepted that chaperones have a diverse array of interactors (e.g., DnaK has been shown to have over 650 clients by co-precipitation, Calloni et al., Cell Rep 1 (3): 251–264.2012), promoting the folding of a wide array of clients. Consistent with this view, in our unbiased photo-crosslinking-based dataset, we find that chaperones indeed have a higher number of interactors than average. Because the coverage in our experiments is spread thin over the whole cell’s interactome, these networks are not comprehensive, but they represent a starting point that could be expanded upon by coupling our approach with pull-down enrichment following cell lysis. In particular, the XL-MS strategy introduces a method to identify clients and their interaction sites with the more elusive E. coli chaperone, HtpG (Hsp90). Chaperones that tightly bind their substrates (such as DnaK) or fully encapsulate them (such as GroL / S) have been particularly amenable to pull-down based proteomics to exhaustively catalog their client substrates. Calloni et al., Cell Rep 1 (3): 251–264.2012; Kerner et al., Cell 2005, 122 (2), 209–220. These methods have been more challenging to apply to HtpG, and consequently our knowledge of its substrate scope is significantly more limited. Mayer et al., Nat Struct Mol Biol 2009, 16 (1), 2–6; Chiosis et al., Nat Struct Mol Biol 2013, 20 (1), 1–4. We note a crosslink between HtpG with the periplasmic chaperone, SurA suggests an intriguing link between the cytosolic and periplasmic proteostasis networks. The crosslink is within the HtpG substrate-binding clamp suggesting that HtpG might be involved in promoting the biogenesis of SurA, (Marx et al., Proceedings of the National Academy of Sciences 2020, 117 (45), 28026–28035) which could explain why HtpG-null knockouts are deficient in their ability to process several periplasmic proteins, β-lactamase and alkaline phosphatase. Grudniak et al., Mutation Research / Fundamental and Molecular Mechanisms of Mutagenesis 2013, 745, 1–5. 1.5.12 Shedding Light on the “Dark” Proteome. We were surprised to identify many crosslinks between chaperones and proteins that are either low abundance or putative / uncharacterized (FIG.10A). Since chaperones tend to 49 42282.601_P17605-02 be highly abundant molecules within the cell, we theorized that crosslink formation between high abundance and low abundance proteins could make it easier to identify peptides from low abundance proteins. We revisited the full in-cell photo-XL-MS dataset, and for each unique interprotein crosslink, assessed whether it bridged between two high-abundance proteins (copy number > 200; roughly the upper two-thirds of the proteome, Karp et al., EcoSal Plus 2018, 8 (1); Li et al., Cell 2014, 157 (3), 624–635), two low-abundance proteins (copy number ≤ 200; the bottom third of the proteome), or between a high- and low- abundance protein. Indeed, we found an unusually high number of crosslinks involving low- abundance proteins (FIG.10B, yellow). Moreover, we found a high number of crosslinks involving proteins with an unreported copy number, which can be qualitatively interpreted as “low-abundance” given the high coverage of the Ribo-Seq experiments on which the abundance data were based. Karp et al., EcoSal Plus 2018, 8 (1); Li et al., Cell 2014, 157 (3), 624–635. The frequency of proteins appearing in this dataset that were not covered by Ribo-Seq is striking, given that proteomics methods typically have lower coverage than those employing Illumina sequencing. Finally, we sought to characterize how “uncharacterized” the in-cell photo-XL-MS dataset is. For model organisms, UniProt proteome files provide a Protein Existence (PE) score for each protein, in which 1 is the highest (protein is experimentally validated) and 5 is the lowest (protein uncertain). Our network included 1243 proteins in total, of which 21 were in the most uncharacterized category. This represents a 6-fold enrichment (FIG.10C) relative to how frequently proteins in this category are observed in a typical search for linear peptides. Indeed, all the PE categories greater than 1 were enriched in our dataset, a difference that was statistically significant by the chi-square test (P < 10–24). We theorize that these uncharacterized proteins may be enriched by crosslinking to high abundance proteins, allowing us to identify transient or unstable interactions that were not identified with affinity purification mass spectrometry (AP-MS) or other interactomic methods. 1.6 Discussion Diazirines have achieved a privileged position in chemical biology because of their versatility as a “universal” crosslinking moiety, reflected in their frequent use as a photoaffinity label in chemoproteomic studies. Conway et al., 2021; Meisner et al., 2022. Nevertheless, transferring this versatility to crosslinking mass spectrometry has been 50 42282.601_P17605-02 challenging. Even though diazirine-containing crosslinkers, including SDA, sulfo-SDA, and photo-amino acids, are commercially available, it has been difficult to apply them to proteome-wide samples because the number of different photo-crosslinked peptides generated is so great and the sequence space against which candidate spectra must be searched even greater. To tame this challenge, we have introduced DizSEC and DizSPC, diazirine- containing NHS-carbamates which react with lysines to generate MS-cleavable urea linkages. These crosslinks are doubly-fissile under HCD, generating characteristic modifications and doublets that can assist spectral assignment and placement of the photo- crosslink site. They have enabled us to confidently assign crosslinked spectra in very large search spaces (i.e., wherein any lysine in any E. coli protein can crosslink to any other residue in the E. coli proteome) including those in which the diazirine reacts with non-acidic residues. Whilst conceptually similar to the previously-described SDASO crosslinker (FIG. 1A), Gutierrez et al., 2021, two potential advantages of DizSEC / DizSPC is the doubly-fissile nature of their linkages (assisting spectral search) and their smaller footprints after attachment to lysine (thereby interfering less with protein interactions). Further despite the commercial availability of SDA and sulfo-SDA, they have yet to be broadly exploited for XL-MS studies because of high quenching rates (lowering the yield of crosslinked peptides), their high reactivity and diverse crosslinking sites (complicating spectral search and limiting the software capable of identification), and their lack of a dedicated MS-cleavable group. In this Example, we demonstrate with the two reported compounds (DizSEC and DizSPC) that by combining diazirenes with NHS-carbamates on a single compound, the complexity associated with identifying photo-crosslinks can be greatly reduced by utilizing the symmetrical MS-cleavability of the urea functional group formed upon acyl transfer to protein targets. We thereby demonstrate the feasibility of proteome- wide identification of photo-crosslinked peptides. Though it was not a part of the original molecular design, the fortuitous water-solubility and membrane permeability of DizSEC and DizSPC enables such experiments to be conducted in cellulo, elucidating protein interactions at residue-level resolution in their native context. In this study, we identified many PPIs that include proteins with unknown functions that have so far not been ascribed interactors or partners. We speculate that photo-XL-MS 51 42282.601_P17605-02 may be particularly well-suited for catching such interactions because the photochemistry is very rapid, thereby trapping potentially transient interactions, and because crosslinks formed between low-abundance uncharacterized proteins and their higher-abundance interactors could help overcome the sensitively challenge of the former. Low-abundance proteins that do not have catalytic activities and are not associated with stable complexes are often challenging to characterize. Even in the highly studied model organism E. coli, there is experimental evidence for only 69.3% of the proteome at the protein level and 72.8% at the transcriptional level based on the PE score system. UniProt, Nucleic Acids Res 2021, 49 (D1), D480–D489. Based on these early photo-XL-MS experiments, we speculate that the method can enrich for low abundance proteins that interact with high abundance proteins, thereby providing a new tool to assign putative functions to uncharacterized proteins. Remarkably, many proteins that we detected in our proteome-wide search for crosslinked peptides were not identified in conventional searches for linear peptides on the same data. This observation is notable because it is common in XL-MS studies to minimize the protein database employed for the crosslink search to include only the proteins that were found in a linear search – the wisdom being that this will ease the more complex crosslink search. In contrast, we show that by combining the efficient MS-cleavability of DizSEC / DizSPC-derived crosslinks with the recently developed “proteome-wide mode” in MeroX (Götze et al., Anal Chem 2019, 91 (15), 10236–10244. https: / / doi.org / 10.1021 / acs.analchem.9b02372). protein database restriction is neither necessary nor desirable, enabling our photo-XL-MS study to identify proteins that were otherwise unidentifiable. Search software is a major consideration to the workflow for photo-crosslinking, and in this Example, we examined two packages, MeroX and XiSearch / XiFDR. We found the risk for false discovery in the large search space spanned by proteome-wide photo- crosslinking is acute, and currently this challenge seems better addressed by the improvements in XiFDR introduced in v2.0. Lenz et al., 2021. On the other hand, an updated version of XiSearch with better memory allocation management (garbage collection) would be helpful for making these searches more accessible to users who lack access to high-performance computation clusters. 52 42282.601_P17605-02 We are excited about the potential for these types of global photo-crosslinking methods. The finding that DizSEC / SPC are cell-permeable and can be rapidly activated in cellulo (in 1 s) with our set-up suggests that they can be useful tools for probing protein structures and interactions at residue-level resolution in their native context (and with high temporal resolution, though we did not explicitly take advantage of that in this Example). Recent years have seen increased interest in characterizing the quinary structure of proteins, Cohen and Pielak, 2017; Guin and Gruebele, 2019, which refers to the weak but specific interactions that cause the cytosol to demix and enable certain sets of proteins to associate with one another with preferential orientations. These interactions are important for the formation of metabolons and biomolecular condensates. There are few techniques, however, that can directly probe protein interaction networks at high structural resolution inside living cells. Currently, XL-MS and cryo-ET are leading contenders, with complementary strengths. Cryo-ET studies have typically focused on large molecular complexes with distinct shapes (such as the ribosome or proteasome). Although the spatial resolution of cryo-ET is remarkable and improving, its “compositional resolution” – that is, its ability to identify which specific protein an arbitrary segment of electron density corresponds to – is more limited. XL-MS excels at resolving molecular identity, and hence close collaboration between them is set to enable “molecular sociology” to characterize smaller proteins that are less easy to identify based on electron density alone. In this vein, photo-activated crosslinkers are germane for providing greater structural and temporal resolution than other XL-MS modalities. On the other hand, there are still major challenges ahead. To be blunt, the number of intermolecular crosslinks we identified (257) from 35 fractions (88 instrument hours) is not enviable, and is modest compared to studies that have used conventional chemical crosslinkers, Lenz et al., 2021; Wheat et al., 2021; Götze et al., 2019; Gonzalez-Lozano et al., 2020; Stieger et al., 2019, or alternative methods. Mishra et al., 2020; Hu et al., 2009; Rajagopala et al., 2014; Roux et al., 2012; Gingras et al., 2019. Hence, further improvements to the workflow will be needed to make photo-crosslinking deliver on these high expectation; we conclude with where we suspect progress could be made. Firstly, given the complexity of photo-crosslinked proteomes, it is likely that even greater fractionation than we employed here would be necessary to mitigate co-isolation 53 42282.601_P17605-02 problems. Lee and O’Reilly, 2023. A likely candidate is 2-D fractionation combining SEC (as done here) with strong-cation exchange, which has proven successful for other high- complexity crosslinking experiments. O’Reilly et al., 2023. Secondly, our experiments would have likely benefitted from tighter instrument calibration and higher resolution on MS2scans. XiFDR is well-equipped at filtering out decoys, but this capability comes with the requirement that candidate spectra match their targets with high precision. Future experiments will need to be more exacting in this choice of parameters. Finally, studies that employ affinity purification or organelle isolation would likely be preferrable to undifferentiated lysates. Lee and O’Reilly, 2023. Although the ability to search a proteome-wide database with photo-crosslinkers is an important milestone demonstrated herein, undifferentiated lysates typically overrepresent high-abundance proteins (such as the ribosome) and is likely not the best path forward to delve deeper into the proteome’s dynamic range. Lee and O’Reilly, 2023. 1.7 Summary In summary, we present new molecular tools, coupled with an experimental and computational workflow, that enables reliable proteome-wide identification of photo- crosslinked peptides. This Example has piloted these tools to produce a small, partial protein-protein interaction network of E. coli, but nevertheless one that is resolved down to the residue-level. We have performed comparisons to known structures of protein complexes and verified short Cα–Cα distances to build confidence in the spectral identifications and have used the dataset as a hypothesis-generating tool to propose three potentially new functional interactions associated with RNAP, ATP synthase, and the degradosome. With the proposed improvements suggested in this final section, we hope photo-crosslinking mass spectrometry will soon contribute meaningfully to our understanding of molecular sociology inside cells. We anticipate photo-XL-MS could have two potential broader impacts on molecularbioscience. As we have emphasized, photo-crosslinks yield information on structure (at theresidue-level of resolution) as well as interaction. Recent years have seen increased interest in characterizing the quinary structure of proteins (Cohen et al., Protein Science 2017, 26 (3), 403–413; Guin et al., Chem Rev 2019, 119 (18), 10691–10717), which refers to the specific stickinginteractions that cause the cytosol to demix and enable 54 42282.601_P17605-02 particular sets of proteins to have higher local concentrations and adopt preferential orientations with respect to one another. Exascale computing has enabled molecular dynamics simulations to be conducted on systems approaching the size of cells (Gupta et al., Curr Opin Struct Biol 2022, 73, 102338; Rickard et al., J Phys Chem Lett 2019, 10 (18), 5667–5673), and with these “computational microscope” studies the structure of transient protein assemblies in a cellular milieu are coming into focus. To date however, there are few experimental techniques that can measure quinary interactions to test the predictions from computational microscopy. We propose that the partial interaction maps based on photo-XL- MS could help refine these models and create dialogue between experiment and simulation. Similarly, the emerging field of molecular sociology aims to provide a census of interactions between biomolecules in their native environment and has been driven primarily by the extraordinary advances in cryo-electron tomography (cryo-ET). Mahamid et al., Science (1979) 2016, 351 (6276), 969–972; Pfeffer et al., Curr Opin Struct Biol 2018, 52, 111–118; Beck et al., Trends Cell Biol 2016, 26 (11), 825–837. These studies have so far focused on large molecular complexes with distinct shapes (such as the ribosome Tegunov et al., Nat Methods 2021, 18 (2), 186–193; or proteasome) and in bacteria with very spacious cytosols, such as Mycoplasma pneumoniae. O’Reilly et al., Nat Struct Mol Biol 2018, 25 (11), 1000–1008. One current roadblock is that whilst the spatial resolution of cryo-ET is remarkable and improving, its “compositional resolution” – that is, the ability to identify which specific protein an arbitrary segment of electron density corresponds to – is more limited. In this regard, XL- MS and cryo-ET have complementary strengths (compositional and structural resolution, respectively) and hence a close collaboration between them is set to enable molecular sociology to break new ground toward characterizing smaller proteins that are less easy to identify based on electron density alone. The advent of photo-XL-MS using the tools and methods we have presented here should be particularly germane toward this effort. 1.7 Data Deposition Raw data corresponding to the crosslinking experiments in extracts (63 LC-MS / MS runs in total) and to the in-cell crosslinking experiments (31 LC-MS / MS runs in total) have been deposited to PRIDE under accession code PXD036884. Also included are MeroX method files (.mxf) and MeroX output files (.zhrm). 55 42282.601_P17605-02 EXAMPLE 2 Supplemental Methods 2.1 Synthetic Methods All the reagents and chemicals for the synthesis were procured from commercial suppliers and used without further purification. Rf values were determined using TLC silica gel G60 aluminum sheets and visualized by UV at wavelength 254 nm or KMnO4. NMR spectra were recorded using Bruker Avance 400 MHz FT-NMR spectrophotometer and chemical shifts are reported as δ (ppm) values. Scheme 1: Synthesis of DizSEC and DizSPCa MeOH, - 20 ºC to RT, stirring, overnight; (III) I2, Et3N, MeOH, 0 ºC to RT, stirring, 2 h; (IV) p-TsCl, pyridine, 0 ºC to RT, stirring, 2 h; (V) NaN3, DMF, RT, stirring, overnight; (VI) PPh3, THF:H2O (9:1); RT, 6 h; (VII) disuccinimidyl carbonate, CH3CN, 0 ºC to RT, stirring, 2 h. 2-(3-methyl-3H-diazirin-3-yl)ethyl 4-methylbenzenesulfonate (5a) and 3-(3-methyl-3H- diazirin-3-yl)propyl 4-methylbenzenesulfonate (5b). Liang et al., 2017; Zheng et al., 2019. In an argon-filled and flame-dried 3-neck round bottom flask, 4-hydroxy-2-butanone (1a, 3.0 g, 34.5 mmol, 1 eq)) or 5-hydroxy-2-pentanone (1b, 3.5 g, 34.5 mmol, 1 eq) was cooled to -20 °C, and 7N methanolic ammonia solution (35 mL) was added very slowly. The reaction mixture was stirred at -20 °C for 3 h under argon to get respective imines (2a or 2b). The solution of hydroxylamine-O-sulfonic acid (4.3 g, 37.9 mmol, 1.1 eq) in 56 42282.601_P17605-02 anhydrous methanol (25 mL) was added dropwise with a syringe pump to the above reaction mixture containing compound 2a or 2b at -20 °C. The reaction mixture was slowly allowed to warm to room temperature and stirred overnight. The insoluble white precipitate was filtered away and washed successively with anhydrous methanol. The filtrate was concentrated in vacuo to get respective diaziridine derivatives (3a and 3b). 2-(3-methyl-3H-diazirin-3-yl)ethyl 4-methylbenzenesulfonate (5a) and 3-(3-methyl-3H- diazirin-3-yl)propyl 4-methylbenzenesulfonate (5b). Liang et al., 2017; Zheng et al., 2019. In an argon-filled and flame-dried 3-neck round bottom flask, 4-hydroxy-2-butanone (1a, 3.0 g, 34.5 mmol, 1 eq)) or 5-hydroxy-2-pentanone (1b, 3.5 g, 34.5 mmol, 1 eq) was cooled to -20 °C, and 7N methanolic ammonia solution (35 mL) was added very slowly. The reaction mixture was stirred at -20 °C for 3 h under argon to get respective imines (2a or 2b). The solution of hydroxylamine-O-sulfonic acid (4.3 g, 37.9 mmol, 1.1 eq) in anhydrous methanol (25 mL) was added dropwise with a syringe pump to the above reaction mixture containing compound 2a or 2b at -20 °C. The reaction mixture was slowly allowed to warm to room temperature and stirred overnight. The insoluble white precipitate was filtered away and washed successively with anhydrous methanol. The filtrate was concentrated in vacuo to get respective diaziridine derivatives (3a and 3b). Compound 5a 1H NMR (400 MHz, CDCl3) δH 7.84 (d, J = 8.3 Hz, 2H), 7.38 (d, J = 8.3 Hz, 2H), 3.97 (t, J = 6.4 Hz, 2H), 2.48 (s, 3H), 1.69 (t, J = 6.4 Hz, 2H), 1.02 (s, 3H). Compound 5b 1H NMR (400 MHz, CDCl3) δH 7.78 (d, J = 8.3 Hz, 2H), 7.36 (d, J = 8.0 Hz, 2H), 3.99 (t, J = 6.1 Hz, 2H), 2.45 (s, 3H), 1.54 – 1.50 (m, 2H), 1.42 – 1.38 (m, 2H), 0.97 (s, 3H). 2-(3-methyl-3H-diazirin-3-yl)ethan-1-amine (7a) and 3-(3-methyl-3H-diazirin-3-yl)propan- 1-amine (7b). Shigdel et al., 2008. To the mixture of compound 5a (1.7 g, 6.8 mmol, 1 eq) or 5b (1.8 g, 6.8 mmol, 1 eq) in DMF (20 mL), sodium azide (1.5 g, 23.7 mmol, 3.5 eq) was added, and the reaction mixture was stirred at room temperature overnight under N2. The resultant mixture was poured on ice water (20 mL) dropwise and extracted with diethyl ether (2 × 20 mL). The 57 42282.601_P17605-02 combined organic layer was dried over anhydrous sodium sulfate and concentrated in vacuo to get compound 6a or 6b, respectively. To a solution of compound 6a or 6b in THF:water (9:1, 10 mL), triphenylphosphine (3.0 g, 11.5 mmol, 1.7 eq) was added, and the reaction mixture was stirred for 6 h at room temperature. The crude reaction mixture was acidified with 1M HCl under an ice-cold condition and washed with diethyl ether (5 × 20 mL). The resulting aqueous solution was further neutralized with 1M NaOH to pH >10.0 and extracted with diethyl ether (4 × 30 mL). The confirmation of desired amine extraction was performed by TLC using DCM:methanol (4:1) as the mobile phase, and ninhydrin was used for visualization. The combined organic layer was dried over anhydrous sodium sulfate and concentrated in vacuo to get the compound as a pale-yellow oil (compound 7a, 0.494 g, 73.3%; compound 7b, 0.585 g, 76.2%). Compound 7a H NMR (400 MHz, CDCl3) δH 2.71 (t, J = 6.4 Hz, 2H), 1.46 – 1.42 (m, 2H), 1.09 (s, 3H). Compound 7b 1H NMR (400 MHz, CDCl3) δH 2.68 (t, J = 6.8 Hz, 2H), 1.43 – 1.38 (m, 2H), 1.36 – 1.31 (m, 2H), 1.03 (s, 3H). 1H NMR (400 MHz, CDCl3) δH 2.68 (t, J = 6.8 Hz, 2H), 1.43 – 1.38 (m, 2H), 1.36 – 1.31 (m, 2H), 1.03 (s, 3H). DizSEC (8a) 1H NMR (400 MHz, CDCl3) δH 5.45 (s, 1H), 3.20 (dd, J = 13.1, 6.9 Hz, 2H), 2.85 (s, 4H), 1.69 (t, J = 7.0 Hz, 2H), 1.10 (s, 3H).13C NMR (126 MHz, CDCl3) δC 169.9, 151.4, 53.4, 37.1, 34.0, 25.5, 25.4. DizSPC (8b) 1H NMR (400 MHz, CDCl3) δH 5.53 (s, 1H), 3.25 (dd, J = 12.6, 6.4 Hz, 2H), 2.84 (s, 4H), 1.56-1.36 (m, 4H), 1.04 (s, 3H).13C NMR (126 MHz, CDCl3) δC 169.9, 151.4, 60.4, 41.4, 31.3, 25.5, 25.3, 19.7. REFERENCES All publications, patent applications, patents, and other references mentioned in the specification are indicative of the level of those skilled in the art to which the presently disclosed subject matter pertains. All publications, patent applications, patents, and other references are herein incorporated by reference to the same extent as if each individual 58 42282.601_P17605-02 publication, patent application, patent, and other reference was specifically and individually indicated to be incorporated by reference. It will be understood that, although a number of patent applications, patents, and other references are referred to herein, such reference does not constitute an admission that any of these documents form part of the common general knowledge in the art. Iacobucci, C.; Götze, M.; Sinz, A. Cross-Linking / Mass Spectrometry to Get a Closer View on Protein Interaction Networks. Curr Opin Biotechnol 2020, 63, 48–53. Mishra, P. K.; Yoo, C.; Hong, E.; Rhee, H. W. Photo‐crosslinking: An Emerging Chemical Tool for Investigating Molecular Networks in Live Cells. ChemBioChem 2020, 21 (7), 924–932. Lenz, S.; Sinn, L. R.; O’Reilly, F. J.; Fischer, L.; Wegner, F.; Rappsilber, J. Reliable Identification of Protein-Protein Interactions by Crosslinking Mass Spectrometry. Nat Commun 2021, 12 (1), 1–11. Wheat, A.; Yu, C.; Wang, X.; Burke, A. M.; Chemmama, I. E.; Kaake, R. M.; Baker, P.; Rychnovsky, S. D.; Yang, J.; Huang, L. Protein Interaction Landscapes Revealed by Advanced in Vivo Cross-Linking–Mass Spectrometry. Proceedings of the National Academy of Sciences 2021, 118 (32), e2023360118. O’Reilly, F. J.; Xue, L.; Graziadei, A.; Sinn, L.; Lenz, S.; Tegunov, D.; Blötz, C.; Singh, N.; Hagen, W. J. H.; Cramer, P. In-Cell Architecture of an Actively Transcribing- Translating Expressome. Science (1979) 2020, 369 (6503), 554–557. Leitner, A.; Walzthoeni, T.; Kahraman, A.; Herzog, F.; Rinner, O.; Beck, M.; Aebersold, R. Probing Native Protein Structures by Chemical Cross-Linking, Mass Spectrometry, and Bioinformatics. Molecular & Cellular Proteomics 2010, 9 (8), 1634– 1649. Sinz, A. Chemical Cross‐linking and Mass Spectrometry for Mapping Three‐ dimensional Structures of Proteins and Protein Complexes. Journal of mass spectrometry 2003, 38 (12), 1225–1237. Trakselis, M. A.; Alley, S. C.; Ishmael, F. T. Identification and Mapping of Protein− Protein Interactions by a Combination of Cross-Linking, Cleavage, and Proteomics. Bioconjug Chem 2005, 16 (4), 741–750. 59 42282.601_P17605-02 Bishop, A. C.; Buzko, O.; Shokat, K. M. Magic Bullets for Protein Kinases. Trends Cell Biol 2001, 11 (4), 167–172. Petrotchenko, E. v; Borchers, C. H. Crosslinking Combined with Mass Spectrometry for Structural Proteomics. Mass Spectrom Rev 2010, 29 (6), 862–876. Tang, X.; Bruce, J. E. A New Cross-Linking Strategy: Protein Interaction Reporter (PIR) Technology for Protein–Protein Interaction Studies. Mol Biosyst 2010, 6 (6), 939– 947. Liu, F.; Rijkers, D. T. S.; Post, H.; Heck, A. J. R. Proteome-Wide Profiling of Protein Assemblies by Cross-Linking Mass Spectrometry. Nat Methods 2015, 12 (12), 1179–1184. Arora, B.; Tandon, R.; Attri, P.; Bhatia, R. Chemical Crosslinking: Role in Protein and Peptide Science. Curr Protein Pept Sci 2017, 18 (9), 946–955. Matzinger, M.; Mechtler, K. Cleavable Cross-Linkers and Mass Spectrometry for the Ultimate Task of Profiling Protein–Protein Interaction Networks in Vivo. J Proteome Res 2020, 20 (1), 78–93. O’Reilly, F. J.; Rappsilber, J. Cross-Linking Mass Spectrometry: Methods and Applications in Structural, Molecular and Systems Biology. Nat Struct Mol Biol 2018, 25 (11), 1000–1008. Götze, M.; Iacobucci, C.; Ihling, C. H.; Sinz, A. A Simple Cross-Linking / Mass Spectrometry Workflow for Studying System-Wide Protein Interactions. Anal Chem 2019, 91 (15), 10236–10244. https: / / doi.org / 10.1021 / acs.analchem.9b02372. De Jong, L.; De Koning, E. A.; Roseboom, W.; Buncherd, H.; Wanner, M. J.; Dapic, I.; Jansen, P. J.; Van Maarseveen, J. H.; Corthals, G. L.; Lewis, P. J. In-Culture Cross- Linking of Bacterial Cells Reveals Large-Scale Dynamic Protein–Protein Interactions at the Peptide Level. J Proteome Res 2017, 16 (7), 2457–2471. Liu, F.; Lössl, P.; Scheltema, R.; Viner, R.; Heck, A. J. R. Optimized Fragmentation Schemes and Data Analysis Strategies for Proteome-Wide Cross-Link Identification. Nat Commun 2017, 8 (1), 15473. Fürsch, J.; Kammer, K.-M.; Kreft, S. G.; Beck, M.; Stengel, F. Proteome-Wide Structural Probing of Low-Abundant Protein Interactions by Cross-Linking Mass Spectrometry. Anal Chem 2020, 92 (5), 4016–4022. 60 42282.601_P17605-02 Xue, L.; Lenz, S.; Zimmermann-Kogadeeva, M.; Tegunov, D.; Cramer, P.; Bork, P.; Rappsilber, J.; Mahamid, J. Visualizing Translation Dynamics at Atomic Detail inside a Bacterial Cell. Nature 2022, 610 (7930), 205–211. Schweppe, D. K.; Chavez, J. D.; Lee, C. F.; Caudal, A.; Kruse, S. E.; Stuppard, R.; Marcinek, D. J.; Shadel, G. S.; Tian, R.; Bruce, J. E. Mitochondrial Protein Interactome Elucidated by Chemical Cross-Linking Mass Spectrometry. Proceedings of the National Academy of Sciences 2017, 114 (7), 1732–1737. Chavez, J. D.; Lee, C. F.; Caudal, A.; Keller, A.; Tian, R.; Bruce, J. E. Chemical Crosslinking Mass Spectrometry Analysis of Protein Conformations and Supercomplexes in Heart Tissue. Cell Syst 2018, 6 (1), 136–141. Klykov, O.; Steigenberger, B.; Pektaş, S.; Fasci, D.; Heck, A. J. R.; Scheltema, R. A. Efficient and Robust Proteome-Wide Approaches for Cross-Linking Mass Spectrometry. Nat Protoc 2018, 13 (12), 2964–2990. Kaake, R. M.; Wang, X.; Burke, A.; Yu, C.; Kandur, W.; Yang, Y.; Novtisky, E. J.; Second, T.; Duan, J.; Kao, A. A New in Vivo Cross-Linking Mass Spectrometry Platform to Define Protein–Protein Interactions in Living Cells. Molecular & Cellular Proteomics 2014, 13 (12), 3533–3543. Jiao, F.; Yu, C.; Wheat, A.; Wang, X.; Rychnovsky, S. D.; Huang, L. Two- Dimensional Fractionation Method for Proteome-Wide Cross-Linking Mass Spectrometry Analysis. Anal Chem 2022, 94 (10), 4236–4242. Moss, R. A. Diazirines: Carbene Precursors Par Excellence. Acc Chem Res 2006, 39 (4), 267–272. Das, J. Aliphatic Diazirines as Photoaffinity Probes for Proteins: Recent Developments. Chem Rev 2011, 111 (8), 4405–4417. Deng, H.; Lei, Q.; Wu, Y.; He, Y.; Li, W. Activity-Based Protein Profiling: Recent Advances in Medicinal Chemistry. Eur J Med Chem 2020, 191, 112151. Hu, W.; Yuan, Y.; Wang, C.-H.; Tian, H.-T.; Guo, A.-D.; Nie, H.-J.; Hu, H.; Tan, M.; Tang, Z.; Chen, X.-H. Genetically Encoded Residue-Selective Photo-Crosslinker to Capture Protein-Protein Interactions in Living Cells. Chem 2019, 5 (11), 2955–2968. Dubinsky, L.; Krom, B. P.; Meijler, M. M. Diazirine Based Photoaffinity Labeling. Bioorg Med Chem 2012, 20 (2), 554–570. 61 42282.601_P17605-02 Zhuo, F.; Guo, Q.; Zheng, Y.; Liu, T.; Yang, Z.; Xu, Q.; Jiang, Y.; Liu, D.; Zeng, K.; Tu, P. Photoaffinity Labeling‐Based Chemoproteomic Strategy Reveals RBBP4 as a Cellular Target of Protopanaxadiol against Colorectal Cancer Cells. ChemBioChem 2022, e202200038. West, A. V; Muncipinto, G.; Wu, H.-Y.; Huang, A. C.; Labenski, M. T.; Jones, L. H.; Woo, C. M. Labeling Preferences of Diazirines with Protein Biomolecules. J Am Chem Soc 2021, 143 (17), 6691–6700. Halloran, M. W.; Lumb, J. Recent Applications of Diazirines in Chemical Proteomics. Chemistry–A European Journal 2019, 25 (19), 4885–4898. Yang, T.; Liu, Z.; Li, X. D. Developing Diazirine-Based Chemical Probes to Identify Histone Modification ‘Readers’ and ‘Erasers.’ Chem Sci 2015, 6 (2), 1011–1017. Yang, T.; Li, X.-M.; Bao, X.; Fung, Y. M. E.; Li, X. D. Photo-Lysine Captures Proteins That Bind Lysine Post-Translational Modifications. Nat Chem Biol 2016a, 12 (2), 70–72. Yang, Y.; Song, H.; He, D.; Zhang, S.; Dai, S.; Lin, S.; Meng, R.; Wang, C.; Chen, P. R. Genetically Encoded Protein Photocrosslinker with a Transferable Mass Spectrometry- Identifiable Label. Nat Commun 2016b, 7 (1), 1–10. Lin, S.; He, D.; Long, T.; Zhang, S.; Meng, R.; Chen, P. R. Genetically Encoded Cleavable Protein Photo-Cross-Linker. J Am Chem Soc 2014, 136 (34), 11860–11863. Kalkhof, S.; Sinz, A. Chances and Pitfalls of Chemical Cross-Linking with Amine- Reactive N-Hydroxysuccinimide Esters. Anal Bioanal Chem 2008, 392 (1), 305–312. Wang, J.; Kubicki, J.; Peng, H.; Platz, M. S. Influence of Solvent on Carbene Intersystem Crossing Rates. J Am Chem Soc 2008, 130 (20), 6604–6609. Blencowe, A.; Hayes, W. Development and Application of Diazirines in Biological and Synthetic Macromolecular Systems. Soft Matter 2005, 1 (3), 178–205. Gutierrez, C.; Salituro, L. J.; Yu, C.; Wang, X.; DePeter, S. F.; Rychnovsky, S. D.; Huang, L. Enabling Photoactivated Cross-Linking Mass Spectrometric Analysis of Protein Complexes by Novel MS-Cleavable Cross-Linkers. Molecular & Cellular Proteomics 2021, 20. 62 42282.601_P17605-02 Häupl, B.; Ihling, C. H.; Sinz, A. Combining Affinity Enrichment, Cross‐linking with Photo Amino Acids, and Mass Spectrometry for Probing Protein Kinase D2 Interactions. Proteomics 2017, 17 (10), 1600459. Iacobucci, C.; Götze, M.; Piotrowski, C.; Arlt, C.; Rehkamp, A.; Ihling, C.; Hage, C.; Sinz, A. Carboxyl-Photo-Reactive MS-Cleavable Cross-Linkers: Unveiling a Hidden Aspect of Diazirine-Based Reagents. Anal Chem 2018, 90 (4), 2805–2809. Yu, C.; Novitsky, E. J.; Cheng, N. W.; Rychnovsky, S. D.; Huang, L. Exploring Spacer Arm Structures for Designs of Asymmetric Sulfoxide-Containing MS-Cleavable Cross-Linkers. Anal Chem 2020, 92 (8), 6026–6033. Nury, C.; Redeker, V.; Dautrey, S.; Romieu, A.; van der Rest, G.; Renard, P.-Y.; Melki, R.; Chamot-Rooke, J. A Novel Bio-Orthogonal Cross-Linker for Improved Protein / Protein Interaction Analysis. Anal Chem 2015, 87 (3), 1853–1860. Mahamid, J.; Pfeffer, S.; Schaffer, M.; Villa, E.; Danev, R.; Kuhn Cuellar, L.; Förster, F.; Hyman, A. A.; Plitzko, J. M.; Baumeister, W. Visualizing the Molecular Sociology at the HeLa Cell Nuclear Periphery. Science (1979) 2016, 351 (6276), 969–972. Pfeffer, S.; Mahamid, J. Unravelling Molecular Complexity in Structural Cell Biology. Curr Opin Struct Biol 2018, 52, 111–118. Freer, R.; McKillop, A. Synthesis of Symmetrical and Unsymmetrical Ureas Using Unsymmetrical Diaryl Carbonates. Synth Commun 1996, 26 (2), 331–349. Höfle, G.; Steglich, W.; Vorbrüggen, H.4‐Dialkylaminopyridines as Highly Active Acylation Catalysts.[New Synthetic Method (25)]. Angewandte Chemie International Edition in English 1978, 17 (8), 569–583. Chambers, M. C.; Maclean, B.; Burke, R.; Amodei, D.; Ruderman, D. L.; Neumann, S.; Gatto, L.; Fischer, B.; Pratt, B.; Egertson, J. A Cross-Platform Toolkit for Mass Spectrometry and Proteomics. Nat Biotechnol 2012, 30 (10), 918–920. Götze, M.; Pettelkau, J.; Schaks, S.; Bosse, K.; Ihling, C. H.; Krauth, F.; Fritzsche, R.; Kühn, U.; Sinz, A. StavroX—a Software for Analyzing Crosslinked Products in Protein Interaction Studies. J Am Soc Mass Spectrom 2011, 23 (1), 76–87. Graham, M.; Combe, C.; Kolbowski, L.; Rappsilber, J. XiView: A Common Platform for the Downstream Analysis of Crosslinking Mass Spectrometry Data. BioRxiv 2019, 561829. 63 42282.601_P17605-02 Honorato, R. V; Koukos, P. I.; Jiménez-García, B.; Tsaregorodtsev, A.; Verlato, M.; Giachetti, A.; Rosato, A.; Bonvin, A. M. J. J. Structural Biology in the Clouds: The WeNMR-EOSC Ecosystem. Front Mol Biosci 2021, 8, 729513. Van Zundert, G. C. P.; Rodrigues, J.; Trellet, M.; Schmitz, C.; Kastritis, P. L.; Karaca, E.; Melquiond, A. S. J.; van Dijk, M.; De Vries, S. J.; Bonvin, A. The HADDOCK2. 2 Web Server: User-Friendly Integrative Modeling of Biomolecular Complexes. J Mol Biol 2016, 428 (4), 720–725. Kolbowski, L.; Lenz, S.; Fischer, L.; Sinn, L. R.; O’Reilly, F. J.; Rappsilber, J. Improved Peptide Backbone Fragmentation Is the Primary Advantage of MS-Cleavable Crosslinkers. Anal Chem 2022, 94 (22), 7779–7786. Iacobucci, C.; Götze, M.; Ihling, C. H.; Piotrowski, C.; Arlt, C.; Schaefer, M.; Hage, C.; Schmidt, R.; Sinz, A. A Cross-Linking / Mass Spectrometry Workflow Based on MS- Cleavable Cross-Linkers and the MeroX Software for Studying Protein Structures and Protein–Protein Interactions. Nat Protoc 2018, 13 (12), 2864–2889. Mendes, M. L.; Fischer, L.; Chen, Z. A.; Barbon, M.; O’Reilly, F. J.; Giese, S. H.; Bohlke‐Schneider, M.; Belsom, A.; Dau, T.; Combe, C. W. An Integrated Workflow for Crosslinking Mass Spectrometry. Mol Syst Biol 2019, 15 (9), e8994–e8994. Graham, M.; Combe, C.; Kolbowski, L.; Rappsilber, J. XiView: A Common Platform for the Downstream Analysis of Crosslinking Mass Spectrometry Data. bioRxiv 2019, 561829. https: / / doi.org / 10.1101 / 561829. Yugandhar, K.; Wang, T.-Y.; Wierbowski, S. D.; Shayhidin, E. E.; Yu, H. Structure- Based Validation Can Drastically Underestimate Error Rate in Proteome-Wide Cross- Linking Mass Spectrometry Studies. Nat Methods 2020, 17 (10), 985–988. Graziadei, A.; Rappsilber, J. Leveraging Crosslinking Mass Spectrometry in Structural and Cell Biology. Structure 2022, 30 (1), 37–54. Klykov, O.; Steigenberger, B.; Pektaş, S.; Fasci, D.; Heck, A. J. R.; Scheltema, R. A. Efficient and Robust Proteome-Wide Approaches for Cross-Linking Mass Spectrometry. Nat Protoc 2018, 13 (12), 2964–2990. Piersimoni, L.; Kastritis, P. L.; Arlt, C.; Sinz, A. Cross-Linking Mass Spectrometry for Investigating Protein Conformations and Protein–Protein Interactions─ A Method for All Seasons. Chem Rev 2021, 122 (8), 7500–7531. 64 42282.601_P17605-02 Li, G.-W.; Burkhardt, D.; Gross, C.; Weissman, J. S. Quantifying Absolute Protein Synthesis Rates Reveals Principles Underlying Allocation of Cellular Resources. Cell 2014, 157 (3), 624–635. Keseler, I. M.; Mackie, A.; Santos-Zavaleta, A.; Billington, R.; Bonavides-Martínez, C.; Caspi, R.; Fulcher, C.; Gama-Castro, S.; Kothari, A.; Krummenacker, M. The EcoCyc Database: Reflecting New Knowledge about Escherichia Coli K-12. Nucleic Acids Res 2017, 45 (D1), D543–D550. Keseler, I. M.; Gama-Castro, S.; Mackie, A.; Billington, R.; Bonavides-Martínez, C.; Caspi, R.; Kothari, A.; Krummenacker, M.; Midford, P. E.; Muñiz-Rascado, L. The EcoCyc Database in 2021. Front Microbiol 2021, 12, 711077. Moura, A.; Savageau, M. A.; Alves, R. Relative Amino Acid Composition Signatures of Organisms and Environments. PLoS One 2013, 8 (10), e77319. Kawashima, T.; Berthet-Colominas, C.; Wulff, M.; Cusack, S.; Leberman, R. The Structure of the Escherichia Coli EF-Tu· EF-Ts Complex at 2.5 Å Resolution. Nature 1996, 379 (6565), 511–518. Williams, R. M.; Rimsky, S.; Buc, H. Probing the Structure, Function, and Interactions of the Escherichia Coli H-NS and StpA Proteins by Using Dominant Negative Derivatives. J Bacteriol 1996, 178 (15), 4335–4343. Rimsky, S. Structure of the Histone-like Protein H-NS and Its Role in Regulation and Genome Superstructure. Curr Opin Microbiol 2004, 7 (2), 109–114. Dorman, C. J.; Hinton, J. C. D.; Free, A. Domain Organization and Oligomerization among H-NS-like Nucleoid-Associated Proteins in Bacteria. Trends Microbiol 1999, 7 (3), 124–128. Stolz, B.; Huber, M.; Marković-Housley, Z.; Erni, B. The Mannose Transporter of Escherichia Coli. Structure and Function of the IIABMan Subunit. Journal of Biological Chemistry 1993, 268 (36), 27094–27099. Liu, X.; Zeng, J.; Huang, K.; Wang, J. Structure of the Mannose Transporter of the Bacterial Phosphotransferase System. Cell Res 2019, 29 (8), 680–682. Erni, B.; Zanolari, B.; Kocher, H. P. The Mannose Permease of Escherichia Coli Consists of Three Different Proteins. Amino Acid Sequence and Function in Sugar 65 42282.601_P17605-02 Transport, Sugar Phosphorylation, and Penetration of Phage Lambda DNA. Journal of Biological Chemistry 1987, 262 (11), 5238–5247. Hu, P.; Janga, S. C.; Babu, M.; Díaz-Mejía, J. J.; Butland, G.; Yang, W.; Pogoutse, O.; Guo, X.; Phanse, S.; Wong, P. Global Functional Atlas of Escherichia Coli Encompassing Previously Uncharacterized Proteins. PLoS Biol 2009, 7 (4), e1000096. Rajagopala, S. V; Sikorski, P.; Kumar, A.; Mosca, R.; Vlasblom, J.; Arnold, R.; Franca-Koh, J.; Pakala, S. B.; Phanse, S.; Ceol, A. The Binary Protein-Protein Interaction Landscape of Escherichia Coli. Nat Biotechnol 2014, 32 (3), 285–290. Ghosh, P.; Ishihama, A.; Chatterji, D. Escherichia Coli RNA Polymerase Subunit ω and Its N‐terminal Domain Bind Full‐length Β′ to Facilitate Incorporation into the Α2β Subassembly. Eur J Biochem 2001, 268 (17), 4621–4627. Guo, X.; Myasnikov, A. G.; Chen, J.; Crucifix, C.; Papai, G.; Takacs, M.; Schultz, P.; Weixlbaumer, A. Structural Basis for NusA Stabilized Transcriptional Pausing. Mol Cell 2018, 69 (5), 816–827. Kar, S.; Edgar, R.; Adhya, S. Nucleoid Remodeling by an Altered HU Protein: Reorganization of the Transcription Program. Proceedings of the National Academy of Sciences 2005, 102 (45), 16397–16402. Sobti, M.; Walshe, J. L.; Wu, D.; Ishmukhametov, R.; Zeng, Y. C.; Robinson, C. V; Berry, R. M.; Stewart, A. G. Cryo-EM Structures Provide Insight into How E. Coli F1Fo ATP Synthase Accommodates Symmetry Mismatch. Nat Commun 2020, 11 (1), 2615. Davies, K. M.; Anselmi, C.; Wittig, I.; Faraldo-Gómez, J. D.; Kühlbrandt, W. Structure of the Yeast F1Fo-ATP Synthase Dimer and Its Role in Shaping the Mitochondrial Cristae. Proceedings of the National Academy of Sciences 2012, 109 (34), 13602–13607. Py, B.; Causton, H.; Mudd, E. A.; Higgins, C. F. A Protein Complex Mediating MRNA Degradation in Escherichia Coli. Mol Microbiol 1994, 14 (4), 717–729. Kido, M.; Yamanaka, K.; Mitani, T.; Niki, H.; Ogura, T.; Hiraga, S. RNase E Polypeptides Lacking a Carboxyl-Terminal Half Suppress a MukB Mutation in Escherichia Coli. J Bacteriol 1996, 178 (13), 3917–3925. Vanzo, N. F.; Li, Y. S.; Py, B.; Blum, E.; Higgins, C. F.; Raynal, L. C.; Krisch, H. M.; Carpousis, A. J. Ribonuclease E Organizes the Protein Interactions in the Escherichia Coli RNA Degradosome. Genes Dev 1998, 12 (17), 2770–2781. 66 42282.601_P17605-02 Tsai, Y.-C.; Du, D.; Domínguez-Malfavón, L.; Dimastrogiovanni, D.; Cross, J.; Callaghan, A. J.; García-Mena, J.; Luisi, B. F. Recognition of the 70S Ribosome and Polysome by the RNA Degradosome in Escherichia Coli. Nucleic Acids Res 2012, 40 (20), 10417–10431. Aït-Bara, S.; Carpousis, A. J.; Quentin, Y. RNase E in the γ-Proteobacteria: Conservation of Intrinsically Disordered Noncatalytic Region and Molecular Evolution of Microdomains. Molecular Genetics and Genomics 2015, 290, 847–862. Conway, L. P.; Jadhav, A. M.; Homan, R. A.; Li, W.; Rubiano, J. S.; Hawkins, R.; Lawrence, R. M.; Parker, C. G. Evaluation of Fully-Functionalized Diazirine Tags for Chemical Proteomic Applications. Chem Sci 2021, 12 (22), 7839–7847. Meissner, F.; Geddes-McAlister, J.; Mann, M.; Bantscheff, M. The Emerging Role of Mass Spectrometry-Based Proteomics in Drug Discovery. Nat Rev Drug Discov 2022, 21 (9), 637–654. Cohen, R. D.; Pielak, G. J. A Cell Is More than the Sum of Its (Dilute) Parts: A Brief History of Quinary Structure. Protein Science 2017, 26 (3), 403–413. Guin, D.; Gruebele, M. Weak Chemical Interactions That Drive Protein Evolution: Crowding, Sticking, and Quinary Structure in Folding and Function. Chem Rev 2019, 119 (18), 10691–10717. Götze, M.; Iacobucci, C.; Ihling, C. H.; Sinz, A. A Simple Cross-Linking / Mass Spectrometry Workflow for Studying System-Wide Protein Interactions. Anal Chem 2019, 91 (15), 10236–10244. Gonzalez-Lozano, M. A.; Koopmans, F.; Sullivan, P. F.; Protze, J.; Krause, G.; Verhage, M.; Li, K. W.; Liu, F.; Smit, A. B. Stitching the Synapse: Cross-Linking Mass Spectrometry into Resolving Synaptic Protein Interactions. Sci Adv 2020, 6 (8), eaax5783. Stieger, C. E.; Doppler, P.; Mechtler, K. Optimized Fragmentation Improves the Identification of Peptides Cross-Linked by MS-Cleavable Reagents. J Proteome Res 2019, 18 (3), 1363–1370. Roux, K. J.; Kim, D. I.; Raida, M.; Burke, B. A Promiscuous Biotin Ligase Fusion Protein Identifies Proximal and Interacting Proteins in Mammalian Cells. Journal of cell biology 2012, 196 (6), 801–810. 67 42282.601_P17605-02 Gingras, A.-C.; Abe, K. T.; Raught, B. Getting to Know the Neighborhood: Using Proximity-Dependent Biotinylation to Characterize Protein Complexes and Map Organelles. Curr Opin Chem Biol 2019, 48, 44–54. Lee, K.; O’Reilly, F. J. Cross-Linking Mass Spectrometry for Mapping Protein Complex Topologies in Situ. Essays Biochem 2023, EBC20220168. O’Reilly, F. J.; Graziadei, A.; Forbrig, C.; Bremenkamp, R.; Charles, K.; Lenz, S.; Elfmann, C.; Fischer, L.; Stülke, J.; Rappsilber, J. Protein Complexes in Bacillus Subtilis by AI-Assisted Structural Proteomics. Mol. Sys. Biol.2023, 19, e11544. Liang, J.; Zhang, L.; Tan, X.; Qi, Y.; Feng, S.; Deng, H.; Yan, Y.; Zheng, J.; Liu, L.; Tian, C. Chemical Synthesis of Diubiquitin‐based Photoaffinity Probes for Selectively Profiling Ubiquitin‐binding Proteins. Angewandte Chemie 2017, 129 (10), 2788–2792. Zheng, Q.; Pang, Z.; Liu, J.; Zhou, Y.; Sun, Y.; Yin, Z.; Lou, Z. Photoaffinity Palladium Reagents for Capture of Protein–Protein Interactions. Org Biomol Chem 2019, 17 (26), 6369–6373. Shigdel, U. K.; Zhang, J.; He, C. Diazirine‐Based DNA Photo‐Cross‐Linking Probes for the Study of Protein–DNA Interactions. Angewandte Chemie 2008, 120 (1), 96–99. Gerfaud, T.; Martin, C.; Bouquet, K.; Talano, S.; Millois-Barbuis, C.; Musicki, B.; Boiteau, J.-G.; Cardinaud, I. Process Development and Good Manufacturing Practice Production of a Tyrosinase Inhibitor via Titanium-Mediated Coupling between Unprotected Resorcinols and Ketones. Org Process Res Dev 2017, 21 (4), 631–640. Diamond, R. Real-Space Refinement of the Structure of Hen Egg-White Lysozyme. J Mol Biol 1974, 82 (3), 371–391. Belsom, A.; Rappsilber, J. Anatomy of a Crosslinker. Curr Opin Chem Biol 2021, 60, 39–46. Iacobucci, C.; Piotrowski, C.; Aebersold, R.; Amaral, B. C.; Andrews, P.; Bernfur, K.; Borchers, C.; Brodie, N. I.; Bruce, J. E.; Cao, Y. First Community-Wide, Comparative Cross-Linking Mass Spectrometry Study. Anal Chem 2019, 91 (11), 6953–6961. Vellucci, D.; Kao, A.; Kaake, R. M.; Rychnovsky, S. D.; Huang, L. Selective Enrichment and Identification of Azide-Tagged Cross-Linked Peptides Using Chemical Ligation and Mass Spectrometry. J Am Soc Mass Spectrom 2010, 21 (8), 1432–1445. Fritzsche, R.; Ihling, C. H.; Götze, M.; Sinz, A. Optimizing the Enrichment of Cross- 68 42282.601_P17605-02 Linked Products for Mass Spectrometric Protein Analysis. Rapid Communications in Mass Spectrometry 2012, 26 (6), 653–658 Schmidt, R.; Sinz, A. Improved Single-Step Enrichment Methods of Cross-Linked Products for Protein Structure Analysis and Protein Interaction Mapping. Anal Bioanal Chem 2017, 409 (9), 2393–2400. Hulsen, T.; de Vlieg, J.; Alkema, W. BioVenn–a Web Application for the Comparison and Visualization of Biological Lists Using Area-Proportional Venn Diagrams. BMC Genomics 2008, 9 (1), 1–6. Keller, A.; Chavez, J. D.; Felt, K. C.; Bruce, J. E. Prediction of an Upper Limit for the Fraction of Interprotein Cross-Links in Large-Scale in Vivo Cross-Linking Studies. J Proteome Res 2019, 18 (8), 3077–3085. Dunkle, J. A.; Wang, L.; Feldman, M. B.; Pulk, A.; Chen, V. B.; Kapral, G. J.; Noeske, J.; Richardson, J. S.; Blanchard, S. C.; Cate, J. H. D. Structures of the Bacterial Ribosome in Classical and Hybrid States of TRNA Binding. Science (1979) 2011, 332 (6032), 981–984. Pulk, A.; Cate, J. H. D. Control of Ribosomal Subunit Rotation by Elongation Factor G. Science (1979) 2013, 340 (6140). Montesano-Roditis, L.; Glitz, D. G.; Traut, R. R.; Stewart, P. L. Cryo-Electron Microscopic Localization of Protein L7 / L12 within the Escherichia Coli 70 S Ribosome by Difference Mapping and Nanogold Labeling. Journal of Biological Chemistry 2001, 276 (17), 14117–14123. Liljas, A.; Sanyal, S. The Enigmatic Ribosomal Stalk. Q Rev Biophys 2018, 51. Wendrich, T. M.; Blaha, G.; Wilson, D. N.; Marahiel, M. A.; Nierhaus, K. H. Dissection of the Mechanism for the Stringent Factor RelA. Mol Cell 2002, 10 (4), 779–788. Zhao, G.; Winkler, M. E. A Novel Alpha-Ketoglutarate Reductase Activity of the SerA-Encoded 3-Phosphoglycerate Dehydrogenase of Escherichia Coli K-12 and Its Possible Implications for Human 2-Hydroxyglutaric Aciduria. J Bacteriol 1996, 178 (1), 232–239. Stepp, L. R.; Bleile, D. M.; McRorie, D. K.; Pettit, F. H.; Reed, L. J. Use of Trypsin and Lipoamidase to Study the Role of Lipoic Acid Moieties in the Pyruvate and. Alpha.- Ketoglutarate Dehydrogenase Complexes of Escherichia Coli. Biochemistry 1981, 20 69 42282.601_P17605-02 (16), 4555–4560. Hammel, M.; Amlanjyoti, D.; Reyes, F. E.; Chen, J.-H.; Parpana, R.; Tang, H. Y. H.; Larabell, C. A.; Tainer, J. A.; Adhya, S. HU Multimerization Shift Controls Nucleoid Compaction. Sci Adv 2016, 2 (7), e1600650–e1600650. Stelmashenko, O.; Lalo, U.; Yang, Y.; Bragg, L.; North, R. A.; Compan, V. Activation of Trimeric P2X2 Receptors by Fewer than Three ATP Molecules. Mol Pharmacol 2012, 82 (4), 760–766 Calloni, G.; Chen, T.; Schermann, S. M.; Chang, H. C.; Genevaux, P.; Agostini, F.; Tartaglia, G. G.; Hayer-Hartl, M.; Hartl, F. U. DnaK Functions as a Central Hub in the E. Coli Chaperone Network. Cell Rep 1 (3): 251–264.2012. Kerner, M. J.; Naylor, D. J.; Ishihama, Y.; Maier, T.; Chang, H.-C.; Stines, A. P.; Georgopoulos, C.; Frishman, D.; Hayer-Hartl, M.; Mann, M. Proteome-Wide Analysis of Chaperonin-Dependent Protein Folding in Escherichia Coli. Cell 2005, 122 (2), 209– 220. Mayer, M. P.; Prodromou, C.; Frydman, J. The Hsp90 Mosaic: A Picture Emerges. Nat Struct Mol Biol 2009, 16 (1), 2–6. Chiosis, G.; Dickey, C. A.; Johnson, J. L. A Global View of Hsp90 Functions. Nat Struct Mol Biol 2013, 20 (1), 1–4. Marx, D. C.; Plummer, A. M.; Faustino, A. M.; Devlin, T.; Roskopf, M. A.; Leblanc, M. J.; Lessen, H. J.; Amann, B. T.; Fleming, P. J.; Krueger, S. SurA Is a Cryptically Grooved Chaperone That Expands Unfolded Outer Membrane Proteins. Proceedings of the National Academy of Sciences 2020, 117 (45), 28026–28035. Grudniak, A. M.; Pawlak, K.; Bartosik, K.; Wolska, K. I. Physiological Consequences of Mutations in the HtpG Heat Shock Gene of Escherichia Coli. Mutation Research / Fundamental and Molecular Mechanisms of Mutagenesis 2013, 745, 1–5. Karp, P. D.; Ong, W. K.; Paley, S.; Billington, R.; Caspi, R.; Fulcher, C.; Kothari, A.; Krummenacker, M.; Latendresse, M.; Midford, P. E. The Ecocyc Database. EcoSal Plus 2018, 8 (1). Li, G.-W.; Burkhardt, D.; Gross, C.; Weissman, J. S. Quantifying Absolute Protein Synthesis Rates Reveals Principles Underlying Allocation of Cellular Resources. Cell 2014, 157 (3), 624–635. 70 42282.601_P17605-02 UniProt: The Universal Protein Knowledgebase in 2021. Nucleic Acids Res 2021, 49 (D1), D480–D489. Gupta, C.; Sarkar, D.; Tieleman, D. P.; Singharoy, A. The Ugly, Bad, and Good Stories of Large-Scale Biomolecular Simulations. Curr Opin Struct Biol 2022, 73, 102338. Rickard, M. M.; Zhang, Y.; Gruebele, M.; Pogorelov, T. v. In-Cell Protein–Protein Contacts: Transient Interactions in the Crowd. J Phys Chem Lett 2019, 10 (18), 5667– 5673. Beck, M.; Baumeister, W. Cryo-Electron Tomography: Can It Reveal the Molecular Sociology of Cells in Atomic Detail? Trends Cell Biol 2016, 26 (11), 825–837. Tegunov, D.; Xue, L.; Dienemann, C.; Cramer, P.; Mahamid, J. Multi-Particle Cryo- EM Refinement with M Visualizes Ribosome-Antibiotic Complex at 3.5 Å in Cells. Nat Methods 2021, 18 (2), 186–193 Although the foregoing subject matter has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be understood by those skilled in the art that certain changes and modifications can be practiced within the scope of the appended claims. 71 42282.601_P17605-02

Claims

THAT WHICH IS CLAIMED:

1. A compound of formula (I): ; wherein n is an integer2. The compound of claim 1, wherein the compound of formula (I) is selected from: .

3. A conjugate comprising a first peptide (P1), a second peptide (P2), and a urea- based linker derived from a compound of formula (I): ; wherein m and n are each1, 2, 3, 4, 5, 6, 7, and 8.

4. A method for detecting protein-protein interactions in a sample, the method comprising: (a) contacting one or more proteins in the sample with a compound of formula (I) comprising a diazirine functional group and an NHS-carbamate functional group: ; wherein n is an integerwhereby the NHS-carbamate functional group reacts with one or more lysine or N-terminal amine groups of the one or more proteins to generate a conjugate comprising a urea linked to the diazirine functional 72 42282.601_P17605-02group and a first peptide residue, wherein the linkage between the urea and the first peptide residue is a first mass spectrometer (MS)-cleavable linkage; (b) irradiating the conjugate with ultraviolet (UV) radiation whereby the diazirine functional group forms a carbene group that reacts with a second peptide residue and links the second peptide residue to the urea, wherein the linkage between the urea and the second peptide residue is a second mass spectrometer (MS)-cleavable linkage; and (c) cleaving the urea at the first MS-cleavable linkage or the second MS-cleavable linkage to generate a first peptide product comprising an amine fragment and a second peptide product comprising an isocyanate fragment, wherein the first peptide product and the second peptide product are on the same protein or different proteins; and (c) analyzing the sample by mass spectrometry to detect protein-protein interactions in the sample, wherein the first peptide product and the second peptide product form characteristic doublets in a MS spectrum.

5. The method of claim 4, further comprising identifying interprotein and / or intraprotein crosslinks.

6. The method of claim 4, wherein the Cα–Cα distance between the first peptide residue and the second peptide residue is below 30 Å, 25 Å, 20 Å, or 15 Å.

7. The method of claim 4, wherein the Cα–Cα distance between the first peptide residue and the second peptide residue is below 15 Å.

8. The method of claim 4, wherein the Cα–Cα distance between the first peptide residue and the second peptide residue is greater than 50 Å.

9. The method of claim 4, wherein the first peptide residue and / or the second peptide reside is a non-acidic residue.

10. The method of claim 4, wherein the first peptide residue and / or the second peptide reside is a Ala, Gly, Val, or Tyr. 73 42282.601_P17605-0211. The method of claim 4, wherein the sample is selected from a lysate, a cell, and a tissue sample.

12. The method of claim 4, wherein steps (a) and (b) are performed in vivo or in cellulo.

13. The method of any one of claims 4 to 12, wherein the mass spectrometry comprises liquid chromatography (LC)-mass spectrometry / mass spectrometry (MS / MS).

14. The method of any one of claims 4 to 13, wherein cleaving comprises higher-energy collision-induced dissociation (HCD) conditions.

15. The method of any one of claims 4 to 14, wherein the UV radiation wavelength is 365 nm.

16. The method of any one of claims 4 to15, wherein the UV radiation is with a UV LED.

17. The method of any one of claims 4 to16, wherein the UV radiation is for 1, 2, or 3 seconds.

18. The method of any one of claims 4 to 17, further comprising adding a crosslinking catalyst during step (a).

19. The method of any one of claims 4 to 18, further comprising adding a quenching reagent prior to step (b).

20. The method of any one of claims 4 to 19, further comprising one or more of the following steps prior to the mass spectrometry analysis of step (c): (a) denaturing the sample; (b) reducing the sample by addition of dithiothreitol (DTT); (c) alkylating the sample; (d) either: (i) digesting the sample with trypsin; or (ii) digesting serially with Lys-C and then trypsin; 74 42282.601_P17605-02(e) desalting the digested samples; and / or (f) fractionating and enriching the crosslinked sample by size exclusion chromatography (SEC).

21. The method of any one of claims 4 to 20, further comprising performing a spectral search for one or more crosslinked peptides. 75 42282.601_P17605-02

Images