Discovery of novel bacterial queuine salvage enzymes and pathways in human pathogens

Edited by Tina M. Henkin, The Ohio State University, Columbus, OH, and approved August 1, 2019 (received for review June 16, 2019)
September 3, 2019
116 (38) 19126-19135


Queuosine (Q) is a tRNA modification found in eukaryotes and bacteria that plays an important role in translational efficiency and accuracy. Queuine (q), the Q nucleobase, is increasingly appreciated as an important micronutrient that contributes to human health. We describe here that q salvage pathways exist in bacteria, including many pathogens and host-associated organisms, suggesting a direct competition for the q precursor in the human gut microbiome. We also show how a rational use of comparative genomics can lead to the discovery of novel types of enzymatic reactions, illustrated by the discovery of the queuine lyase enzyme.


Queuosine (Q) is a complex tRNA modification widespread in eukaryotes and bacteria that contributes to the efficiency and accuracy of protein synthesis. Eukaryotes are not capable of Q synthesis and rely on salvage of the queuine base (q) as a Q precursor. While many bacteria are capable of Q de novo synthesis, salvage of the prokaryotic Q precursors preQ0 and preQ1 also occurs. With the exception of Escherichia coli YhhQ, shown to transport preQ0 and preQ1, the enzymes and transporters involved in Q salvage and recycling have not been well described. We discovered and characterized 2 Q salvage pathways present in many pathogenic and commensal bacteria. The first, found in the intracellular pathogen Chlamydia trachomatis, uses YhhQ and tRNA guanine transglycosylase (TGT) homologs that have changed substrate specificities to directly salvage q, mimicking the eukaryotic pathway. The second, found in bacteria from the gut flora such as Clostridioides difficile, salvages preQ1 from q through an unprecedented reaction catalyzed by a newly defined subgroup of the radical-SAM enzyme family. The source of q can be external through transport by members of the energy-coupling factor (ECF) family or internal through hydrolysis of Q by a dedicated nucleosidase. This work reinforces the concept that hosts and members of their associated microbiota compete for the salvage of Q precursors micronutrients.
The gut microbiome provides a variety of micronutrients important for human health (1). Among them is queuine (q), recently designated “a longevity vitamin” (2). In eukaryotes, q is the precursor base to the queuosine (Q) tRNA modification (3). Mammals solely rely on q salvage from dietary sources and from their gut microbiota to synthesize Q-modified tRNAs (46). To obtain Q34-tRNA, q is exchanged with the guanine (G) at the wobble position of tRNAs containing G34U35N36 anticodons (Asp, Asn, Tyr, and His) (7), in a reaction catalyzed by a eukaryotic type tRNA(34) guanine transglycosylase (TGT) (8, 9). The physiological importance of the Q modification is not fully understood. Although it is not essential in tested cells under normal conditions (4, 1014), Q plays a role in regulation of translation, cell proliferation, stress response, and cell signaling (6, 1517). Deficiency of Q-tRNA level correlates with diseases including tumor growth (reviewed in ref. 6), encephalomyelitis (18), and leukemia (19). Recently, Q-tRNA levels in human and mice cells have been shown to control translational speed of Q‐decoded codons as well as near‐cognate codons (20). Taken together, q is a micronutrient that links nutrition to translation efficiency.
Unlike eukaryotes, most bacteria perform de novo synthesis of Q via a pathway similar to the one elucidated in Escherichia coli, recently summarized by Hutinet et al. (21) and presented in Fig. 1A. The TGT enzyme, which is responsible for the base exchange, is the signature enzyme in the Q biosynthesis pathway. Major differences are found between bacterial and eukaryotic TGT enzymes (22). In eubacteria, TGT functions as a homodimer to incorporate preQ1 into the anticodon wobble of tRNAs (23, 24), and maturation to Q is then needed. In contrast, the eukaryotic TGT (eTGT) inserts the q base in tRNAs, yielding Q directly. eTGTs are heterodimeric enzymes composed of a catalytic subunit (queuine tRNA-ribosyltransferase catalytic subunit 1 or QTRT1) and an accessory subunit (queuine tRNA-ribosyltransferase accessory subunit 2 or QTRT2) (25). Key differences defining substrate specificities of the prokaryotic and eukaryotic TGTs are in the substrate binding pocket where Val233 (Zymomonas mobilis numbering) is replaced by a glycine and Cys158 is replaced by a valine. These changes allow for the accommodation of q in eTGT as opposed to preQ1 (24, 26, 27).
Fig. 1.
Queuosine tRNA modification biosynthesis and predicted salvage pathways. (A) Biosynthesis of the Q modification at position 34 (Q34-tRNA) and preQ0/preQ1 salvage pathway in E. coli. (B) Predicted Q34-tRNA biosynthesis and queuine salvage pathway in C. trachomatis D/UW-3/CX. Red dashed arrows represent uncharacterized reactions. Molecule abbreviations and protein names are described in the main text.
Q synthesis is costly: it starts from GTP (Fig. 1A) and requires many cofactors (reviewed by Hutinet et al. [21]). Additionally, 6 of 8 enzymes in the pathway are metalloenzymes. Hence, even in bacteria that can synthesize Q de novo, salvaging precursors is advantageous, as it reduces metabolic demand. Many organisms have TGT encoding genes but lack those necessary for preQ0 and preQ1 synthesis (28). In those organisms, precursors must be salvaged to obtain Q-modified tRNAs. Significant amounts of free q have been detected in various plant and animal food products (29), and preQ0 is certainly available in natural environments (30). Specific transporters are expected to be involved for salvage to occur. We previously demonstrated that a member of the COG1738 protein family (YhhQ) can transport preQ0 and preQ1 in E. coli (28). Transporters of the energy-coupling factor (ECF) family (QueT and QrtT) have been predicted to be involved in 7-deazapurine salvage, mainly in Gram-positive organisms, but have never been experimentally validated (31). While tRNA turnover is ubiquitous and constant (32) and releases the Q nucleoside and/or its monophosphate derivatives (6), the identities of the nucleotidases, nucleosidases, and other enzymes involved in tRNA degradation for Q precursors salvage remain elusive.
Riboswitches are regulatory elements of messenger RNA molecules. Three classes of preQ1 riboswitches have been described upstream of known Q synthesis genes (33, 34). Exploring genes downstream of predicted preQ1 riboswitches played a key role for the identification of YhhQ as preQ1/preQ0 transporter (28). In the RegPrecise database (35), genes downstream of predicted preQ1 riboswitches include queC, queD, queE, queF, yhhQ, and genes previously associated with Q metabolism but without experimental validation: members of COG1957 (inosine-uridine nucleoside N-ribohydrolase; IunH), COG4708 (predicted membrane protein, QueT), and QrtT (substrate-specific component STY3230 of queuosine-regulated ECF transporter) families (33, 36).
In summary, while the de novo Q biosynthesis pathway is well characterized, many open questions remain regarding the possibility of Q salvage in bacteria. The work presented here elucidates 2 salvage pathways in pathogenic bacteria.


The Chlamydia trachomatis TGT and YhhQ Homologs Are Involved in Q Salvage.

The distribution of Q metabolism genes in more than 10,000 genomes is shown in the PubSEED subsystem (37) “q_salvage_in_Bacteria” ( Previous phylogenomic analyses suggested that an eukaryotic-type salvage pathway, i.e., the ability to directly use q, must exist in bacteria such as Chlamydia, Wolbachia, Corynebacterium, Actinomyces, and Bifidobacterium species (28, 38), as they possess a TGT encoding gene and sometimes predicted Q precursor transporters but lack all of the other genes encoding Q biosynthetic enzymes. This is the case for the intracellular human pathogen Chlamydia trachomatis D/UW-3/CX, even if the presence of Q has never been experimentally validated. The only predicted Q synthesis genes in C. trachomatis are the TGT homolog TGTCt (CT193; UniProt ID O84196) and the YhhQ homolog YhhQCt (CT140; UniProt ID O84142; Fig. 1B). Indeed, its environment (the mammalian cell) does not contain preQ0/preQ1, but q or Q are supposedly available.
We predict that the TGTCt substrate specificity must have switched from preQ1 (classically observed for bacterial TGTs) to q (observed for eTGTs). A structure-based multiple sequence alignment was performed by using TGTs from C. trachomatis, Chlamydophila caviae, and Chlamydia psittaci (harboring the proposed eukaryotic type salvage pathway), typical bacterial TGTs (preQ1 specific), and eTGTs (q specific). While sequences aligned well overall, the residues that accommodate the 7-substituent group of the substrate differ for the organisms predicted to salvage q (Fig. 2A). Docking q in the active site of a typical bacterial TGT (from Z. mobilis) is not possible: the V233 side chain restricts the space to allow only the small preQ1 substrate in the binding pocket (Fig. 2B). The C. trachomatis enzyme is atypical in this aspect, as it allows docking of q in its active site: G235 allows the cyclopentenediol ring present in q to be accommodated, similarly to what is seen for eTGTs (Fig. 2B) (26, 27).
Fig. 2.
C. trachomatis salvages queuine in 2 steps. (A) Amino acid sequence alignment of select TGT proteins using PROMALS3D (74). The catalytic residues are shown in bold. The residues that accommodate the 7-substituent group of the substrate are shown in red. Dots indicate regions intentionally deleted for this figure. Dashes indicate gaps in the sequence alignment. UniProt IDs for proteins included in multiple alignment are as follows: Zymomonas mobilis (P28720), E. coli (P0A847), Shigella flexneri (Q54177), Homo sapiens QTRT1 (Q9BXR0), Caenorhabditis elegans (Q23623), C. trachomatis (A0A0E9DEF3), C. caviae (Q822U8), and C. psittaci (A0A2D2DY33). (B) Comparison of substrate-binding pockets of TGT proteins. The binding pocket of TGTCt is modeled and docked with q (purple), compared with that of Z. mobilis TGT (green; crystal structure with docked q) and H. sapiens QTRT1 (orange; crystal structure bound to q). The residues that accommodate the substrate’s 7-substituent moiety are shown in a stick model; S233LG235 in TGTCt, L231AVG234 in Z. mobilis TGT and L230SGG233 in H. sapiens QTRT1. Queuine is colored bright green. A steric clash in Z. mobilis TGT precludes binding of q (dashed circle). (C) Protein sequence similarity network of 6,187 YhhQ sequences that were retrieved from the PubSEED subsystem “q_salvage_in_Bacteria” and colored based on the predicted salvaged molecule in the organism from which they originate: red for preQ0, yellow for preQ1, and dark blue and sky blue for queuine. Red and blue arrows indicate YhhQ homologs from E. coli and C. trachomatis, respectively. (D) Scheme of Q metabolism in the E. coli derivatives used to test the function of CT140 and CT193. Dashed arrows represent reactions that are being tested. Precursors in gray are not synthesized de novo in these strains. (E and F) Detection of Q-tRNA by the APB assay in tRNAAspGUC in the presence of exogenous queuine (q) while expressing CT140/yhhQCt and/or CT193/tgtCt in different E. coli derivatives.
If the TGTCt substrate is q, YhhQCt must, unlike its E. coli homolog, be transporting q and not preQ0/preQ1. No structure is available for any member of the YhhQ family, and the residues involved in substrate recognition and transport are yet to be identified. However, sequence similarity networks (SSNs) analyses using the EFI webtools (39, 40) allowed us to define subgroups within the YhhQ family that correlate well with the various pathway configurations (Fig. 2C). This suggests a specialization of the transport activity for salvage of q, preQ0, or preQ1. The C. trachomatis YhhQCt transporter (Fig. 2C, blue arrow) is in a subgroup from organisms that harbor TGT but not QueA or QueG/QueH homologs, and is hence predicted to be involved in q salvage.
To test our hypothesis in vivo, we expressed predicted C. trachomatis q salvage genes in an E. coli derivative auxotrophic for Q (Fig. 2D). While E. coli is among the organisms that harbor a complete Q de novo pathway, it can also salvage preQ0 and preQ1, imported by YhhQEc (Fig. 1A) (28). E. coli is not predicted to salvage q based on the substrate specificity of its TGTEc enzyme (41). We set out to test whether Q could be detected in tRNAs extracted from an E. coli ΔqueD strain after feeding with exogenous q (10 nM or 100 nM; SI Appendix, Fig. S1). In this Northern-based assay, tRNAs containing Q migrate more slowly on a polyacrylamide gel containing 3-(acrylamido)phenylboronic acid (APB) than tRNAs lacking the modification (42). After transfer on a nylon membrane, a biotinylated probe specific for tRNAAspGUC is used for detection (17, 43). On the obtained blot, tRNAs that are modified with Q (SI Appendix, Fig. S1, WT control) appear in a band located higher than unmodified tRNAs (SI Appendix, Fig. S1, Δtgt control). The presence of this specific band is used as a proxy for the presence of the Q modification in the tRNAs extracted. As shown in SI Appendix, Fig. S1, no Q-modified tRNAs were detected in bulk tRNA extracted from the ΔqueD strain in the presence of q. The same result was obtained in the E. coli ΔqueD ΔqueF strain (SI Appendix, Fig. S1). As a positive control, the formation of Q was observed when preQ1 (10 nM) was added (SI Appendix, Fig. S1). These results confirmed that E. coli MG1655 salvages preQ0 and preQ1, but not q, a condition necessary to test our hypothesis.
To test whether the YhhQ homolog (CT140/YhhQCt) and TGT homolog (CT193/TGTCt) of C. trachomatis use q as a substrate, the corresponding genes were expressed in the E. coli ΔqueD strain (Fig. 2D) and the presence of Q was evaluated in extracted bulk tRNAs. Q was detected in the presence of exogenous q (10 nM) only when both genes were expressed (Fig. 2E). This result supports the prediction that YhhQCt transports q and that TGTCt catalyzes the base exchange between q and the target guanine in tRNA. To further explore the substrate specificities of YhhQCt and TGTCt, we deleted yhhQ or tgt in the E. coli ΔqueD ΔqueF strain expressing one or both Chlamydia genes and probed for the presence of Q when different precursors (preQ0, preQ1, and q) are supplemented during growth (Fig. 2F). tRNA extracted from the ΔqueD ΔqueF Δtgt pBAD24:tgtCt strain grown in the presence of any of the 3 precursors lacks Q (Fig. 2 F, Left), demonstrating that TGTCt does not use preQ1 or preQ0. However, YhhQCt can transport preQ1, as tRNA extracted from ΔqueD ΔqueF ΔyhhQ pBAD33:yhhQCt cells fed with preQ1 shows a low but detectable signal for Q (Fig. 2 F, Middle).
This set of genetic experiments not only validated the functional hypothesis about the C. trachomatis q salvage genes but also provided tools to test the function of predicted q salvage genes from other species.

Identification and Experimental Validation of Clostridioides difficile Q Salvage Genes.

Clostridioides difficile is an enteropathogen that can develop in the colon after an antibiotic exposure, leading to dysbiosis of the gut microbiota. C. difficile is an example of a subgroup of organisms that lack the preQ1 synthesis genes (queD, queE, queC, and queF) but harbor genes encoding the remaining pathway enzymes TGT, QueA, and QueG or QueH (28). A parsimonious prediction is that preQ1 must be salvaged in these organisms (Fig. 3A) and that they harbor bacterial-type TGT enzymes. The prediction about the substrate specificity of the TGTCd enzyme of C. difficile (CD2802) was first tested. When expressing the corresponding gene in an E. coli ΔqueD ΔqueF Δtgt pBAD::yhhQCt strain that can transport all Q precursors, Q was detected in tRNAs only in the presence of preQ1, confirming that C. difficile TGTCd, like its E. coli ortholog, incorporates preQ1 only in tRNA (SI Appendix, Fig. S2). The role of TGTCd in Q synthesis was confirmed by deleting the corresponding gene in the C. difficile 630 genome (SI Appendix, Fig. S3). tRNA extracted from the tgt+ strain contained Q, while the Δtgt strain lacked the modification (Fig. 3B).
Fig. 3.
CD1682, CD1683, and CD1684 are required for queuosine modification in tRNA of C. difficile. (A) Predicted Q-tRNA biosynthesis pathway in C. difficile strain 630. The magnified subfigure shows a model of the ECF transporters that include 4 subunits: S, the substrate-specific transmembrane component (S component); T, the energy-coupling module consisting of a transmembrane protein (T component); and A and A′, pairs of ABC ATPases (A proteins). Dashed arrows represent uncharacterized reactions. Molecule abbreviations and protein names are described in the text. (B) Detection of Q-tRNA by the APB assay in tRNAAspGUC of C. difficile 630 WT, 630 Δtgt, and 630 ΔCD1682-CD1684 strains. tRNA extracted from E. coli WT and Δtgt strains was used as control. (C) Representation of the genomic context of the radical SAM cluster. C. difficile 630 (accession: NC_009089.1), C. perfringens ATCC 13124 (accession: NC_008261.1), Clostridium botulinum E1 str. “BoNT E Beluga” (accession: NZ_ACSC00000000.1), R. gnavus ATCC 29149 (accession: NZ_AAYG02000018.1), and L. bacterium 2_1_58FAA (accession: ACTO00000000.1). Each gene is colored according to Pfam domain. Predicted promoters and rho-independent terminators are indicated by dashed arrows and dots, respectively. PreQ1 riboswitches are indicated by stem loops.
Because C. difficile does not encode a YhhQ homolog, another transporter must be involved in salvaging Q precursors. One candidate is CD1683 (UniProt ID Q186P1), annotated as a substrate-specific component of the queuosine ECF transporter qrtT in the RegPrecise database (35). CD1683 is the second gene of a 3-gene operon predicted to be under the control of a preQ1 riboswitch [Rfam accession RF00522 in RegPrecise (35)], as reported in a previous bioinformatic study (36). This riboswitch is located ∼140 bp upstream the start site of CD1682 (UniProt ID Q186N9), the first gene of the operon, predicted to encode a nucleoside hydrolase (member of the IunH family, COG1957 and IPR036452). The last gene of the operon, CD1684 (UniProt ID Q186P0), is predicted to encode for an enzyme belonging to the radical SAM enzyme family (IPR006638; Fig. 3C). Identical operons were also identified in other bacteria from the gut microbiome such as Clostridium perfringens, Ruminococcus gnavus, and Lachnospiraceae bacterium (Fig. 3C). The presence of a preQ1 riboswitch suggested a role of the CD1682-CD1684 operon in Q salvage. It was confirmed genetically in C. difficile, as tRNA extracted from the ΔCD1682-CD1684 strains lacked Q (Fig. 3B and SI Appendix, Fig. S3).

The Nucleoside Hydrolase CD1682 Hydrolyzes Queuosine to Queuine.

CD1682 is a protein belonging to a nucleoside hydrolase family, enzymes typically cleaving the N-glycosidic bond connecting nucleobases to ribose (44) with a preference for inosine and uridine (IunH family, also COG1957 and IPR036452). Nucleoside hydrolase family members harbor a conserved motif DxDxxxDD, located toward their N termini (44). The second and last Asp in the motif are involved in the chelation of a Ca2+ ion at the active site. Upon binding, the ribose moiety of substrate nucleobases is coordinated by the metal ion, allowing for hydrolysis. Within the family, while the residues involved in binding the Ca2+ ion and the positioning of the ribose moiety are conserved, the residues interacting with the nucleobase are variable. Initially, several nucleoside hydrolase subclasses were defined on the basis of their preferred substrate, i.e., the purine-specific inosine-adenosine-guanosine–preferring nucleoside hydrolases (IAGNHs), the 6-oxopurine–specific inosine-guanosine–preferring nucleoside hydrolases (IGNHs), and the base-aspecific inosine-uridine–preferring nucleoside hydrolases (IUNHs). However, with the increasing number of sequences becoming available, it appears that this classification is not in accordance with what has been identified at the sequence level (44).
In order to place the CD1682 protein within the nucleoside hydrolase family (IPR036452), we performed SSN analyses (39, 40). In the UniRef90 SSN generated at a low alignment score threshold of 58, CD1682 (UniProt ID Q186N9) falls into cluster 1 (the largest), containing SwissProt-annotated sequences RihA, B, and C, IUNH, and uridine nucleosidase (SI Appendix, Fig. S4A). This confirmed that CD1682 is related to sequences known to hydrolyze nucleobases. When subjected to a higher alignment score threshold of 75, CD1682 and highly homologous sequences are segregated into their own cluster (SI Appendix, Fig. S4B). This suggests that sequences constituting the CD1682 cluster are different enough from the rest of the family that they could be active against a different substrate, namely Q.
Genome Neighborhood Network (GNN) analysis (39, 40) of a local high-resolution SSN generated around CD1682 identified a set of closely related nucleoside hydrolases that are highly likely dedicated for Q salvage (SI Appendix, Fig. S5). A multiple sequence alignment for UniRef90 representative sequences from the CD1682 cluster (156 sequences) showed the absolute conservation of residues involved in Ca2+ chelation (SI Appendix, Fig. S6). However, the typical conserved residues known to be involved in the coordination of the pyrimidine or purine moieties in the characterized enzymes among the family are not present (45). As queuosine is a 7-deazaguanine carrying a cyclopentenediol ring linked at position 7 through an aminomethyl linkage, a more sizeable substrate pocket to accommodate the queuine moiety is required. The divergence detected at the sequence level for the nucleobase recognition is thus consistent with the proposed divergence in activity carried: Q hydrolysis.
Bioinformatic evidence associates CD1682 (and closely related sequences) with Q salvage. To directly test the hypothesis that CD1682 catalyzes the hydrolysis of Q to q, CD1682 was expressed in E. coli with a C-terminal hexahistidine tag, purified by Ni++-affinity chromatography followed by size-exclusion chromatography (SI Appendix, Fig. S7), and assayed for activity. Analytical size-exclusion chromatography performed on the purified recombinant CD1682 reveals a molecular weight of 118 kDa, while the molecular weight predicted from the coding sequence is 37.1 kDa, suggesting that a trimeric complex forms in solution. Other nucleoside hydrolases have been reported to exist as dimers or tetramers in solution (44). Activity was evaluated quantitatively by following the time-dependent elution of substrate and product by LC-MS. As visualized with absorption at 260 nm (Fig. 4A) and confirmed with MS for their specific masses (Fig. 4 B and C), over time, Q is consumed concomitantly with q being produced, providing in vitro support for the proposed hypothesis. The hydrolysis capacity of the purified hydrolase against other purines nucleosides was also evaluated (SI Appendix, Fig. S8). In identical conditions regarding substrate concentration, enzyme amount, and time units, Q appears consumed more rapidly than guanosine or inosine, while adenosine is nearly not consumed. Future investigations will explore the detailed kinetic parameters and substrate preferences for this nucleoside hydrolase subgroup often identified embedded in Q salvage operonic structures.
Fig. 4.
CD1682 is a queuosine hydrolase (QueK). (A) The queuosine hydrolysis activity of purified recombinant CD1682 was assayed and analyzed at different time points by injection of the reaction mixture into an HPLC system and measurement of the absorbance at 260 nm. The assay was performed with 100 µM queuosine and 100 nM CD1682. A control incubated without enzyme is included. (B and C) The identities of the substrate and product, with corresponding retention times, were verified by extracting the ion counts for the expected masses [M+H+] = m/z 278 for queuine and 410 for queuosine, respectively.
In vivo, the source of Q could be intracellular from degradation of tRNA or exogenous if a Q transporter is present in C. difficile. We therefore set out to characterize the substrate specificity of predicted Q precursor transporters in this organism.

Substrate Specificity Analysis of the Q Precursor Transporters Suggested C. difficile Salvages Not Only preQ1 but Also Queuine and Queuosine.

The strain 630 of C. difficile does not encode a homolog of YhhQ but encodes 3 homologs of QrtT/QueT (31) (CD1683, CD2097, and CD3073; UniProt IDs Q186P1, Q188G5, and Q184R1, respectively; Fig. 3A). These ECF S components interact with the core components of ECF transporter encoded in C. difficile by CD0100, CD0101, and CD0102 (46). We showed here earlier that deletion of the operon containing CD1683 abolished Q presence in this organism (Fig. 3B). To further characterize these transporters, we constructed 3 pBAD33 derivatives each expressing the 3 core components genes in 1 operon with 1 of the 3 predicted ECF S components and transformed the plasmids in the E. coli ΔqueD ΔyhhQ strain that expresses the C. trachomatis tgt gene (tgtCt). This strain can salvage both provided q and preQ0/preQ1 only if an appropriate transporter gene is expressed in trans (Fig. 5A). The presence of Q in tRNA was monitored in that strain after feeding with 10 nM preQ0, preQ1, or q (Fig. 5B). We found that preQ1 but not preQ0 was transported both by CD1683 and CD2097 and that CD1683 is able to transport q, albeit with a lower efficiency. No transport of any precursor was observed when expressing CD3073.
Fig. 5.
The ECF substrate specificity component CD1683 transports preQ1, queuine, and queuosine. (A) Scheme of Q metabolism in the E. coli derivatives used to test substrate specificity (queuine, preQ1, and preQ0) of the ECF transporter genes. Genes encoding C. difficile ECF core components (CD100, CD101, CD102) with those encoding different substrate-specific components (CD1683, CD2097, and CD3073) were expressed in E. coli ΔqueD ΔyhhQ pBAD24::tgtCt strain. Dashed arrows represent reactions being tested. (B) Q-tRNA levels were detected by the ABP assay in tRNAAspGUC extracted 60 min after supplementing with different precursors. (C) Scheme of Q metabolism in the E. coli strains used to test queuosine transport and hydrolysis activity. Genes coding different transporters, including YhhQCt and ECF complex with S component (CD1683, CD2097, or CD3073), were coexpressed with the predicted queuosine hydrolase (CD1682) in E. coli ΔqueD ΔyhhQ strains expressing both tgtEc and tgtCt. Red dashed arrows represent reactions being tested. Precursors in gray are not synthesized de novo in these strains. (D) Detection of Q-tRNA by the APB assay in tRNAAspGUC extracted 60 min after supplementing with exogenous queuosine (Q; 10 or 500 nM).
The same strains were used to test if these ECF transport proteins as well as the previously characterized YhhQ proteins could be involved in Q import (Fig. 5C). Q-tRNA was detected when CD1683 (S component) and genes of ECF core components were coexpressed with CD1682 (Fig. 5D) in the presence of Q (10 nM), confirming the nucleoside hydrolase activity of CD1682 in vivo. We propose to name members from this subgroup of the nucleoside hydrolase family: queuosine hydrolase or QueK. It is worth noting that a small amount of Q was detected in tRNA in the presence of high Q concentrations (500 nM) when yhhQCt or CD1683 was expressed while CD1682 is absent (SI Appendix, Fig. S9). This observation could result from a nonspecific Q nucleoside hydrolase activity in E. coli, potentially carried out by an endogenous member of the nucleoside hydrolase family.

The C. difficile q Salvage Pathway Uses a Queuine Lyase of the Radical SAM Enzyme Family.

In combination, the experiments described here so far suggest that Q can be transported in C. difficile by an ECF transporter using CD1683 as a specificity component and hydrolyzed to q by CD1682 (Fig. 3A). As the substrate of TGTCd (CD2802) is preQ1 (SI Appendix, Fig. S2), the conversion of q to preQ1, a reaction never described previously to our knowledge, is needed to complete the pathway. The third gene of the CD1682 operon encodes for a protein of the radical-SAM (RS) enzyme family, known for its remarkable and versatile enzymology (4750). We hypothesized that CD1684 (UniProt ID Q186P0) could act as a lyase, breaking a C-N bond to generate preQ1 from q.
The hypothesis that CD1684 is a q lyase was tested by using the E. coli ΔqueD ΔqueF pBAD33:yhhQCt strain, which is able to import q but cannot catalyze its insertion in tRNAs (Fig. 6A). Remarkably, when the CD1684 gene was expressed in this strain, q could be salvaged (Fig. 6B). TGTEc is preQ1-specific, and the q salvage activity is abolished by deletion of the downstream gene queA (Fig. 6 B, Right). Thus, the insertion of Q in tRNAs when q is provided externally required the production of the preQ1 intermediate and is not the result of a direct q incorporation. These results taken together suggest that, indeed, CD1684 is an enzyme capable of producing preQ1 from q.
Fig. 6.
CD1684 generates preQ1 from queuine in vivo. (A) Scheme of Q metabolism in the E. coli derivatives used to test the activity of CD1684. Red dashed arrows represent reactions being tested. Precursors in gray are not synthesized de novo in these strains. (B) Detection of Q-tRNA by the APB assay in tRNAAspGUC extracted 60 min after supplementing with different precursors. The C. difficile CD1684 gene was expressed in E. coli ΔqueD ΔqueF pBAD33::yhhQCt strain and ΔqueD ΔqueF ΔqueA pBAD33::yhhQCt strain.
SNNs were built to place CD1684 within the RS superfamily (composed of overlapping IPR006638 and IPR007197). As detailed in SI Appendix, Fig. S10, SSNs analyses showed that this protein was part of a set of closely related radical SAM enzymes that form a separate group that is likely to be involved in Q salvage based on gene neighborhood information (SI Appendix, Fig. S11).
To directly test the hypothesis that CD1684 could act as a lyase, breaking a C-N bond to generate preQ1 from q, CD1684, with a C-terminal hexahistidine tag, was expressed in E. coli, purified in aerobic conditions by Ni++-affinity chromatography (SI Appendix, Fig. S7), followed by reconstitution of its Fe/S cluster and then by size-exclusion chromatography in anaerobic conditions. UV/Vis absorption spectra of CD1684 as purified and after reconstitution reveal characteristics of proteins possessing a Fe/S center (SI Appendix, Fig. S12). Estimation of the molar content of Fe and S before and after reconstitution reveals 2.17 ± 0.27 Fe and 3.19 ± 0.05 S for the as-purified protein and 1.92 ± 0.15 Fe and 4.45 ± 0.09 S for the reconstituted protein. The reconstituted enzyme was assayed for activity under anaerobic conditions. Activity was evaluated at the qualitative level in the presence of sodium dithionite (chemical electron donor) by following the evolution of substrates and products by LC-MS over time. As visualized with absorption at 260 nm (Fig. 7A) and confirmed with MS (Fig. 7 BE), q and SAM were consumed concomitantly with production of preQ1 and 5′dA, validating in vitro the proposed hypothesis, and confirming that CD1684 is indeed a radical SAM enzyme, because of its utilization of SAM to produce 5′dA (47). Because we have not identified the coproduct produced by CD1684 during preQ1 production, we cannot unambiguously assign the type of reaction catalyzed. However, based on our genetic and biochemical characterization, we preliminarily named this enzyme queuine lyase or QueL.
Fig. 7.
Queuine lyase (QueL) activity and structure. (A) The queuine lyase activity of purified recombinant CD1684 was analyzed by separation of quenched reaction mixture by an HPLC system with monitoring at 260 nm absorbance. Data from a representative assay are presented from an assay performed under anaerobic conditions with 100 µM queuine, 200 µM sodium dithionite, 66.67 µM S-adenosyl-l-methionine (SAM limited to allow for visualization of preQ1; otherwise, a high concentration of SAM was used), and 10 µM of purified CD1684. The control is a reaction lacking enzyme. (BE) The identities of the substrate and product, with corresponding retention times, were verified by extracting the ion counts for the expected masses [M+H+] = m/z 278, 180, 399, and 252 for queuine, preQ1, SAM, and 5′dA (5′deoxyadenosine), respectively. The UV signal and mass corresponding to queuine and SAM are reduced over time, while signal and mass corresponding to preQ1 and 5′dA are increased, demonstrating that CD1684 is an RS enzyme with queuine lyase activity. (F) Overall view of QueLCs with secondary structural elements assigned numerically. The α6 extension (blue) caps the active site terminating in Asp229. SAM (light gray) and queuine (dark gray) are depicted as ball-and-stick illustrations (oxygen, red; nitrogen, blue). The RS cluster is depicted as spheres (iron, burnt orange; sulfur, yellow). (G) A top-down view of the active site showing highly conserved amino acids (depicted as ball and sticks in pink) and hydrogen bonds (dashed lines) formed between substrates (coloring as in A) and the QueLCs active site. The red dashed line denotes the distance (3.5 Å) between the 5′-carbon of SAM and the 5′-oxygen of queuine. (H) Mechanistic proposal for catalysis by QueL.

Crystal Structure of Queuine Lyase from Clostridium spiroforme DSM 1552.

To gain insight into the reaction catalyzed by QueL, we cloned several homologous members from the CD1684 SSN cluster, heterologously expressed and purified them as previously described (51), and subjected them to sparse matrix screening crystallization trials by using the sitting-drop vapor diffusion method. The QueL from Clostridium spiroforme DSM 1552 (QueLCs; UniProt ID B1C2R2) formed hexagonal rod-like crystals in the presence of SAM and q, which exhibited strong diffraction at a synchrotron X-ray source. We determined the structure by using single-wavelength anomalous dispersion by collecting data at a wavelength of 1.378 Å, which takes advantage of Fe absorption from the native [4Fe-4S] cluster. The resulting 1.73-Å–resolution structure of QueLCs (deposited in the PDB, ID 6P78) contains all 229 amino acids from the native sequence, as well as a [4Fe-4S] cluster, SAM, and q (SI Appendix, Fig. S13 A and B), in a single molecule in the asymmetric unit. A search of the PDB with the Dali server (52) found that QueLCs has structural similarly with pyruvate formate-lyase activating enzyme (PFL-AE; RMSD of 3.9 Å for 192 Cα atoms; PDB ID code 3CB8), even though these enzymes share only 12% sequence identity. Overall, QueLCs adopts the same (β/α)6 partial TIM barrel as PFL-AE (Fig. 7F), but unlike the “splayed” core of PFL-AE, its core is compressed by ∼2 Å, resulting in a more complete active site (53). This difference is likely because, while the substrate for PFL-AE is a large folded protein (PFL) that contributes to the active site, QueLCs acts on the much smaller q molecule. SAM binds to QueLCs mostly through hydrogen bonding with protein backbone functionalities, as is the case for most RS enzymes (54). The only sequence-specific interactions coming from QueLCs are Glu123 and Asn184, which directly hydrogen-bond with the dihydroxycyclopentene moiety (Fig. 7G). Interestingly, QueLCs contains a highly conserved active-site cys residue (C154) within ∼4 Å of the SAM methyl group. When this residue was modeled as cysteine, extra electron density was apparent, which was best filled by a model that included a methyl-cys (mCys) residue (SI Appendix, Fig. S13C), similar to mCys355 found in the RS methyltransferase RlmN (55). Currently, the importance of this modification is unclear, even if it does form part of the SAM binding pocket. Studies are under way to determine how this modification forms and its importance for the catalyzed reaction. Unlike SAM, q is coordinated by an extensive network of hydrogen bonds provided by several conserved amino acids, including Ser67, Glu96, Lys119, and Asp229 (SI Appendix, Fig. S14). Interestingly, the side chain of Asp229 does not actually participate in the hydrogen bonding; instead, it is the C-terminal carboxylate of Asp229 that H-bonds to the N9 of deazaguanine in q (Fig. 7G). This interaction is stabilized by Arg187, which plays dual roles by also providing H-bonds to the adenosine moiety of SAM with its amide and carbonyl backbone atoms. In addition to H-bonds, q is also further stabilized by a pi-stacking interaction with Tyr33. H-bonding reinforces this interaction between the Tyr33 carbonyl atom and the exocyclic N in q, with additional binding interactions provided by an ordered water molecule that is positioned between the hydroxyl of Tyr33 and the bridging amine in q. The dihydroxycyclopentene moiety of queuine has 2 primary interactions: Glu96 forms H-bonds with the 4′- and 5′-position hydroxyl groups, and Lys119 H-bonds to the 4′-position hydroxyl (Fig. 7G). These interactions orient the dihydroxycyclopentene to place the 5′-position hydroxyl proton in direct alignment and distance (3.5 Å) from the 5′ carbon of SAM, which indicates that this hydrogen atom is likely abstracted by the 5′deoxyadenosyl-radical (5′dA•) to start catalysis, which is consistent with the distances found in other RS enzymes crystalized in the presence of SAM and their requisite substrate (54, 56). The remainder of the binding interactions with q involve van der Waals interactions with several Val, Leu, and Ile residues.


This work illustrates that bacteria adapt their Q salvage strategies to their environments. Bacteria that are located inside a human cell such as C. trachomatis have streamlined their metabolism and rely on import for many nutrients (5760). Only q or Q are supposedly available in the intracellular environment, and we show here that Chlamydiae species have evolved to salvage the q base: a YhhQ family member (CT140) imports q, and a bacterial TGT homolog (CT193) catalyzes the base exchange between q and the target guanine in tRNAs. Those transport and base-exchange activities are not the historically identified and canonical ones but are related. Clearly, depending on the organism, the actual substrates for YhhQ and bacterial TGT homologs vary. This illustrates the difficulty in defining the boundaries between functions and thus annotations within protein families. Defining accurate annotation therefore requires context information and/or signature motifs (61). For example, in order to differentiate the subfamilies of bacterial TGT that incorporate q from those that incorporate preQ1, one can combine the absence of QueA with the difference in residues in the binding pocket for the substrate’s 7-substituent moiety (Fig. 2A and SI Appendix, Supplementary Text). Similarly, analysis of the specificity of YhhQ subfamilies requires combination of the presence of Q biosynthetic genes from the organism in the same cluster with the specificity of their TGT enzymes (SI Appendix, Fig. S15).
Gut microbiota represent a complex ecosystem that develops in close interaction with the host. In terms of Q metabolism, this environment is particularly complex, as specific microbes can be sources but also sinks of q, Q, preQ0, or preQ1 (62) with the additional competition for the q precursor by the human host. Indeed, a recent analysis of Q metabolism genes in 2,216 representative bacteria of the gut microbiome found that ∼50% of these bacteria must salvage a Q precursor (63). The final numbers are given in SI Appendix, Fig. S16 and Dataset S1. To summarize, 51% of the bacteria analyzed are predicted to synthesize Q de novo. While a few strains such as Streptococcus pneumoniae are predicted to import preQ0 (2%), a large proportion (29%) must salvage preQ1 (based on the absence of queF and the presence of tgt and queA), while 13% salvage q. Finally, only 5% of the genomes analyzed do not encode TGT homologs and are hence predicted not to modify their tRNA with queuosine.
The characterization of the Q salvage pathway in C. difficile now allows the study of the physiological role of this tRNA modification in the important enteropathogen. The absence of Q does not seem to affect growth rate, but we found that WT tRNAs are only partially modified in BHI, presumably because the Q source is limiting (Fig. 3B). Further studies are required in more relevant physiological conditions, as Q might be important for fitness under stress conditions and during colonization. Indeed, the expression of genes from the CD1682 operon was reported to be up-regulated in biofilm compared with planktonic cells (64), when iron is low (65), during exponential growth compared with stationary phase, and when the sigma factor of the stationary phase, SigH, is deleted (66).
Finally, this study is another example of the power of comparative genomics approaches to discover novel enzymatic reactions. The q lyase activity had never been described in the literature to our knowledge, and the discovery that a subgroup of the radical-SAM enzyme superfamily catalyzes this reaction is unexpected. The crystal structure of QueLCs bound to its substrates allows us to propose how preQ1 is generated from q. As proposed in Fig. 7H, the reaction starts by abstraction of the 5-position hydroxyl H-atom by 5′dA•, producing intermediate 1. This is the most likely outcome because all other hydrogen atoms in q are greater than 5 Å away and would require large movements in the active site to place them in resonance with the 5′dA•. The presence of the 5-position hydroxyl radical would allow fission of the 4′-5′ C-C bond of the dihydroxycyclopentene ring, generating a vinylic-stabilized radical intermediate 2 (Fig. 7H). The redox potential of this radical would be sufficiently oxidizing to accept an electron from a reduced RS cluster to yield anion 3 (Fig. 7H). The anion could then isomerize, in the process eliminating preQ1 to produce the divinyl ketone 5 (Fig. 7H). With the assistance of a proton from Glu96, a cationic 4π-electrocyclic-ring closure would produce 7, which would be followed by a bulk solvent-catalyzed tautomerization to yield cyclopentenone 8, which is reminiscent of the Nazarov cyclization (67) (Fig. 7H).



For sequence analyses, the BLAST tools (68) and resources at NCBI (69) (, UniProt (70) (, and PATRICBRC (71) ( were routinely used. Further details on all bioinformatic analyses are provided in SI Appendix.

Strains, Media, and Growth Conditions.

Strains and plasmids used in this study are listed in SI Appendix, Table S1. Oligonucleotides used for mutant construction and plasmid construction are listed in SI Appendix, Table S2. Further details are provided in SI Appendix.


PreQ0 was purchased from ArkPharm (AK-32535). PreQ1 was purchased from Sigma-Aldrich (SML0807-5MG). Queuine was purchased from Toronto Research Chemicals (Q525000). Queuosine was synthesized as previously described (72) and detailed in SI Appendix.

Exogenous q Precursors Feeding.

As previously described (73), cells were cultured in M9 defined media with glycerol as carbon source and arabinose for induction. After supplementing with queuosine precursors, the transport reaction was stopped, followed by tRNA extraction. Details are provided in SI Appendix.

tRNA Extraction.

As previously described (28), small RNAs of E. coli cells were extracted by using a PureLink miRNA Isolation kit (Life Technologies) according to manufacturer protocol. For C. difficile, small RNAs were extracted by using a FastRNA Pro Blue Kit (MP Biomedicals) and FastPrep instrument according to manufacturer protocol. Details are provided in SI Appendix.

Detection of Queuosine in Bulk tRNA.

Detection of the presence of Q in tRNA is based on a method originally developed by Igloi and Kössel (42) and later improved (73), as detailed in SI Appendix.

Enzyme Expression, Purification, and Assays.

C-terminal hexahistidine-tagged CD1682 (QueK, Q186N9) and CD1684 (QueL, Q186P0) were overexpressed in E. coli Rosetta DE3 pLysS cells and E. coli BL21 DE3 cells, respectively, followed by purification and enzyme assay, as detailed in SI Appendix. The CD1684 homolog in C. spiroforme DSM 1552, QueLCs (UniProt ID code B1C2R2), with a N-ter hexahistidine tag, was overexpressed in E. coli BL21 DE3 cells, followed by purification and structure determination. Details are provided in SI Appendix.

X-Ray Diffraction and Structure Determination.

QueLCs was crystallized followed by diffraction and structure determination as detailed in SI Appendix.

Data Availability

Data deposition: The data reported in this paper have been deposited in the Protein Data Bank, (ID code 6P78).


This work was funded by the National Institutes of Health (R01 GM70641 to V.d.C.-L.; P01 GM118303 to J.A.G. and S.C.A.; U54-GM093342 to J.A.G. and S.C.A.; R21-AI133329 to T.L.G. and S.C.A.; U54-GM094662 to S.C.A.; GM110588 to M.A.S.), the Price Family Foundation (S.C.A.), and the California Metabolic Research Foundation (M.A.S.). We acknowledge the Albert Einstein Anaerobic Structural and Functional Genomics Resource ( We thank Martin I. H. McLauglin and Wilfred A. van der Donk for providing the guidance and the plasmid isc-pBADCDF for in vivo maturation of metalloenzymes. The LC-MS analyses were performed by Furong Sun at the Mass Spectrometry Laboratory, a Service Facility from the School of Chemical Sciences at University of Illinois at Urbana–Champaign. The Einstein Crystallographic Core X-Ray diffraction facility is supported by NIH Shared Instrumentation Grant S10 OD020068. This research used resources of the Advanced Photon Source, a US Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. Use of the Lilly Research Laboratories Collaborative Access Team (LRL-CAT) beamline at Sector 31 of the Advanced Photon Source was provided by Eli Lilly Company, which operates the facility.

Supporting Information

Appendix (PDF)
Dataset_S01 (XLSX)


C. M. Guinane, P. D. Cotter, Role of the gut microbiota in health and chronic gastrointestinal disease: Understanding a hidden metabolic organ. Therap. Adv. Gastroenterol. 6, 295–308 (2013).
B. N. Ames, Prolonging healthy aging: Longevity vitamins and proteins. Proc. Natl. Acad. Sci. U.S.A. 115, 10836–10844 (2018).
S. Yokoyama et al., Three-dimensional structure of hyper-modified nucleoside Q located in the wobbling position of tRNA. Nature 282, 107–109 (1979).
W. R. Farkas, Effect of diet on the queuosine family of tRNAs of germ-free mice. J. Biol. Chem. 255, 6832–6835 (1980).
J. P. Reyniers, J. R. Pleasants, B. S. Wostmann, J. R. Katze, W. R. Farkas, Administration of exogenous queuine is essential for the biosynthesis of the queuosine-containing transfer RNAs in the mouse. J. Biol. Chem. 256, 11591–11594 (1981).
C. Fergus, D. Barnes, M. A. Alqasem, V. P. Kelly, The queuine micronutrient: Charting a course from microbe to man. Nutrients 7, 2897–2929 (2015).
F. Harada, S. Nishimura, Possible anticodon sequences of tRNA His, tRNA Asm, and tRNA Asp from Escherichia coli B. Universal presence of nucleoside Q in the first postion of the anticondons of these transfer ribonucleic acids. Biochemistry 11, 301–308 (1972).
P. F. Crain, S. K. Sethi, J. R. Katze, J. A. McCloskey, Structure of an amniotic fluid component, 7-(4,5-cis-dihydroxy-1-cyclopenten-3-ylaminomethyl)-7-deazaguanine (queuine), a substrate for tRNA: Guanine transglycosylase. J. Biol. Chem. 255, 8405–8407 (1980).
Y. C. Chen, V. P. Kelly, S. V. Stachura, G. A. Garcia, Characterization of the human tRNA-guanine transglycosylase: Confirmation of the heterodimeric subunit structure. RNA 16, 958–968 (2010).
K. B. Jacobson, W. R. Farkas, J. R. Katze, Presence of queuine in Drosophila melanogaster: Correlation of free pool with queuosine content of tRNA and effect of mutations in pteridine metabolism. Nucleic Acids Res. 9, 2351–2366 (1981).
G. Jänel, U. Michelsen, S. Nishimura, H. Kersten, Queuosine modification in tRNA and expression of the nitrate reductase in Escherichia coli. EMBO J. 3, 1603–1608 (1984).
G. M. Kirtland et al., Novel salvage of queuine from queuosine and absence of queuine synthesis in Chlorella pyrenoidosa and Chlamydomonas reinhardtii. J. Bacteriol. 170, 5633–5641 (1988).
W. Langgut, T. Reisser, Involvement of protein kinase C in the control of tRNA modification with queuine in HeLa cells. Nucleic Acids Res. 23, 2488–2491 (1995).
R. Gaur, G. R. Björk, S. Tuck, U. Varshney, Diet-dependent depletion of queuosine in tRNAs in Caenorhabditis elegans does not lead to a developmental block. J. Biosci. 32, 747–754 (2007).
M. Vinayak, C. Pathak, Queuosine modification of tRNA: Its divergent role in cellular machinery. Biosci. Rep. 30, 135–148 (2009).
E. M. Novoa, M. Pavon-Eternod, T. Pan, L. Ribas de Pouplana, A role for tRNA modifications in genome structure and codon usage. Cell 149, 202–213 (2012).
J. M. Zaborske et al., A nutrient-driven tRNA modification alters translational fidelity and genome-wide protein coding across an animal genus. PLoS Biol. 12, e1002015 (2014).
S. Varghese et al., In vivo modification of tRNA with an artificial nucleobase leads to full disease remission in an animal model of multiple sclerosis. Nucleic Acids Res. 45, 2029–2039 (2017).
S. Ishiwata et al., Increased expression of queuosine synthesizing enzyme, tRNA-guanine transglycosylase, and queuosine levels in tRNA of leukemic cells. J. Biochem. 129, 13–17 (2001).
F. Tuorto et al., Queuosine-modified tRNAs confer nutritional control of protein translation. EMBO J. 37, e99777 (2018).
G. Hutinet, M. A. Swarjo, V. de Crécy-Lagard, Deazaguanine derivatives, examples of crosstalk between RNA and DNA modification pathways. RNA Biol. 14, 1175–1184 (2017).
C. Romier, J. E. W. Meyer, D. Suck, Slight sequence variations of a common fold explain the substrate specificities of tRNA-guanine transglycosylases from the three kingdoms. FEBS Lett. 416, 93–98 (1997).
C. Romier, K. Reuter, D. Suck, R. Ficner, Crystal structure of tRNA-guanine transglycosylase: RNA modification by base exchange. EMBO J. 15, 2850–2857 (1996).
I. Biela et al., Investigation of specificity determinants in bacterial tRNA-guanine transglycosylase reveals queuine, the substrate of its eucaryotic counterpart, as inhibitor. PLoS One 8, e64240 (2013).
C. Boland, P. Hayes, I. Santa-Maria, S. Nishimura, V. P. Kelly, Queuosine formation in eukaryotic tRNA occurs via a mitochondria-localized heteromeric transglycosylase. J. Biol. Chem. 284, 18218–18227 (2009).
B. Stengl, K. Reuter, G. Klebe, Mechanism and substrate specificity of tRNA-guanine transglycosylases (TGTs): tRNA-modifying enzymes from the three different kingdoms of life share a common catalytic mechanism. ChemBioChem 6, 1926–1939 (2005).
Y. C. Chen et al., Evolution of eukaryal tRNA-guanine transglycosylase: Insight gained from the heterocyclic substrate recognition by the wild-type and mutant human and Escherichia coli tRNA-guanine transglycosylases. Nucleic Acids Res. 39, 2834–2844 (2011).
R. Zallot, Y. Yuan, V. de Crécy-Lagard, The Escherichia coli COG1738 member yhhq is involved in 7-cyanodeazaguanine (preQ0) transport. Biomolecules 7, E12 (2017).
J. R. Katze, B. Basile, J. A. McCloskey, Queuine, a modified base incorporated posttranscriptionally into eukaryotic transfer RNA: Wide distribution in nature. Science 216, 55–56 (1982).
D. Xu et al., PreQ0 base, an unusual metabolite with anti-cancer activity from Streptomyces qinglanensis 172205. Anti Cancer Agents Med. Chem. 15, 285–290 (2015).
D. A. Rodionov et al., A novel class of modular transporters for vitamins in prokaryotes. J. Bacteriol. 191, 42–51 (2009).
A. K. Hopper, D. A. Pai, D. R. Engelke, Cellular dynamics of tRNAs and their genes. FEBS Lett. 584, 310–317 (2010).
P. J. McCown, J. J. Liang, Z. Weinberg, R. R. Breaker, Structural, functional, and taxonomic diversity of three preQ1 riboswitch classes. Chem. Biol. 21, 880–889 (2014).
A. Roth et al., A riboswitch selective for the queuosine precursor preQ1 contains an unusually small aptamer domain. Nat. Struct. Mol. Biol. 14, 308–317 (2007).
P. S. Novichkov et al., RegPrecise 3.0–A resource for genome-scale exploration of transcriptional regulation in bacteria. BMC Genomics 14, 745 (2013).
E. I. Sun et al., Comparative genomics of metabolic capacities of regulons controlled by cis-regulatory RNA motifs in bacteria. BMC Genomics 14, 597 (2013).
R. Overbeek et al., The subsystems approach to genome annotation and its use in the project to annotate 1000 genomes. Nucleic Acids Res. 33, 5691–5702 (2005).
D. Iwata-Reuyl, V. de Crécy-Lagard, “Enzymatic formation of the 7-deazaguanosine hypermodified nucleosides of tRNA” in DNA and RNA Modification Enzymes: Structure, Mechanism, Function and Evolution, H. Grosjean, Ed. (Landes Bioscience, 2009), pp. 377–391.
J. A. Gerlt et al., Enzyme function initiative-enzyme similarity tool (EFI-EST): A web tool for generating protein sequence similarity networks. Biochim. Biophys. Acta 1854, 1019–1037 (2015).
R. Zallot, N. O. Oberg, J. A. Gerlt, ‘Democratized’ genomic enzymology web tools for functional assignment. Curr. Opin. Chem. Biol. 47, 77–85 (2018).
S. Noguchi, Y. Nishimura, Y. Hirota, S. Nishimura, Isolation and characterization of an Escherichia coli mutant lacking tRNA-guanine transglycosylase. Function and biosynthesis of queuosine in tRNA. J. Biol. Chem. 257, 6544–6550 (1982).
G. L. Igloi, H. Kössel, Affinity electrophoresis for monitoring terminal phosphorylation and the presence of queuosine in RNA. Application of polyacrylamide containing a covalently bound boronic acid. Nucleic Acids Res. 13, 6881–6898 (1985).
J. J. Thiaville et al., Novel genomic island modifies DNA with 7-deazaguanine derivatives. Proc. Natl. Acad. Sci. U.S.A. 113, E1452–E1459 (2016).
W. Versées, J. Steyaert, Catalysis by nucleoside hydrolases. Curr. Opin. Struct. Biol. 13, 731–738 (2003).
R. K. Singh, J. Steyaert, W. Versées, Structural and biochemical characterization of the nucleoside hydrolase from C. elegans reveals the role of two active site cysteine residues in catalysis. Protein Sci. 26, 985–996 (2017).
M. Monot et al., Reannotation of the genome sequence of Clostridium difficile strain 630. J. Med. Microbiol. 60, 1193–1199 (2011).
H. J. Sofia, G. Chen, B. G. Hetzler, J. F. Reyes-Spindola, N. E. Miller, Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: Functional characterization using new analysis and information visualization methods. Nucleic Acids Res. 29, 1097–1106 (2001).
P. A. Frey, A. D. Hegeman, F. J. Ruzicka, The radical SAM superfamily. Crit. Rev. Biochem. Mol. Biol. 43, 63–88 (2008).
Q. Zhang, W. Liu, Complex biotransformations catalyzed by radical S-adenosylmethionine enzymes. J. Biol. Chem. 286, 30245–30252 (2011).
G. L. Holliday et al., Atlas of the radical SAM superfamily: Divergent evolution of function using a “plug and play” domain. Methods Enzymol. 606, 1–71 (2018).
T. L. Grove et al., Structural insights into thioether bond formation in the biosynthesis of sactipeptides. J. Am. Chem. Soc. 139, 11734–11744 (2017).
L. Holm, L. M. Laakso, Dali server update. Nucleic Acids Res. 44, W351–W355 (2016).
J. L. Vey et al., Structural basis for glycyl radical formation by pyruvate formate-lyase activating enzyme. Proc. Natl. Acad. Sci. U.S.A. 105, 16137–16141 (2008).
J. B. Broderick, B. R. Duffus, K. S. Duschene, E. M. Shepard, Radical S-adenosylmethionine enzymes. Chem. Rev. 114, 4229–4317 (2014).
A. K. Boal et al., Structural basis for methyl transfer by a radical SAM enzyme. Science 332, 1089–1092 (2011).
T. A. Stich, W. K. Myers, R. D. Britt, Paramagnetic intermediates generated by radical S-adenosylmethionine (SAM) enzymes. Acc. Chem. Res. 47, 2235–2243 (2014).
T. Goldfarb et al., BREX is a novel phage resistance system widespread in microbial genomes. EMBO J. 34, 169–183 (2015).
R. E. Ley, C. A. Lozupone, M. Hamady, R. Knight, J. I. Gordon, Worlds within worlds: Evolution of the vertebrate gut microbiota. Nat. Rev. Microbiol. 6, 776–788 (2008).
T. M. Fuchs, W. Eisenreich, J. Heesemann, W. Goebel, Metabolic adaptation of human pathogenic and related nonpathogenic bacteria to extra- and intracellular habitats. FEMS Microbiol. Rev. 36, 435–462 (2012).
A. Best, Y. Abu Kwaik, Nutrition and bipartite metabolism of intracellular pathogens. Trends Microbiol. 27, 550–561 (2019).
R. Zallot, K. J. Harrison, B. Kolaczkowski, V. de Crécy-Lagard, Functional annotations of paralogs: A blessing and a curse. Life (Basel) 6, E39 (2016).
C. M. Theriot, V. B. Young, Interactions between the gastrointestinal microbiome and Clostridium difficile. Annu. Rev. Microbiol. 69, 445–461 (2015).
D. A. Rodionov et al., Micronutrient requirements and sharing capabilities of the human gut microbiome. Front. Microbiol. 10, 1316 (2019).
I. Poquet et al., Clostridium difficile biofilm: Remodeling metabolism and cell surface to build a sparse and heterogeneously aggregated architecture. Front. Microbiol. 9, 2084 (2018).
J. L. Hastie, P. C. Hanna, P. E. Carlson, Transcriptional response of Clostridium difficile to low iron conditions. Pathog. Dis. 76, fty009 (2018).
L. Saujet, M. Monot, B. Dupuy, O. Soutourina, I. Martin-Verstraete, The key sigma factor of transition phase, SigH, controls sporulation, metabolism, and virulence factor expression in Clostridium difficile. J. Bacteriol. 193, 3186–3196 (2011).
M. A. Tius, Allene ether Nazarov cyclization. Chem. Soc. Rev. 43, 2979–3002 (2014).
S. F. Altschul, W. Gish, W. Miller, E. W. Myers, D. J. Lipman, Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).
E. W. Sayers et al., Database resources of the National Center for Biotechnology Information. Nucleic Acids Res. 47, D23–D28 (2019).
UniProt Consortium, UniProt: A worldwide hub of protein knowledge. Nucleic Acids Res. 47, D506–D515 (2019).
A. R. Wattam et al., Improvements to PATRIC, the all-bacterial bioinformatics database and analysis resource center. Nucleic Acids Res. 45, D535–D542 (2017).
F. Klepper, E.-M. Jahn, V. Hickmann, T. Carell, Synthesis of the transfer-RNA nucleoside queuosine by using a chiral allyl azide intermediate. Angew. Chem. Int. Ed. Engl. 46, 2325–2327 (2007).
Y. Yuan et al., Identification of the minimal bacterial 2′-deoxy-7-amido-7-deazaguanine synthesis machinery. Mol. Microbiol. 110, 469–483 (2018).
J. Pei, B.-H. Kim, N. V. Grishin, PROMALS3D: A tool for multiple protein sequence and structure alignments. Nucleic Acids Res. 36, 2295–2300 (2008).

Information & Authors


Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 116 | No. 38
September 17, 2019
PubMed: 31481610


Data Availability

Data deposition: The data reported in this paper have been deposited in the Protein Data Bank, (ID code 6P78).

Submission history

Published online: September 3, 2019
Published in issue: September 17, 2019


  1. queuosine
  2. nucleoside transport
  3. sequence similarity network
  4. comparative genomics
  5. rSAM


This work was funded by the National Institutes of Health (R01 GM70641 to V.d.C.-L.; P01 GM118303 to J.A.G. and S.C.A.; U54-GM093342 to J.A.G. and S.C.A.; R21-AI133329 to T.L.G. and S.C.A.; U54-GM094662 to S.C.A.; GM110588 to M.A.S.), the Price Family Foundation (S.C.A.), and the California Metabolic Research Foundation (M.A.S.). We acknowledge the Albert Einstein Anaerobic Structural and Functional Genomics Resource ( We thank Martin I. H. McLauglin and Wilfred A. van der Donk for providing the guidance and the plasmid isc-pBADCDF for in vivo maturation of metalloenzymes. The LC-MS analyses were performed by Furong Sun at the Mass Spectrometry Laboratory, a Service Facility from the School of Chemical Sciences at University of Illinois at Urbana–Champaign. The Einstein Crystallographic Core X-Ray diffraction facility is supported by NIH Shared Instrumentation Grant S10 OD020068. This research used resources of the Advanced Photon Source, a US Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. Use of the Lilly Research Laboratories Collaborative Access Team (LRL-CAT) beamline at Sector 31 of the Advanced Photon Source was provided by Eli Lilly Company, which operates the facility.


This article is a PNAS Direct Submission.



Yifeng Yuan1
Department of Microbiology and Cell Science, University of Florida, Gainesville, FL 32611;
Rémi Zallot1
Institute for Genomic Biology, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
Tyler L. Grove1
Department of Biochemistry, Albert Einstein College of Medicine, Bronx, NY 10461;
Daniel J. Payan
Institute for Genomic Biology, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
Isabelle Martin-Verstraete
Laboratoire de Pathogénèse des Bactéries Anaérobies, Institut Pasteur et Université de Paris, F-75015 Paris, France;
Sara Šepić
Department of Microbiology and Cell Science, University of Florida, Gainesville, FL 32611;
Seetharamsingh Balamkundu
Singapore-MIT Alliance for Research and Technology, Infectious Disease Interdisciplinary Research Group, 138602 Singapore, Singapore;
Ramesh Neelakandan
Singapore-MIT Alliance for Research and Technology, Infectious Disease Interdisciplinary Research Group, 138602 Singapore, Singapore;
Vinod K. Gadi
Singapore-MIT Alliance for Research and Technology, Infectious Disease Interdisciplinary Research Group, 138602 Singapore, Singapore;
Chuan-Fa Liu
Singapore-MIT Alliance for Research and Technology, Infectious Disease Interdisciplinary Research Group, 138602 Singapore, Singapore;
Manal A. Swairjo
Department of Chemistry and Biochemistry, San Diego State University, San Diego, CA 92182;
The Viral Information Institute, San Diego State University, San Diego, CA 92182;
Peter C. Dedon
Singapore-MIT Alliance for Research and Technology, Infectious Disease Interdisciplinary Research Group, 138602 Singapore, Singapore;
Department of Biological Engineering and Chemistry, Massachusetts Institute of Technology, Cambridge, MA 02139;
Center for Environmental Health Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139;
Steven C. Almo
Department of Biochemistry, Albert Einstein College of Medicine, Bronx, NY 10461;
John A. Gerlt
Institute for Genomic Biology, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
Department of Biochemistry, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
Department of Chemistry, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
Valérie de Crécy-Lagard2 [email protected]
Department of Microbiology and Cell Science, University of Florida, Gainesville, FL 32611;
University of Florida Genetics Institute, Gainesville, FL 32610


To whom correspondence may be addressed. Email: [email protected].
Author contributions: Y.Y., R.Z., I.M.-V., S.C.A., J.A.G., and V.d.C.-L. designed research; Y.Y., R.Z., T.L.G., D.J.P., and I.M.-V. performed research; S.B., R.N., V.K.G., C.-F.L., and P.C.D. contributed new reagents/analytic tools; Y.Y., R.Z., T.L.G., S.Š., M.A.S., and V.d.C.-L. analyzed data; and Y.Y., R.Z., T.L.G., I.M.-V., P.C.D., and V.d.C.-L. wrote the paper.
Y.Y., R.Z., and T.L.G. contributed equally to this work.

Competing Interests

The authors declare no conflict of interest.

Metrics & Citations


Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.

Citation statements



If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by


    View Options

    View options

    PDF format

    Download this article as a PDF file


    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    Discovery of novel bacterial queuine salvage enzymes and pathways in human pathogens
    Proceedings of the National Academy of Sciences
    • Vol. 116
    • No. 38
    • pp. 18747-19210







    Share article link

    Share on social media

    Further reading in this issue