Origin and novel transport pathways of bacterial hopanoids

Origin and transport of bacterial hopanoids | 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 Origin and transport of bacterial hopanoids Evan C. Lawrence , Huiqiao Pan , Noah Ollikainen , View ORCID Profile Brittany J. Belin doi: https://doi.org/10.1101/2025.07.29.667474 Evan C. Lawrence 1 Department of Embryology, Carnegie Science , Baltimore, MD 21218 2 Department of Biology, Johns Hopkins University , Baltimore, MD 21218 Find this author on Google Scholar Find this author on PubMed Search for this author on this site Huiqiao Pan 1 Department of Embryology, Carnegie Science , Baltimore, MD 21218 Find this author on Google Scholar Find this author on PubMed Search for this author on this site Noah Ollikainen 3 Laboratory of Molecular Biology and Immunology, National Institute on Aging, National Institutes of Health , Baltimore, MD 21224 Find this author on Google Scholar Find this author on PubMed Search for this author on this site Brittany J. Belin 1 Department of Embryology, Carnegie Science , Baltimore, MD 21218 2 Department of Biology, Johns Hopkins University , Baltimore, MD 21218 Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Brittany J. Belin For correspondence: belin{at}carnegiescience.edu Abstract Full Text Info/History Metrics Supplementary material Preview PDF ABSTRACT Hopanoids are a class of steroid-like lipids that fortify bacterial membranes. Derivatives of hopanoids known as “geo-hopanes” are abundant in sediments and are signatures of ancient bacteria, yet there are conflicting views on their use as markers of specific taxa or environments. Here we analyze conservation of hopanoid biosynthesis across bacterial genomes using modern taxonomic tools. We find that hopanoids most likely originated in an ancestor of marine alphaproteobacteria, and that tolerance of high osmolarity is the most common feature of hopanoid-producing strains. Additionally, microsynteny with hopanoid-related loci revealed new hopanoid-associated gene families, which we term hpnT and hpnS . Structural predictions and the restriction of these gene families to diderms suggests they participate in hopanoid trafficking, potentially forming a pathway analogous to Mla proteins in E. coli . INTRODUCTION Hopanoids are steroid-like isoprenoid lipids produced by bacteria 1 , 2 . In sediments, ancient hopanoid derivatives (“geo-hopanes”) are abundant, with the oldest biomarkers dating to >1.6 Ga 3 . Their longevity in geologic time is due to their highly stable, pentacyclic hydrocarbon core, generated by consolidation and cyclization of six 5-carbon isoprene units. The simplest hopanoids contain only this 30-carbon core, sometimes with minor modifications (methylation or C-C bond desaturation), and are known as the “short” or C 30 class. Hopanoids with ribose-derived side chains form the “extended” or C 35 class. The C 35 hopanoids are structurally diverse, including the addition of hydroxyl or amine moieties or covalent attachment to other biomolecules. Many studies have examined the extent to which hopanoid biomarkers can be used to infer the identities of ancient microbes and their environments (reviewed in 4 ). Such inferences rely on understanding which bacteria produce which hopanoids, with the assumption that the distribution of biological processes across phyla – and their compatibilities with certain environments ( e.g. , oxic vs. anoxic habitats) – are conserved throughout time. Early in vivo studies hypothesized that geo-hopanes are markers of the advent of oxygenic photosynthesis. Specifically, hopanoids methylated at the C2 position (2Me-hopanoids) by the enzyme HpnP have been used as markers for marine cyanobacteria 3 , and the proliferation of 2Me-hopanoid derivatives during oceanic anoxic events (OAEs) led to the suggestion that marine cyanobacteria expanded in these periods 5 . Increased availability of bacterial genomes and metagenomes, and the discovery of hopanoid biosynthesis enzymes ( Table 1 ), have broadened our understanding of where and by which species hopanoids can be produced 6 , 7 . We now know that hopanoid-producing organisms are taxonomically diverse, making it difficult to classify geo-hopanes as markers of specific metabolisms or environmental contexts. The degree to which hopanoids can be used as markers of marine cyanobacteria remains an active subject of debate. Some genomics-based studies argue against use of hopanoids as cyanobacterial markers, suggesting that 2-Me hopanoids originated in the alphaproteobacteria 8 and that hopanoids are more generally markers for nitrogen cycling 7 , 9 – 12 . Other works have reinforced the connection between hopanoids and cyanobacteria 13 . Most recently it was suggested that 2-Me hopanoids can be used as cyanobacterial markers prior to 750 Ma 14 , based on the authors’ conclusions that (i) the evolutionary relationship between cyanobacterial hopanoid C2-methylases ( hpnP ) closely matches that of cyanobacteria genomes, indicating that hpnP was present in the last common cyanobacterial ancestor, and (ii) hpnP -encoding proteobacteria evolved much later than hpnP -encoding cyanobacteria 15 . View this table: View inline View popup Download powerpoint Table 1. Hopanoid-associated protein domain profiles in the Pfam/InterPro, TIGRFam, Kegg Ontology, and MetaCyc databases. *IPR005814 contains other aminotransferases and is not specific for HpnO/HpnQ. In the TIGRFAM database, an entry under hopanoid lipid biosynthesis pathway (GenProp0761) named HpnO is annotated with “VacJ-like lipoprotein”/PF04333. These terms do not refer to the same protein; use of HpnO/HpnQ to refer to the enzyme responsible for aminobacteriohopanetetrol production is the more common designation. The HpnO protein associated with GenProp0761 in TIGRFAM is likely the protein we name HpnS in this manuscript. Here we analyze the distribution of the hopanoid biosynthesis enzymes SHC, HpnH, HpnP, and HpnR across sequenced bacteria, incorporating an updated understanding of bacterial taxonomic relationships and comprehensive experimental data. Our results suggest that the capacity to produce hopanoids likely originated in an ancestor of marine alphaproteobacteria, and that tolerance of high osmolarity environments, nitrogen cycling, and single-carbon metabolism are common to hopanoid-producing species. Additionally, by evaluating microsynteny between known hopanoid-related genes and uncharacterized bacterial genes, we identify two new hopanoid-associated families, which we term hpnT and hpnS . Structural predictions and the restriction of these gene families to diderms suggests they participate in hopanoid trafficking, potentially forming a pathway analogous to Mla proteins in E. coli and other species. RESULTS Benchmarking annotation approaches for hopanoid-associated genes Determining the distribution of hopanoids across bacteria requires a robust approach for annotation of hopanoid biosynthesis genes. Prevailing methods for gene annotation include comparison to reference gene sets and, for protein-coding genes, identification of conserved protein domains. The latter approach relies on profile hidden Markov models (HMM), and there are multiple domain profiles for most known hopanoid biosynthesis proteins across the InterPro, Pfam, TIGRFAM, and Kegg Ontology databases ( Table 1 ). Domain assignments based on these profile HMMs are not manually curated and have not been benchmarked against experimental data. To evaluate the accuracy of each hopanoid-associated profile HMM, we assembled a set of experimentally validated control sequences. First, we reviewed the literature to identify all bacterial strains that have been shown to produce hopanoids in pure culture and thus must encode the SHC enzyme. We also inferred whether HpnP, HpnR, or HpnH also were encoded by these strains based on identification of 2Me-, 3Me- or C 35 hopanoids, respectively ( Table S1 ). Though other hopanoid biosynthesis enzymes that are involved in C 35 hopanoid modification (HpnG, HpnI, HpnJ, HpnK, and HpnO) have associated profile HMMs, there are few experimental studies on C 35 sidechain diversity across bacteria. Because we could identify few experimentally validated genes for these enzymes, we excluded them from benchmarking. We collected genomes for all hopanoid-producing strains in list Table S1 , as available, and we identified their SHC, HpnP, HpnR, and HpnH homologs by BLAST using hopanoid biosynthesis proteins validated by knockout (KO) or in vitro (IV) study as queries. These homologs were assembled into a “positive control” sequence dataset ( Table S2 ). KO/IV-verified sequences were also used to query the manually reviewed Swiss-Prot reference collection to generate a “negative control” dataset containing members of related but functionally distinct gene families. We used these control sequences to benchmark the identification of SHC, HpnH, HpnP, and HpnR using each of their associated profile HMMs in Table 1 . For HpnP, we also evaluated annotation via a combination of the Pfam vitamin B12-binding (PF02310), radical SAM (PF04055), and DUF4070 (PF13282) domains, following the approach used in Hoshino et al. 14 . In the first stage of benchmarking, we assessed whether each profile could successfully identify all and only the positive control sequences ( Table S2 ). We also examined each profile’s accuracy by collecting a diverse set of domain-associated sequences in public protein databases, aligning these sequences with our control datasets, and generating maximum likelihood trees ( Figures S1 - S4 ). These trees allowed us to assess the extent to which proteins annotated with each domain are related to verified sequences versus negative control sequences from families with distinct functions. Download figure Open in new tab Figure S1. Benchmarking tree for SHC-associated domain profiles. Unrooted maximum likelihood phylogenetic trees of SHC benchmarking control sequences plus proteins associated with PF13249+PF13243 ( A ), TIGR01507 ( B ), IPR006400 ( C ), K06045 ( D ) in UniProt or IMG. Negative control sequences (“Distinct family”, blue), knockout- or biochemically-validated positive control sequences (“Verified by KO”, red), and positive control sequences verified by presence of hopanoid product in culture (“Verified by Product”, orange) are indicated by box labels at branch tips. Dotted black line indicates position of separation between positive and negative control-associated clades. All four SHC-associated domains (PF13249+PF13243, TIGR01507, IPR006400, and K06045) were found in >99% of the positive controls, but these domains were also assigned to terpene cyclase negative controls ( Table S2 ). In phylogenetic trees of domain-associated proteins, an average of 26.5% of proteins annotated with SHC-related domains are positioned within negative control branches ( Figure S1 ). Most of these false positive sequences are homologs of the Bacillus subtilis sporulenol synthase, a squalene cyclase that does not produce hopanoids and thus does not belong in the SHC class. With the exception of TIGR01507, the SHC-related domain profiles also cannot distinguish squalene-hopene cyclases from the oxidosqualene cyclases that produce sterols. Of the hopanoid-modifying enzymes, HpnH-associated domain profiles were most reliable. Both the TIGR03470 and IPR017833 profiles for HpnH detected all positive control sequences and excluded all negative control proteins ( Table S2 ). This high accuracy was further supported by phylogenetic trees; proteins annotated with either domain all are within branches including at least one positive control, with no annotated proteins positioned in the negative control branch ( Figure S2A-B ). Domain profiles for the methylases HpnP and HpnR yielded more variable results. Among HpnP-associated domains, the overly stringent K22705 term matched only a minority (32%) of positive controls ( Table S2 ); this stringency is also evident in the phylogenetic tree ( Figure S3A ), where most of the positive control sequences are on a separate branch from K22705-annotated proteins. The HpnP annotation method used by Hoshino et al. (PF02310+PF04055+PF13282) failed to capture roughly one third of the positive controls ( Table S2 ), and in the phylogenetic tree, proteins annotated with the three Pfam domains do not fully populate the positive control branches ( Figure S3B ). These triply annotated proteins also include HpnP-like proteins that cannot be clearly assigned to either the negative control radical SAM methylases nor true HpnP branches. The IPR034530 domain for HpnP was the only profile we tested that matched all positive and no negative control sequences, though it similarly includes HpnP-like proteins ( Figure S3C ). Download figure Open in new tab Figure S2. Benchmarking tree for HpnH-associated domain profiles. Unrooted maximum likelihood phylogenetic trees of SHC benchmarking control sequences plus proteins associated with TIGR03470 ( A ) or IPR017833 ( B ) in UniProt or IMG. Negative control sequences (“Distinct family”, blue), knockout- or biochemically-validated positive control sequences (“Verified by KO”, red), and positive control sequences verified by presence of hopanoid product in culture (“Verified by Product”, orange) are indicated by box labels at branch tips. Dotted black line indicates position of separation between positive and negative control-associated clades. Download figure Open in new tab Figure S3. Benchmarking tree for HpnP-associated domain profiles. Unrooted maximum likelihood phylogenetic trees of SHC benchmarking control sequences plus proteins associated with PF13282+PF02310+PF04055 ( A ), IPR034530 ( B ), or K22705 ( C ) in UniProt or IMG. Negative control sequences (“Distinct family”, blue), knockout- or biochemically-validated positive control sequences (“Verified by KO”, red), and positive control sequences verified by presence of hopanoid product in culture (“Verified by Product”, orange) are indicated by box labels at branch tips. Dotted black line indicates position of separation between positive and negative control-associated clades. Finally, for HpnR, the IPR027564, TIGR04367, and K22704 domains were absent in negative controls, but these profiles did not all detect the positive controls ( Table S2 ). We found that the HpnR-associated KO term is the least stringent of the three domain profiles, such that most K22704-annotated proteins do not appear to be true HpnR homologs ( Figure S4A ). Both IPR027564 and TIGR04367 profiles also match numerous uncharacterized proteins in a branch with NirJ proteins ( Figure S4B-C ). Download figure Open in new tab Figure S4. Benchmarking tree for HpnR-associated domain profiles. Unrooted maximum likelihood phylogenetic trees of SHC benchmarking control sequences plus proteins associated with K22704 ( A ), IPR027564 ( B ), or TIGR04367 ( C ) in UniProt or IMG. Negative control sequences (“Distinct family”, blue), knockout- or biochemically-validated positive control sequences (“Verified by KO”, red), and positive control sequences verified by presence of hopanoid product in culture (“Verified by Product”, orange) are indicated by box labels at branch tips. Dotted black line indicates position of separation between positive and negative control-associated clades. Collectively, our results demonstrate that domain assignments based on pre-computed profile HMMs alone are not a reliable metric for hopanoid gene annotation, and that the source of the optimal domain profile (Intepro/Pfam, TIGRFAM, Kegg Ontology, etc .) varies on a case-by-case basis. These findings led us to refine our strategy for identification of hopanoid biosynthesis genes families across bacterial genomes: (i) select HMM profiles that match 100% of the positive control sequences, using the most comprehensive set of experimentally validated sequences possible; (ii) apply these profiles to the genome set; and (iii) remove false positive sequences based on their similarity to positive vs. negative control sequences, as determined using phylogenetic tools. Hopanoid biosynthesis genes are distributed across diverse bacterial phyla To assess hopanoid distribution across bacteria, we collected complete bacterial genomes in the Joint Genome Institute’s Integrated Microbial Genomes (IMG) database 16 , filtered them by quality, and classified them into taxa as described in the Methods. This dataset included 15,541 unique species and a total of 62,143 genomes ( Table S3 ). Within each species, we calculated the percentage of genomes encoding SHC, HpnH, HpnP, and HpnR, using the benchmarked annotation approach described in the previous section ( Tables S4-S5 ); phylogenetic trees used to identify and remove false positives are shown in Figure S5 . We also annotated the genomes with C 35 hopanoid-modifying enzymes HpnG, HpnI, HpnJ, HpnK, and HpnO based on presence of their associated TIGRFAM domains ( Table 1 ), though we note that these annotations were not benchmarked for reasons outlined above. Download figure Open in new tab Figure S5. Benchmarking tree for identifying SHC, HpnH, HpnP, and HpnR family members in IMG genomes. Unrooted maximum likelihood phylogenetic trees of benchmarking control sequences plus proteins associated with ( A ) SHC (Pfam13249+Pfam13243), ( B ) HpnH (TIGR03470), ( C ) HpnP (IPR034530), and ( D ) HpnR (TIGR04367) in genomes from Table S3. Negative control sequences (“Distinct family”, blue), knockout-or biochemically-validated positive control sequences (“Verified by KO”, red), and positive control sequences verified by presence of hopanoid product in culture (“Verified by Product”, orange) are indicated by box labels at branch tips. Dotted black line indicates position of separation between positive and negative control-associated clades. Scale bars indicate substitutions per 1000 amino acids. By this metric, we calculated that ∼10% (1645/15541) of species within our dataset have the genetic capacity to produce hopanoids ( Table S4 ; Figure 1 ). SHC-containing species are most abundant within the pseudomonadota, with high conservation within multiple families of alphaproteobacteria and the clade containing Burkholderia and Paraburkholderia spp. (classified as gammaproteobacteria in the Genome Taxonomy Database (GTDB) 17 and as betaproteobacteria in the List of Prokaryotic names with Standing in Nomenclature (LPSN) 18 ). SHC is also highly conserved within acidobacteriota and the GTDB desulfobacterota (a.k.a.deltaproteobacteria and thermodesulfobacteriota in LPSN), as well as within the cyanobacteriota and actinomycetota. Lower levels of SHC conservation are found in planctomycetota, verrucomicrobiota, and chloroflexota. SHC is absent or extremely rare within bacillota, bacteroidota, campylobacteriota, deinococcota, fusobacteriota, synergistota, spirochoaetota, and thermotogota. Download figure Open in new tab Figure 1. Conservation of hopanoid biosynthesis in bacteria. Maximum likelihood cladogram of bacterial taxa using bac120 sequences. Each branch represents a group of phylogenetically related organisms; the number of genomes represented by each branch and the taxonomic classification of those genomes are provided as branch tip labels and branch names, respectively. Phyla assigned to genomes in each branch using the GTDB-Tk program are shown in the outer ring, with corresponding color-coding of the individual branches. Inner rings with grayscale rectangles indicate the degree of conservation within each branch of various hopanoid-associated domains (see legend for domain/pathway represented by each ring). Scale bar indicates substitutions per 1000 nucleotide positions. Of the hopanoid-modifying enzymes, HpnH is most common; nearly all SHC-containing genomes also encode HpnH. Domains for HpnG, HpnI, and HpnJ were only identified in genomes that contain the SHC and HpnH domains required for the first steps of C 35 hopanoid production. HpnK also is involved in downstream modification of C 35 hopanoids, but its distribution includes some likely spurious annotations within the bacillota. We infer from these distributions that C 35 hopanoid production and modification is most common in the pseudomonadota, desulfobacterota, acidobacteriota, and cyanobacteriota. Actinomycetota appear to have the capacity to produce ribosylhopane via HpnH but lack the enzymes for downstream modification. Hopanoid methylases HpnP and HpnR have the sparsest taxon distributions. HpnP and HpnR are both present in pseudomonadota and desulfobacterota, whereas only HpnP is found in cyanobacteriota and planctomycetota, and only HpnR is present in actinomycetota. We note that within cyanobacteria, conservation of HpnP is not universal. Within the cyanobacteriia class – which contains all of the complete, high quality cyanobacterial genomes in IMG – hopanoid production is absent in both the synechococcales order (dominated by Synechococcus elongatus genomes) and the PCC-6307 order (dominated by Prochlorococcus spp. genomes) ( Table S5 ). These results differ from the HpnP distribution in cyanobacteria reported by Hoshino et al . 14 . We hypothesized this discrepancy could arise from differences in the genomes examined, the taxonomic classification of those genomes, or in our HpnP identification strategies. We assessed these hypotheses by collecting all cyanobacterial genomes evaluated in the other study, assigning taxa to those genomes using GTDB-Tk, and re-calculating their HpnP conservation levels using the other group’s HpnP identification strategy. We also evaluated their list of cyanobacterial HpnP proteins against our benchmarking strategy. When using the same taxon classification strategy, our results generally agree ( Table S6 ); our observation that HpnP homologs are absent in synechococcales arises from the re-classification of many genomes annotated as synechococcales members in LPSN as members of the thermosynechococcales order in GTDB. When HpnP is present, we do observe higher levels of conservation than the other study, which we attribute to the limitations of the other group’s HpnP identification strategy (co-annotation with PF02310+PF04055+PF13282). This is supported by our phylogenetic analyses of the HpnP sequences identified by Hoshino et al. 14 , which do not capture the full diversity of verified HpnP sequences, even when comparing only to the subset of experimentally verified HpnP homologs from cyanobacteria ( Figure S6 ). Download figure Open in new tab Figure S6. Comparison of IMG HpnP domain-associated proteins with Hoshino et al. HpnP homologs. Unrooted maximum likelihood phylogenetic tree of HpnP benchmarking control sequences, IPR034530-annotated proteins in IMG, and cyanobacterial HpnP homologs (annotated with green rectangles at branch tips) identified by Hoshino et al . (annotated with green rectangles at branch tips) 14 . Control sequences are annotated as in Fig. S1-S5. Experimentally validated HpnP proteins in the control sequences are indicated by black circles, demonstrating that cyanobacterial HpnP proteins form two clades; only one clade is captured by the Hoshino et al . annotation method. Hopanoid production likely arose in marine alphaproteobacteria Our genome analysis ( Figure 1 ) suggested two possible origins for hopanoid biosynthesis: an ancestor of the clade containing the cyanobacteriota, chloroflexota, deinococcota, and actinomycetota, or an acestor of the clade containing the pseudomonadota, acidobacteriota, and desulfobacterota. To assess these possibilities, we generated maximum likelihood phylogenetic trees containing all SHC homologs in our genome dataset identified the most likely root of the tree by rootstrapping. We also assessed the origins of C 35 hopanoid production, 2Me-hopanoid production, and 3Me-hopanoid production by generating trees for HpnH, HpnP, and HpnR. The maximum likelihood phylogenetic tree for SHC homologs ( Figure 2 ) suggests that SHC originated in marine alphaproteobacteria, including members of the parvularculaceae and CGMCC-115125 ( Pyruvatibacter spp.) families, and was present in other (non-marine) aquatic alphaproteobacterial families before arising in other lineages. From these aquatic alphaproteobacteria, SHC proteins form three distinct clades. One clade includes terrestrial alphaproteobacteria and a subset of SHC-encoding burkholderiaceae, whereas the SHC proteins from the majority of hopanoid-producing gamma-proteobacteria form a separate clade originating in the marine nitrococcaceae and salinisphaeraceae families. All non-proteobacterial SHC proteins are present in a separate clade that originates in the newly cultivated methylaequoraceae/TMED127 family of alphaproteobacteria, including the SHC proteins from desulfobacterota, acidobacteriota, cyanobacteria, and actinobacteria. Download figure Open in new tab Figure 2. Hopanoid biosynthesis originated in marine alphaproteobacteria. Maximum likelihood phylogenetic tree generated for protein sequences associated with SHC. Branch names include IMG genome ID, GTDB family and genus assignments, and IMG protein ID. Numbers at internal nodes (highlighted in white) indicate branch support values using ultrafast bootstrap approximation (UFBoot), based on 1000 bootstraps; 100 = highest confidence, 0 = no confidence. Scale bar indicates substitutions per 1000 amino acids. Branch backgrounds are color-coded by phyla according to the color scheme (outer ring and phylum text labels) used in Figure 1 . The maximum likelihood tree of HpnH homologs is surprisingly similar ( Figure S7 ), containing clades with similar composition to the SHC tree that also appear to originate in the parvularculaceae, CGMCC-115125, and methylaequoraceae families of marine alphaproteobacteria. Given the similarity of the SHC and HpnH trees and their highly overlapping distributions, we conclude that the capacity to produce C 35 hopanoids occurred early in the evolution of hopanoid biosynthesis. That the roots of both SHC and HpnH trees were identified the same families of organisms provides strong evidence – beyond the individual trees’ root confidence values (>85% in both cases) – that hopanoids originated in marine alphaproteobacteria. Download figure Open in new tab Figure S7. Extended hopanoids originated in marine alphaproteobacteria. Maximum likelihood phylogenetic tree generated for protein sequences associated with HpnH. Branch names include IMG genome ID, GTDB family and genus assignments, and IMG protein ID. Numbers at internal nodes (highlighted in white) indicate branch support values using ultrafast bootstrap approximation (UFBoot), based on 1000 bootstraps; 100 = highest confidence, 0 = no confidence. Scale bar indicates substitutions per 1000 amino acids. Branch backgrounds are color-coded by phyla according to the color scheme (outer ring and phylum text labels) used in Figure 1 . The origins of methylated hopanoids are less clear. In the maximum likelihood tree for HpnP proteins, two possible origins are suggested: one major clade potentially originating in the acidobacteriota (though we note our dataset contained few acidobacterial proteins), from which the cyanobacterial and proteobacterial HpnP proteins evolved, and a minor clade originating in the desulfobacterota containing HpnP proteins from all other 2Me-hopanoid-producing phyla ( Figure 3 ). However, this root location is weakly supported by rootstrapping. Statistical analyses of all possible root positions via tree topology tests identified 37 potential root positions resulting in trees with a statistically similar likelihood (p > 0.05 threshold) as the maximum likelihood tree ( Dataset S1 ). Interestingly, none of these trees suggest that 2-Me hopanoids originated in cyanobacteria. A plurality of likely root positions indicate that HpnP first appeared in both the acidobacteriota and the desulfobacterota, and alternative origins in the planctomycetota or chloroflexota were also common. However, we cannot rule out long-branch effects as the explanation for these potential root positions, especially given the paucity if hpnP sequenced from these phyla. Download figure Open in new tab Figure 3. Phylogenetic relationships of HpnP homologs. Maximum likelihood phylogenetic tree generated for protein sequences associated with HpnP. Branch names include IMG genome ID, GTDB family and genus assignments, and IMG protein ID. Numbers at internal nodes (highlighted in white) indicate branch support values using ultrafast bootstrap approximation (UFBoot), based on 1000 bootstraps; 100 = highest confidence, 0 = no confidence. Scale bar indicates substitutions per 1000 amino acids. Branch backgrounds are color-coded by phyla according to the color scheme (outer ring and phylum text labels) used in Figure 1 . The origin of 3Me-hopanoids is also ambiguous. The maximum likelihood tree of HpnR sequences tree indicates that HpnR proteins may have evolved from a shared ancestor of pseudomonadota and acidobacteriota ( Figure S8 ) and were transferred to actinomycetota through these groups. The HpnR proteins from zetaproteobacteria, myxococcota, and methylomirabilota may have originated independently as they are present on a separate clade from the majority of HpnR sequences. Because the predicted root position of the tree is weakly supported by rootstrapping, we also performed tree topology tests as described for HpnP ( Dataset S1 ). The 17 equally likely root positions (p > 0.05 threshold) generally agree with interpretations based on the maximum likelihood tree, though some root positions nest the zetaproteobacteria, myxococcota, and methylomirabilota HpnR proteins within one of the two major clades. Download figure Open in new tab Figure S8. Phylogenetic relationships of HpnR homologs. Maximum likelihood phylogenetic tree generated for protein sequences associated with HpnR. Branch names include IMG genome ID, GTDB family and genus assignments, and IMG protein ID. Numbers at internal nodes (highlighted in white) indicate branch support values using ultrafast bootstrap approximation (UFBoot), based on 1000 bootstraps; 100 = highest confidence, 0 = no confidence. Scale bar indicates substitutions per 1000 amino acids. Branch backgrounds are color-coded by phyla according to the color scheme (outer ring and phylum text labels) used in Figure 1 . Hopanoid biosynthesis genes are correlated with nitrogen cycling and high osmolarity tolerance To find shared features of hopanoid-producing bacteria, we identified all protein domain families that are enriched within genomes encoding SHC and other hopanoid biosynthesis genes ( Table S7 ). This analysis revealed several protein functional categories that are positively correlated with hopanoid production ( Table 2 ). The most enriched domains in SHC-containing genomes were associated with synthesis and modification of hopanoids, as well as of squalene, isoprenoid precursors (farnesyl and geranyl diphosphate), and other cyclic isoprenoids such as carotenoids or sterols. An enrichment of domains associated with production of B vitamins and S-adenosyl-methionine (SAM) metabolism also was apparent, likely relating to the role of vitamin B12 and SAM as co-factors for hopanoid biosynthesis proteins HpnH, HpnP, and HpnR. View this table: View inline View popup Download powerpoint Table 2. Protein domain categories positively correlated with SHC. Manually annotated domain categories correlated with SHC across all genomes. Center column represents average correlation coefficient (r) for domains within each category. A full list of correlated domains is provided in Table S7 . The second most significant category of correlated proteins includes those involved in osmotic pressure sensing and response. Domains involved in biosynthesis of the osmolyte trehalose and other saccharides involved in osmotic stress ( e.g. , osmoregulated periplasmic glucans) are enriched in hopanoid-producers to a similar level as other isoprenoid biosynthesis domains. This category also includes the osmotic stress-regulated two-component system KdpPE, which controls potassium flux across the membrane via the potassium pump KdpFABC, and various transporters involved in divalent cation homeostasis. Interestingly, osmo-responsive protein domains are correlated with SHC across nearly all hopanoid-producing taxa. Specific bacterial metabolisms were also highlighted by our analysis. Domains involved in utilization of single-carbon substrates and in nitrogen cycling are common to SHC-encoding genomes, and the biggest group of SHC-correlated metabolic domains are involved in aerobic respiration, such as the high efficiency terminal oxidases required for nitrogen fixation. To test this relationship between SHC and the nitrogen cycle, we measured the growth of a hopanoid-deficient strain of Bradyrhizobium diazoefficiens 19 , 20 on several nitrogen cycle intermediates ( Figure S9 ). We find that hopanoid-deficient mutants grow slowly in the presence of nitrite but have no growth defect in the presence of ammonia or nitrate. Our results support other recent work linking hopanoids to nitrite metabolism 10 – 12 . Download figure Open in new tab Figure S9. Hopanoid-deficient B. diazoefficiens have slow growth on nitrite. Growth curves of WT (WT+E), hopanoid-deficient (delta shc + E), and rescue (delta shc + C) strains of B. diazoefficiens on minimal media supplemented with indicated nitrogen sources. Microsynteny suggests novel hopanoid transport factors Roughly 16% of the top 1000 domains positively correlated with SHC were annotated as uncharacterized or domains of unknown function (DUFs) ( Table S7 ); collectively, this group was the most enriched category in SHC-containing genomes. We hypothesized that any DUFs directly involved in hopanoid biosynthesis would be adjacent to other hopanoid-associated genes within the genome. To identify such DUFs, we performed a targeted analysis of protein domains found near the shc gene locus. We collected domain annotations for all genes within 20 ORFs up- or down-stream of shc in each genome, categorizing these genes into families based on shared domain annotations ( Table S8 ). Our analysis revealed four uncharacterized gene families that are commonly located near the shc locus and have potentially hopanoid-related functions ( Figure 4 ). Two of these families are annotated with domains that already were computationally predicted to be related to hopanoids. The hpnL gene family includes genes annotated with the uncharacterized “HpnL” domain in TIGRFAM (TIGR03476); this domain does not exist in other protein family or pathway databases, and in InterPro is instead annotated as a member of the lysylphosphatidylglycerol synthetase/glycosyltransferase AglD (IPR022791) family. The AglD InterPro domain includes proteins involved in addition of lysine addition to the phospholipid phosphatidylglycerol and in microbial S-layer modification. The hpnM gene family is annotated with the InterPro “Hopanoid biosynthesis-associated membrane protein HpnM” domain (IPR017842), a subset of “Toluene tolerance Ttg2/phospholipid-binding protein MlaC” (IPR008869) domain proteins. Two new candidate hopanoid gene families, which we name here hpnS and hpnT , have not previously been associated with hopanoids. The hpnS family genes share the domain annotation MlaA lipoprotein (IPR007428). The hpnT genes are annotated with Domain of unknown function DUF2147 (IPR019223); a crystal structure of the DUF2147 domain-containing protein HP1028 from Helicobacter pylori 21 suggests these domains belong to the lipocalin family, a group of proteins that transport hydrophobic small molecules 22 . Download figure Open in new tab Figure 4. Candidate hopanoid-associated genes hpnL , hpnM , hpnS , and hpnT are enriched near shc loci. Genes adjacent to the shc locus in selected bacterial genomes. Color-coding of ORFs indicates functional categories for conserved gene families: known hopanoid-associated genes (Hopanoid, yellow); known isoprenoid biosynthesis genes (Isoprenoid, pink), and candidate gene families hpnM , hpnS , hpnT , and hpnL (purple, green, blue, and orange, respectively). ORFs in the Other category (grey) are conserved gene families with unknown functions; number labels correspond to specific protein domain annotations provided in Table S3. Gene names labeled with asterisks indicate likely gene famly members that lack one or more domain annotation relative to other genes in the same family ( e.g., ‘ hpnM * ’ = ‘hpnM-like ’). All unlabeled ORFs (white) are non-conserved genes. In E. coli and other diderms, MlaC and MlaA are components of an intermembrane phospholipid transport system 23 , 24 , whereas lipocalins can serve as extracellular lipid carriers in eukaryotes 25 . We thus hypothesized that HpnM, HpnS, and HpnT may form components of inner-to-outer membrane transport systems for hopanoids, in concert with the inner membrane hopanoid transporter HpnN. This possibility is supported by the restriction of these proteins to diderms and the presence of predicted signal peptides, indicating a likely periplasmic localization ( Figure S10 ). Our subsequent analysis of the conservation of hpnM , hpnS , and hpnT genes within the vicinity of hpnN loci ( Table S8 ) suggests that they evolved in distinct taxa: the lipocalin-like HpnT protein is conserved in the hyphomicrobiales/rhizobiales order of alphaproteobacteria, whereas both HpnM and HpnS co-occur in the sphingomonadales, rhodospirillales, and azospirillales orders of alphaproteobacteria and across gammaproteobacteria ( Figure 5 ). Download figure Open in new tab Figure S10. Signal peptides are prevalent in HpnL, HpnM, HpnS, and HpnT proteins. Fraction of proteins each family predicted to have a signal peptide for periplasmic secretion by Phobius 78 . Download figure Open in new tab Figure 5. The HpnM/S module and HpnT appear in distinct HpnN-encoding taxa. Maximum likelihood tree of bac120 sequences from selected HpnN-encoding, Gram-negative bacterial genera. Branch tip labels (shaded rectangles) indicate conservation of HpnM, HpnS, and HpnT across all HpnN-encoding genomes within each genus; conservation levels corresponding to each shade are provided in the legend. Presence of these proteins was not annotated in the archaeon outgroup, Thermococcus . Scale bar indicates substitutions per 1000 nucleotides. MlaFEDB proteins form a complex 26 that moves phospholipids through the inner membrane, delivered by the periplasmic shuttle protein MlaC, which in turn ferries phospholipids from the MlaA lipoprotein that associates with outer membrane proteins 24 , 27 , 28 ( Figure 6A ). We speculated that HpnM and HpnS might also form a shuttle and OM anchor protein system, which would require them to bind one another at the OM ( Figure 6B ). To test this hypothesis, we used AlphaFold 3 29 , which predicted a high probability that the Burkholderia multivorans HpnM protein binds both the hopanoid diploptene and HpnS in a manner similar to the predicted MlaA/MlaC dimer 30 ( Figure 6C ). Bradyrhizobium diazoefficiens HpnT is predicted to bind diploptene and may also act as a periplasmic shuttle, though it has no obvious cognate OM binding partner. These structural predictions support the hypothesis that HpnM, HpnS, and HpnT are novel hopanoid trafficking factors. Download figure Open in new tab Figure 6. Candidate hopanoid transport pathways in proteobacteria. ( A ) Schematic of the Mla phospholipid (PL) transport system from E. coli . The inner membrane transport complex MlaBDEF transports PLs to the periplasmic shuttle protein MlaC, which interacts with OmpF/C-associated lipoprotein MlaA in the outer membrane. ( B ) Candidate hopanoid transport pathways in diderm bacteria. In most proteobacteria (left), the inner membrane transporter HpnN may interact with the periplasmic shuttle HpnM to deliver hopanoids to the outer membrane lipoprotein HpnS. In the Rhizobiales order (right), hopanoids trafficked by HpnN are delivered to the periplasmic HpnT protein. ( C ) AlphaFold 3-predicted structures of the Burkholderia multivorans HpnSM dimer (top) and Bradyrhizobium diazoefficiens HpnT (bottom) in association with the hopanoid diploptene (grey). MATERIALS & METHODS Assembly of control sequence datasets Strains experimentally verified to produce hopanoids were identified by literature review, and strain names were updated manually to their preferred LPSN names by searching the strains’ culture collection IDs in the DMZ StrainInfo database 31 . The LPSN strain names were used to identify corresponding genome assemblies from NCBI. Homologs of SHC, HpnH, HpnP, and HpnR were identified in these genomes by BLAST search (e-value < 1e-50) using knockout- or biochemically-verified enzymes as queries. Genome accession numbers and relevant citations for these queries are indicated with asterisks in Table S1 , and the accession numbers of corresponding BLAST hits are provided in Table S2 . Negative control sequences were determined by BLAST search of the manually curated Swiss-Prot database using knockout- or biochemically-verified SHC, HpnH, HpnP, and HpnR queries. Selected negative control sequences (accessions provided in Table S2 ) represent the top BLAST hits with experimental evidence to demonstrate that they are involved in functions distinct from hopanoid biosynthesis. Protein domain profile benchmarking Protein domain profiles matching negative and positive control sequences were determined using the InterProScan command line tool (version 5.69-101.0, https://github.com/ebi-pf-team/interproscan ) 32 for Pfam, InterPro, and TIGRFAM domains and KofamKOALA (ver. 2026-01-01, KEGG release 117.0, https://www.genome.jp/tools/kofamkoala/ ) 33 for Kegg Ontology domains. For phylogenetic analysis-based benchmarking, all proteins matching each SHC, HpnH, HpnP, and HpnR-associated domains in Table 1 were collected from UniProt (for Pfam and InterPro domains) or Joint Genome Institute’s Integrated Microbial Genomes (IMG) database ( https://img.jgi.doe.gov/ ) 16 (for TIGRFAM and Kegg Ontology domains). Matching proteins were filtered by similarity using a 75% identity cutoff CD-Hit (cd-hit, https://github.com/weizhongli/cdhit/wiki ) 34 and were aligned, along with their corresponding positive and negative control sequences, using the super5 algorithm of the MUSCLE v. 5.1 sequence alignment program ( https://www.drive5.com/muscle/muscle.html ) 35 . This alignment was used to generate a maximum likelihood phylogenetic tree with IQ-TREE 2 ( https://iqtree.github.io/ ) 36 using the Q.pfam amino acid substitution model. Trees were visualized and annotated with control sequence positions using TreeViewer 37 . Bacterial genome selection Genomes were sourced from IMG and filtered by quality using CheckM2 v1.0.2 38 . We selected only genomes meeting the following criteria: (i) completeness estimate >50%, (ii) contamination estimate 50, (iv) 5 kb, and (vi) presence of >40% of the bac120 gene regions used for taxon assignment by GTDB. Genomes passing these criteria were further filtered to remove accessions where <75% of genes were associated with one or more TIGRFAM, Pfam, or Kegg Ontology annotation(s), indicating incomplete annotation. We also removed genomes with a total gene count < 1000, as these genomes are predominantly from obligate endosymbionts with heavily reduced genomes. Complete sequences for accessions that met these criteria (totaling 62,143 genomes, listed in Table S3 ) were downloaded from IMG along with their associated ORF, protein domain, and functional pathway annotations. Genome classification Taxonomic classification of the genomes was performing using the GTDB Toolkit v2.4.0+ ( https://github.com/Ecogenomics/GTDBTk ) 39 . For the majority of genomes, GTDB-Tk was sufficient to provide annotations through the species level. However, a subset of the genomes did not match one of the GTDB reference strains. For these accessions, tentative species were assigned by 16S rRNA gene similarity, using the canonical 97% sequence identity cutoff using CD-Hit. Because genomes in our dataset contained an average of three 16S genes, 16S-based assignments were not always consistent. In 1.2% of cases, 16S sequences from the same genome did not share at least 97% identity, and in ∼9% of genomes, 16S-based genus assignments did not match assignments based on whole-genome annotations ( Table S9 ). We addressed this by discarding taxon assignments from 16S sequences for which the associated genome had a different genus assignment than the associated 16S cluster. Generation of bac120 phylogenetic tree For each species in our dataset, a single genome was selected as a species representative based on completeness of its corresponding bac120 region. These representative bac120 sequences were pre-aligned by GTDB-Tk, and a maximum likelihood phylogenetic tree was generated using IQ-TREE 2. First, the optimal substitution model was determined by ModelFinder 40 , and a maximum likelihood tree using that model was generated using 1,000 replicates each for calculation of branch support values using ultrafast bootstrap approximation (UFBoot) 41 . To improve readability, species representatives were further consolidated into higher level groupings such that each group contained at least 5 genomes (which the exception of phyla for which fewer than 5 sequenced genomes are available, which remained single groups), resulting in 445 higher order taxon groups. Representatives for the 445 groups were selected based on the number of genomes, and only these representatives were shown on the final bac120 tree in Figure 1 . This final tree was annotated to include the corresponding taxon for each branch and the number of genomes represented by each branch in TreeViewer. Calculation and visualization of hopanoid biosynthesis gene conservation We annotated the protein-coding sequences of all genomes in our dataset using command-line InterProScan. We then extracted the sequences for all proteins matching the broadest protein domain profiles for SHC (Pfam13249+Pfam13243), HpnH (TIGR03470), HpnP (IPR034530), and HpnR (TIGR04367). These proteins were aligned with the appropriate set of positive and negative control sequences, and a maximum likelihood tree was generated as described in the Protein domain profile benchmarking section. We calculated the percentage of genomes within each species with at least one protein sequence within the positive control branches of the SHC, HpnH, HpnP, and HpnR trees ( Table S4 ), respectively, as well as the percentages of genomes annotated with domains associated with the extended hopanoid-modifying proteins listed in Table 1 . These percentages were recalculated at the order rank ( Table S5 ) and the 445 higher order taxon groups described in the previous section by taking the average percentages for all species within each order/taxon group. Percentages were mapped onto the bac120 tree and visualized using TreeViewer. Protein tree generation For protein trees, the SHC, HpnH, HpnP, and HpnR sequences identified above were aligned without control or outgroup sequences in MUSCLE. For each protein family, we generated 500 diversified alignments using the align algorithm, each with a unique seed, and the best scoring alignment was identified by maximum column confidence score using maxcc . To increase downstream processing time, columns in which >95% of sequences contained gaps were removed from the best-scoring alignment prior to tree generation. Maximum likelihood trees were generated using IQ-TREE 2 with 1,000 bootstrap replicates using UFBoot and root positions were inferred using the general non-reversible amino acid model NON-REV 42 . For maximum likelihood trees with low rootstrapping support values ( 0.05) likelihoods as the maximum likelihood tree using at least one of the following significance tests: one sided Kishino-Hasegawa test, Shimodaira-Hasegawa test, Expected Likelihood Weight (Strimmer & Rambaut), approximately unbiased (AU) test 42 . As above, all maximum likelihood trees were visualized and annotated using TreeViewer. Species tree – gene tree reconciliation for HpnP A distribution of 1,000 gene trees for HpnP was generated using MrBayes v3.2.7a 43 using a mixed amino acid substitution model for 1 million generations; all other parameters were set to default values. A corresponding species tree was generated in IQ-TREE 2 using aligned bac120 sequences from GTDB-Tk. The tree included bac120 sequences from all HpnP-encoding species as well as group representatives from Figure 1 , which served as outgroups. The root of the species tree was inferred by rootstrapping as described in the previous section. Reconciliation was performed with AleRax v1.4.0 using all 1,000 gene trees for calculation of duplication, transfer, and loss (DTL) probabilities. Predicted DTL events with >50% probability were extracted from AleRax and used to annotate the HpnP species tree in TreeViewer via custom Phyton scripts. Calculation of protein domain correlations We calculated correlation coefficients between SHC, HpnH, HpnP, HpnR and all other bacterial protein domains by first determining whether each Pfam, TIGRFAM, KO, and InterPro domain was present in the genome of each species representative. We then calculated the Pearson correlation coefficient and corresponding p-value for every pair of domains using binary presence (1) and absence (0) values, either across all genomes in the data or within phyla ( Table S7 ). Correlation coefficients were calculated in custom Python scripts using the scipy.stats.pearsonr function. B. diazoefficiens nitrogen utilization assays Wild-type Bradyrhizobium diazoefficiens strain USDA110 spc4 and inducible hopanoid mutant strain Pcu-shc::Δshc (abbreviated as Δshc) were cultured as described previously 20 . Briefly, an individual colony of each strain was incubated in liquid AG liquid media aerobically in an incubating shaker at 30°C and 250 rpm until mid-log phase (OD 600 = 0.4–0.8). WT, Δ shc, and Δ shc with cumate supplementation (Δ shc + c) were diluted into AG minimal media lacking a nitrogen source (2.3 mM sodium gluconate, 3.3 mM arabinose, 5.6 mM MES monohydrate, 5 mM HEPES, 1 mM Na2HPO4, 1.76 mM Na2SO4, 88 µM CaCl2, 25 µM FeCl3, and 0.73 mM MgSO4, pH 6.6) to an initial OD 600 = 0.008. Aliquots of this dilute subculture (100 µl) were dispensed into a 96-well Biolog PM3 Microplates containing 95 different nitrogen sources and a negative control. Each well was overlaid gently with 30 µL of mineral oil to prevent evaporation. The plate was placed in a BioTek microplate reader with continuous shaking, and the OD 600 was measured every 2 h for 168 h. The growth curve data were plotted using Prism GraphPad. Microsynteny analysis of shc- and hpnN-adjacent genomic regions To compare regions surrounding shc across genomes, we selected the subset of genomes we found to contain SHC. We defined shc -adjacent regions (“neighborhoods”) as the nearest 20 upstream and 20 downstream neighboring ORFs. We used InterProScan to identify the protein domains encoded by each neighboring gene, and then we assigned each gene a numerical ID based on the set of InterPro domains that it contained. For some genes, such as those known to be involved in hopanoid or isoprenoid biosynthesis, we replaced the numerical ID with a gene name such as “hpnN” or “hpnM”. In some cases, genes contained some but not all of the domains of a known gene, and we assigned these genes the name of the known gene followed by an asterisk ( e.g. , “hpnN*” or “hpnM*”). InterPro domains that occurred in fewer than 5% of the neighboring genes were excluded from this analysis, and genes containing only those domains were not assigned a name. To visualize microsynteny in these neighborhoods, we obtained the coordinates and orientation of each shc -neighboring gene and used a custom Python script to draw the genomic regions. To align the regions, the starting coordinate of shc was set to 0 for each genome. We colored each gene based on its assigned name/numerical ID and illustrated gene orientation using triangles pointed toward the 3’ end. To evaluate genes in the hpnN neighborhood, we selected genomes containing hpnN and removed any genomes that were not assigned to a Gram-negative species. We then calculated the percentage of genomes with hpnM , hpnS , and hpnT in the hpnN neighborhoods across genera. These values were visualized on a maximum likelihood tree of bac120 sequences for species representatives of hpnN -containing, Gram-negative genera generated as described in the Generation of bac120 phylogenetic tree section and visualized in TreeViewer. HpnT, HpnM and HpnS structure and binding predictions To predict the three-dimensional structures of HpnT, HpnM, and HpnS with diploptene, we used AlphaFold 3 (AF3, https://github.com/google-deepmind/alphafold3 ) 29 . For these predictions, we used the protein sequences of HpnM and HpnS from Burkholderia multivorans ATCC 17616 (BMULJ_05342 / Uniprot A0A0H3KU80 and BMULJ_05340 / Uniprot ID A0A0H3KPM6) and the protein sequence of HpnT from Bradyrhizobium diazoefficiens USDA110 ( bll3009 , Uniprot ID Q89QW5). The SMILES string of diploptene was obtained from PubChem. For HpnT, we predicted the structure of a HpnT monomer in complex with diploptene. For HpnM and HpnS, we predicted the structure of a HpnM / HpnS heterodimer in complex with diploptene. We used the default AF3 parameters for all predictions. DISCUSSION Gene annotation relies on the detection of conserved sequence patterns, typically using probabilistic models such as profile HMMs. The first protein databases relied on manually curated seed sequences to ensure model fidelity, but most of the >30,000 bacterial protein profile HMMs now available in InterPro and other databases were generated without strict oversight and with limited experimental findings. Our benchmarking of profile HMMs for hopanoid biosynthesis enzymes demonstrates that the accuracy of such models is variable and difficult to predict. This finding allowed us to establish a higher fidelity annotation pipeline, and it also provides a cautionary note on the limitations of machine learning that as applications of generative artificial intelligence become more common. An annotation pipeline is necessarily limited by the breadth and quality of the experimental data on which it is trained, and such pipelines must be continuously re-evaluated as new experimental data become available. Our benchmarked annotation approach allowed us to refine the understanding of the distribution of hopanoid biosynthesis in bacteria. The major revision relative to prior studies on hopanoid distribution 14 , 44 – 46 is that we observe minimal SHC conservation within bacillota. The only SHC-producing species we identified in this phylum are members of the alicyclobacillaceae family such as Alicyclobacillus acidocaldarius , which is the only experimentally verified hopanoid-producing firmicute 47 , 48 . The SHC-like enzymes from bacillota that were classified as SHC in previous works appear to be sporulenol cyclases (sqhC) that do not produce hopanoids; these squalene cyclase families appear to have evolved early in triterpene cyclase evolution but represent functionally distinct groups 49 . We similarly distinguish hopanoid-producing strains from sterol-producing strains, which is particularly relevant to recent analyses of the planctomycetota. New work has proposed that a novel planctomycete isolate from marine algae, Crateriforma conspicua Mal65, produces hopanoids to support oxygen-tolerant symbiotic nitrogen fixation 50 . However, we do not observe SHC in any members of the pirellulales order to which C. conspicua belongs and instead observed only terpene cyclases involved in sterol production. Given that sterol production is known to occur and is essential in the model planctomycete Gemmata obscuriglobus 51 , our results suggest it is unlikely that C. conspicua produces hopanoids. Our conclusions regarding the distribution of methylated hopanoids largely agree with prior work. For HpnR, our analysis slightly expands the range of 3Me-hopanoid-producing organisms to include members of the Paraburkholderia genus; however, 3Me-hopanoid production has never been reported in these species. As there is only one hpnR gene validated by knockout or biochemical analysis in the literature, limiting the benchmarking confidence, whether Paraburkholderia are true 3Me-hopanoid-producers will await future experimental analysis. Regarding 2Me-hopanoids generated by HpnP, our benchmarked distribution appears similar to that of Hoshino et al. 14 in cyanobacteria; however, comparing these distributions required extensive re-analysis due to the use of different taxon nomenclatures and taxon classification criteria. Though whole-genome classification approaches such as GTDB-Tk are widely accepted as more accurate, the resulting classifications can be substantially different from those supported by bacterial taxonomic authorities such as LPSN. Hundreds of new and re-classifications have been formally proposed based on whole-genome definitions 52 , but these name changes have yet to be widely adopted by researchers. We encourage future hopanoid researchers to prioritize the more accurate whole genome-based taxonomies to facilitate comparison of results across publications. In addition to understanding their distribution, we analyzed the phylogenetic relationships of hopanoid-related enzymes to predict their origins. These relationships suggested that unmethylated hopanoids, including their C 35 varieties, originated in an ancestor of marine alphaproteobacteria. This was further supported by our determination that hopanoids are strongly associated with adaptations to high osmolarity environments, methanotrophy, and nitrogen cycling. The link between hopanoids and osmotic stress tolerance is expected given the similarity of hopanoids to eukaryotic sterols, which are known to fortify cell membranes to reduce permeability and resist force-mediated lysis. It is well established that hopanoids similarly condense and thicken in vitro lipid bilayers 1 , 2 , and though the relevance of these studies to living membranes is not completely clear, sensitivity to osmotic stress has been reported in nearly every hopanoid-deficient mutant tested 20 , 53 – 56 . We speculate that survival of high salinity environments was a major driver of hopanoid evolution. The enrichment of single carbon- and nitrogen-metabolism in hopanoid-producing organisms is more difficult to parse. Many of these processes rely on isoprenoid quinones or radical SAM enzymes, and it is possible that their correlation with hopanoids arises from the use of shared substrates rather than a functional relationship. Presence of hopanoids in marine nitrite-oxidizing bacteria (NOB) has been recently emphasized in hopanoid geobiology literature 10 – 12 , and in our growth assays, SHC is required for efficient nitrite utilization in the terrestrial nitrogen-fixing and denitrifying 57 alphaproteobacterium Bradyrhizobium diazoefficiens . Whether hopanoids participate directly in nitrite utilization will require future experiments. Though previous work identified an enrichment of hopanoids in plant-associated nitrogen-fixing bacteria 9 , subsequent work suggests that hopanoids do not play a direct role in nitrogen fixation 20 , 58 . Rather, we believe that hopanoids and nitrogen fixation are correlated because they are both common to low oxygen environments, given that hopanoids (unlike sterols) do not require oxygen for their biosynthesis and that oxygen is inhibitory to the nitrogen fixation machinery. The degree to which unmethylated geo-hopane deposition can be attributed to ancestors of marine alphaproteobacteria will be an exciting future area for hopanoid geobiology research. We are only beginning to appreciate the diversity and significance of non-cyanobacterial marine hopanoid producers. Though such species are dominant in marine animal microbiomes 59 – 61 , and symbiotic alphaproteobacteria appear to be major contributors to modern marine nitrogen cycling 62 , relatively few complete marine alphaproteobacterial genomes are available. Incorporating our growing understanding of the biology marine alphaproteobacteria, ideally coupled with genetic tools, will aid our knowledge of the role of hopanoids in these origins. The origins of the 2Me- and 3Me-geo-hopanes are less clear. We could not identify a single phylum of origin for either HpnR or HpnP proteins, though all of the statistically supported potential root positions for both protein families were all within the clade of Figure 1 containing the pseudomonadota, acidobacteriota, and desulfobacterota. Our HpnP tree is similar to the HpnP tree published by Hoshino and colleagues 14 , as both trees indicates that the earliest branching HpnP sequences are from phyla other than cyanobacteriota or alphaproteobacteria. The other group’s conclusion that 2Me-geo-hopanes dated prior to ∼750 Ma were deposited by cyanobacteria is based on a study on the origin of the rhizobiales order of alphaproteobacteria that proposed that the hopanoid-producing families of alphaproteobacteria did not emerge until 750 Ma or later 15 . In our analyses, we observe high levels of HpnP conservation in the following alphaproteobacterial families: xanthobacteraceae, acetobacteraceae, beijerinckaceae, and reyranellaceae; we note the latter family does not appear to be represented in the Hoshino et al. dataset, perhaps due to differences in HpnP annotation approaches. A new geologic timescale of evolution across all bacteria suggests the earliest-branching of these families is the reyranellaceae, originating between 783-1212 Ma (95% confidence interval) 63 . Though this supports the conclusion that the oldest 2Me-geo-hopanes were more likely deposited by cyanobacteria than alphaproteobacteria, we restrict that conclusion to geo-hopanes dated prior to ∼1200 Ma. Whether cyanobacteria were the first of any phyla to produce hopanoids, however, cannot be resolved by molecular dating. The same pan-bacterial evolution timeline indicates that the desulfuromonadia class of the desulfobacterota F phlyum originated 937-2040 Ma 63 . HpnP is conserved highly in this class, and this phylum was the most highly supported root position for HpnP in our phylogenetic analyses. Beyond implications for geo-hopane interpretation, our work suggests new avenues for hopanoid cell biology. We identify four new uncharacterized but likely hopanoid-related proteins: the previously annotated HpnL and HpnM families and candidate proteins HpnS and HpnT, which are new to this study. Using the latest structure prediction tools, we predict that HpnS, HpnM, and HpnT participate in hopanoid trafficking across proteobacterial cell envelopes, interacting with the inner membrane hopanoid transporter HpnN in a manner analogous to the MlaC-MlaA-MlaBDEF phospholipid transport system. New analysis of mlaC -like gene copy numbers across bacteria 64 supports this notion. Tripathi et al. identified a subset of pseudomonadota that contain multiple copies of mlaC -like genes. In these species, which we find to be SHC-containing organisms, one mlaC -like copy is adjacent to other mla genes, whereas additional mlaC -like copies typically are adjacent to hopanoid-associated genes. The bacterial genera containing secondary, hpnN -adjacent mlaC -like genes in Tripathi et al. largely overlap with the genera in which we’ve identified hpnM genes, suggesting that we are referring to the same loci. This opens up future directions for cellular, structural, and biochemical assays to test this hypothetical transport pathway. If our functional predictions are verified, it will be interesting to study how these protein families originated, both relative to the Mla proteins and in the alpha- versus gamma-proteobacteria. The function of HpnL remains unclear. Its closest relatives are the AglD family of lysylphosphatidylglycerol (lysyl-PG) synthases, which covalently link the carboxyl group of lysine to the terminal hydroxyl in the glycerol moiety of PG 65 , 66 , increasing the positive charge of the bilayer to improve resistance to cationic antimicrobial peptides. It is possible that HpnL might use a similar chemistry to join amino acids or other carboxylated compounds to the polyol groups of extended hopanoids such as bacteriohopanetetral (BHT), though the structure of BHT is quite distinct from the PG substrate of AglD. Covalent attachment of BHT to the amino acid derivative phenyl-acetic acid (PAA), an auxin-like compound and weak phytohormone, has been reported in actinobacterial Frankia spp. 67 , 68 , suggesting other amino acid-hopanoid conjugates exist. However, the HpnL-associated protein domain (TIGR03476) is found predominantly in pseudomonadota, and to our knowledge PAA-BHT not been found in this phylum. The function of HpnL awaits future genetic characterization. ACKNOWLEGEMENTS This work was supported by NIH grants R00GM126141 and R35GM147015 (to B.J.B) and Carnegie Institution for Science endowment funds. Computational analyses were performed using Caltech’s Resnick High Performance Computing Center and the NIH Biowulf HPC. We thank members of the Belin lab, Dr. Karina Gutiérrez-García (University of Arizona), Dr. Will Ludington (Johns Hopkins University), Dr. Damian Ekiert (Johns Hopkins University), and Dr. Dianne Newman (Caltech) for helpful discussions and feedback. We are grateful to Carnegie Embryology’s IT, front office, and facilities support staff for making our work possible. The contributions of NIH authors were made as a part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered works of the United States Government. However, the findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the US Department of Health and Human Services. Funder Information Declared National Institutes of Health , R00GM126141 , National Institutes of Health , R35GM147015 Carnegie Institution for Science, https://ror.org/04jr01610 Footnotes New analyses performed; Figures 1-3 revised; supplemental files added; supplemental files updated. REFERENCES ↵ Belin , B. J. et al. Hopanoid lipids: from membranes to plant-bacteria interactions . Nat Rev Microbiol 16 , 304 – 315 ( 2018 ). doi: 10.1038/nrmicro.2017.173 OpenUrl CrossRef PubMed ↵ Mitchison-Field , L. M. & Belin , B. J . Bacterial lipid biophysics and membrane organization . Curr Opin Microbiol 74 , 102315 ( 2023 ). doi: 10.1016/j.mib.2023.102315 OpenUrl CrossRef ↵ Summons , R. E. , Jahnke , L. L. , Hope , J. M. & Logan , G. A . 2-Methylhopanoids as biomarkers for cyanobacterial oxygenic photosynthesis . Nature 400 , 554 – 557 ( 1999 ). doi: 10.1038/23005 OpenUrl CrossRef PubMed Web of Science ↵ Newman , D. K. , Neubauer , C. , Ricci , J. N. , Wu , C.-H. & Pearson , A . Cellular and Molecular Biological Approaches to Interpreting Ancient Biomarkers . Annual Review of Earth and Planetary Sciences 44 , 493 – 522 ( 2016 ). doi: 10.1146/annurev-earth-050212-123958 OpenUrl CrossRef ↵ Kuypers , M. M. M. , van Breugel , Y. , Schouten , S. , Erba , E. & Damsté , J. S. S . N2-fixing cyanobacteria supplied nutrient N for Cretaceous oceanic anoxic events . Geology 32 , 853 – 856 ( 2004 ). doi: 10.1130/g20458.1 OpenUrl Abstract / FREE Full Text ↵ Pearson , A. & Rusch , D. B . Distribution of microbial terpenoid lipid cyclases in the global ocean metagenome . The ISME Journal 3 , 352 – 363 ( 2009 ). doi: 10.1038/ismej.2008.116 OpenUrl CrossRef PubMed Web of Science ↵ Naafs , B. D. A. , Bianchini , G. , Monteiro , F. M. & Sánchez-Baracaldo , P . The occurrence of 2-methylhopanoids in modern bacteria and the geological record . Geobiology 20 , 41 – 59 ( 2022 ). doi: 10.1111/gbi.12465 OpenUrl CrossRef PubMed ↵ Ricci , J. N. , Michel , A. J. & Newman , D. K . Phylogenetic analysis of HpnP reveals the origin of 2-methylhopanoid production in Alphaproteobacteria . Geobiology 13 , 267 – 277 ( 2015 ). doi: 10.1111/gbi.12129 OpenUrl CrossRef GeoRef PubMed ↵ Ricci , J. N. et al. Diverse capacity for 2-methylhopanoid production correlates with a specific ecological niche . Isme j 8 , 675 – 684 ( 2014 ). doi: 10.1038/ismej.2013.191 OpenUrl CrossRef PubMed ↵ Mueller , A. J. , Daebeler , A. , Herbold , C. W. , Kirkegaard , R. H. & Daims , H . Cultivation and genomic characterization of novel and ubiquitous marine nitrite-oxidizing bacteria from the Nitrospirales . The ISME Journal 17 , 2123 – 2133 ( 2023 ). doi: 10.1038/s41396-023-01518-6 OpenUrl CrossRef PubMed Elling , F. J. et al. Vitamin B 12 -dependent biosynthesis ties amplified 2-methylhopanoid production during oceanic anoxic events to nitrification . Proceedings of the National Academy of Sciences 117 , 32996 – 33004 ( 2020 ). doi: 10.1073/pnas.2012357117 OpenUrl Abstract / FREE Full Text ↵ Elling , F. J. et al. Marine and terrestrial nitrifying bacteria are sources of diverse bacteriohopanepolyols . Geobiology 20 , 399 – 420 ( 2022 ). doi: 10.1111/gbi.12484 OpenUrl CrossRef PubMed ↵ Cornejo-Castillo , F. M. & Zehr , J. P . Hopanoid lipids may facilitate aerobic nitrogen fixation in the ocean . Proceedings of the National Academy of Sciences 116 , 18269 – 18271 ( 2019 ). doi: 10.1073/pnas.1908165116 OpenUrl Abstract / FREE Full Text ↵ Hoshino , Y. et al. Genetics re-establish the utility of 2-methylhopanes as cyanobacterial biomarkers before 750 million years ago . Nature Ecology & Evolution 7 , 2045 – 2054 ( 2023 ). doi: 10.1038/s41559-023-02223-5 OpenUrl CrossRef PubMed ↵ Wang , S. , Meade , A. , Lam , H.-M. & Luo , H . Evolutionary Timeline and Genomic Plasticity Underlying the Lifestyle Diversity in Rhizobiales . mSystems 5 , 10.1128/msystems.00438–00420 ( 2020 ). doi: 10.1128/msystems.00438-20 OpenUrl CrossRef ↵ Chen , I.-M. A. et al. The IMG/M data management and analysis system v.6.0: new tools and advanced capabilities . Nucleic Acids Research 49 , D751 – D763 ( 2020 ). doi: 10.1093/nar/gkaa939 OpenUrl CrossRef ↵ Parks , D. H. et al. GTDB: an ongoing census of bacterial and archaeal diversity through a phylogenetically consistent, rank normalized and complete genome-based taxonomy . Nucleic Acids Research 50 , D785 – D794 ( 2021 ). doi: 10.1093/nar/gkab776 OpenUrl CrossRef PubMed ↵ Freese , H. M. , Meier-Kolthoff , J. P. , Sardà Carbasse , J. , Afolayan , Ayorinde O. & Göker , M. TYGS and LPSN in 2025: a Global Core Biodata Resource for genome-based classification and nomenclature of prokaryotes within DSMZ Digital Diversity . Nucleic Acids Research 54 , D884 – D891 ( 2025 ). doi: 10.1093/nar/gkaf1110 OpenUrl CrossRef ↵ Tookmanian , E. et al. Hopanoids Confer Robustness to Physicochemical Variability in the Niche of the Plant Symbiont Bradyrhizobium diazoefficiens . J Bacteriol 204 , e0044221 ( 2022 ). doi: 10.1128/jb.00442-21 OpenUrl CrossRef PubMed ↵ Pan , H. , Shim , A. , Lubin , M. B. & Belin , B. J . Hopanoid lipids promote soybean-Bradyrhizobium symbiosis . mBio 15 , e0247823 ( 2024 ). doi: 10.1128/mbio.02478-23 OpenUrl CrossRef PubMed ↵ Barison , N. et al. Protein HP1028 from the human pathogen Helicobacter pylori belongs to the lipocalin family . Acta Crystallogr D Biol Crystallogr 69 , 1387 – 1394 ( 2013 ). doi: 10.1107/s0907444913008160 OpenUrl CrossRef PubMed ↵ Yang , H. et al. Lipocalin family proteins and their diverse roles in cardiovascular disease . Pharmacology & Therapeutics 244 ( 2023 ). doi: 10.1016/j.pharmthera.2023.108385 OpenUrl CrossRef ↵ Malinverni , J. C. & Silhavy , T. J . An ABC transport system that maintains lipid asymmetry in the gram-negative outer membrane . Proc Natl Acad Sci U S A 106 , 8009 – 8014 ( 2009 ). doi: 10.1073/pnas.0903229106 OpenUrl Abstract / FREE Full Text ↵ Ekiert , D. C. et al. Architectures of Lipid Transport Systems for the Bacterial Outer Membrane . Cell 169 , 273 – 285 .e217 ( 2017 ). doi: 10.1016/j.cell.2017.03.019 OpenUrl CrossRef PubMed ↵ Al Jaberi , S. et al. Lipocalin-2: Structure, function, distribution and role in metabolic disorders . Biomedicine & Pharmacotherapy 142 ( 2021 ). doi: 10.1016/j.biopha.2021.112002 OpenUrl CrossRef PubMed ↵ Coudray , N. et al. Structure of bacterial phospholipid transporter MlaFEDB with substrate bound . Elife 9 ( 2020 ). doi: 10.7554/eLife.62518 OpenUrl CrossRef PubMed ↵ Abellón-Ruiz , J. et al. Structural basis for maintenance of bacterial outer membrane lipid asymmetry . Nat Microbiol 2 , 1616 – 1623 ( 2017 ). doi: 10.1038/s41564-017-0046-x OpenUrl CrossRef PubMed ↵ Chong , Z. S. , Woo , W. F. & Chng , S. S . Osmoporin OmpC forms a complex with MlaA to maintain outer membrane lipid asymmetry in Escherichia coli . Mol Microbiol 98 , 1133 – 1146 ( 2015 ). doi: 10.1111/mmi.13202 OpenUrl CrossRef PubMed ↵ Abramson , J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3 . Nature 630 , 493 – 500 ( 2024 ). doi: 10.1038/s41586-024-07487-w OpenUrl CrossRef PubMed ↵ Yeow , J. , Luo , M. & Chng , S. S . Molecular mechanism of phospholipid transport at the bacterial outer membrane interface . Nat Commun 14 , 8285 ( 2023 ). doi: 10.1038/s41467-023-44144-8 OpenUrl CrossRef PubMed ↵ Lissin , A. et al. StrainInfo—the central database for linked microbial strain identifiers . Database 2025 ( 2025 ). doi: 10.1093/database/baaf059 OpenUrl CrossRef ↵ Jones , P. et al. InterProScan 5: genome-scale protein function classification . Bioinformatics 30 , 1236 – 1240 ( 2014 ). doi: 10.1093/bioinformatics/btu031 OpenUrl CrossRef PubMed Web of Science ↵ Aramaki , T. et al. KofamKOALA: KEGG Ortholog assignment based on profile HMM and adaptive score threshold . Bioinformatics 36 , 2251 – 2252 ( 2019 ). doi: 10.1093/bioinformatics/btz859 OpenUrl CrossRef PubMed ↵ Fu , L. , Niu , B. , Zhu , Z. , Wu , S. & Li , W . CD-HIT: accelerated for clustering the next-generation sequencing data . Bioinformatics 28 , 3150 – 3152 ( 2012 ). doi: 10.1093/bioinformatics/bts565 OpenUrl CrossRef PubMed Web of Science ↵ Edgar , R. C . Muscle5: High-accuracy alignment ensembles enable unbiased assessments of sequence homology and phylogeny . Nature Communications 13 , 6968 ( 2022 ). doi: 10.1038/s41467-022-34630-w OpenUrl CrossRef PubMed ↵ Minh , B. Q. et al. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era . Molecular Biology and Evolution 37 , 1530 – 1534 ( 2020 ). doi: 10.1093/molbev/msaa015 OpenUrl CrossRef PubMed ↵ Bianchini , G. & Sánchez-Baracaldo , P . TreeViewer: Flexible, modular software to visualise and manipulate phylogenetic trees . Ecol Evol 14 , e10873 ( 2024 ). doi: 10.1002/ece3.10873 OpenUrl CrossRef PubMed ↵ Chklovski , A. , Parks , D. H. , Woodcroft , B. J. & Tyson , G. W . CheckM2: a rapid, scalable and accurate tool for assessing microbial genome quality using machine learning . Nature Methods 20 , 1203 – 1212 ( 2023 ). doi: 10.1038/s41592-023-01940-w OpenUrl CrossRef PubMed ↵ Chaumeil , P. , Mussig , A. , Hugenholtz , P. & Parks , D . GTDB-Tk v2: memory friendly classification with the genome taxonomy database . Bioinformatics 38 , 5315 – 5316 ( 2022 ). doi: 10.1093/bioinformatics/btac672 OpenUrl CrossRef PubMed ↵ Kalyaanamoorthy , S. , Minh , B. Q. , Wong , T. K. F. , von Haeseler , A. & Jermiin , L. S . ModelFinder: fast model selection for accurate phylogenetic estimates . Nature Methods 14 , 587 – 589 ( 2017 ). doi: 10.1038/nmeth.4285 OpenUrl CrossRef PubMed ↵ Hoang , D. T. , Chernomor , O. , von Haeseler , A. , Minh , B. Q. & Vinh , L. S . UFBoot2: Improving the Ultrafast Bootstrap Approximation . Molecular Biology and Evolution 35 , 518 – 522 ( 2017 ). doi: 10.1093/molbev/msx281 OpenUrl CrossRef PubMed ↵ Naser-Khdour , S. , Quang Minh , B. & Lanfear , R . Assessing Confidence in Root Placement on Phylogenies: An Empirical Study Using Nonreversible Models for Mammals . Systematic Biology 71 , 959 – 972 ( 2021 ). doi: 10.1093/sysbio/syab067 OpenUrl CrossRef ↵ Ronquist , F. et al. MrBayes 3.2: Efficient Bayesian Phylogenetic Inference and Model Choice Across a Large Model Space . Systematic Biology 61 , 539 – 542 ( 2012 ). doi: 10.1093/sysbio/sys029 OpenUrl CrossRef PubMed ↵ Racolta , S. , Juhl , P. B. , Sirim , D. & Pleiss , J . The triterpene cyclase protein family: A systematic analysis . Proteins: Structure, Function, and Bioinformatics 80 , 2009 – 2019 ( 2012 ). https://doi-org.proxy1.library.jhu.edu/10.1002/prot.24089 OpenUrl Gong , L. et al. pH Shaping the Composition of sqhC-Containing Bacterial Communities . Geomicrobiology Journal 32 , 433 – 444 ( 2015 ). doi: 10.1080/01490451.2014.950363 OpenUrl CrossRef ↵ Mukherjee , A. , Tikariha , H. , Bandla , A. , Pavagadhi , S. & Swarup , S . Global analyses of biosynthetic gene clusters in phytobiomes reveal strong phylogenetic conservation of terpenes and aryl polyenes . mSystems 8 , e00387 – 00323 ( 2023 ). doi: 10.1128/msystems.00387-23 OpenUrl CrossRef PubMed ↵ Poralla , K. , Hartner , T. & Kannenberg , E . Effect of temperature and pH on the hopanoid content of Bacillus acidocaldarius . FEMS Microbiology Letters 23 , 253 – 256 ( 1984 ). OpenUrl CrossRef Web of Science ↵ Ochs , D. , Kaletta , C. , Entian , K. D. , Beck-Sickinger , A. & Poralla , K . Cloning, expression, and sequencing of squalene-hopene cyclase, a key enzyme in triterpenoid metabolism . J Bacteriol 174 , 298 – 302 ( 1992 ). doi: 10.1128/jb.174.1.298-302.1992 OpenUrl Abstract / FREE Full Text ↵ McShea , H. S. , Viens , R. A. , Olagunju , B. O. , Giner , J.-L. & Welander , P. V . The Evolutionary History and Modern Diversity of Triterpenoid Cyclases . Molecular Biology and Evolution 42 ( 2025 ). doi: 10.1093/molbev/msaf203 OpenUrl CrossRef ↵ Zhang , Z. , Wang , Z. , Teng , P. , Yu , T. & Zhang , Y. Oxygen-tolerant nitrogen fixation in a marine alga-colonizing Planctomycetota . Applied and Environmental Microbiology 91 , e01316 – 01325 ( 2025 ). doi: 10.1128/aem.01316-25 OpenUrl CrossRef PubMed ↵ Rivas-Marin , E. et al. Essentiality of sterol synthesis genes in the planctomycete bacterium Gemmata obscuriglobus . Nature Communications 10 , 1 – 6 ( 2019 ). doi: 10.1038/s41467-019-10983-7 OpenUrl CrossRef PubMed ↵ Chuvochina , M. et al. Proposal of names for 329 higher rank taxa defined in the Genome Taxonomy Database under two prokaryotic codes . FEMS Microbiol Lett 370 ( 2023 ). doi: 10.1093/femsle/fnad071 OpenUrl CrossRef ↵ Tookmanian , E. et al. Hopanoids Confer Robustness to Physicochemical Variability in the Niche of the Plant Symbiont Bradyrhizobium diazoefficiens . Journal of Bacteriology 204 , e00442 – 00421 ( 2022 ). doi: 10.1128/jb.00442-21 OpenUrl CrossRef PubMed Welander , P. V. et al. Hopanoids play a role in membrane integrity and pH homeostasis in Rhodopseudomonas palustris TIE-1 . J Bacteriol 191 , 6145 – 6156 ( 2009 ). doi: 10.1128/jb.00460-09 OpenUrl Abstract / FREE Full Text Silipo , A. et al. Covalently linked hopanoid-lipid A improves outer-membrane resistance of a Bradyrhizobium symbiont of legumes . Nature Communications 5 , 5106 ( 2014 ). doi: 10.1038/ncomms6106 OpenUrl CrossRef PubMed ↵ Garby Tamsyn , J ., et al. Lack of Methylated Hopanoids Renders the Cyanobacterium Nostoc punctiforme Sensitive to Osmotic and pH Stress . Applied and Environmental Microbiology 83 , e00777 – 00717 ( 2017 ). doi: 10.1128/AEM.00777-17 OpenUrl Abstract / FREE Full Text ↵ Bedmar , E. J. , Robles , E. F. & Delgado , M. J . The complete denitrification pathway of the symbiotic, nitrogen-fixing bacterium Bradyrhizobium japonicum . Biochemical Society Transactions 33 , 141 – 144 ( 2005 ). doi: 10.1042/BST0330141 OpenUrl Abstract / FREE Full Text ↵ Belin , B. J. , Tookmanian , E. M. , de Anda , J. , Wong , G. C. L. & Newman , D. K . Extended Hopanoid Loss Reduces Bacterial Motility and Surface Attachment and Leads to Heterogeneity in Root Nodule Growth Kinetics in a Bradyrhizobium-Aeschynomene Symbiosis . Molecular Plant-Microbe Interactions® 32 , 1415 – 1428 ( 2019 ). doi: 10.1094/MPMI-04-19-0111-R OpenUrl CrossRef PubMed ↵ Sehnal , L. et al. Microbiome Composition and Function in Aquatic Vertebrates: Small Organisms Making Big Impacts on Aquatic Animal Health . Frontiers in Microbiology Volume 12 - 2021 ( 2021 ). doi: 10.3389/fmicb.2021.567408 OpenUrl CrossRef PubMed Bik , E. M. et al. Marine mammals harbor unique microbiotas shaped by and yet distinct from the sea . Nature Communications 7 , 10516 ( 2016 ). doi: 10.1038/ncomms10516 OpenUrl CrossRef PubMed ↵ Nguyen , V. H. et al. Identification, classification, and functional characterization of novel sponge-associated acidimicrobiial species . Systematic and Applied Microbiology 46 , 126426 ( 2023 ). doi: 10.1016/j.syapm.2023.126426 OpenUrl CrossRef ↵ Tschitschko , B. et al. Rhizobia–diatom symbiosis fixes missing nitrogen in the ocean . Nature 630 , 899 – 904 ( 2024 ). doi: 10.1038/s41586-024-07495-w OpenUrl CrossRef ↵ Davín , A. A. et al. A geological timescale for bacterial evolution and oxygen adaptation . Science 388 , eadp1853 ( 2025 ). doi: 10.1126/science.adp1853 OpenUrl CrossRef PubMed ↵ Tripathi , R. , Ayekpam , D. & Kanaujia , S. P . Unveiling multiple copies of MlaC highlights its multifaceted nature . Archives of Microbiology 207 , 107 ( 2025 ). doi: 10.1007/s00203-025-04308-0 OpenUrl CrossRef ↵ Staubitz , P. , Neumann , H. , Schneider , T. , Wiedemann , I. & Peschel , A . MprF-mediated biosynthesis of lysylphosphatidylglycerol, an important determinant in staphylococcal defensin resistance . FEMS Microbiol Lett 231 , 67 – 71 ( 2004 ). doi: 10.1016/s0378-1097(03)00921-2 OpenUrl CrossRef PubMed Web of Science ↵ Oku , Y. , Kurokawa , K. , Ichihashi , N. & Sekimizu , K . Characterization of the Staphylococcus aureus mprF gene, involved in lysinylation of phosphatidylglycerol . Microbiology 150 , 45 – 51 ( 2004 ). doi: 10.1099/mic.0.26706-0 OpenUrl CrossRef PubMed Web of Science ↵ Berry , A. M. , Moreau , R. A. & Jones , A. D . Bacteriohopanetetrol: Abundant Lipid in Frankia Cells and in Nitrogen-Fixing Nodule Tissue 1 . Plant Physiology 95 , 111 – 115 ( 1991 ). doi: 10.1104/pp.95.1.111 OpenUrl Abstract / FREE Full Text ↵ Kleemann , G. , Alskog , G. , Berry , A. M. & Huss-Danell , K . Lipid composition and nitrogenase activity of symbioticFrankia (Alnus incana) in response to different oxygen concentrations . Protoplasma 183 , 107 – 115 ( 1994 ). doi: 10.1007/BF01276818 OpenUrl CrossRef Perzl , M. , Müller , P. , Poralla , K. & Kannenberg , E. L . Squalene-hopene cyclase from Bradyrhizobium japonicum: cloning, expression, sequence analysis and comparison to other triterpenoid cyclases . Microbiology (Reading ) 143 ( Pt 4 ) , 1235 – 1242 ( 1997 ). doi: 10.1099/00221287-143-4-1235 OpenUrl CrossRef PubMed Welander , P. V. et al. Identification and characterization of Rhodopseudomonas palustris TIE-1 hopanoid biosynthesis mutants . Geobiology 10 , 163 – 177 ( 2012 ). doi: 10.1111/j.1472-4669.2011.00314.x OpenUrl CrossRef PubMed Web of Science Liu , W. et al. Ribosylhopane, a novel bacterial hopanoid, as precursor of C35 bacteriohopanepolyols in Streptomyces coelicolor A3(2) . Chembiochem 15 , 2156 – 2161 ( 2014 ). doi: 10.1002/cbic.201402261 OpenUrl CrossRef PubMed Bradley , A. S. , Pearson , A. , Sáenz , J. P. & Marx , C. J . Adenosylhopane: The first intermediate in hopanoid side chain biosynthesis . Organic Geochemistry 41 , 1075 – 1081 ( 2010 ). doi: 10.1016/j.orggeochem.2010.07.003 OpenUrl CrossRef GeoRef Schmerk , C. L. et al. Elucidation of the Burkholderia cenocepacia hopanoid biosynthesis pathway uncovers functions for conserved proteins in hopanoid-producing bacteria . Environ Microbiol 17 , 735 – 750 ( 2015 ). doi: 10.1111/1462-2920.12509 OpenUrl CrossRef Doughty , D. M. et al. The RND-family transporter, HpnN, is required for hopanoid localization to the outer membrane of Rhodopseudomonas palustris TIE-1 . Proc Natl Acad Sci U S A 108 , E1045 – 1051 ( 2011 ). doi: 10.1073/pnas.1104209108 OpenUrl Abstract / FREE Full Text Kumar , N. et al. Crystal structures of the Burkholderia multivorans hopanoid transporter HpnN . Proceedings of the National Academy of Sciences 114 , 6557 – 6562 ( 2017 ). doi: 10.1073/pnas.1619660114 OpenUrl Abstract / FREE Full Text Welander , P. V. , Coleman , M. L. , Sessions , A. L. , Summons , R. E. & Newman , D. K . Identification of a methylase required for 2-methylhopanoid production and implications for the interpretation of sedimentary hopanes . Proceedings of the National Academy of Sciences 107 , 8537 – 8542 ( 2010 ). doi: 10.1073/pnas.0912949107 OpenUrl Abstract / FREE Full Text Welander , P. V. & Summons , R. E . Discovery, taxonomic distribution, and phenotypic characterization of a gene required for 3-methylhopanoid production . Proc Natl Acad Sci U S A 109 , 12905 – 12910 ( 2012 ). doi: 10.1073/pnas.1208255109 OpenUrl Abstract / FREE Full Text ↵ Käll , L. , Krogh , A. & Sonnhammer , E. L . A combined transmembrane topology and signal peptide prediction method . J Mol Biol 338 , 1027 – 1036 ( 2004 ). doi: 10.1016/j.jmb.2004.03.016 OpenUrl CrossRef PubMed Web of Science View the discussion thread. Back to top Previous Next Posted March 25, 2026. 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 Origin and transport of bacterial hopanoids 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 Origin and transport of bacterial hopanoids Evan C. Lawrence , Huiqiao Pan , Noah Ollikainen , Brittany J. Belin bioRxiv 2025.07.29.667474; doi: https://doi.org/10.1101/2025.07.29.667474 Share This Article: Copy Citation Tools Origin and transport of bacterial hopanoids Evan C. Lawrence , Huiqiao Pan , Noah Ollikainen , Brittany J. Belin bioRxiv 2025.07.29.667474; doi: https://doi.org/10.1101/2025.07.29.667474 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Microbiology Subject Areas All Articles Animal Behavior and Cognition (7637) Biochemistry (17705) Bioengineering (13899) Bioinformatics (41968) Biophysics (21460) Cancer Biology (18603) Cell Biology (25526) Clinical Trials (138) Developmental Biology (13385) Ecology (19910) Epidemiology (2067) Evolutionary Biology (24327) Genetics (15614) Genomics (22513) Immunology (17741) Microbiology (40423) Molecular Biology (17193) Neuroscience (88646) Paleontology (667) Pathology (2835) Pharmacology and Toxicology (4825) Physiology (7647) Plant Biology (15160) Scientific Communication and Education (2046) Synthetic Biology (4302) Systems Biology (9825) Zoology (2271)