Promiscuous structural cross-compatibilities between major shell components of Klebsiella pneumoniae bacterial microcompartments

Promiscuous structural cross-compatibilities between major shell components of Klebsiella pneumoniae bacterial microcompartments | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results Promiscuous structural cross-compatibilities between major shell components of Klebsiella pneumoniae bacterial microcompartments Lucie Barthe , View ORCID Profile Damien Balestrino , Bessam Azizi , View ORCID Profile Delphine Dessaux , View ORCID Profile Vanessa Soldan , View ORCID Profile Jeremy Esque , View ORCID Profile Thomas Schiex , View ORCID Profile Sophie Barbe , View ORCID Profile Luis Fernando Garcia-Alles doi: https://doi.org/10.1101/2025.01.01.631032 Lucie Barthe 1 TBI, Université de Toulouse , CNRS, INRAE, INSA, 31077 Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site Damien Balestrino 2 UFR Pharmacie, LMGE UMR CNRS6023 , 63001 Clermont-Ferrand, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Damien Balestrino Bessam Azizi 1 TBI, Université de Toulouse , CNRS, INRAE, INSA, 31077 Toulouse, France 3 MIAT, Université Fédérale de Toulouse , INRAE, UR 875, ANITI 31326, Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site Delphine Dessaux 1 TBI, Université de Toulouse , CNRS, INRAE, INSA, 31077 Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Delphine Dessaux Vanessa Soldan 4 Plateforme de microscopie électronique intégrative METi, Centre de Biologie Intégrative , CNRS, 118 Route de Narbonne, 31062 Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Vanessa Soldan Jeremy Esque 1 TBI, Université de Toulouse , CNRS, INRAE, INSA, 31077 Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jeremy Esque Thomas Schiex 3 MIAT, Université Fédérale de Toulouse , INRAE, UR 875, ANITI 31326, Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Thomas Schiex Sophie Barbe 1 TBI, Université de Toulouse , CNRS, INRAE, INSA, 31077 Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Sophie Barbe Luis Fernando Garcia-Alles 1 TBI, Université de Toulouse , CNRS, INRAE, INSA, 31077 Toulouse, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Luis Fernando Garcia-Alles For correspondence: lgarciaa{at}insa-toulouse.fr Abstract Full Text Info/History Metrics Supplementary material Preview PDF ABSTRACT Bacterial microcompartments (BMC) are submicrometric reactors that encapsulate dedicated metabolic activities. BMC-H hexamers, the most abundant components of BMC shells, play major roles for shell plasticity and permeability. In part, chemical exchanges between the BMC lumen and the cellular cytosol will be defined by the disposition of amino acids lining the central BMC-H pores. Current models attribute to BMC-H a homo-oligomeric nature. The hexagonal symmetry of corresponding pores, however, would break down if hetero-hexamers formed, a possibility suggested by the frequent presence of multiple paralogs within BMC operons. Here, we gauged the degree of structural promiscuity between the 11 BMC-H paralogs from Klebsiella pneumoniae , a potential human pathogen endowed with the capacity to express three different BMC types. Concomitant activation of transcription of several BMC operons was first shown to be possible. By leveraging an adapted tripartite GFP technology, all possible BMC-H pair combinations were screened in E. coli . Multiple structural cross-compatibilities were pinpointed between homologs arising not only from the same BMC operon, but also from different BMC types, results supported by Alphafold and ESMFold predictions. The structural stability and assembly propensity of selected hetero-associations was established by biochemical means. In light of these results, we reinterpreted published lysine cross-linking mass spectrometry data to demonstrate that one of these hetero-hexamers, involving PduA and PduJ, was already detected to form in the shell of a recombinantly-expressed 1,2-propanediol utilization compartment from Salmonella enterica . Altogether, this study points to the need to embrace an augmented structural complexity in BMC shells. INTRODUCTION Bacterial microcompartments (BMC) are complex macrostructures composed of a protein shell encompassing an enzymatic set that defines its precise function. Polyhedral shapes have been inferred for several BMC types using techniques like electron cryotomography or transmission electron microscopy (TEM) 1 – 4 . BMC sizes spread from about 40 nm up till 600 nm 5 , 6 . Best characterized cases are the α- and β-carboxysomes, which contribute to CO 2 -fixation by cyanobacteria and some autotrophs, or the 1,2-propanediol (PD) utilization (PDU) and ethanolamine (EA) utilization (EUT) compartments 7 , 8 . However, the catalog of BMC-mediated activities/functions is steadily expanding 9 – 11 . Indeed, bioinformatic surveys of available genomes uncovered BMC loci in about 20% of all sequenced bacterial genomes 11 – 13 . Compelling evidence supports that BMC confer competitive advantages to some bacteria under physiological contexts 14 . For instance, by processing EA or PD released in the gastro-intestinal tract, pathogenic microorganisms as well as some pro-pathogenic commensals, might out-compete other strains 15 – 17 . BMCs would therefore contribute to boost the virulence of some pathogens 14 . Not surprisingly, non-pathogenic commensal strains have also developed BMC-related counteracting strategies 18 . Remarkably, about 22% of BMC-endowed organisms carry genetic information coding for two or more BMC types, with most extreme cases reaching up to 6 BMC loci 11 . A relatively well-characterized case is Salmonella enterica LT2 ( Sal ), which contains both EUT and PDU compartments. Most often, the set of genes coding for shell components, auxiliary proteins and cargo enzymes are gathered in a single locus or operon. In some cases, the genetic information is split in several loci . This happens, for instance, in β-cyanobacteria, which distribute carboxysome genes in a main operon and satellite loci 8 , 12 . Shell proteins are often classified in three major classes: i) the most abundant subunits, stoichiometrically, adopt a PF00936 structural fold and associate as hexamers (BMC-H) 19 ; ii) in the second group, each subunit is a fusion of two PF00936 domains. Consequently, the final pseudo-hexamer is in fact a trimer (BMC-T) 20 , 21 . BMC-H and BMC-T can be further subclassified depending on other criteria, such as the presence of permuted secondary structural elements, the existence of N-terminal or C-terminal extensions or the capacity to form hexagonal stacks 22 , 23 ; iii) components that oligomerize as pentamers (BMC-P), each monomer obeying a PF003319 fold. Several lines of evidence 23 – 25 , including high resolution structures of synthetic 26 , 27 and natural compartments 28 support a shell model with polyhedral facets and edges composed of BMC-H and, when present, BMC-T, while vertices are capped by BMC-P. About 2.6 BMC-H, 0.9 BMC-T and 1.7 BMC-P gene paralogs are found on average per BMC locus 11 . These homologs generally share sufficient sequence identity as to envision that hetero-oligomers combining several homologs could form. Indeed, structural compatibilities were proved within a subgroup of BMC-H. CcmK3 and CcmK4 homologs from Halothece sp. PCC 7418, Synechococcus elongatus PCC 7942 or Synechocystis. sp. PCC 6803 ( Syn6803 ), as well as CcmK1 and CcmK2 from the last species formed hetero-hexamers after recombinant co-expression in E. coli 29 , 30 . Exploring whether this is a widespread phenomenon is important, since heteromerization is anticipated to impact features like shell permeability 31 – 33 , the capacity to recruit other BMC components 34 , 35 , or the kinetics of compartment biosynthesis, degradation and repartition during cell division 36 – 38 . Herein, we selected for our investigation Klebsiella pneumoniae (strain 342, hereafter Kpe ), an important human pathogen that harbors three different BMC operons ( eut1 , pdu1a and grm2 ), and demonstrated that hetero-hexamerization also concerns catabolic compartments. Transcriptomic data first supported that expression of several BMC might happen simultaneously. Subsequently, and mostly leveraging an adapted protocol to monitor protein-protein interactions (PPI) between BMC-H 39 , based on the tripartite-GFP technology (tGFP) 40 , we could pinpoint abundant structural compatibilities between BMC-H. A majority of combinations between the 11 Kpe BMC-H monomers indicated positive PPI, both when considering components from the same or from different BMC types. A systematic study of the different combinations with artificial intelligence (AI)-based multimeric structure prediction methods corroborated structural compatibilities. Finally, we reinterpreted reported cross-linking data 41 , demonstrating the occurrence of hetero-hexamers in BMC shells. Our study thus extends previous conclusions established from studies of carboxysome components to other BMC types, arguing for the strong necessity to characterize hetero-oligomerization extent in natural hosts and to evaluate the consequences for BMC function and eventually for the organism pathogenicity/virulence. RESULTS Genetic organization and regulation of Klebsiella pneumoniae BMC Together with K. quasipneumoniae , and K. variicola, K. pneumoniae ( Kpe ) is responsible for many nosocomial and community-acquired pneumonia, urinary tract, and bloodstream infections in healthcare settings. Kpe harbors three different BMC loci ( Fig. 1 ). The first one closely resembles the pdu1a operon from S. enterica 42 . The signature pathway enzyme is the diol dehydratase (DDH) that converts PD to propionaldehyde and is composed of three subunits (PduC/PduD/PduE). The operon codes for 4 BMC-H (PduA/PduJ/PduK/PduU), 2 BMC-T (PduB/T) and 1 BMC-P (PduN). It is preceded by two adjacent and divergently transcribed genes, pocR and pduF , which code for a positive transcription factor directly acting on the pdu1a promoter and a PD transporter, respectively. The last two genes are followed by the cob operon that participates to the de novo synthesis of adenosyl-cobalamin (Ado-B 12 ), a cofactor required for PD and EA utilization and implicated in transcriptional regulation mechanims 7 , 43 , 44 . Download figure Open in new tab Figure 1. Genetic organization of Kpe BMC operons The three BMC-coding operons of Klebsiella pneumoniae (strain 342) are shown, together with flanking regulatory elements. Promoters are represented by arrows, in black for those which directly control BMC transcription. Numbers beneath BMC-H coding sequences indicate the chromosomal location of the first nucleotide of the translation start codon. The eut1 operon also globally recapitulates the organization found in S. enterica . The key enzyme is the EA ammonia lyase, composed of EutB and EutC proteins, which activity also depends on Ado-B 12 . Among the 16 CDS that constitute the Kpe eut1 operon, 3 of them correspond to BMC-H proteins (EutS/EutM/EutK), 1 to BMC-T (EutL) and 1 to BMC-P (EutN). As in S. enterica , the transcriptional regulator (EutR) is coded at the operon downstream end. S. enterica EutR was shown to activate the eut operon when both EA and Ado-B are present 43 . Importantly, however, such activation was repressed by PD 42 , something required to prevent the detrimental mixing of components from the two compartments. Accordingly, PD was preferentially consumed over EA as carbon source during growth. Kpe is also endowed with a group-2 glycyl-radical enzyme-associated microcompartment (GRM2) that participates to choline (CL) metabolism. Its signature enzyme is the CL-trimethylamine (TMA) lyase, CutC, which catalyzes the decomposition of CL into acetaldehyde and TMA 10 , 45 . The corresponding grm2 locus of Kpe codes for 4 BMC-H (CmcA/CmcB/CmcC/CmcE) and a single BMC-P (CmcD). GRM2 are deprived of BMC-T 11 , 46 . Upstream the grm2 promoter is found a small operon that codes for regulatory CutW-CutX-CutY. By similarity to what happens in uropathogenic E. coli 536 strain 31 , CutX and CutY are expected to activate grm2 transcription in the presence of CL under anaerobic conditions. BMC transcription The comparable organization of eut1 and pdu1a operons in S. enterica and Kpe , with flanking eutR and pocR regulatory genes, suggested that similar transcriptional regulation schemes might apply. Thus, we investigated whether similar repression schemes operate in Kpe . Preliminarily, we sought for conditions in which this organism was able to grow on PD, EA or CL. Under aerobic conditions, EA could be utilized as carbon and nitrogen sources in minimal medium, provided that vitamin B12 was added, in agreement with the literature 43 . CL stimulated Kpe growth on minimal medium that also contained small amounts of yeast extract (YE, 0.1-0.2%). Something similar occurred under anaerobic conditions with EA or CL. However, no growth benefit could be attributed to the presence of PD, both under aerobic or anaerobic conditions and regardless of supplementation with diverse components: presence or absence of cyanocobalamin (vit-B 12 ), YE (0.1 to 0.5 %), fumarate, amino acids, metal salts (including ferric citrate), or bile salts 16 . It is important to mention that metabolite utilization is not necessarily associated to BMC. This is illustrated by a recent report, which attributed EA utilization as carbon source under aerobic conditions to the main Kpe eut operon, while nitrogen utilization was basically dependent on a second short locus that includes EutB/EutC ammonia-lyase subunits but lacks genes coding for other EUT components, including those necessary for shell assembly 15 . Short locus enzyme sequences are 51 to 33% identical to those from the eut1 operon. Similarly, PduC/PduD/PduE homologs are encoded in a small locus and exhibit even stronger resemblance to the PDU counterparts (75%/63%/61% identities). Two cutC/cutD-containing stretches are also present in Kpe , though more distant from GRM2 sequences (32-34% identities). Taking into account such complexifying possibilities, we focused our investigation on the measurement of transcription shifts induced by PD, EA and/or CL. This was ascertained using reverse-transcription quantitative PCR (RT-qPCR) on total RNA extracted from cells growing under anaerobic conditions in NCE minimal medium supplemented or not with each individual metabolite (PD, EA or CL), or with combinations of two of them. We monitored upstream, middle and downstream regions from each operon (see M&M section). In that manner, a strong upregulation of the three grm2 sections was induced by CL. Almost 500- and 100-fold higher transcript levels were measured for cmcA or cutC ( Table 1 , see also Tables S1-S2). GRM2-induction by CL was repressed by EA, although not completely. PD addition was without effect. To our surprise, none of the screened eut or pdu transcript levels was significantly modified by the corresponding substrate, the highest upregulation being a highly dispersed 6-fold increase measured for pduU . Only CL continued to promote higher transcript levels, as estimated for eutM (17-fold), pduJ (36-fold) and pduU (27-fold). Intriguingly, EA or PD promoted a clear augmentation of transcript levels of their cognate BMC genes when combined to CL (a ca. 500-fold strongly dispersed average for eutM or a 80- and 50-fold for pduJ or pduU , respectively). Noticeably, when CL and EA were supplemented together, values suffered from a considerably high dispersion, something particularly evident for eut genes. Experimental manipulation errors were ruled out by correct readings on proC and rpoD housekeeping gene transcripts. Variability deriving from inadequate primer selection was also excluded as amplification reactions of the same stretches were reproducible for cultures grown in the presence of other substrates. View this table: View inline View popup Download powerpoint Table 1. Induction of Kpe BMC transcription in response to metabolite presence Globally, these data supported the possibility that components from different BMC can be expressed simultaneously in Kpe cells. Kpe BMC-H sequence and structural predictions Sequence considerations permitted to classify the 11 Kpe BMC-H into three groups (Fig. S1A). The first one includes all canonical BMC-H: CmcA, CmcB, CmcC, EutM, PduA and PduJ. Identities range within this group between 56% (CmcA vs EutM) and 95% (CmcA vs CmcB), and are even higher (60 to 98 %) when considering only residues at the interface between monomers (Fig. S1B). The second group comprises BMC-H with long C-terminal extensions (40-70 residues) compared to canonical BMC-H. This group includes CmcE, EutK and PduK. Their sequences show greater divergence from each other than from the canonical BMC-Hs within their respective BMCs. Finally, EutS and PduU, which belong to the third BMC-H group, carry circular permutations of secondary structure elements, compared to canonical proteins. They also possess N-terminal extensions that associate in the context of an hexamer to build a narrow β-barrel above the central pore, as demonstrated in crystal structures for other homologs 47 – 49 . As already known 11 , these proteins have more in common together (56% identity) than with homologs of their respective BMC (19 to 25%). The difference becomes more pronounced when only interfacial residues are considered (73% for EutS-PduU, as compared to 15-27%, Fig. S1B), which suggests that different evolutionary constraints might apply on this BMC-H group. The potential formation of Kpe BMC-H homo-hexamers was explored with two AI-based multimeric structure prediction methods: AlphaFold2-Multimer (AF2) 50 , which primarily infers residue-residue relationships through multi-sequence alignments (MSAs), and ESMFold 51 , which predicts structures directly from amino acid sequences using a transformer-based language model, ESM-2. The two algorithms proposed homo-hexamers for canonical BMC-H, each monomer adopting the characteristic PF00936 fold (Table S3, Fig. S2A-B). Prediction quality scores pointed to high-confidence associations (global and interface_pLDDT > 75 and interchain and interface_PAEs < 3.5, see M&M for the detailed description of prediction quality metrics). These values were similar to those obtained in predictions for the positive control based on the single BMC-H from M. smegmatis (RMM), and differed from data for negative controls based on the monomeric buckwheat trypsin inhibitor (BWI) or the insoluble Syn6803 CcmK3. In fact, AF2 metrics suggested homo-hexamer formation for the later, pointing to a possible bias caused by the presence of similar sequences that would impact the MSA and eventually the confidence scores and the comparison between models of distinct proteins. Interaction energies (ΔE) estimated for relaxed structures were also in the range of those calculated for RMM and available crystallographic structures of BMC-H homo-hexamers (Table S4). Models predicted by AF2 for BMC from the second group, CmcE, EutK and PduK, were associated to relatively low-confidence quality scores, though energies seemed correct. Since low quality metrics for CmcE, EutK and PduK might be caused be the long C-ter flexible extensions, which were predicted with large unstructured parts by both methods (Fig. S2), we also evaluated PAE values for only residues from the PF00936 core (c_PAE). These values were correct for CmcE and EutK. Conversely, ESMFold directly rejected the oligomerization of PduK, and pointed to unlikely CmcE and EutK hexamer occurrence, even when considering only core residues. pLDDT and PAE indicators were in the range of those for negative controls. Remarkably, the two methods predicted the presence in C-terminal extensions of a small structured Cys-rich motif in PduK, which could harbor a [Fe-S] cluster domain (Fig. S2E-F), something partly demonstrated for the S. enterica homolog 52 , and a fold in EutK typical of nucleic acid-binding proteins (PF16365) (Fig. S2I-J), as proposed for the monomeric E. coli homolog too 47 . Predictions for EutS and PduU diverged considerably depending on the AI-algorithm. ESMFold indicated an absence of oligomerization for EutS or a hexamer model with low-confidence metrics for PduU. The disposition of the first β-strand of the EutS monomer (corresponding to the canonical β4, Fig. S1A) deviated in fact from the PF00936 fold (Fig. S2D), adopting a conformation that would collapse with a neighbor monomer if an hexamer formed. ESMFold also attributed helical structures to a portion of the EutS and PduU N-terminal extensions, in disagreement with expectations based on the known structures of PduU and EutS homologs 47 – 49 . Conversely, AF2 predicted for both proteins high-confidence homo-hexamers displaying favorable binding energies (Fig. S2C) and N-terminal extensions reproducing faithfully the short β-barrels modelled in crystal structures. Experimental structural viability and oligomeric state of recombinant Kpe BMC-H The experimental behavior, oligomerization state and assembly properties of individual Kpe BMC-H was next inspected. SDS-PAGE analysis confirmed that all 11 His 6 -tagged proteins could be over-expressed in BL21(DE3) cells ( Fig. 2 ). Only EutM was produced in moderate yields. After cellular lysis and centrifugation, most of them were retained in pellets, pointing to protein aggregation or assembly into supramolecular structures. As exceptions, EutM and CmcB remained soluble and could be purified directly. All other BMC-H could be recovered from pellets after protein disassembly with 1 M urea at 4°C (which does not permit resolubilization of aggregated proteins). These fractions remained soluble after purification. Download figure Open in new tab Figure 2. Expression of Kpe BMC-H in E. coli SDS-PAGE characterization of His 6 -tagged Kpe BMC-H over-expressed in BL21(DE3). From top to bottom are presented total cellular fractions, soluble fractions retained in supernatants after lysis and centrifugation, and material purified by cobalt-based affinity chromatography. The bottom panel corresponds to fractions that were purified similarly after load into the same affinity resins of material resolubilized from pellets treated with 1 M urea. Only the of interest is shown. An asterisk preceding attachement to the protein N-terminus; C-t ortion of the gel showing the proteins the protein name is to denote tag rminal tags are present for all other cases. The total and soluble fractions, likewise purified and urea-purified fractions, allow a direct comparative view of protein levels since gel loads were identical while fraction volumes remained constant throughout the experiment. Note however that Coomassie staining/destaining might slightly vary from gel to gel. The approximate migration of protein ladder species of 10 and 15 kDa sizes is indicated on the left side. See M&M for further details. Urea purified His 6 -tagged BMC-H were analyzed by size-exclusion chromatography (SEC-HPLC). EutM and CmcB samples purified from first soluble fractions, or the two purified EutK fractions were also chromatographed. Species eluting at volumes expected for hexamers were revealed for a majority of cases (Table S5). Exceptions were PduK and EutK, both behaving as monomers, as predicted by ESMFold (AF2 at least failed for EutK). Indeed, EutK eluted at a time expected for a protein approximately half its size (9 kDa) when the first purified fraction was injected, while estimations were closer to a monomer size (17 kDa) for the urea-recovered sample, something unexpected considering their similar SDS-PAGE migration ( Fig. 2 ). Highest intensity peaks with CmcE (both N- or C-ter His 6 -tagged) and PduU occurred at volumes expected for aggregated/assembled species (> 500 kDa), hexamer-compatible peaks being faint. High molecular-weight species were also abundant in EutM, EutS, PduA, and PduK samples, which suggested that urea treatments did not fully dismantle the pelleted macro-assemblies. BMC-H assembly propensity was inspected by TEM after overexpression in E. coli ( Fig. 3 ). Only the three BMC-H with long C-ter tails failed to reveal structured organizations. Four other cases, EutM, PduA, PduJ and CmcC, resulted in mixtures of varied motifs, some compatible with piled sheets (stripped motifs with variable separations in the 3-8 nm range), others with nanotube bundles (16-24 nm interspacing) that resembled motifs reported for PduA from S. enterica or Cit. freundii 53 , 54 . Nanotubes were wider (20-43 nm) for PduJ, which also displayed piled and rolled sheets (8 to 12 nm spacing) and potential hollow spheroids (60-80 nm). Possibly consequence of the lower expression level, EutM nanotubes were less evident than for the other 3 BMC-H ( Fig. 3B ), appearing in the center of cells as short rods. Finally, EutS, PduU and CmcB expressing cells displayed 2D patches that looked like fingerprints: 4-6 nm periodicities for EutS and 5-6 nm for CmcB and PduU. CmcA structures were less clear, possibly combining different kind of motifs. Download figure Open in new tab Figure 3. Electron microscopy of Kpe BMC-H homomers E. coli BL21(DE3) cells over-expressing next cases were imaged by transmission electron microscopy: EutS ( A ), EutM ( B ), PduA ( C ), PduJ ( D ), PduU ( E ), CmcA ( F ), CmcB (G) and CmcC (H). All BMC-H carried C-terminal His 6 tags, except for PduU (N-terminal). The arrows point to motifs interpreted to be nanotubes. Overall, these data indicated that Kpe BMC-H are correctly produced in E. coli and form oligomers that further assemble into higher order supramolecular organizations. Indirectly, the occurrence of organized assemblies supported oligomerization for otherwise doubtful cases like PduU. On the contrary, EutK monomeric behavior and failure to assemble seemed in conflict with a retention in cellular pellets. The apparent discrepancy might stem from interactions with other sedimented components, such as DNA. The tripartite GFP as tool to study BMC-H oligomerization Our major purpose was to characterize potential interactions between Kpe BMC-H, as a means to anticipate the occurrence of non-canonical hetero-combinations in the natural host, and, by extension, in other microorganisms harboring multiple BMC-H paralogs. Our screen was based on the tGFP technology 40 , since amenable for high-throughput screening of protein interactions and well-suited for the study of BMC shell proteins 39 . All possible paired combinations between the 11 Kpe BMC-H were covered. Briefly, necessary DNA fragments coding for a first BMC-H in fusion to the GFP10 β-strand and a second BMC-H connected to the GFP11 β-strand were mounted in a pET-based expression vector that also coded for the remaining part of GFP (GFP1-9 fragment) (Fig. S3A). Subsequent GFP reconstitution and fluorescence signal boosting is expected if the two proteins associate (Fig. S3B). Bearing in mind that tags often impact expression levels and could interfere with interactions, all the 8 possible tagging attribution/orientations were considered: both GFP10 and GFP11 in fusion to each protein C-terminus (C/C), N-terminus (N/N) or the two other combinations (C/N or N/C). This led to a total of 484 plasmid constructs. Forty-four were for combinations of GFP10 and GFP11-fusions of the same BMC-H, each one with the 4 different tag orientations (11 homo-pairs). The remaining 440 cases covered all 55 possible combinations of different BMC-H (hetero-pairs), each case comporting 8 combinations that included different tag orientations and exchange of the identity of the GFP10- and GFP11-carrying BMC-H. Maximal fluorescence readings (F max ) averaged over two independent experiments (comprising > 6 clones) are represented in Figure 4 (numerical values with standard deviations, as well as the estimated times of half-fluorescence are provided in supplementary File S1). Download figure Open in new tab Figure 4. Relative F measured in tGFP assays for Kpe BMC-H pairs Screens consisted in a combined expression of a first BMC-H partner in fusion to the GFP10 (indicated horizontally at the top of the table) and a second BMC-H tagged with the GFP11 (vertical labels), together with the GFP1-9 necessary for fluorescence recons itution. Tags were at the N-terminus of BMC-H (indicated by the asterisk), or at C-terminus (all others). F max values expressed as percentages of the RMM/RMM reference case (100%) are represented following the color scale shown on the right. Each represented value corresponds to the average of at least 6 measurements, generated in 2 independent experiments. NM: not measured, as the corresponding vector could not be prepared. The combinations of homo-pairs were first considered (diagonal of Fig. 4 ), as fluorescence readings could be confronted with data presented above. To facilitate the visualization of effects of tag orientation, F max were also plotted ( Fig. 5A and S4). Globally speaking, the highest signals were detected with C/C orientations. Within this configuration, canonical BMC-H resulted in comparable (CmcC and EutM) or considerably higher fluorescence (CmcA, CmcB, PduA and PduJ) than the reference case, which combined two copies of RMM with GFP10 and GFP11 tags in C/C orientation. Similar values were also recorded for CmcE. On the contrary, PduU fluorescence was under the negative signal threshold established by two controls that corresponded to homo-pair combinations of either CcmK3 from Syn6803 or BWI. F max values with EutK, EutS and PduK fell between those for RMM and negative controls. The introduction of N-tags progressively induced a decline of fluorescence with all BMC-H (Fig. S4), suggesting that either expression levels or the screened interaction might be impacted by the tag. Download figure Open in new tab Figure 5. Monitoring interactions between Kpe BMC-H couples A – tGFP reconstitution fluorescence signals were recorded for the indicated Kpe BMC-H, constructed as homo-pairs carrying GFP10 and GFP11 tags on the C-termini (C/C). Data recorded in similar experiments for other tag configurations are presented in Fig. S4. B – Protein-protein interactions between BMC-H from the same BMC. Studied hetero-pair combinations are indicated using a BMC-H1/BMC-H2 format below the X-axis. F max values plotted here re from screens of either BMC-H1-GFP10/BMC-H2-GFP11 (dark grey) or BMC-H2-GFP10/BMC-H1-GFP11 (light grey) combinations. For the two panels, maxi al fluorescence signals (F max ) are expressed as percentage of the value measured for the RMM homo-pair with C/C orientation (dashed line). The dotted line re measured for negative-cases based on hom from Syn6803 or of BWI. resents the highest threshold signal -pair combinations of either CcmK3 To better understand these data, the expression and solubility of individual BMC-H with GFP10 or GFP11 tags at N- or C-terminus were monitored (Fig. S5). Excluding CmcC, BMC-H band intensities in cellular and soluble fractions looked similar, indicating that GFP10/GFP11 constructs might be less prone to form macro-assemblies than His 6 -tagged constructs. N-ter tagging frequently led to reduced expression levels, especially with the GFP11. In spite of such differences, weak F max readouts with EutK, EutS and PduU could not be justified by weak protein levels, instead supporting oligomerization failure. Of note, smears of lower sized bands were noticed for EutS, PduU, and PduK samples, suggesting proteolysis. With the latter, this might reveal auto-proteolytic damage caused by the predicted Fe-S cluster-harboring domain. On the opposite side, CmcE oligomerization was supported by the comparatively high F max values despite faint protein levels. In summary, the tGFP screen correctly informed on homo-oligomer formation by CmcA, CmcB, CmcC, CmcE, EutM, PduA and PduJ, as well as on non-associating cases EutK and PduK, but failed to detect EutS and PduU oligomerization characterized here for the two His 6 -tagged BMC-H, and reported for other homologs. The overall data also indicated that AF2 and ESMFold perform well as predictive tools. The quality metrics and interaction energies of proposed models were consistent with the majority of canonical BMC-H experimental behaviors (Table S3). AF2, but not ESMFold, also succeeded with EutS and PduU. However, predictions for BMC-H with long C-terminal extensions most often seemed to fail, both when using AF2 (EutK, PduK) or ESMFold (CmcE and EutK). Abundant structural compatibilities between BMC-H from the same BMC Fluorescence readings for hetero-pairs that combined BMC-H coming from the same BMC pointed to abundant cross-interactions ( Fig. 4 & 5B ). Structural promiscuity would be higher within the GRM2 and PDU shells than between EUT BMC-H, as can be visually inferred from data represented in Fig. 4 . EUT BMC-H combinations were also more sensitive to tag orientations. Some heteromeric associations could be expected on the basis of a high sequence resemblance. Our data supported that this happened for combinations of CmcA, CmcB and CmcC homologs, also between PduA and PduJ. Some other pinpointed potential interactions were more surprising. For instance, PduK (N-oriented) resulted in high F max values when combined to PduA or PduJ. In fact, according to tGFP data, PduA would be able to associate with all homologs from the PDU1A, in agreement with a proposed role as protein interaction hub 13 . Less clear, PduK/PduU led to positive signals in only one of the eight screened configurations. Concerning the GRM2 group, a noticeable result was the strong F max augmentation for CmcE combinations with CmcA or CmcB, when compared to the CmcE homo-pair, something that was reproduced in 6 of the 8 configurations of each hetero-pair. In fact, all GRM2 BMC-H combinations conducted to strong fluorescence, hinting at a remarkable promiscuity and raising the question of whether oligomers composed by 3 or 4 BMC-H homologs might occur in GRM2 shells. Also unanticipated, EutK/EutS fluorescence (C/C orientation) was considerably stronger than when screening the respective homo-pairs. Something similar was obtained for EutS/EutM combinations, though the result was strongly impacted by the inversion of the BMC-H in fusion to GFP10 or GFP11. Many of the hetero-associations pointed out by the tGFP data appeared to be supported by AF2 and ESMFold, which globally concurred to similar results (Table S6). Thus, associations with alternating monomers (organization mode ABABABA, A and B representing the two monomer types) were invariantly recovered as top ranked models, with the exception of all combinations including EutS or PduU (excluding AF2-prediction for PduK/PduU), which resulted in models where each type of monomer clustered together (AAABBB), overall hinting at a lack of cross-interaction for these cases that is partial disagreement with tGFP measurements. Thus, the two approaches validated an structural viability for pairs of canonical BMC-H (CmcA/CmcB, CmcA/CmcC, CmcB/CmcC, and PduA/PduJ), which resulted in quality metrics and binding energy scores (ΔE) similar to values obtained for a positive control based on the combination between CcmK1 and CcmK2 from Syn6803 (Table S6) or to estimations based on crystal structure models (Table S4). ESMFold also supported interactions implying BMC-Hs from the first and second groups (CmcA/CmcE, CmcB/CmcE, CmcC/CmcE, EutK/EutM, PduA/PduK and PduJ/PduK), which, excluding EutK/EutM, caused high tGFP F max readings. Conversely, core_PAEs associated to similar ABABAB predictions with AF2 remained exceedingly high [please note that the quality metrics scores seem to differ depending on the method, as higher (better) pLDDT values in AF2 predictions were not often accompanied by lower (better) PAE values, when compared to ESMFold metrics, altogether arguing for a need to apply different thresholds depending on the algorithm]. Interestingly, the two algorithms suggested that formation of hetero-hexamers between CmcE and CmcA/B/C monomers might be reinforced by a disulfide bridge between the CmcE Cys80 and the Cys24 of the other homologs (Fig. S6). Co-purification experiments validate tGFP-unveiled cross-interactions Like any other approach intended for mapping PPI interactions 55 , the tGFP was presumed to rise both false-positive and false-negative readouts. In this section, we verified some selected combinations using a different biochemical approach, similar to the one applied in previous studies 29 , 30 . Briefly, the first BMC-H was connected to a FLAG peptide, while the potential partner was His 6 -tagged. The FLAG construct should co-purify in association with the His 6 -tagged protein in cobalt-based affinity resins provided that their interaction/association is strong enough as to endure necessary treatments and column washing steps. The next cases were selected among those revealing a potential positive interaction in the tGFP assay: EutK/EutS, EutM/EutS, PduA/PduJ, PduJ/PduK, PduU/PduK, CmcA/CmcC, CmcE/CmcA and CmcB/CmcE. Two presumed non-interacting cases, EutK/EutM and PduU/PduJ, were also included. Good expression in E. coli and solubility of the two partners occurred for the majority of combinations ( Fig. 6A ). The presence of two bands in purified fractions directly supported cross-associations between PduA/PduJ, PduJ/PduK, CmcE/CmcA and CmcB/CmcE, something confirmed by western blots (WB) using a FLAG-specific antibody ( Fig. 6B ). WB analysis also permitted to confirm the CmcA/CmcC interaction. Less clear-cut was EutS/EutM, which bands could not be resolved and WB pointed to low level of co-purification of EutS-FLAG. Formation of EutK/EutS was not confirmed under this assay. Purification led to an enrichment of EutS-His 6 , whereas EutK-FLAG bands were practically absent, contrasting with an inversed intensity ratio detected in the total cellular or soluble fractions. Indeed, this also happened with EutK/EutM, which was confirmed not to interact, as indicated the tGFP assay. PduU/PduK and PduU/PduJ combinations gave unclear results, observations being strongly influenced by the supplementation of β-mercaptoethanol (βME, 1 mM) prior to purification. Namely, the two partners were clearly visible in purified fractions in the absence of βME (Fig. S7), but were lost under reducing conditions (though still revealing surprisingly strong in WB, Fig. 6B ). Although unspecific binding of PduU driven by cysteine residues was suspected, we could not conclude on whether PduU/PduK or even PduU/PduJ interacted together or not. Download figure Open in new tab Figure 6. Verification of Kpe BMC-H heteromerization A. Characterization by SDS-PAGE of fractions collected from E. coli cells overexpressing indicated Kpe BMC-H couples. The first component is FLAG-tagged (green label), second His 6 -tagged (black). Soluble fractions were supplemented with 1 mM β-mercaptoethanol prior to load into purification resins. Green and black arrows indicate the approximate migration position of each partner, theoretical molecular weights being given at th bottom of the panel, following the same color code. Other details are like in Figure 2 . B. Detection of co-purified FLAG-tagged species by western blot (WB). Two development techniques were applied: based on alkaline phosphatase reaction on (top) or HRP reaction on a chemiluminescent substrate (bottom), as explained in M&M. Comparison of SEC-HPLC elution profiles with those observed for respective homo-pairs directly supported some cross-interactions. Most remarkable, PduJ/PduK eluted with an apparent molecular weight (MW) of 110 kDa, clearly above species characterized for individual partners (Table S5). Something similar occurred for CmcA/CmcE, CmcB/CmcE, CmcA/CmcC and PduA/PduJ. EutK/EutS and EutS/EutM resulted in HPLC profiles with minor peaks detected within the 5-500 kDa resolving range of the SEC-column. Purified samples implying combinations with PduU, and EutK/EutM, gave peaks exclusively at MW higher > 500 kDa. A notable difference between data collected for homo- or heteromers concerned the distribution of BMC-H between soluble and assembled pools. With the exception of PduA/PduJ, the expression of hetero-pairs resulted in materials that remained basically soluble despite high expression levels, contrasting with data for homo-pairs, where a majority of cases was found abundantly retained in the pellet ( Fig. 6 vs Fig. 2 ). PduA/PduJ assembly was confirmed by TEM (Fig. S8). Motifs recapitulated arrangements formed by individual proteins. Nanotube bundles displayed however sharper straight edges than those visualized for individual PduA or PduJ. As a counter-example, CmcA/CmcC overexpressing cells did not reveal organized structures (Fig. S8). The absence of sedimentation for most cases might be considered as symptomatic of the occurrence of cross-associations (or structural interferences), in view of strong assembly tendencies of homomers. It might also denote an impossibility of heteromers to build higher order structures. Structural compatibilities between BMC-H from different BMC types Fluorescence measurements presented in Fig. 4 also suggested that BMC-H from different compartments might cross-interact together. According to tGFP signal levels, 12 out of 16 possible hetero-pair combinations would be positive when combining BMC-H from the GRM2 and PDU1A groups, 10 over 12 for GRM2/EUT1 combinations and 5 over 12 for PDU1A/EUT1 (see also Fig. S9). Again, all cases mixing canonical BMC-H conducted to high F max , once more confirming the high degree of structural promiscuity within this particular BMC-H cluster. A few combinations between canonical proteins and BMC-H with long C-terminal extensions revealed highly fluorescent too ( e.g., CmcB/EutK, CmcA/PduK, CmcE/EutS or CmcE/PduJ). Co-purification experiments with FLAG/His 6 constructs clearly validated two of the three screened interactions within these BMC-H groups (EutM/PduJ and CmcE/PduJ, but not CmcB/EutK) ( Fig. 6 ). The later also eluted in SEC-HPLC experiments as a single peak, attributed to EutK monomers (Table S5). Worth-mentioning, EutS/PduU gave a comparable F max to the RMM reference, which contrasted with readings for EutS or PduU homo-pairs ( Fig. 5A ). However, the interaction remained less certain in co-purification experiments. Despite the two components seemed correctly expressed and soluble, the EutS-FLAG and His 6 -PduU partners eluted together only in the absence of βME ( Fig. 6 and S7). Nonetheless, FLAG peptides were detected with significant intensities in WB for the two experiments. Irrespective of βME presence, only high MW species were detected in SEC-HPLC, resembling PduU chromatograms (Table S5). AF2 and ESMFold predicted ABABAB hetero-hexamers for all 11 combinations between canonical BMC-H, in excellent agreement with empirical observations. ABABAB models were also generated for 15 remaining possible combinations between members of the two first BMC-H groups (Tables S7A-S7B), rising a notable discordance for weakly fluorescent combinations EutK/PduA, EutK/PduK, EutM/PduK, and CmcC/PduK (Fig. S9). Again, quality scores accompanying some of the AF2-predicted ABABAB models revealed unsatisfactory, and we privileged ESMFold predictions. The 14 heteromeric combinations involving a BMC-H with an extended N-ter region were either not predicted to form PF00936 hexamers or were organized as AAABBB hexamers. The only exception was the combination of the two (EutS/PduU), which led to an atypical ABBAAB model proposition by ESMFold, or to an ABABAB with AF2 displaying good quality metrics. These results are not consistent with the moderate to good fluorescence measurements for 5 out of the 14 combinations. Despite noticed discrepancies depending on the assay, the overall conclusion is that cross-interactions between homologs from different BMC are structurally feasible. Cross-linking data support presence of hetero-hexamers in BMC shells Abovementioned results disclosed a strong degree of structural compatibility between Kpe BMC-H homologs, extending to catabolic BMC the conclusions established before for carboxysomal components 29 , 30 . All these studies have been carried out recombinantly. It becomes now mandatory to ascertain whether such mixed associations are integrated in BMC shells in natural contexts. In that respect, a valuable source of information was found within mass spectrometry data reported by Trettel et al , which we reinterpreted here, in the light of our results 41 . In this study, S. enterica PDU microcompartments were purified from E. coli , after transfer of the whole operon. Lysine residues lying sufficiently close in the shell context were then identified using bifunctional N-hydroxysuccinimide (NHS) crosslinkers coupled to mass spectrometry analysis. Among the many contacts detected in this impressive study, those implicating PduA Lys90 (PduA 90 ) and either PduJ 36 or PduJ 89 called our attention. As the authors recognized, such reactions were unanticipated, since K90-K36 and K90-K89 couples would be too distant if PduA and PduJ assembled together as adjacent homo-hexamers ( ca. 34 and 43 Å, respectively, Fig. 7A ), since the two reactive NHS moieties of the cross-linker were connected by a 12.5-Å long arm. To explain these reactions, authors speculated on the possibility that the short PduA C-terminus was floopy, supposedly revolving towards the (PduJ) neighboring hexamer. The high mobility in molecular dynamics (MD) simulations of the C-terminal portion of CcmK2 proteins 56 , and the detection of a reaction between PduA 12 and PduA 90 , which lie at opposite sides of the hexamer, were presented in support of the argumentation. Download figure Open in new tab Figure 7. Cross-linking mass spectrometry data support PduA/PduJ hetero-hexamerization A. Localization of cross-linked residues in a hypothetical assembly formed by homo-hexamers of PduA (green) and PduJ (pink, violet). Only one PduA and wo PduJ neighboring monomers are shown in cartoon representation. Lysine sidechains are depicted as sticks, blue balls indicating the position of their reactive amino group. B. Similar representation to illustrate an hypothetical PduA/PduJ hetero-hexamer, as predicted by AF2. Measured distances are compatible with the detection by mass spectrometry of PduA 90 /PduJ 36 and PduA 90 /PduJ 89 crosslinked peptides. Regardless of whether homo or hetero-hexamer formed, we propose that the reaction between the K12/K90 couple of PduA might imply the transfer of a NHS moiety through the central pore, rather tha edge. a long-distance displacement towards the hexamer edge. From our point of view, it seems unlikely that PduA 90 (and the preceding region) could be as flexible as to cross the estimated > 50 Å-length required to cross-react with the K12 primary amine. The probe would also have to trans-pierce the BMC shell. Undetected reactions with K55 and K72, which lie closer to the edge than K12 on the convex face, would be expected too. Several additional arguments might contribute to discredit Trettel’s et al argumentation (see Fig. S10), including the analysis of local movements during MD simulations on S. enterica PduA trihexameric assesmblies, which were performed before with a different purpose 57 . Instead, reported cross-links could be better justified within a framework with PduA/PduJ hetero-hexamers composing PDU shells. As illustrated in Fig. 7B , distances of 13 to 17 Å would be fully compatible with the cross-linker probe radius, not only for PduA 86 -PduJ 89 or PduA 90 -PduJ 89 , but most notably for PduA 90 -PduJ 36 . As explanation to the PduA 12 -PduA 90 reaction, we propose that one of the NHS moieties of the cross-linker could be transferred through the central pore. However, this is independent of whether homo- or hetero-hexamers occurred. In the same study, cross-links were also identified between PduJ and PduK or PduA and PduK residues. Such reactions were dominated by K144 or K156 that compose/surround the predicted Fe-S cluster-harboring domain at the end of PduK C-terminal extension. The only mapped PduK-PduK cross-reaction also implicated these two residues. Unfortunately, the preceding portion of this long tail is predicted fully unstructured, thus precluding structural interpretations. In spite of that, the lack of identified reactions on any of the 6 other PduK lysines, some of them corresponding to positions that were reactive in PduA or PduJ, might support an absence of PduK homo-oligomers in BMC shells, indirectly suggesting an integration in the form of hetero-associations. DISCUSSION Understanding BMC assembly and function is fundamental. This particular natural mode of organization contributes to essential biological processes like carbon fixation. Besides, BMC provide new catabolic opportunities to about 20% of bacterial species 11 . Genomic surveys evidenced the existence of a great variety of BMC-mediated processes, something of special interest for synthetic biology purposes. A hallmark of catabolic BMC is the encapsulation of pathways that imply volatile or toxic (aldehyde) intermediates. Accordingly, the major role attributed to the shell is that of a chemical insulator 58 . In this respect, a few experimental studies indicated that shell permeability is defined by the distribution of amino acid residues that surround the central pores of shell constituents 31 , 32 , 59 , something supported also by theoretical simulations 33 , 56 . A good example is the Ser40His mutation of S. enterica PduA, which increased the apparent sensitivity of the encapsulated diol dehydratase to inhibition by glycerol, when compared to wild-type BMC 32 , which evidenced a more permissive access of glycerol to the BMC lumen. This mutation was selected for being present in a PduA subgroup from PDU compartments that are capable of processing glycerol as alternative substrate. These studies prove the importance of characterizing hetero-association occurrence in BMC shells, as their physicochemical properties will differ from those of homomers. Several lines of evidence suggested that hybrid subunits might assemble in BMC shells. On one hand, the multiplicity of shell subunit paralogs inventoried for BMC operons. Thus, about 78% of the 68 BMC types classified by Sutter et al. have several BMC-H 11 . On average, 2.6 BMC-H copies were found per BMC locus , 2.2 after excluding those with permuted secondary structure elements. Furthermore, arrangements of gene paralog neighbors are not rare. On the other hand, data presented here for catabolic BMCs, or before for carboxysomal CcmK from different origins 29 , 30 , proved the formation of hybrid subunits when different BMC-H were simultaneously expressed. Although these studies were performed in E. coli , and varied mechanisms could mitigate the extent of hetero-association occurrence in the natural context (see below), demonstrated structural compatibilities hold true. Results from tGFP fluorescence assays presented here consistently revealed a strong structural promiscuity between BMC-H. Although not all interactions were confirmed in co-purification experiments, a quality indicator was that strongest F max values and lowest sensitivity to tagging schemes occurred for combinations between canonical BMC-H. Strong signals also manifested when canonical and C-terminal extended BMC-H were co-expressed together, in agreement with expectations drawn from sequence identities. Hetero-hexamers consisting of combinations between CmcA/CmcB/CmcC/CmcE or between PduA/PduJ/PduK were clearly pinpointed by the tGFP screen, in accordance with the structural models predicted by ESMFold or AF2. Some of these combinations were also confirmed to co-purify. Corresponding associations behaved in solution much like homo-hexamers, although basically remaining in soluble fractions after cellular lysis, pointing to weaker macro-assembly trends. Experimental results also supported structural compatibilities between homologs from different BMC types, such as EutM/PduJ or combinations between CmcA, CmcB, CmcC or CmcE with PduA or PduJ, conclusions that were also inferred by the two AI-predictive approaches. Indeed, a remarkable resemblance emerged when considering such predictions (Tables SYY and SS). Thus, almost invariably, combinations considered directly by ESMFold as incompatible interactors, or leading to AAABBB clusters, also led to AAABBB associations with AF2. In that manner, the two algorithms ruled out any entity implying EutS or PduU. These combinations frequently resulted in low fluorescence readings or signals below negative thresholds for a majority of the 8 screened bicistronic configurations tested. Moreover, none of the 6 selected combinations that included PduU and/or EutS convincingly demonstrated a co-purification of the two partners. In several occasions, however, structural propositions seemed erroneous, like for instance for Syn6803 CcmK3 or Kpe EutK, considered to form homo-hexamers by AF2 or by both AF2 and ESMFold, respectively, or for Kpe EutS, which according to ESMFold would not associate into homo-hexamers. In spite of demonstrated structural compatibilities, several phenomena might prevent hetero-association occurrence in physiological contexts. The first one operates at the transcriptional level, and is presumed to be especially important in preventing the occurrence of chimeric and/or aberrant BMCs constituted by mixtures of components from different BMC types. Indeed, several studies not only proved that assembly of chimeric BMC is feasible 60 , 61 , but also that properties like shell permeability, and consequently metabolic function, would be altered 42 , 59 . Regions inside operons might also be subject to regulation, for instance by means of 2-component regulatory sensors and/or riboswitches like those controlling eut and pdu expression in E. faecalis 62 or L. monocytogenes 44 . According to RT-qPCR data, only grm2 genes were activated among all three BMCs in the presence of their cognate substrate. To a lower extent, CL also induced the transcription of pdu and eutM regions. To our surprise, EA and PD were without effect on their respective metabolomes. Transcription changes only manifested when combined with CL. On one hand, EA repressed considerably the CL-activation of grm2, resembling the repression by PD on EA-activated eut transcription (and concomitant pdu activation) in S. enterica that was demonstrated to prevent deleterious effects caused by the appearance of hybrid BMCs 42 . On the other hand, PD combined with CL activated Kpe pdu transcription, while still maintaining relatively high grm2 levels. Accordingly, GRM2/PDU hybrid structures might occur. This could signify that co-expression of some compartments is less detrimental than presumed, at least in some organisms. Bearing in mind that intestinal CL basically derives from our diet 63 , whereas EA or PD might be more abundantly released from cellular membranes during inflammation 64 , 65 , we hypothesize that these data could fit with a constitutive expression of Kpe grm2 in the gut, whereas the PDU or EUT would provide novel metabolic opportunities to the organism under inflammatory contexts. As to why each metabolite needs to be combined with CL to exert an effect, it could be part of an environmental sensing mechanism, as apart from being a human commensal, Kpe colonizes other organs. At a second level, structural pairing efficiency might be decoupled as a consequence of the chromosomal separation between paralogs. If this parameter was determinant, only demonstrated cross-associations between CmcA/CmcB/CmcC or between PduJ/PduK might be physiologically meaningful in Kpe , as other BMC-H genes are distant from any other paralog. Yet, other players might be the presence of structured mRNA regions or the density of ribosomes simultaneously acting on each coding segment 66 . At the protein level, formation of hetero-hexamers might not necessarily guarantee an integration within BMC shells. In fact, our data pointed to a decline of assembly trends for BMC-H hetero-associations, when compared to homo-hexamers. This could reflect the transmission of small structural deviations towards the hexamer edges, which would be caused by a slightly altered packing of residues at the interface between monomers. In spite of all mentioned mechanisms, compelling arguments support that hybrid compartments may form. We showed that transcription of Kpe BMC operons might be less tightly regulated than in S. enterica . Similarly, bile salts were reported to augment transcription of both pdu and eut operons in adherent-invasive E coli strains 16 . The combination of EA with Vit-B also increased eut and pdu readouts in commensal L. brevis ATCC 14869 18 . BMC structures formed in vivo , although RT-qPCR readings suggested that shells might be basically PDU-like. The characterization of hybrid PDU/EUT compartments when natural transcriptional regulation was bypassed in Salmonella would also suggest that the second and third levels of control might be inoperative 42 . Yet, hybrid compartments might integrate only homomeric subunits from each BMC type. In fact, the most straightforward argument supporting an incorporation of hetero-associations within BMC shells was provided by MS cross-linking data collected for purified S. enterica PDU compartments that were produced in an E. coli strain harboring the full pdu1a operon 41 . We demonstrated that signals implicating cross-reactions between several couples of PduA and PduJ lysines could be easily justified in the context of a S. enterica PduA/PduJ mixed oligomer. In line with this, we observed that Kpe PduA/PduJ was hetero-hexameric in solution and continued to assemble inside cells. These data are complementary to the likely observation by cryo-EM of CsoS1A/CsoS1B hetero-hexamers in a study of synthetic recombinant α-carboxysomes from H. neapolitanus that was published while submission of this work for publication 67 . In conclusion, our study demonstrates a strong structural compatibility between BMC-H and presents compelling evidence in favor of the occurrence of hetero-associations in BMC shells. Future studies will be necessary to evaluate the extent and physiological importance of this phenomenon in natural contexts. MATERIALS AND METHODS Quantification of RNA transcripts by RT-qPCR Precultures of Klebsiella pneumoniae (strain 342) were prepared from single colonies grown anaerobically in lysogeny broth (LB), ON at 37°C. Final cultures were grown statically in glass tubes (15 × 150 mm) under anaerobic conditions at 37°C for 48 hours by inoculation of precultures (40 µL) in modified no-carbon-E (NCE) medium (10 mL) 68 . This medicum was priorly degassed for 3 days under nitrogen atmosphere in an anaerobic chamber (PureEvo T4, Jacomex). The modified NCE medium was composed of: 0.82 mM MgSO₄, 57.4 mM K₂HPO₄, 16.74 mM Na(NH₄)HPO₄, 175 mM NaCl, 20 mM NH₄Cl, 50 µM ferric citrate and 5 mg/L thiamine. It also included 0.2% (w/v) yeast extract, 0.0015% (w/v) each of valine, isoleucine, threonine, and leucine, 100 nM cyanocobalamin (vit. B 12 , ref V2876), 50 mM disodium fumarate and trace metals (0.3 mM CaCl₂, 0.1 mM ZnSO₄, 0.045 mM FeSO₄, 0.2 mM Na₂Se₂O₃, 0.2 mM Na₂MoO₄, 2 mM MnSO₄, 0.1 mM CuSO₄, 3 mM CoCl₂, and 0.1 mM NiSO₄). When indicated, the next metabolites were added separately or combined (final conc.): EA hydrochloride (30 mM, pH 7.0, Sigma-Aldrich 411000); PD (55 mM, Sigma-Aldrich 82280); or CL chloride (1%, Sigma-Aldrich, C7017). An equivalent of 3 units of OD 620nm were pelleted by centrifugation at 6000 g for 5 min at 4°C. Pellets were resuspended in 0.5 mL of PBS and 1 mL of RNA later ® solution (Thermo Fisher Scientific) to avoid transcriptional changes and RNA degradation. Samples were stored at −80 °C until RNA extraction. Total RNA was extracted as described 69 . The absence of DNA contamination was verified by qPCRs performed with the primer pair qPCR-proC-Fw/qPCR-proC-Rv (Table S8) and the SsoAdvanced SYBR® Green Supermix (Bio-Rad, Hercules, California, USA), according to the manufacturer’s recommendations. Reverse transcription was performed with 500 ng of total RNA using the iScript cDNA Synthesis kit (Bio-Rad) following provided indications. qPCRs were carried out in the CFX96 Real Time System (Bio-Rad) with the SsoAdvanced SYBR® Green Supermix (Bio-Rad) in 10 µL total volume per well with 2 µL of 10X diluted cDNA. Primers targeting different Kpe genomic regions are listed in Table S8. Melting curve analysis was used to verify the specific single-product amplification. The gene expression levels were normalized relative to the expression level of the proC and rpoD housekeeping genes 70 . Relative quantifications were determined with CFX Maestro software (Bio-Rad), following provider’s instructions. The amplification efficiency (E) of each primer pair used for the quantification was calculated from a standard amplification curve obtained by five dilution series of genomic DNA. All assays were performed in technical triplicates with three independently isolated RNA samples. Preparation of plasmid constructs for recombinant studies Nucleotide sequences of the 11 BMC-H of Kpe 342 (GenBank CP000964.1 ) were codon-optimized for E. coli while maintaining codon frequency profiles along the sequence. For studies of expression of individual BMC-H, resulting DNA sequences, connected to segments coding for His 6 tags, were synthesized and provided cloned between NdeI/XhoI sites of a pET29b (Kan R ) vector (Twist). Plasmids permitting expression of His 6 /FLAG BMC-H combinations as bicistrons were also prepared by Twist, cloned similarly in the same vector. However, it was necessary to adapt some codons to permit DNA synthesis. DNA and proteins sequences are compiled in two sheets in an excel file (Table S9). The preparation of the library of BMC-H pairs for tGFP screens was carried out following a described protocol 39 . Each codon-optimized BMC-H sequence was synthetized by Twist as DNA fragments in 4 different conformations: GFP10-linker29-BMC-H (N-10 vector), BMC-H-linker31-GFP10 (C-10 vector), GFP11-linker29-BMC-H (N-11 vector) or BMC-H-linker33-GFP11(C-11 vector), with flanking sequences necessary for individual assembly in a pET26b (Kan R ) vector opened with XbaI/XhoI. After being PCR amplified (see Table S10 for DNA sequences of primers, BMC-H and vector DNA), each couple of fragments was Gibson-assembled into a BglII/HindIII-opened tGFP vector (pET26b, Kan R ) using robotic means provided by a local platform (Toulouse White Biotechnology) (schematized in Fig. S3A). Transformation was performed directly in T7 express competent cells (NEB), following the provider protocol. After robotic plating on LB agar (40 µg.mL -1 Kan), cells were stored for 2 to 3 days at 4°C. Fluorescent clones were then selected on a QPix 460 (Molecular Devices), and the sequence of corresponding plasmid preparation confirmed to be correct. Biochemical characterization of individual BMC-H or combinations After transformation of BL21(DE3) cells with corresponding pET29 vectors, expression of His 6 tagged BMC-H or combinations of His 6 - and FLAG-tagged BMC-H was carried out in ZYM-5052 auto-induction medium (8 mL) overnight at 37°C, following a described protocol 29 . Cells pellets collected by centrifugation (4000 g) were lysed with Lysis solution (1 mL): BugBuster extraction reagent (Sigma) supplemented with lysozyme (0.03 mg/mL final conc.), Benzonase Nuclease (Merck, 268 U/mL), and PMSF (1 mM). For some experiments, βME (0.75 mM) was also included. Small aliquots were withdrawn and prepared for SDS-PAGE analysis (denatured at 95°C), corresponding to cellular fractions. Soluble fractions were prepared similarly from material remaining in supernatants after centrifugation at 21000 g, 4°C for 15 min (pellets were kept at 4°C for studies of protein disassembly). Purification was performed on Vivapure 8 96-well cobalt-chelate microplate columns (VivaScience), following provider instructions. After 4 washes with 500 mL of solution A (20 mM NaPi/300 mM NaCl/10 mM imidazole, pH 8.0), bound proteins were eluted with 300 µL of a 300 mM imidazole in solution A. EDTA (5 mM) and βME (5 mM) were added immediately after elution and purified fraction aliquots prepared for SDS-PAGE analysis. Material remaining in pellets were resuspended in Lysis solution (1mL), sonicated for 20 sec at 4°C and centrifuged at 21000 g. After discarding supernatants, 1 M urea in solution A (1 mL, ± 2.5 mM βME) was added to each pellet. After vigorous resuspension, the solution was shaken for 30 min at 4°C. After 21000 g, urea-solubilized fraction aliquots were prepared for SDS-PAGE as other fractions, and the material was purified as indicated above on microplate columns, leading to urea-purified fractions. Samples were run on 17-18% polyacrylamide gels, which were stained with Coomassie Brilliant Blue R-250 (Bio-Rad). Cellular and soluble fractions corresponding to the analysis of individual GFP10/GFP11 constructs were prepared as indicated above for His 6 -tagged BMC-H. Studies were performed with BL21(DE3) cells transformed with N-10, C-10, N-11 or C-11 vectors. The only difference was that ON cultures were grown in LB (2 mL) supplemented with IPTG (10 µM) from the start. βME was omitted during treatments. Western blots of different fractions were performed as described in 29 , after running samples on 17-18% polyacrylamide gels. For experiment presented in Fig. 6 , an identical second western blot was performed on purified fractions replacing the selected secondary antibody by an antimouse-HRP (Thermo Fisher Scientific, A16066) and the development reactant by a chemiluminescent substrate (TS, ref 34577). Tripartite GFP assay An overnight (ON) preculture (2 μL) of each tGFP vector-transformed T7 express strain was seeded in Luria-Bertani broth (LB, 200 μL, 40 µg.mL -1 Kan), dispensed in a 96-well glass-bottomed black plate (Greiner). IPTG (10 μM, final conc.) was added from the start. Fluorescence signals and optical density at 600nm were monitored continuously on a CLARIOstar Plus (BMC Labtech), as described in 39 . Data were processed with GraphPad Prism 6: the signals of fluorescence obtained were fitted to a sigmoidal function of equation: with t half Fmax the time necessary to reach half of the F max value. The F max values for cases displaying inadequate fits (generally low signal data) were extracted manually. Values were normalized by the F max measured for C-terminally-tagged RMM/RMM in the same assay. Plotted values are the mean F max and ± standard deviations deriving from at least 2 independent repetitions of experiments performed on 3 clones per library member. Size-exclusion HPLC Protein sizes were estimated by SEC using a Beckman Ultraspherogel SEC2000 column (7.5 x 300 mm) mounted on a Waters 2690 HPLC separation module. Purified proteins (100 µL) were dialyzed for 5 h (4°C) in Pur-A-Lyzer dialysis columns (3500 Da cut-off, ThermoFisher), against 500 mL of a 10 mM Tris/200mM NaCl/0.5mM EDTA/pH 7.8 solution. Samples (10-20 µL) were injected at 1 mL/min flowrate after conditioning the column in 20 mM Tris-HCl, 300 mM NaCl at pH 7. Elution was monitored with a Waters 996 Photodiode Array Detector. Elution volumes (280 nm absorption peaks) were used to estimate protein MW by comparison to next calibration standards run under identical conditions: Ferritin (440 kDa), Aldolase (158 kDa), Conalbumin (75 kDa), Ovalbumin (43 kDa) and Ribonuclease (13.7 kDa). Transmission electron microscopy Cells permitting the overexpression of His 6 -tagged BMC-H, in combination or not with a FLAG-fused partner, were precultured ON at 37°C in LB (Kan). Next morning, 40 µL of precultures was seeded in 4 mL of LB (Kan). Induction with 200 µM IPTG was triggered in exponential phase [OD (600 nm) of 0.6]. Cultures were continued for 6 h longer at 37°C (200 rpm). After 1000 g centrifugation, and discarding supernatants, cellular pellets were gently resuspended in 1 mL of fixation mixture: 2,5% glutaraldehyde and 2% paraformaldehyde in cacodylate buffer (0.1 M, pH 7.2). After 15 min, the cells were sedimented again (800 g) and the pellet resuspended in 2 mL of the fixation mixture, which was kept at 4°C ON. Next morning, the cells were washed in 3 cycles of pelleting (800 g) and gentle resuspension in 1 mL of cacodylate buffer. Fixed cells were post-fixed with 1% OsO 4 in cadodylate buffer. Three washings were performed before inlaying the cells in 2% low-melting point agarose. Then, samples were treated with 1% uranyl acetate for 1h. They were dehydrated using an ethanol gradient: 25, 50, 70 and 90% for 15min, plus 3 times 30min at 100%. They were then transferred in Epon resin baths (Embed 812, EMS) of increasing concentration (25, 50, 75% Epon in ethanol for 1h and twice 2h in 100% Epon at 37°C). Finally, they were embedded in Epon resin by a 48h polymerization at 60°C. An ultramicrotome UCT (Leica) was used in order to prepare 80nm-thick sections of the embedded cells. The sections were then mounted onto formvar/carbon-coated copper grids of 200-mesh and stained with Uranyless ( EM-grade.com ) and Reynolds lead citrate 3% ( EM-grade.com ). TEM acquisitions were made on a JEM-1400 electron microscope (JEOL Inc, Peabody, MA, USA) operating at 80 kV, equipped with Gatan Orius or Rio 9 digital cameras (Gatan Inc, Pleasanton, CA, USA). AF2 and ESMFold predictions 3D structures of homohexamers and heterohexamers formed by combinations of BMC-H proteins were predicted using two AI-based protein structure prediction tools: the ColabFold implementation of AlphaFold2-Multimer (AF2) version 3 and ESMFold version 1 50 , 51 . For all homomeric and heteromeric combinations, the standard AF2 mode in ColabFold, which relies on multiple sequence alignments (MSAs), was used. For each combination, 20 structural models were generated and the top-ranked model was selected for further analysis. Additionally, for homomeric assemblies, the single-sequence mode of AF2, which does not use MSAs, was explored. In this mode, 5 models were generated for each homomeric assembly prediction, and the top-ranked model was retained. ESMFold, which does not rely on MSAs but predicts proteins directly from primary sequences using a transformer-based language model (ESM-2) was used to generate a structural model for each homomeric and heteromeric assembly. For all structural models, the fold, hexameric geometry, and, for heteromeric assemblies, the organization between the two types of monomers were examined using Pymol. In addition to common protein prediction quality metrics provided by AlphaFold and ESMFold such as the predicted local difference distance test (pLDDT, reflecting local structural accuracy), specific scores dedicated to multimeric assemblies were considered: the average predicted aligned error of all interchain residue pairs (interchain_PAE, usually denoted iPAE), the average predicted aligned error of interface residue pairs within a 4 Å distance cutoff (interface_PAE) and the average predicted local distance difference test of interface residues within a 4 Å distance cutoff (interface_pLDDT). A high-confidence model is characterized by low interface_PAE and interchain_PAE values and high interface_pLDDT values. In addition, for BMC-H with C-ter or N-ter extensions, the interchain_PAE was also computed by excluding these regions and named core_PAE. Relaxation of the structural models was performed using the AMBER protocol from ColabFold followed by the Rosetta FastRelax protocol 71 , based on the beta_nov16 energy function, to resolve unfavorable local geometries or clashes. Following relaxation, binding energy scores (ΔE) were computed using Rosetta’s InterfaceAnalyzer. SUPPORTING INFORMATION Supplementary figures S1 to S10, Tables S1 to S8 are provided, as well as two excel files with all DNA and protein sequences. A supplementary file S1 is also given which contains all averaged F max values presented in Fig. 4 , as well accompanying standard deviations and culture times to reach half of F max fluorescence. NOTES The authors declare no conflict of interest ACKNOWLEDGEMENTS The French National Research Agency (ANR) financially supported this work: ANR-19-CE09-0032-01. This work also benefited from a grant managed by the same agency, under the “Investissements d’Avenir” programme: ANR-18-EURE-0021. This work was granted access to the HPC resources of CALMIP supercomputing center. We thank Prof. Thomas A. Bobik for helpful discussions concerning Klebsiella pneumoniae culturing conditions. REFERENCES 1. ↵ Dou , Z. et al. CO2 Fixation Kinetics of Halothiobacillus neapolitanus Mutant Carboxysomes Lacking Carbonic Anhydrase Suggest the Shell Acts as a Diffusional Barrier for CO2 . J. Biol. Chem . 283 , 10377 – 10384 ( 2008 ). OpenUrl Abstract / FREE Full Text 2. Iancu , C. V. et al. Organization, Structure, and Assembly of α-Carboxysomes Determined by Electron Cryotomography of Intact Cells . J. Mol. Biol . 396 , 105 – 117 ( 2010 ). OpenUrl CrossRef PubMed 3. Iancu , C. V. et al. The Structure of Isolated Synechococcus Strain WH8102 Carboxysomes as Revealed by Electron Cryotomography . J. Mol. Biol . 372 , 764 – 773 ( 2007 ). OpenUrl CrossRef PubMed 4. ↵ Schmid , M. F. et al. Structure of Halothiobacillus neapolitanus Carboxysomes by Cryo-electron Tomography . J. Mol. Biol . 364 , 526 – 535 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 5. ↵ de Araujo , C. et al. Identification and characterization of a carboxysomal γ-carbonic anhydrase from the cyanobacterium Nostoc sp. PCC 7120 . Photosynth. Res . 121 , 135 – 150 ( 2014 ). OpenUrl CrossRef PubMed 6. ↵ Lassila , J. K. , Bernstein , S. L. , Kinney , J. N. , Axen , S. D. & Kerfeld , C. A . Assembly of Robust Bacterial Microcompartment Shells Using Building Blocks from an Organelle of Unknown Function . J. Mol. Biol . 426 , 2217 – 2228 ( 2014 ). OpenUrl CrossRef PubMed 7. ↵ Chowdhury , C. , Sinha , S. , Chun , S. , Yeates , T. O. & Bobik , T. A . Diverse Bacterial Microcompartment Organelles . Microbiol. Mol. Biol. Rev . 78 , 438 – 468 ( 2014 ). OpenUrl Abstract / FREE Full Text 8. ↵ Rae , B. D. , Long , B. M. , Badger , M. R. & Price , G. D. Functions, Compositions, and Evolution of the Two Types of Carboxysomes: Polyhedral Microcompartments That Facilitate CO 2 Fixation in Cyanobacteria and Some Proteobacteria . Microbiol. Mol. Biol. Rev . 77 , 357 – 379 ( 2013 ). OpenUrl Abstract / FREE Full Text 9. ↵ Doron , L. , Sutter , M. & Kerfeld , C. A . Characterization of a novel aromatic substrate-processing microcompartment in Actinobacteria . mBio e01216 – 23 ( 2023 ) doi: 10.1128/mbio.01216-23 . OpenUrl CrossRef PubMed 10. ↵ Kalnins , G. et al. Structure and Function of CutC Choline Lyase from Human Microbiota Bacterium Klebsiella pneumoniae . J. Biol. Chem . 290 , 21732 – 21740 ( 2015 ). OpenUrl Abstract / FREE Full Text 11. ↵ Sutter , M. , Melnicki , M. R. , Schulz , F. , Woyke , T. & Kerfeld , C. A . A catalog of the diversity and ubiquity of bacterial microcompartments . Nat. Commun . 12 , 3809 ( 2021 ). OpenUrl CrossRef PubMed 12. ↵ Axen , S. D. , Erbilgin , O. & Kerfeld , C. A . A Taxonomy of Bacterial Microcompartment Loci Constructed by a Novel Scoring Method . PLoS Comput. Biol . 10 , e1003898 ( 2014 ). OpenUrl CrossRef PubMed 13. ↵ Jorda , J. , Liu , Y. , Bobik , T. A. & Yeates , T. O . Exploring Bacterial Organelle Interactomes: A Model of the Protein-Protein Interaction Network in the Pdu Microcompartment . PLOS Comput. Biol . 11 , e1004067 ( 2015 ). OpenUrl CrossRef PubMed 14. ↵ Prentice , M. B . Bacterial microcompartments and their role in pathogenicity . Curr. Opin. Microbiol . 63 , 19 – 28 ( 2021 ). OpenUrl CrossRef PubMed 15. ↵ Barnes , A. J. et al. Ethanolamine metabolism through two genetically distinct loci enables Klebsiella pneumoniae to bypass nutritional competition in the gut . PLOS Pathog . 20 , e1012189 ( 2024 ). OpenUrl CrossRef PubMed 16. ↵ Delmas , J. et al. Metabolic adaptation of adherent-invasive Escherichia coli to exposure to bile salts . Sci. Rep . 9 , 2175 ( 2019 ). OpenUrl CrossRef PubMed 17. ↵ Dogan , B. et al. Inflammation-associated Adherent-invasive Escherichia coli Are Enriched in Pathways for Use of Propanediol and Iron and M-cell Translocation: Inflamm . Bowel Dis . 20 , 1919 – 1932 ( 2014 ). OpenUrl CrossRef 18. ↵ Akouris , P. P. et al. Ethanolamine enhances adhesion, promotes microcompartment formation, and modulates gene expression in Levilactobacillus brevis ATCC 14869 . Gut Microbes 16 , 2350778 ( 2024 ). 19. ↵ Yang , M. et al. Decoding the stoichiometric composition and organisation of bacterial metabolosomes . Nat. Commun . 11 , 1976 ( 2020 ). OpenUrl CrossRef PubMed 20. ↵ Klein , M. G. et al. Identification and Structural Analysis of a Novel Carboxysome Shell Protein with Implications for Metabolite Transport . J. Mol. Biol . 392 , 319 – 333 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 21. ↵ Sagermann , M. , Ohtaki , A. & Nikolakakis , K . Crystal structure of the EutL shell protein of the ethanolamine ammonia lyase microcompartment . Proc. Natl. Acad. Sci . 106 , 8883 – 8887 ( 2009 ). OpenUrl Abstract / FREE Full Text 22. ↵ Ochoa , J. M. & Yeates , T. O . Recent structural insights into bacterial microcompartment shells . Curr. Opin. Microbiol . 62 , 51 – 60 ( 2021 ). OpenUrl CrossRef PubMed 23. ↵ Kerfeld , C. A. , Aussignargues , C. , Zarzycki , J. , Cai , F. & Sutter , M. Bacterial microcompartments . Nat. Rev. Microbiol . 16 , 277 – 290 ( 2018 ). OpenUrl CrossRef PubMed 24. Kerfeld , C. A. et al. Protein Structures Forming the Shell of Primitive Bacterial Organelles . Science 309 , 936 – 938 ( 2005 ). OpenUrl Abstract / FREE Full Text 25. ↵ Mills , C. E. et al. Vertex protein PduN tunes encapsulated pathway performance by dictating bacterial metabolosome morphology . Nat. Commun . 13 , 3746 ( 2022 ). OpenUrl CrossRef PubMed 26. ↵ Sutter , M. , Greber , B. , Aussignargues , C. & Kerfeld , C. A . Assembly principles and structure of a 6.5-MDa bacterial microcompartment shell . Science 356 , 1293 – 1297 ( 2017 ). OpenUrl Abstract / FREE Full Text 27. ↵ Kalnins , G. et al. Encapsulation mechanisms and structural studies of GRM2 bacterial microcompartment particles . Nat. Commun . 11 , 388 ( 2020 ). 28. ↵ Zhou , R.-Q. et al. Structure and assembly of the α-carboxysome in the marine cyanobacterium Prochlorococcus . Nat. Plants 10 , 661 – 672 ( 2024 ). OpenUrl CrossRef PubMed 29. ↵ Garcia-Alles , L. F. et al. Occurrence and stability of hetero-hexamer associations formed by β-carboxysome CcmK shell components . PLOS ONE 14 , e0223877 ( 2019 ). OpenUrl CrossRef PubMed 30. ↵ Sommer , M. et al. Heterohexamers Formed by CcmK3 and CcmK4 Increase the Complexity of Beta Carboxysome Shells . Plant Physiol . 179 , 156 – 167 ( 2019 ). OpenUrl Abstract / FREE Full Text 31. ↵ Ochoa , J. M. et al. A single shell protein plays a major role in choline transport across the shell of the choline utilization microcompartment of Escherichia coli 536 . Microbiology 169 , ( 2023 ). 32. ↵ Chowdhury , C. et al. Selective molecular transport through the protein shell of a bacterial microcompartment organelle . Proc. Natl. Acad. Sci . 112 , 2990 – 2995 ( 2015 ). OpenUrl Abstract / FREE Full Text 33. ↵ Mahinthichaichan , P. , Morris , D. M. , Wang , Y. , Jensen , G. J. & Tajkhorshid , E . Selective Permeability of Carboxysome Shell Pores to Anionic Molecules . J. Phys. Chem. B 122 , 9110 – 9118 ( 2018 ). OpenUrl CrossRef PubMed 34. ↵ Fan , C. et al. Short N-terminal sequences package proteins into bacterial microcompartments . Proc. Natl. Acad. Sci . 107 , 7509 – 7514 ( 2010 ). OpenUrl Abstract / FREE Full Text 35. ↵ Kinney , J. N. , Salmeen , A. , Cai , F. & Kerfeld , C. A . Elucidating Essential Role of Conserved Carboxysomal Protein CcmN Reveals Common Feature of Bacterial Microcompartment Assembly . J. Biol. Chem . 287 , 17729 – 17736 ( 2012 ). OpenUrl Abstract / FREE Full Text 36. ↵ Chen , A. H. , Robinson-Mosher , A. , Savage , D. F. , Silver , P. A. & Polka , J. K . The Bacterial Carbon-Fixing Organelle Is Formed by Shell Envelopment of Preassembled Cargo . PLoS ONE 8 , e76127 ( 2013 ). OpenUrl CrossRef PubMed 37. Cameron , J. C. , Wilson , S. C. , Bernstein , S. L. & Kerfeld , C. A . Biogenesis of a Bacterial Organelle: The Carboxysome Assembly Pathway . Cell 155 , 1131 – 1140 ( 2013 ). OpenUrl CrossRef PubMed 38. ↵ Hill , N. C. , Tay , J. W. , Altus , S. , Bortz , D. M. & Cameron , J. C . Life cycle of a cyanobacterial carboxysome . Sci. Adv . 6 , eaba1269 ( 2020 ). OpenUrl FREE Full Text 39. ↵ Barthe , L. , Soldan , V. & Garcia-Alles , L. F . Assessment of oligomerization of bacterial micro-compartment shell components with the tripartite GFP reporter technology . PLOS ONE 18 , e0294760 ( 2023 ). OpenUrl CrossRef PubMed 40. ↵ Cabantous , S. et al. A New Protein-Protein Interaction Sensor Based on Tripartite Split-GFP Association . Sci. Rep . 3 , 2854 ( 2013 ). OpenUrl CrossRef PubMed 41. ↵ Trettel , D. S. , Resager , W. , Ueberheide , B. M. , Jenkins , C. C. & Winkler , W. C . Chemical probing provides insight into the native assembly state of a bacterial microcompartment . Structure 30 , 537 – 550 .e5 ( 2022 ). OpenUrl CrossRef 42. ↵ Sturms , R. , Streauslin , N. A. , Cheng , S. & Bobik , T. A. In Salmonella enterica, Ethanolamine Utilization Is Repressed by 1,2-Propanediol To Prevent Detrimental Mixing of Components of Two Different Bacterial Microcompartments . J. Bacteriol . 197 , 2412 – 2421 ( 2015 ). OpenUrl Abstract / FREE Full Text 43. ↵ Roof , D. M. & Roth , J. R . Autogenous Regulation of Ethanolamine Utilization by a Transcriptional Activator of the eut Operon in Salmonella typhimunium . J BACTERIOL 174 , ( 1992 ). 44. ↵ Mellin , J. R. et al. A riboswitch-regulated antisense RNA in Listeria monocytogenes . Proc. Natl. Acad. Sci . 110 , 13132 – 13137 ( 2013 ). OpenUrl Abstract / FREE Full Text 45. ↵ Martínez-del Campo , A. , et al. Characterization and Detection of a Widely Distributed Gene Cluster That Predicts Anaerobic Choline Utilization by Human Gut Bacteria . mBio 6 , e00042 – 15 ( 2015 ). OpenUrl CrossRef PubMed 46. ↵ Ferlez , B. , Sutter , M. & Kerfeld , C. A . A designed bacterial microcompartment shell with tunable composition and precision cargo loading . Metab. Eng . 54 , 286 – 291 ( 2019 ). OpenUrl CrossRef PubMed 47. ↵ Tanaka , S. , Sawaya , M. R. & Yeates , T. O . Structure and Mechanisms of a Protein-Based Organelle in Escherichia coli . Science 327 , 81 – 84 ( 2010 ). OpenUrl Abstract / FREE Full Text 48. Pitts , A. C. , Tuck , L. R. , Faulds-Pain , A. , Lewis , R. J. & Marles-Wright , J . Structural Insight into the Clostridium difficile Ethanolamine Utilisation Microcompartment . PLoS ONE 7 , e48360 ( 2012 ). OpenUrl CrossRef PubMed 49. ↵ Crowley , C. S. , Sawaya , M. R. , Bobik , T. A. & Yeates , T. O . Structure of the PduU Shell Protein from the Pdu Microcompartment of Salmonella . Structure 16 , 1324 – 1332 ( 2008 ). OpenUrl CrossRef PubMed 50. ↵ Mirdita , M. et al. ColabFold: making protein folding accessible to all . Nat. Methods 19 , 679 – 682 ( 2022 ). OpenUrl CrossRef PubMed 51. ↵ Lin , Z. et al. Evolutionary-scale prediction of atomic-level protein structure with a language model . Science 379 , 1123 – 1130 ( 2023 ). OpenUrl CrossRef PubMed 52. ↵ Crowley , C. S. et al. Structural Insight into the Mechanisms of Transport across the Salmonella enterica Pdu Microcompartment Shell . J. Biol. Chem . 285 , 37838 – 37846 ( 2010 ). OpenUrl Abstract / FREE Full Text 53. ↵ Pang , A. , Frank , S. , Brown , I. , Warren , M. J. & Pickersgill , R. W . Structural Insights into Higher Order Assembly and Function of the Bacterial Microcompartment Protein PduA . J. Biol. Chem . 289 , 22377 – 22384 ( 2014 ). OpenUrl Abstract / FREE Full Text 54. ↵ Young , E. J. et al. Engineering the Bacterial Microcompartment Domain for Molecular Scaffolding Applications . Front. Microbiol . 8 , 1441 ( 2017 ). OpenUrl CrossRef PubMed 55. ↵ Choi , S. G. et al. Maximizing binary interactome mapping with a minimal number of assays . Nat. Commun . 10 , 3907 ( 2019 ). OpenUrl CrossRef PubMed 56. ↵ Faulkner , M. et al. Molecular simulations unravel the molecular principles that mediate selective permeability of carboxysome shell protein . Sci. Rep . 10 , 17501 ( 2020 ). 57. ↵ Garcia-Alles , L. F. , Fuentes-Cabrera , M. , Truan , G. & Reguera , D . Inferring assembly-curving trends of bacterial micro-compartment shell hexamers from crystal structure arrangements . PLOS Comput. Biol . 19 , e1011038 ( 2023 ). OpenUrl CrossRef PubMed 58. ↵ Sampson , E. M. & Bobik , T. A . Microcompartments for B12-dependent 1,2-propanediol degradation provide protection from DNA and cellular damage by a reactive metabolic intermediate . J. Bacteriol . 190 , 2966 – 2971 ( 2008 ). OpenUrl Abstract / FREE Full Text 59. ↵ Slininger Lee , M. F. , Jakobson , C. M. & Tullman-Ercek , D. Evidence for Improved Encapsulated Pathway Behavior in a Bacterial Microcompartment through Shell Protein Engineering . ACS Synth. Biol . 6 , 1880 – 1891 ( 2017 ). OpenUrl CrossRef PubMed 60. ↵ Cai , F. , Sutter , M. , Bernstein , S. L. , Kinney , J. N. & Kerfeld , C. A . Engineering Bacterial Microcompartment Shells: Chimeric Shell Proteins and Chimeric Carboxysome Shells . ACS Synth. Biol . 4 , 444 – 453 ( 2015 ). OpenUrl CrossRef PubMed 61. ↵ Fang , Y. et al. Engineering and Modulating Functional Cyanobacterial CO2-Fixing Organelles . Front. Plant Sci . 9 , 739 ( 2018 ). 62. ↵ Fox , K. A. et al. Multiple posttranscriptional regulatory mechanisms partner to control ethanolamine utilization in Enterococcus faecalis . Proc. Natl. Acad. Sci . 106 , 4435 – 4440 ( 2009 ). OpenUrl Abstract / FREE Full Text 63. ↵ Arias , N. et al. The Relationship between Choline Bioavailability from Diet, Intestinal Microbiota Composition, and Its Modulation of Human Diseases . Nutrients 12 , 2340 ( 2020 ). OpenUrl CrossRef PubMed 64. ↵ Mishra , S. P. et al. A mechanism by which gut microbiota elevates permeability and inflammation in obese/diabetic mice and human gut . Gut 72 , 1848 – 1865 ( 2023 ). OpenUrl Abstract / FREE Full Text 65. ↵ Sugihara , K. & Kamada , N . Metabolic network of the gut microbiota in inflammatory bowel disease . Inflamm. Regen . 44 , 11 ( 2024 ). 66. ↵ Bertolini , M. et al. Interactions between nascent proteins translated by adjacent ribosomes drive homomer assembly . Science 371 , 57 – 64 ( 2021 ). OpenUrl Abstract / FREE Full Text 67. ↵ Wang , P. et al. Molecular principles of the assembly and construction of a carboxysome shell . Sci. Adv . 10 , eadr4227 ( 2024 ). OpenUrl CrossRef PubMed 68. ↵ Berkowitz , D. , Hushon , J. M. , Whitfield , H. J. , Roth , J. & Ames , B. N . Procedure for identifying nonsense mutations . J. Bacteriol . 96 , 215 – 220 ( 1968 ). OpenUrl Abstract / FREE Full Text 69. ↵ Dramé , I. et al. FabR, a regulator of membrane lipid homeostasis, is involved in Klebsiella pneumoniae biofilm robustness . mBio 15 , e0131724 ( 2024 ). OpenUrl CrossRef PubMed 70. ↵ Gomes , A. É. I. et al. Selection and validation of reference genes for gene expression studies in Klebsiella pneumoniae using Reverse Transcription Quantitative real-time PCR . Sci. Rep . 8 , 9001 ( 2018 ). OpenUrl CrossRef PubMed 71. ↵ Khatib , F. et al. Algorithm discovery by protein folding game players . Proc. Natl. Acad. Sci. U. S. A . 108 , 18949 – 18953 ( 2011 ). OpenUrl Abstract / FREE Full Text View the discussion thread. Back to top Previous Next Posted January 02, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Promiscuous structural cross-compatibilities between major shell components of Klebsiella pneumoniae bacterial microcompartments Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Promiscuous structural cross-compatibilities between major shell components of Klebsiella pneumoniae bacterial microcompartments Lucie Barthe , Damien Balestrino , Bessam Azizi , Delphine Dessaux , Vanessa Soldan , Jeremy Esque , Thomas Schiex , Sophie Barbe , Luis Fernando Garcia-Alles bioRxiv 2025.01.01.631032; doi: https://doi.org/10.1101/2025.01.01.631032 Share This Article: Copy Citation Tools Promiscuous structural cross-compatibilities between major shell components of Klebsiella pneumoniae bacterial microcompartments Lucie Barthe , Damien Balestrino , Bessam Azizi , Delphine Dessaux , Vanessa Soldan , Jeremy Esque , Thomas Schiex , Sophie Barbe , Luis Fernando Garcia-Alles bioRxiv 2025.01.01.631032; doi: https://doi.org/10.1101/2025.01.01.631032 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Biochemistry Subject Areas All Articles Animal Behavior and Cognition (7640) Biochemistry (17706) Bioengineering (13902) Bioinformatics (41978) Biophysics (21465) Cancer Biology (18611) Cell Biology (25528) Clinical Trials (138) Developmental Biology (13387) Ecology (19920) Epidemiology (2067) Evolutionary Biology (24332) Genetics (15615) Genomics (22519) Immunology (17747) Microbiology (40424) Molecular Biology (17194) Neuroscience (88662) Paleontology (667) Pathology (2838) Pharmacology and Toxicology (4827) Physiology (7650) Plant Biology (15160) Scientific Communication and Education (2046) Synthetic Biology (4302) Systems Biology (9826) Zoology (2271)