Document Detail

An enzyme-trap approach allows isolation of intermediates in cobalamin biosynthesis.
Jump to Full Text
MedLine Citation:
PMID:  23042036     Owner:  NLM     Status:  MEDLINE    
The biosynthesis of many vitamins and coenzymes has often proven difficult to elucidate owing to a combination of low abundance and kinetic lability of the pathway intermediates. Through a serial reconstruction of the cobalamin (vitamin B(12)) pathway in Escherichia coli and by His tagging the terminal enzyme in the reaction sequence, we have observed that many unstable intermediates can be isolated as tightly bound enzyme-product complexes. Together, these approaches have been used to extract intermediates between precorrin-4 and hydrogenobyrinic acid in their free acid form and permitted the delineation of the overall reaction catalyzed by CobL, including the formal elucidation of precorrin-7 as a metabolite. Furthermore, a substrate-carrier protein, CobE, that can also be used to stabilize some of the transient metabolic intermediates and enhance their onward transformation, has been identified. The tight association of pathway intermediates with enzymes provides evidence for a form of metabolite channeling.
Evelyne Deery; Susanne Schroeder; Andrew D Lawrence; Samantha L Taylor; Arefeh Seyedarabi; Jitka Waterman; Keith S Wilson; David Brown; Michael A Geeves; Mark J Howard; Richard W Pickersgill; Martin J Warren
Related Documents :
1717266 - Monoclonal antibodies against core and cellulose-binding domains of trichoderma reesei ...
19640646 - Influence of organic and inorganic compounds on oxidoreductive decolorization of sulfon...
24249346 - Subcellular localization of the common shikimate-pathway enzymes in pisum sativum l.
9726856 - Purification and characterization of laccase from chaetomium thermophilium and its role...
12844556 - Alzheimer's disease, neuropeptides, neuropeptidase, and amyloid-beta peptide metabolism.
16233206 - An investigation of the ph-activity relationships of cex, a family 10 xylanase from cel...
Publication Detail:
Type:  Journal Article; Research Support, Non-U.S. Gov't     Date:  2012-10-07
Journal Detail:
Title:  Nature chemical biology     Volume:  8     ISSN:  1552-4469     ISO Abbreviation:  Nat. Chem. Biol.     Publication Date:  2012 Nov 
Date Detail:
Created Date:  2012-10-18     Completed Date:  2013-02-12     Revised Date:  2013-07-11    
Medline Journal Info:
Nlm Unique ID:  101231976     Medline TA:  Nat Chem Biol     Country:  United States    
Other Details:
Languages:  eng     Pagination:  933-40     Citation Subset:  IM    
School of Biosciences, University of Kent, Canterbury, Kent, UK.
Data Bank Information
Bank Name/Acc. No.:
PDB/2W6K;  3NJR;  4AU1;  4FDV; PubChem-Substance/144100029;  144100030
Export Citation:
APA/MLA Format     Download EndNote     Download BibTex
MeSH Terms
Escherichia coli / enzymology,  metabolism
Methyltransferases / chemistry,  metabolism*
Models, Molecular
Molecular Structure
Uroporphyrins / chemistry,  isolation & purification,  metabolism
Vitamin B 12 / biosynthesis*,  chemistry,  metabolism
Grant Support
091163/Z/10/Z//Wellcome Trust; BB/E024203//Biotechnology and Biological Sciences Research Council; BB/E024203/1//Biotechnology and Biological Sciences Research Council; BB/I013334//Biotechnology and Biological Sciences Research Council
Reg. No./Substance:
0/Uroporphyrins; 0/precorrin 4; 23599-55-5/hydrogenobyrinic acid; 68-19-9/Vitamin B 12; EC 2.1.1.-/Methyltransferases

From MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine

Full Text
Journal Information
Journal ID (nlm-journal-id): 101231976
Journal ID (pubmed-jr-id): 32624
Journal ID (nlm-ta): Nat Chem Biol
Journal ID (iso-abbrev): Nat. Chem. Biol.
ISSN: 1552-4450
ISSN: 1552-4469
Article Information
Download PDF

nihms-submitted publication date: Day: 18 Month: 9 Year: 2012
Electronic publication date: Day: 07 Month: 10 Year: 2012
Print publication date: Month: 11 Year: 2012
pmc-release publication date: Day: 01 Month: 5 Year: 2013
Volume: 8 Issue: 11
First Page: 933 Last Page: 940
PubMed Id: 23042036
ID: 3480714
DOI: 10.1038/nchembio.1086
ID: UKMS49908

An enzyme-trap approach allows isolation of intermediates in cobalamin biosynthesis
Evelyne Deery1
Susanne Schroeder1*
Andrew D. Lawrence1*
Samantha L. Taylor1
Arefeh Seyedarabi2
Jitka Waterman3
Keith S. Wilson3
David Brown1
Michael A. Geeves1
Mark J. Howard1
Richard W. Pickersgill2
Martin J. Warren1
1School of Biosciences, University of Kent, Canterbury, Kent, CT2 7NJ, UK.
2School of Biological and Chemical Sciences, Queen Mary University of London, Mile End Road, London E1 4NS, UK.
3Structural Biology Laboratory, Department of Chemistry, University of York, Heslington, York YO10 5DD, UK.
Correspondence: Correspondence to: M. J. Warren, tel 00 44 (0)1227 824690, or R. W. Pickersgill, tel 00 44 (0)20 7882 8444,
*Made an equal contribution to the work
Author contributions ED designed and performed most of the experiments and analysis with support of ADL and SS; ADL performed mass spectroscopy analysis. SLT performed all NMR data acquisition, which was analysed with MJH. SS, AS and RWP contributed the CobLC and CobH-HBA structures and CobL and CobE-HBA models. JV and KSW contributed the CobE structure. DB and SS determined the CobH-5-desmethyl HBA structure. MAG provided insight into substrate channeling. MJW directed all aspects of the project. ED and MJW wrote the manuscript.


The isolation and characterization of biochemical pathways for vitamins and coenzymes has often proved technically challenging, mainly due to the low abundance and kinetic lability of the metabolic intermediates1. Such is the case with the intermediates generated en route to the biosynthesis of adenosylcobalamin (vitamin B12) (Fig. 1), the antipernicious factor. This biosynthesis requires around thirty enzyme-mediated steps for the complete de novo construction of the coenzyme and represents one of the most intricate syntheses found in nature2,3. The biosynthesis of the corrin-ring component is mediated via distinct aerobic and anaerobic pathways that oversee the transformation of the tetrapyrrole primogenitor, uroporphyrinogen III, into cobyric acid.

Corrin ring synthesis involves significant decoration of the tetrapyrrole framework and includes the chemically challenging process of ring contraction to expel the C20 meso position to allow for the tighter coordination of the centrally chelated cobalt ion2-4. Overall, the transformation of uroporphyrinogen III into hydrogenobyrinic acid (HBA), the first stable intermediate along the pathway5,6, involves 8 S-adenosyl-L-methionine (SAM) -dependent methylations, extrusion of an integral macrocyclic carbon atom (ring contraction), decarboxylation of an acetate side chain and methyl migration3 (Fig. 1). There are a number of specific questions still to be addressed on the mechanisms underpinning many of these transformations as well as general questions about how pathways handle large numbers of labile intermediates and how such intricate pathways evolved.

This paper addresses some of these questions through a study of the aerobic cobalamin pathway, which is outlined in Fig.1. The intermediates of the pathway are termed precorrin-n, where n refers to the number of methyl groups that have been added to the framework7. The pathway is initiated by the action of a bis-methylase, CobA, which methylates uroporphyrinogen III at positions 2 and 7 to generate precorrin-2 8. A further methylation at C20 by CobI synthesizes precorrin-3A9. This acts as the substrate for the monooxygenase CobG, which generates a tertiary alcohol at C20 and forms a γ-lactone with the acetate side chain on ring A10,11 (precorrin-3B). Another methyltransferase, CobJ, methylates at C17 and mediates ring contraction by catalyzing a masked pinacol rearrangement11,12. Additional methylations by CobM and CobF at C11 and C1, the latter with loss of the extruded C20 position as acetic acid, results in the synthesis of precorrin-5 and precorrin-6A, respectively13,14. The macrocyclic ring system is reduced by the action of CobK, producing precorrin-6B15. This is the substrate for CobL, a multifunctional enzyme, that promotes the decarboxylation of the acetic acid side chain at C12 and methylation at the northern (C5) and southern (C15) meso positions and in so doing makes precorrin-816. Finally, CobH catalyzes a 1,5-sigmatropic rearrangement that sees the migration of the methyl group from C11 to C12, generating HBA17.

Most of the aforementioned intermediates have only been isolated after esterification and in very small quantities. Yet to investigate further many of these highly unusual reactions simple reproducible protocols need to be established for the production of these compounds. To achieve this we attempted building the pathway recombinantly in E. coli, an organism that does not house the cobalamin biosynthetic pathway. We demonstrated that it was possible to generate HBA in E. coli6 by overproducing enzymes (CobA-I-G-J-M-F-K-L-H) that manage the transformation of uroporphyrinogen III into the comparatively stable orange-colored HBA. This ring-contracted corrin intermediate can be purified from a cell lysate of this strain by ion exchange and reverse phase chromatography6. However, attempts to produce the intervening intermediates between uroporphyrinogen III and HBA using the same strategy of building the pathway one enzyme at a time were not successful. It appears that the intermediates are too unstable to be purified.

In this paper we outline a strategy for the isolation of five pathway intermediates between precorrin-4 and HBA in their free acid forms, stabilized as protein complexes. This approach resulted in the isolation and characterization of a new intermediate along the pathway, precorrin-7 (1), the use of this template for the combinatorial synthesis of a new pathway analogue and the identification of pathway carrier protein. The results have implications not only for understanding how complex pathways enhance and control the stability of labile intermediates by substrate channeling but also for how pathways can be manipulated for the synthesis of cofactor analogues.

Enzymes as traps for pathway intermediates

In the absence of a downstream enzyme in a pathway we wondered if labile intermediates could be stabilized as enzyme-product complexes. To investigate this we constructed a number of plasmids where the corrin pathway was built up in an ordered fashion but where the terminally-encoded enzyme also contained a His-tag. In this way it should be possible to isolate quickly the terminal enzyme in the pathway and determine if it contains any bound intermediate. Thus, the following constructs were made: cobA-I-G-J*, cobA-I-G-J-M*, cobA-I-G-J-M-F*, cobA-I-G-J-M-F-K*, cobA-I-G-J-M-F-K-L* and cobA-I-G-J-M-F-K-L-H* (where the asterisk represents a His-tagged encoded version of the gene).

After E. coli was transformed with these plasmids and the His-tagged protein purified from the appropriate individual strains it was found that CobJ*, CobM*, CobF*, CobK* and CobH* all purified with an associated chromophore that varied in color from blue to yellow and red (Fig. 2). Only CobL* failed to purify with a bound pigment. An SDS gel of the isolated His-tagged proteins is shown in Supplementary Results, Supplemental Fig. 1.

Isolation of precorrin-4 and factor IV with CobJ*

When CobJ* was isolated aerobically from the recombinant strain harboring cobA-I-G-J*, the protein purified with a strong blue coloration. The spectrum obtained from this blue pigment-protein complex had maxima at 378, 477, 510, 593, and 641 nm (Fig. 2a) whilst the free pigment had a mass of m/z 907 (M+H), all of which is consistent with the presence of factor IV12, the oxidized form of precorrin-4.

When the purification was repeated under anaerobic conditions CobJ* was found to purify with a yellow color, consistent with the presence of precorrin-4 (Fig. 2a). Exposure of the anaerobically isolated CobJ to oxygen led to a rapid transformation into the blue pigment. Evidence that the yellow pigment was precorrin-4 came from incubating this protein-bound complex with SAM, NADPH and purified CobF, CobM, CobK, CobL and CobH, which resulted in the appearance of the orange-colored HBA and its characteristic spectrum (data not shown).

Isolation of complexes with CobF* and CobM*

A very light yellow pigment was isolated with CobM* after extraction and purification of the protein by immobilized metal affinity chromatography from the recombinant strain harboring cobA-I-G-J-M* (Fig. 2b) while a dark yellow pigment was observed associated with CobF* after its purification from the strain containing cobA-I-G-J-M-F* (Fig. 2c). The spectra obtained from CobM* and CobF* were not very defined and the attached chromophores were not characterized in detail. Moreover, incubation of these complexes with the downstream enzymes for HBA synthesis did not result in the production of HBA. Thus the compounds bound to CobM* and CobF* are not pathway intermediates and are likely to represent either oxidized products or derailed pathway analogues. Such derailed analogues have been observed previously and result from either overmethylation or mis-methylation18,19, a consequence of a particularly high concentration of biosynthetic enzyme.

Isolation of precorrin-5 and -6A

Although precorrin-5 and -6A were not found associated with CobM* and CobF* these intermediates could be generated by adding the enzymes sequentially to the CobJ-precorrin-4-complex. Thus, addition of SAM and CobM resulted in the appearance of precorrin-5 (Fig. 2b), a pale yellow compound that was very unstable and which turned purple within minutes inside the anaerobic chamber in the absence of the next enzyme of the pathway. This purple material appears to be a dead-end compound as addition of further enzymes did not result in any further spectral changes. Addition of SAM plus CobM and CobF to the CobJ-precorrin-4-complex generated precorrin-6A (Fig. 2c). This could be rapidly reduced to precorrin-6B when it was incubated with NADPH and CobK, while the further addition of CobL lead to the appearance of precorrin-8, from which HBA was produced by the addition of CobH (data not shown).

Isolation of precorrin-6B

When CobK* was produced and purified from the recombinant strain containing cobA-I-G-J-M-F-K*, it was found associated with a yellow pigment that had an absorption maximum at 348 nm (Fig. 2d). This enzyme-product complex could be converted into HBA by incubation with the downstream enzymes CobL and CobH (data not shown). When released from CobK the intermediate was found to have a mass of 897.4 (M+H), which is that expected for precorrin-6B.

CobL* does not form a stable enzyme-product complex

The only protein within this section of the cobalamin biosynthetic pathway that does not appear to bind anything substantial is CobL. Thus, CobL* isolated from the strain harboring cobA-I-G-J-M-F-K-L* gave a very bland spectrum with only minor absorbances (Fig. 2e, f).

CobH* binds HBA

When CobH* was purified from the strain harboring cobA-I-G-J-M-F-K-L-H* it was found to purify with a strong orange chromophore (Fig. 2g and Supplemental Fig. 2). A UV-visible scan of the protein solution showed that it contained a slightly altered UV-visible spectrum of HBA, with absorption maxima at 330 and 484 nm (Fig. 2g). The CobH*-HBA complex could be dissociated by boiling, which resulted in a concomitant change in the spectrum to that of free HBA with absorption maxima at 329, 496 and 522 nm (Fig. 2g) and a mass of m/z 881.4 (M+H) was observed by mass spectroscopy. The structure of this free acid form of HBA was further confirmed by NMR (Supplemental Table 1a) and by the crystallization (Supplemental Fig. 2) and subsequent X-ray structure determination of the CobH-HBA complex. Previous reports have highlighted a tight association between CobH and HBA17,20.

CobE binds a number of different intermediates

The lack of a binding protein for precorrin-8 led us to think that this role could be played by CobE, whose function within the aerobic B12 pathway has been the subject of some speculation4. In the current work we have shown that CobE is not required for the in vitro transformation of precorrin-4 into HBA as this conversion is mediated by CobM, F, K, L and H alone. To investigate the possibility that CobE may act as a substrate carrier cobE was modified in such a manner so as to include a fusion encoding a hexahistidine tag (CobE*) and then cloned in tandem into a range of plasmids with the genes responsible for the production of precorrin-4, precorrin-5, precorrin-6A, precorrin-6B, precorrin-8 or HBA. This resulted in the following constructs: cobA-I-G-J-E*, cobA-I-G-J-M-E*, cobA-I-G-J-M-F-E*, cobA-I-G-J-M-F-K-E*, cobA-I-G-J-M-F-K-L-E* and cobA-I-G-J-M-F-K-L-H-E*. Encouragingly, when CobE* was isolated from the resulting strains transformed with these plasmids the protein was found to purify associated with a range of different pigments (Fig. 2). An SDS gel of CobE* isolated from these strain is shown in Supplemental Fig. 1.

When CobE* was isolated from the strain harboring cobA-I-G-J-E* there was little evidence of anything binding to the protein (Fig. 2a). However, when CobE* was isolated from strains with the genes required for precorrin-5 and precorrin-6A (i.e. from strains containing cobA-I-G-J-M-E* and cobA-I-G-J-M-F-E* respectively) the protein was found to purify with strong absorption spectra (Fig. 2b, c). The spectra did not correspond to free precorrin-5 or precorrin-6A but did have some similarity to the spectra observed with CobM* and CobF*. Unfortunately, as with CobM* and CobF*, incubation of the isolated CobE*s from the precorrin-5 and precorrin-6A strains with the downstream enzymes for HBA synthesis did not result in a change in the spectra. More encouragingly, CobE* isolated from the strain with the genetic components for precorrin-6B synthesis (i.e. containing cobA-I-G-J-M-F-K-E*) had a broad absorption spectrum around 350 nm, although it is not as defined as the product-complex spectrum that was observed when CobK* was purified (Fig. 2d). Significantly, incubation of this CobE* protein with CobL and CobH resulted in the production of HBA. Thus CobE* has the ability to bind precorrin-6B.

CobE* was also found to bind precorrin-8 and HBA since isolation of CobE* from strains harboring cobA-I-G-J-M-F-K-L-E* and cobA-I-G-J-M-F-K-L-H-E* resulted in yellow and orange colored protein, respectively (Fig. 2f, g). The yellow color was shown to be precorrin-8 as incubation of this protein with CobH resulted in HBA production. The structure of precorrin-8 was confirmed by NMR (Supplemental Table 1b). It is interesting to note that CobE is able to bind precorrin-8 whereas CobL does not. Overall, therefore, CobE is able to bind precorrin-6B, precorrin-8 and HBA. As a result, we propose that its role in the aerobic pathway is to act as a substrate carrier protein or substrate chaperone, binding excess labile compounds and ensuring that they are delivered to the next enzyme in the pathway.

To support this theory we observed that CobE was able to stabilize precorrin-8. When left free in solution precorrin-8 bleaches with a half-life of 3-4 hours (Fig. 3a and Supplemental Fig. 3). In contrast, when bound to CobE, comparatively little precorrin-8 is lost (Fig. 3a). Moreover, the yield of HBA from precorrin-6B through incubation with CobL, CobH and SAM, is approximately doubled in the presence of CobE (Fig 3b and Supplemental Fig. 3). In order to understand how CobE, which is a comparatively small protein of around 150 amino acids, may act to stabilize the various pathway intermediates that it binds, the structure of the protein was solved by X-ray crystallography (Fig. 3c). This revealed that CobE has a globular α/β fold with dimensions approximately 30 × 35 × 35 Å3, with a single domain containing a five-stranded β-sheet with five associated α-helices and one 310-helix (Fig. 3c). The β-sheet and α-helices α5 and α6 create the walls of a cleft of approximate dimensions 8 × 15 × 17 Å3. The cleft is likely to be the binding site of the protein and in the present structure several non-protein ligand molecules are bound therein. The presence of conserved basic residues lining this cleft strongly suggests their involvement in binding groups, such as the negative charges on the side chain groups of the corrin ring. Modeling of HBA in the cavity suggests that CobE can easily accommodate molecules the size of HBA in a slightly bent conformation both in terms of shape and electrostatic complementary (Fig. 3d), though further experimental evidence is needed to confirm the scope of CobE function.

Although crystallography of CobE-bound complexes has not yet been achieved, NMR spectra of CobE bound to a range of compounds were recorded. 15N,1H HSQC spectra of CobE before and after the addition of HBA and precorrin-6B (Supplemental Fig. 4a-b) support binding of the intermediates by this protein. Changes in chemical shifts further suggest that a significant structural change in CobE must occur upon interaction. Closer inspection of the data (Supplemental Fig. 4) confirm both similarities and differences between the chemical shift changes observed for CobE with each intermediate. Apo CobE 15N HSQC data contains 62% of the resonances expected for the protein with this number rising to 80% in the presence of ligands. To probe the specificity of binding of HBA to CobE further, quenching fluorescence assays were performed that confirmed a 1:1 stoichiometric HBA-CobE interaction with a Kd = 0.08 μM (Supplemental Fig. 5). Additionally, 15N HSQC NMR titration curve datasets were created using variable final concentrations between 0.0 and 1.0 mM of HBA with a final constant 0.2 mM concentration of CobE. Tracking of HBA addition using chemical shift changes confirmed the tight interaction observed by fluorescence quenching (Kd = 0.12 μM; Supplemental Fig. 6a). Fitting this data also yielded a second event with a lower Kd of 309 μM (Supplemental Fig. 6b), though the basis for this is unknown.

CobL dissection

With the basis for CobE function established, we returned to the one enzyme that had proven uninformative in the reconstruction assays, CobL. The precorrin-6B/CobK* complex was used to study the reaction catalyzed by CobL, which is responsible for the C5 and C15 methylations and the decarboxylation of the acetate side chain attached to the C12. When precorrin-6B-bound CobK was incubated with SAM and CobL, the solution turned a brighter yellow shade associated with the generation of precorrin-8 (Supplemental Fig. 7). Analysis of the resulting compound revealed that it had a mass of m/z of 881.4 (M+H) in agreement with such a transformation.

Comparative studies suggest that CobL has arisen from a fusion between two separate classes of SAM methyltransferases, the canonical B12-biosynthetic class III type and a class I methyltransferase. Indeed, in the anaerobic pathway the orthologue of CobL is sometimes found as two separate genes, which is the case in Salmonella enterica where they are termed CbiE and CbiT. In order to investigate the role played by the two regions within CobL the gene was dissected to give cobLN, which encodes the N-terminal part of CobL, and cobLC, which encodes the C-terminal domain of CobL (Fig. 4a, and Supplemental Fig. 8 for sequence alignment). The two fragmented genes were cloned individually into pET14b and transformed into E. coli BL21 cells to allow the production of their encoded dissected protein products. Incubation of the precorrin-6B/CobK* complex with CobLN and SAM did not alter the original spectrum (Supplemental Fig. 7) whereas, in contrast, the same incubation with CobLC and SAM resulted in the appearance of a yellow pigment with a spectrum that was slightly shifted in comparison to that of precorrin-8 (Supplemental Fig. 7). Mass spectral analysis of this compound was consistent with the decarboxylation and monomethylation of precorrin-6B, precorrin-7 (m/z of 867.4 (M+H)). This is the first time this compound had been isolated. Precorrin-7 could subsequently be transformed into precorrin-8 by the addition of full length CobL, but not CobLN (Supplemental Fig. 7). It would thus appear that CobLN was inactivated by the dissection procedure.

An insight into the structure/function of CobL was initiated via X-ray crystallography. Although CobL did crystallize the diffraction was poor and reliable protein phases have yet to be determined. However, CobLC also crystallized and this structure was determined to 2.7 Å resolution (Fig. 4b). The SAH and putative precorrin binding sites are away from the subunit interface and open to solvent. The SAH binding site is at the C-terminal end of β-strands 1, 2, and 3, sandwiched between two glycine rich loops and adjacent to the putative precorrin-6B binding cleft (Fig. 4c). The structure of CobL can now be modeled using the solved structure of the C-terminal domain, which forms a tetramer, plus two dimers of the closest canonical methyltransferase, Archaeoglobus fulgidus CbiE (Supplemental Fig. 9). Each of the four protein chains in the tetramer has two SAH binding sites, one in the C-terminal non-canonical and one in the N-terminal canonical domain. The tetramer is a dimer of dimers with the active centers apparently independent in this model. The active sites are not juxtaposed, so there is apparently no simple transfer of precorrin-7 from the active centre of the C-terminal domain to the N-terminal domain. However, this role could be played by CobE, and such a model would predict that CobE would also bind precorrin-7. The positioning of the C-terminal domain relative to the N-terminal domain of CobL, while not bringing the active centers of the two domains close together, could influence the accessibility of the N-terminal domain’s active centre to CobE or CobH.

Isolation and characterization of precorrin-7

To investigate the ability of CobE to bind precorrin-7, CobE* was produced in a strain harboring cobA-I-G-J-M-F-K-LC-E*. The protein was, indeed, found to purify with precorrin-7 attached (Fig. 2e). As with precorrin-8, CobE also appears to stabilize precorrin-7 since the bound complex has an extended half life in comparison to the unbound intermediate, where the binding can also be monitored by NMR (Supplemental Fig. 4c). After the protein was gently denatured NMR analysis of the liberated precorrin-7 was undertaken. The data support the presence of major and minor tautomeric states (Supplemental Table 1c) that differ in the presence of a proton attached at the C8 position in the minor state (Supplemental Fig. 10). This was confirmed from 13C- 1H HSQC (Heteronuclear Single Quantum Correlation spectroscopy), HMBC (Heteronuclear Multiple Bond Correlation spectroscopy) and TOCSY (TOtal Correlation SpectroscopY) datasets (Supplemental Fig. 11). Such resonances are broad, suggesting strong evidence for chemical exchange in this state. 1H-13C HSQC spectra were also collected at 5°C and 40°C in an attempt to minimize the presence of the minor form but further deconvolution of the data was not possible.

The major state of precorrin-7 is consistent with the data shown in Supplemental Table 1c. The NMR confirms the presence of methyl groups at positions C1, C2, C7, C11, C12, C15 and C17. This demonstrates that CobLC is responsible for the C15 methylation and decarboxylation of the acetate side chain on C12. From a mechanistic perspective it is likely that the decarboxylation and methylation are concerted reactions (Fig. 5).

To generate a form of CobL that only has C5 methylase activity, with no C15 methylase/decarboxylase activity, two separate point mutations, G257R and E276A, were introduced in the SAM binding site of the CobLC domain of full-length CobL (Fig. 4c and Supplemental Fig. 8). Incubation of CobLG257R or CobLE276A with precorrin-6B-bound CobK and SAM did not alter the spectrum of the substrate. However, when CobLC was added to the reaction, the spectrum of precorrin-8 was produced (Supplemental Fig. 7), which changed to HBA when CobH was added. CobLG257R or CobLE276A are therefore able to use precorrin-7 as a substrate and functional solely as C5 methyltransferases (Supplemental Fig. 7). These series of experiments clearly delineate the reaction sequence for the transformation of precorrin-6B into precorrin-8.

Synthesis of C5-desmethyl-HBA

A further surprising result was obtained by incubating precorrin-7 with CobH. Here the compound with an absorption maximum at 397 nm was readily transformed into a compound with a spectrum that was similar to, but distinct from, HBA (Supplemental Fig. 12). Mass spectral analysis of this new compound revealed that it is 14 mass units smaller than HBA, consistent with the formation of C5-desmethyl-HBA (2) (m/z of 867.4 (M+H)). It would thus appear that precorrin-7 acts as an out-of-turn substrate for CobH. The structure of the compound was confirmed by crystallizing the CobH-desmethyl-HBA complex, from which the resulting 1.4 Å data clearly revealed the absence of the C-5 methyl group (Fig. 5).


This report describes an approach for trapping intermediates involved in the cobalamin biosynthetic pathway in their natural (free acid) form as enzyme-product complexes. CobJ, CobK and CobH, were found to bind their products tightly and protect them from the external environment. Furthermore, CobE, a protein to which no function had been formally assigned, was found to associate with a number of intermediates including precorrin-6B, precorrin-7, precorrin-8 and HBA. Our in vitro data supports a role for CobE in the stabilization of intermediates such as precorrin-7 and precorrin-8, most likely through binding to the large crevice found at the opposite end of the protein to the N and C termini. The inclusion of CobE in ex vivo assays of precorrin-6B to HBA also leads to an enhancement in overall yield. Thus, CobE plays a role as a substrate carrier or chaperone protein, shuttling labile intermediates between different Cob enzymes.

The observation that several stable enzyme-product complexes can be isolated from the aerobic pathway is unusual for a metabolic pathway and is probably linked to the low stability of the pathway intermediates. The high affinity of the intermediates for the enzymes require very low dissociation rate constants that are incompatible with the expected flux through the pathway. This implies that the enzymes retain the unstable intermediates until some signal or interaction allows the release of the product to be efficiently used by the next enzyme in the pathway. The established hypothesis for such behaviour is called metabolite channelling – that is the enzyme holds onto its product to pass it efficiently if not directly onto the succeeding step in the biosynthesis. At one time this was a contentious area21,22 but there are now several clear-cut examples where substrate channelling does occur (for reviews see23-25). We believe that this is the first report of a series of enzymes that form stable product complexes, in the absence of a stable multienzyme complex. Channelling requires a series of transient interactions between the enzymes along the pathway to allow intermediates to be passed from enzyme to enzyme. Given that the cytoplasm of E. coli appears as a sieve that permits the diffusion of small molecules while limiting the movement of larger ones26,27 the observed formation of stable enzyme-product complexes would suggest that the enzymes of the cobalamin pathway must loosely associate in the form of a metabolon28 to allow the pathway to operate.

However, our research does not differentiate between “real channelling” and “enhanced transfer” between enzymes, which can occur between two enzymes if there is a higher probability of substrate rebinding to the downstream enzyme. This will require further research to measure the loss of substrate to side reactions as a function of enzyme concentration in pairs of enzymes and in the complete pathway.

The research describes a simple but highly efficient method for the isolation of pathway intermediates, thereby allowing the elucidation of metabolic pathways in a fairly rapid timescale. This method will be particularly applicable to pathways involving labile compounds. This report also describes the combinatorial synthesis of a non-physiological corrin analogue, 5-desmethyl-HBA. Such an approach could be used to make a range of cobaltoporphyrin and corrin derivatives, molecules that are of increasing interest in the bioenergy field since they have been shown to have the ability to oxidize water29,30. It can also be used to make useful probes for the study of cobalamin-dependent reactions and derivatives of the cofactor for enhanced drug delivery31.

Construction of plasmids

All cobalamin biosynthetic genes were amplified by PCR using genomic DNA from either Rhodobacter capsulatus SB1003 (cobA, I, J, F, M, K, L and H), Brucella melitensis 16M (cobG, cobE), Pseudomonas denitrificans (cobG) (gift from Dr C. Roessner, Texas A & M University) or Pseudomonas aeruginosa (cobE). The cob genes were amplified individually with primers containing either AseI or NdeI restriction sites at the 5′ end and a SpeI site at the 3′ end to allow them to be cloned contiguously within the plasmid as previously described6. Genes were cloned into modified pET3a and pET14b plasmids as highlighted in Supplementary Table 2.

Recombinant protein overproduction and purification

E. coli strain BL21 star (DE3) pLysS was transformed with the relevant plasmids. The recombinant strains were grown in LB at 37°C. The protein overproduction was induced with isopropyl 1-thio-β-D-galactopyranoside (0.4 mM). The cells were resuspended in buffer A (500 mM NaCl, 5 mM imidazole, 20 mM HEPES, pH 7.5) and sonicated. The protein was purified using Chelating SepharoseTM charged with NiSO4. Unbound proteins were washed off with buffer A and with buffer A containing 60 mM imidazole. Proteins were eluted with buffer A containing 400 mM imidazole and passed through a PD10 column equilibrated with buffer B (20 mM HEPES pH 7.5 and 100 mM NaCl).

Purification of proteins with trapped intermediates

The different E. coli strains transformed with a plasmid encoding multiple proteins of which a unique histidine-tagged protein binding the trapped intermediate were grown in 2YT at 28°C for 24 hours. The protein purification procedure was identical to the method above except that NaCl was lowered to 100 mM.

Site-directed mutagenesis

The CobL G257R mutant was created using the QuickChange II site-directed mutagenesis (Agilent). The CobL E276A was constructed following the site-specific mutagenesis by overlap extension protocol32.

Assays from PC4 to HBA or C5-desmethyl-HBA

All assays were performed in buffer B. SAM was added at a final concentration of 0.58 mM, NADPH at 15 μM and all proteins were added at 2 μM. The reactions were incubated at room temperature. Anaerobic conditions were obtained using an anaerobic glove-box (Belle Technology).

Mass spectrometry

Samples were separated on an Agilent 1100 series HPLC coupled to a micrOTOF-Q II (Bruker) mass spectrophotometer using an Ace 5 AQ column (2.1 × 150 mm; Advanced Chromatography Technologies) maintained at 30°C and with a flow rate of 0.2 ml/min. The mobile phase consisted of 0.1% acetic acid (v/v) in water (solvent A) and 100% methanol (solvent B). The initial conditions were set at 95% A, 5% B. The concentration of solvent B was increased with a linear gradient to 80% over 30 min and back to the initial conditions at 40 min. The mass spectrometer was operated in the positive electrospray mode.

NMR of precorrin-7, precorrin-8 and HBA

Samples of precorrin-7, precorrin-8 and HBA for NMR analysis were prepared from their respective strains bound to either CobE* or CobH*. Precorrin-7 and precorrin-8 were purified trapped with CobE as described above whereas HBA was trapped with CobH. The complexes were then heated to denature the protein and leave the intermediate free in solution. After lowering the pH to around 5, the pigments were applied to LiChroprep RP18 columns. The compounds were briefly washed with water and eluted in 50% ethanol. After lyophilization, they were finally dissolved in deuterium oxide (99.9% atom D). All NMR experiments were carried out using a 600 MHz (1H) Varian UnityINOVA spectrometer a 5 mm z-PFG HCN NMR probe or a 600 MHz (1H) Bruker Avance III spectrometer with a 5 mm QCI-F cryoprobe.

NMR details are given in the supplementary information. 1H Chemical shift referencing was based on the position of the water resonance, 13C referencing used 1H/13C gyromagentic ratios to define indirect carrier position33 and all data were obtained at 25°C unless otherwise stated. Data processing was completed using NMRPipe34 and resonance assignments were completed using CCPN analysis 235.


CobE crystals were obtained from 0.1M MES buffer pH 6.9, 2 M ammonium sulphate and 5% dioxane36 and the structure solved as outlined in the Supplemental information. The final R-factor and R-free are 0.194 and 0.268 respectively for a model with good stereochemistry (Supplemental Table 3a). CobLC crystals were produced as described previously37 and solved at 2.7 Å by molecular replacement using the C. diphtheriae putative precorrin-6Y C5,15-methyltransferase targeted domain structure (PDB code: 3hm2) as a search model. The C. diphtheria methyltransferase sequence is 41 % identical to that of R. capsulatus CobLC and MOLREP38 gave a model with a score of 0.314 that could be successfully rebuilt and refined using CNS39 and BUCANEER40/REFMAC41 with manual rebuilding using COOT42 (Supplemental Fig. 9). The final R-factor and R-free are 0.185 and 0.254 respectively for a model with good stereochemistry (Supplemental Table 3b). The CobH complexes were crystallized using 25-30% PEG 8K, 0.2 M ammonium sulfate and 0.1 M cacodylate pH 6.5. The structures were solved by molecular replacement using 1I1H20 and refined using similar methods to those described for CobLC (Supplemental Table 3c)43-45. Coordinates and structure factor amplitudes have been deposited in the protein databank with accession codes: 2BSN (CobE), 3NJR (CobLC), 4FDV (CobH-HBA) and 4AU1 (CobH-desmethyl-HBA).

Molecular modeling

Docking was used to fit HBA into the binding cavity of CobE. There was a single orientation of HBA that fitted well with the acidic side chains mapping to basic regions of CobE. No adjustments of the conformation of HBA or CobE were necessary. The CobL structure was modeled by docking the N-terminal domains of CbiE onto the C-terminal domains of CobLC. Again this was a straightforward modeling procedure achieved using molecular docking in PyMOL.

Supplementary Material 1


FN3Competing Financial Interests Statement. The authors have no competing financial interest.


This work was supported by grants from the Biotechnology and Biological Sciences Research Council (BBSRC) BB/E024203 and BB/I013334 and the Wellcome Trust 091163/Z/10/Z. We thank Dr. Michelle Rowe for additional NMR technical support. Diffraction data were collected at the European Synchrotron Radiation Facility, Grenoble, France (for CobE) and the Diamond Light Source, Oxfordshire, UK (for CobL & CobH).

1. Webb ME,Marquet A,Mendel RR,Rebeille F,Smith AG. Elucidating biosynthetic pathways for vitamins and cofactorsNat. Prod. RepYear: 200724988100817898894
2. Battersby AR. How nature builds the pigments of life: the conquest of vitamin B12ScienceYear: 1994264155115578202709
3. Warren MJ,Raux E,Schubert HL,Escalante-Semerena JC. The biosynthesis of adenosylcobalamin (vitamin B12)Nat. Prod. RepYear: 20021939041212195810
4. Blanche F,et al. Vitamin B12: How the problem of its biosynthesis was solvedAngew. Chem. Int. Ed. EnglYear: 199534383411
5. Blanche F,et al. Hydrogenobyrinic Acid - Isolation, Biosynthesis, and FunctionAngew. Chem. Int. EditYear: 199029884886
6. McGoldrick HM,et al. Identification and characterization of a novel vitamin B12 (cobalamin) biosynthetic enzyme (CobZ) from Rhodobacter capsulatus, containing flavin, heme, and Fe-S cofactorsJ. Biol. ChemYear: 20052801086109415525640
7. Uzar HC,Battersby AR,Carpenter TA,Leeper FJ. Biosynthesis of Porphyrins and Related Macrocycles .28. Development of a Pulse Labeling Method to Determine the C-Methylation Sequence for Vitamin-B12J. Chem. Soc. Perk. TYear: 1987116891696
8. Blanche F,Debussche L,Thibaut D,Crouzet J,Cameron B. Purification and characterization of S-adenosyl-L-methionine: uroporphyrinogen III methyltransferase from Pseudomonas denitrificansJ. BacteriolYear: 1989171422242312546914
9. Warren MJ,et al. Enzymatic synthesis and structure of precorrin-3, a trimethyldipyrrocorphin intermediate in vitamin B12 biosynthesisBiochemistryYear: 1992316036091731915
10. Debussche L,et al. Biosynthesis of vitamin B12: Structure of precorrin-3B, the trimethylated substrate of the enzyme catalysing ring contractionJ. Chem. Soc., Chem. CommunYear: 199311001103
11. Scott AI,et al. Biosynthesis of vitamin B12. Discovery of the enzymes for oxidative ring contraction and insertion of the fourth methyl groupFEBS LettsYear: 19933311051088405386
12. Thibaut D,Debussche L,Frechet D,Herman F,Vuilhorgne M,Blanche F. Biosynthesis of vitamin-B12 - the structure of factor-IV, the oxidized form of precorrin-4J. Chem. Soc., Chem. CommunYear: 1993513515
13. Min C,Atshaves BA,Roessner CA,Stolowich NJ,Spencer JB,Scott AI. Isolation, structure and genetically engineered synthesis of precorrin-5J. Am. Chem. SocYear: 19931151038010381
14. Thibaut D,Blanche F,Debussche L,Leeper FJ,Battersby AR. Biosynthesis of vitamin B12: structure of precorrin-6x octamethyl esterProc. Nat.l Acad. Sci. USAYear: 19908788008804
15. Blanche F,et al. Precorrin-6x reductase from Pseudomonas denitrificans: purification and characterization of the enzyme and identification of the structural geneJ. BacteriolYear: 1992174103610421732193
16. Blanche F,et al. Biosynthesis of vitamin B12 in Pseudomonas denitrificans: the biosynthetic sequence from precorrin-6y to precorrin-8x is catalyzed by the cobL gene productJ. BacteriolYear: 1992174105010521732195
17. Thibaut D,et al. The final step in the biosynthesis of hydrogenobyrinic acid is catalyzed by the cobH gene product with precorrin-8x as the substrateJ. BacteriolYear: 1992174104310491732194
18. Santander PJ,Kajiwara Y,Williams HJ,Scott AI. Structural characterization of novel cobalt corrinoids synthesized by enzymes of the vitamin B-12 anaerobic pathwayBioorg. Med. ChemYear: 20061472473116198574
19. Scott AI,Warren MJ,Roessner CA,Stolowich NJ,Santander PJ. Development of an Overmethylation Strategy for Corrin Synthesis - Multienzyme Preparation of PyrrocorphinsJ. Chem. Soc., Chem. CommYear: 1990593597
20. Shipman LW,Li D,Roessner CA,Scott AI,Sacchettini JC. Crystal Structure of Precorrin-8x Methyl MutaseStructure with Folding & DesignYear: 2001958759611470433
21. Pettersson G. No Convincing Evidence Is Available for Metabolite Channeling between Enzymes Forming Dynamic ComplexesJ. Theor. BiolYear: 199115265691753770
22. Wu XM,Gutfreund H,Lakatos S,Chock PB. Substrate channeling in glycolysis: a phantom phenomenonProc. Natl. Acad. Sci. USAYear: 1991884975011988948
23. Huang XY,Holden HM,Raushel FM. Channeling of substrates and intermediates in enzyme-catalyzed reactionsAnn. Rev. BiochemYear: 20017014918011395405
24. Jorgensen K,et al. Metabolon formation and metabolic channeling in the biosynthesis of plant natural productsCurr. Opin. Plant BiolYear: 2005828029115860425
25. Miles EW,Rhee S,Davies DR. The molecular basis of substrate channelingJ. Biol. ChemYear: 1999274121931219610212181
26. McGuffee SR,Elcock AH. Diffusion, crowding & protein stability in a dynamic molecular model of the bacterial cytoplasmPLoS Comput. BiolYear: 20106e100069420221255
27. Mika JT,van den Bogaart G,Veenhoff L,Krasnikov V,Poolman B. Molecular sieving properties of the cytoplasm of Escherichia coli and consequences of osmotic stressMol. MicroYear: 201077200207
28. Robinson JB Jr.,Inman L,Sumegi B,Srere PA. Further characterization of the Krebs tricarboxylic acid cycle metabolonJ. Biol. ChemYear: 1987262178617902433288
29. Dogutan DK,et al. Hangman corroles: efficient synthesis and oxygen reaction chemistryJ. Am. Chem. SocYear: 201113313114021142043
30. Lee CH,Dogutan DK,Nocera DG. Hydrogen generation by hangman metalloporphyrinsJ. Am. Chem. SocYear: 20111338775877721557608
31. Waibel R,et al. New derivatives of vitamin B12 show preferential targeting of tumorsCancer ResYear: 2008682904291118413759
32. Sambrook J,Russell D. Molecular cloning: A laboratory manualYear: 2001Cold Spring Harbor Laboratory Press3rd edn
33. Wishart DS,Sykes BD. Chemical-Shifts as a Tool for Structure DeterminationMethod. EnzymolYear: 1994239363392
34. Delaglio F,et al. NMRpipe - a Multidimensional Spectral Processing System Based on Unix PipesJ. Biomol. NMRYear: 199562772938520220
35. Vranken WF,et al. The CCPN data model for NMR spectroscopy: development of a software pipelineProteinsYear: 20055968769615815974
36. Vevodova J,Graham RM,Raux E,Warren MJ,Wilson KS. Crystallization and preliminary structure analysis of CobE, an essential protein of cobalamin (vitamin B12) biosynthesisActa Crystallogr. Sect. F Struct. Biol. Cryst. CommunYear: 200561442444
37. Seyedarabi A,et al. Cloning, purification and preliminary crystallographic analysis of cobalamin methyltransferases from Rhodobacter capsulatusActa Crystallogr. Sect. F Struct. Biol. Cryst. CommunYear: 20106616521656
38. Vagin A,Teplyakov A. Molecular replacement with MOLREPActa Crystallogr. DYear: 201066222520057045
39. Brunger AT. Version 1.2 of the Crystallography and NMR systemNat ProtocYear: 200722728273318007608
40. Cowtan K. The Buccaneer software for automated model building. 1. Tracing protein chainsActa Crystallogr. DYear: 2006621002101116929101
41. Murshudov GN,Vagin AA,Dodson EJ. Refinement of macromolecular structures by the maximum-likelihood methodActa Crystallogr. DYear: 19975324025515299926
42. Emsley P,Lohkamp B,Scott WG,Cowtan K. Features and development of CootActa Crystallogr. DYear: 20106648650120383002
43. Battye TGG,Kontogiannis L,Johnson O,Powell HR,Leslie AGW. iMOSFLM: a new graphical interface for diffraction-image processing with MOSFLMActa Crystallogr. DYear: 20116727128121460445
44. Blanc E,et al. Refinement of severely incomplete structures with maximum likelihood in BUSTER-TNTActa Crystallogr. DYear: 2004602210222115572774
45. Winter G. xia2: an expert system for macromolecular crystallography data reductionJ. Appl. CrystallogrYear: 201043186190

Article Categories:
  • Article

Previous Document:  A RubisCO-like protein links SAM metabolism with isoprenoid biosynthesis.
Next Document:  Dually cationic and anionic pH/temperature-sensitive injectable hydrogels and potential application ...