Bae et al. 2000 — Synthetic phytochelatins displayed on E. coli enhance Cd²⁺ bioaccumulation

Bae, Chen, Mulchandani, and Mehra (Department of Chemical and Environmental Engineering and Environmental Toxicology Program, University of California, Riverside) report the construction and characterisation of recombinant Escherichia coli JM105 strains that display synthetic phytochelatin analogs of defined chain length (EC8, EC11, EC20 with 8, 11, and 20 cysteines respectively) on the cell surface via a Lpp-OmpA fusion anchor, and demonstrate that whole cells displaying the longest analog (EC20) accumulate ≈15-fold more Cd²⁺ on a dry-weight basis than the wild-type control and ≈2-fold more than cells expressing EC20 periplasmically as an MBP fusion. Purified MBP-EC20 is shown to bind Cd²⁺ at a stoichiometry of ≈10 Cd²⁺ per peptide — about 40 % higher than the canonical 7 Cd²⁺ per peptide reported for mammalian metallothioneins. The paper reports no measurements in any food or environmental matrix; its value to the wiki is conceptual background for the cadmium page and for the mitigation/remediation-evidence section on engineered-microbial biosorbents.

Why this matters

• It is one of the foundational primary studies for the surface-displayed-phytochelatin strategy of heavy-metal bioremediation, alongside the Sousa et al. (1996, 1998) LamB-MT fusion line of work that the authors benchmark against. The Lpp-OmpA-EC fusion system reported here goes on to seed a substantial downstream literature on engineered microbial biosorbents for cadmium and mercury remediation.
• It demonstrates a clean chain-length dependence: EC8 → ≈18, EC11 → ≈29, EC20 → ≈60 nmol Cd²⁺/mg dry cells, consistent with the predicted increase in metal-binding centres per peptide as cysteine content rises. The chain-length dependence is direct mechanistic support for the design principle that longer cysteine-rich peptides give higher whole-cell bioaccumulation capacity.
• It documents an unexpected localisation effect — surface-anchored EC20 outperforms periplasmically-expressed EC20 by a factor of ≈2 in whole-cell Cd²⁺ accumulation despite the periplasmic construct having higher per-cell expression — and attributes the gap to cysteine binding sites in the MBP-EC20 fusion being vacant or oxidised in the periplasmic context, with less than one Cd²⁺ associated per fusion protein in the unprocessed periplasmic preparation versus ≈10 Cd²⁺ per fusion after reduction with DTT.
• It provides a numerical anchor (10 Cd²⁺/EC20 vs 7 Cd²⁺/MT) for the comparative-stoichiometry literature on cysteine-rich heavy-metal-binding peptides, of relevance whenever the wiki’s mitigation page contrasts phytochelatin-based and metallothionein-based engineering routes.

Key numbers

• EC20 Cd²⁺ binding stoichiometry, purified MBP-EC20 fusion after DTT reduction: 9.9, 10.1, 9.8 Cd²⁺ per EC20 in Sephadex G-25 fractions 6, 7, 8 respectively (Fig. 4, p. 521). Reported in the Abstract as a stoichiometry of 10 Cd²⁺ per peptide.
• Mammalian metallothionein (MT) reference stoichiometry the authors compare against: 7 Cd²⁺ per MT (Hamer 1986; Stillman et al. 1992 — cited values, not measured here). Authors conclude EC20 has ≈40 % higher Cd²⁺ binding capacity than MTs on a per-peptide basis (Discussion, p. 522).
• Whole-cell Cd²⁺ accumulation, E. coli JM105 strains grown in MJS medium with 0.1 mM CdSO4, 1 mM IPTG induction, monitored to 16 h (Fig. 6, p. 522):
  • pUC18 negative control: ≈4 nmol Cd²⁺/mg dry cell weight (Fig. 5; the exact bar value is not numerically tabulated and is read from the graph).
  • pLO8 (Lpp-OmpA-EC8, 8 cysteines): ≈18 nmol Cd²⁺/mg dry cells.
  • pLO11 (Lpp-OmpA-EC11, 11 cysteines): ≈29 nmol Cd²⁺/mg dry cells.
  • pLO20 (Lpp-OmpA-EC20, 20 cysteines, surface-displayed): ≈60 nmol Cd²⁺/mg dry cells (called out in the Abstract as “a maximum of 60 nmoles of Cd²⁺/mg dry cells”).
  • pM20 (MBP-EC20, periplasmic expression): ≈35 nmol Cd²⁺/mg dry cells (Fig. 5).
  • Surface-displayed pLO20 / periplasmic pM20 ratio: ≈2-fold higher whole-cell Cd²⁺ accumulation for the surface-displayed construct (Abstract; Fig. 5).
• Periplasmic MBP-EC20 native binding (before DTT): less than 1 Cd²⁺ per MBP-EC20 fusion in the unprocessed periplasmic preparation (p. 521).
• CdSO4 challenge concentration in the bioaccumulation assays: 100 µM CdSO4 (Materials and Methods, p. 519); whole-cell experiments used 0.1 mM CdSO4 (Results, p. 522) — these are the same value reported in two unit conventions.
• Replicate counts: n = 5 independent experiments for the whole-cell Cd²⁺ accumulation data in Figs. 5 and 6 (figure legends, p. 521–522).
• Fusion-protein expression levels (SDS-PAGE autoradiograph, Fig. 2): Lpp-OmpA-EC8 18.5 kDa, Lpp-OmpA-EC11 19 kDa, Lpp-OmpA-EC20 21 kDa; MBP-EC20 47 kDa. MBP-EC20 expression is more than 10× higher than the surface Lpp-OmpA-EC fusions despite the surface display giving higher whole-cell Cd²⁺ accumulation (p. 520–521).
• Proteinase K accessibility experiment (Fig. 3): Lpp-OmpA-EC20 signal is “no longer detectable after 2 h” of proteinase K treatment, whereas MBP-EC20 signal shows “no observable decline… even after 21 h of incubation” — confirming surface vs periplasmic localisation respectively (p. 521).

The paper reports no Pb, Hg, As, Ni, Al, Cr, Sn, or other-metal measurements. Mercury is mentioned only as an analogous limitation case from Chen and Wilson (1997) in the Discussion (p. 522), not measured in this study.

Methods (brief)

The authors used E. coli JM105 (genotype recited in Materials and Methods, p. 519) as the recipient strain throughout. Synthetic genes encoding (Glu-Cys)nGly phytochelatin analogs of three chain lengths (EC8, EC11, EC20) were prepared from two complementary oligonucleotides synthesised by Research Genetics (Huntsville, AL); codon usage was deliberately varied between GAA/GAG (Glu) and TGT/TGC (Cys) to prevent unwanted hybridisation. Double-strand synthesis used the Klenow fragment (Promega, Madison, WI); the synthetic gene was cloned into the yeast–E. coli shuttle vector pVT102-U (Vernet et al. 1987) via BamHI and HindIII to generate pVT20, sequence-verified. The ec20 fragment was PCR-amplified from pVT20 and cloned downstream of a pre-existing 481-bp lpp-ompA cassette (Richins et al. 1997) in pUC18 to generate plasmid pLO20; analogous procedures produced pLO8 and pLO11. For periplasmic expression, the BamHI–HindIII ec20 fragment was cloned into pMAL-p2x (New England BioLabs) to generate pM20, fusing EC20 to the maltose-binding protein (MBP).

Cultures were grown in low-phosphate MJS medium (Sambrook et al. 1989) supplemented with 50 µg/mL ampicillin at 30 °C to OD600 ≈0.3 with 1 mM IPTG induction; 100 µM CdSO4 was added to allow EC expression in the presence of Cd²⁺. Radiolabelling for SDS-PAGE-autoradiograph confirmation of fusion-protein expression used [35S]-cysteine (1075 Ci/mmol, ICN) at a final concentration of 5 µCi/mL. Samples were separated on 12.5 % (w/v) SDS-PAGE per Laemmli (1970). Protease accessibility experiments used proteinase K (Sigma) at 10 µg/mL after resuspension in incubation buffer (15 % w/v sucrose, 15 mM Tris-HCl pH 7.8), with samples taken at 3 min, 10 min, 1 h, 2 h, 5 h, and 21 h and visualised by SDS-PAGE autoradiography.

For purified MBP-EC20 Cd²⁺ binding stoichiometry, MBP-EC20 was purified through an amylose affinity column (New England BioLabs); purity was confirmed by SDS-PAGE. Ten nmol of purified fusion protein was treated with 5 mM DTT in 50 mM Tris-Cl pH 7.4 for 2 h at 37 °C to reduce thiol groups, then 300 nmol Cd²⁺ was added for 1 h at 37 °C. The mixture was fractionated through a Sephadex G-25 column to separate the Cd²⁺-protein complex from free Cd²⁺ and DTT. Cd²⁺ and protein concentrations in each fraction were analysed by flame atomic absorption spectrophotometry (PerkinElmer AAS3100, Norwalk, CT) for cadmium and thiol assay per Grassetti and Murray (1967) for the protein component.

For whole-cell bioaccumulation, JM105 cultures expressing each plasmid construct were grown in MJS medium with 0.1 mM CdSO4 monitored to 16 h after IPTG induction. Cells were harvested, washed twice with double-distilled water, and disrupted overnight by treatment with concentrated nitric acid. The disrupted cells were diluted with double-distilled water and centrifuged at 4 °C for 10 min; Cd²⁺ in the soluble fraction was measured directly by atomic absorption spectrophotometry (PerkinElmer AAS3100).

Funding: UC Biotechnology Research and Education Program; US EPA contract grant R827227. No conflict-of-interest declaration is reported on the front page; the article appears under the standard Wiley copyright for Biotechnology and Bioengineering Vol. 70, No. 5, 5 December 2000, pp. 518–524.

Implications

• Certification: This paper contributes no occurrence data and no exposure data, so it does not move any HMTc threshold-setting work for Cd or any other certified analyte. Its value to HMTc is indirect — it is one of the primary engineering studies behind the broader question of whether engineered-microbial biosorbents are a feasible route for Cd reduction in agricultural or industrial waste streams that ultimately enter HMTc-certified categories. The supply-chain relevance is upstream of HMTc certification (irrigation water, soil-amendment compost, aquaculture effluent), not on the product side.
• App: No routing to any food or personal-care product page. The paper does not measure Cd in any consumer-facing matrix.
• Courses: Useful as an early primary citation in a remediation-evidence module on engineered-microbial biosorbents for Cd. The EC8 → EC11 → EC20 chain-length series provides a clean teaching example of the design-principle gradient (more cysteines per peptide → more Cd²⁺ binding sites per peptide → higher whole-cell capacity), and the surface-vs-periplasmic comparison (pLO20 ≈ 2× pM20) illustrates why localisation matters as much as expression level. Should not be cited as authority for any specific quantitative Cd-reduction claim in a real-world food or water matrix; the work is bench-scale on cell suspensions, not pilot- or process-scale on actual contaminated streams.
• Microbiome: Not relevant. The work uses recombinant E. coli JM105 as a bench heterologous host; it does not engage the gut microbiome, the rhizosphere microbiome, or any natural microbial community. WikiBiome federation is not implicated.

Limitations

The work is bench-scale on whole cells grown in defined laboratory medium with a single Cd²⁺ challenge concentration (100 µM CdSO4 ≈ 11.2 mg/L Cd as element), well above the concentrations relevant to most environmental remediation contexts (industrial-effluent Cd is typically 0.1–10 mg/L; agricultural-irrigation Cd is typically 0.001–0.05 mg/L). The authors do not report a concentration-response curve, so the relevance of the bench stoichiometry to lower-Cd remediation streams is not characterised. The whole-cell uptake numbers (18 / 29 / 60 nmol Cd²⁺/mg dry cells across the EC8/EC11/EC20 series) are reported only as bar-graph readings in Fig. 5 and Fig. 6 with no explicit numerical table, so values in the Key numbers section are approximations within ≈±5 nmol/mg of the graphed bar tops. Cells were harvested after a single fixed induction-and-incubation interval (16 h post-induction); time-course data on Cd²⁺ uptake kinetics are not reported. The downstream-process question (recovery and recycling of the Cd-loaded biomass; cell stability under operating conditions) is acknowledged in the Discussion as “currently under investigation” but not addressed in this paper. The paper covers Cd²⁺ exclusively; the title’s plural “heavy metals” is not matched by experimental data on lead, mercury, or any other analyte, and the wiki’s metals frontmatter is restricted to Cd accordingly. No replicate-level standard deviations are tabulated for the stoichiometry measurements (only n=3 fraction values are given for MBP-EC20). The 2000 publication date predates the current understanding of metallothionein and phytochelatin proteomics, so the comparative-stoichiometry framing against MTs (“40 % higher”) rests on the canonical 7 Cd²⁺/MT value of Hamer (1986) and Stillman et al. (1992); subsequent literature may have revised those reference values.

Wiki pages this source may touch

• cadmium
• remediation-evidence

Verification notes

Existing-page check. DOI grep on 10.1002/1097-0290(20001205)70 over wiki/sources/ on 2026-06-08 returned no matches. raw_handle grep on MFK_52-enhanced-bioaccumulation returned no matches. Cite-key glob bae2000-* over wiki/sources/ returned no matches. The four phytochelatin-keyword hits (cobbett2002-phytochelatins-metallothioneins-review, grill1989-phytochelatins-heavy-metal-binding-peptides-plants, marques2025-phytochelatins-cadmium-mitigation, seregin2023-phytochelatins-sulfur-metal-chelating) are all different papers. This is a NEW source page; no prior version to merge-enhance.

Evidence tier. A (primary peer-reviewed quantitative study with explicit methodology, sample sizes, and statistical reporting). The work is bench bioengineering on recombinant E. coli with clearly described constructs, growth conditions, replicate counts, and instrumentation. Per CLAUDE.md Part 13, primary peer-reviewed studies are A-tier regardless of whether the contribution is occurrence/exposure or mitigation/remediation; tier reflects methodological reliability of what the source reports, not HMI-scope fit.

Metals frontmatter. [Cd] only. The paper measures cadmium exclusively. Mercury appears in the Discussion only as a one-sentence reference to Chen and Wilson (1997)‘s analogous limitation in mercury transport; no Hg measurement is reported here. Lead appears only as a citation to Mehra et al. (1996a) in the reference list; no Pb measurement is reported here. The title’s plural “heavy metals” is forward-looking framing for the broader bioremediation programme the authors situate themselves in, not a description of this paper’s experimental scope.

Ingredients, products, matrices, jurisdictions frontmatter. All empty. The paper measures nothing in any food or personal-care matrix. The matrix is E. coli whole cells in defined laboratory medium (MJS); per CLAUDE.md Part 5b the routing layer should not fan this paper out to any consumer product or ingredient page. The work is geographically institutional (UC Riverside) but conceptually international; no national regulatory or occurrence frame applies, so jurisdictions: remains empty.

Sample size. Null. There is no human, animal, or food-sample sampling frame. The figure replicate counts (n=5 independent experiments for Figs. 5–6 whole-cell uptake; n=3 Sephadex fractions for the binding-stoichiometry measurement) are recorded in the sample_population block but sample_n is reserved for human/biological/food-sample N and is null here.

Brand firewall (Part 12). No consumer-product brand names appear in the source body or in the wiki page. The Methods section names scientific-instrument and reagent vendors: PerkinElmer AAS3100 (Norwalk, CT) atomic absorption spectrophotometer, Research Genetics (Huntsville, AL) for oligonucleotide synthesis, Promega (Madison, WI) for Klenow fragment, New England BioLabs for the pMAL-p2x vector and amylose affinity column, ICN for [35S]-cysteine, Sigma for proteinase K. Per the verification checklist’s Exception 2 (locked 2026-05-17), scientific-method vendor and material names are permitted in Methods context for reproducibility purposes. No firewall action required.

HMTc firewall (Part 2). The paper contains no HMI/HMTc-threshold language, no claims about HMI certification levels, and no consumer-audience risk advisories. It does contain forward-looking framing about engineered-microbial biosorbents as a route to “heavy metal removal” from waste streams; this is engineering-research framing, not a wiki-side synthesis or threshold proposal, and is preserved in the Implications section without escalation. No firewall action required.

Date arithmetic. Received 23 March 2000, accepted 30 June 2000, published Biotechnology and Bioengineering Vol. 70, No. 5, 5 December 2000, pp. 518–524 — all consistent with the year: 2000 frontmatter.

DOI form. The Wiley-assigned 2000-era DOI uses the angle-bracket form 10.1002/1097-0290(20001205)70:5<518::AID-BIT6>3.0.CO;2-5. The brackets and semicolons are part of the DOI per Wiley’s pre-2002 convention. The access_url uses https://doi.org/<doi> per the wiki’s standard convention and resolves to the article landing page at Wiley Online Library.

Reviewer’s note on scope fit. This paper is in the “Black Market Peptide Metal Survey / heavy_metals_peptides” Manual Fetch Kimi June 8 folder alongside marques2025-phytochelatins-cadmium-mitigation and similar peptide-mediated mitigation/remediation sources. Per the 2026-06-02 commit 3f47f95 — scope: mitigation/remediation is in-scope, not a skip, peptide-mediated mitigation/remediation papers are in scope as background for the mitigation-evidence chapter. This paper is the foundational primary-engineering study for the surface-displayed-EC strategy of Cd biosorption and is the bench-side counterpart to the Marques 2025 plant-side review of PCS genetic manipulation.

Slug-vocabulary note. [[mitigation/remediation-evidence]] is not in the 2026-05-18 taxonomy snapshot. This is the same snapshot-coverage gap noted in marques2025-phytochelatins-cadmium-mitigation; the wikilink points to the wiki’s mitigation/remediation-evidence section and is in-scope per the cited 2026-06-02 scope commit. No correction applied; the snapshot will catch up in a future refresh.

Audit subagent (2026-06-08) verdict: PROMOTE. Five checks (numerical fidelity, slug vocabulary, speciation/methods, brand firewall, HMTc firewall) returned four ✅ and one ⚠️.

• Check 1 numerical-fidelity ✅. Subagent independently verified against the PDF: the 9.9 / 10.1 / 9.8 Cd²⁺-per-MBP-EC20 fractions 6/7/8 (p. 521 Fig. 4 caption), the 10 Cd²⁺/peptide and 60 nmol/mg pLO20 maximum (Abstract p. 518), the 7 Cd²⁺/MT reference (p. 522 citing Hamer 1986; Stillman et al. 1992), the 40 % higher EC20-vs-MT capacity (p. 522 Discussion), the EC8 → 18, EC11 → ≈29, EC20 → ≈60 nmol/mg chain-length series (Abstract p. 518; Fig. 6 p. 522), the pUC18 control ≈4 nmol/mg and pM20 ≈35 nmol/mg approximations from Fig. 5 (p. 522), the 2-fold surface-vs-periplasmic ratio (“almost twice the amount” p. 522 Discussion), the <1 Cd²⁺ per MBP-EC20 native pre-DTT (“less than one equivalent” p. 521), the 100 µM CdSO4 challenge / 0.1 mM whole-cell concentrations (pp. 519, 522), the n = 5 independent experiments (Figs. 5–6 captions p. 522), the 18.5 / 19 / 21 / 47 kDa fusion sizes and the >10× higher MBP-EC20 expression level (p. 520), and the proteinase K time-course endpoints (Lpp-OmpA-EC20 “no longer detectable after 2 h”; MBP-EC20 “no observable decline… even after 21 h”, p. 521). No invented numbers, transpositions, or mis-attributions detected.
• Check 2 slug-vocabulary ⚠️ on [[mitigation/remediation-evidence]] not in the 2026-05-18 snapshot — same snapshot-coverage gap as the Marques 2025 sibling, already disclosed in the slug-vocabulary note above and accepted per Marques precedent. No content correction applied.
• Checks 3 (speciation/methods), 4 (Part 12 brand firewall), and 5 (Part 2 wiki/HMTc firewall) all ✅. Subagent verified the Methods (brief) section against the source row-by-row (JM105 host, Lpp-OmpA fusion via pUC18, pMAL-p2x for periplasmic MBP, MJS medium with 50 µg/mL ampicillin at 30 °C, 1 mM IPTG, [35S]-cysteine at 5 µCi/mL from ICN at 1075 Ci/mmol, 12.5 % SDS-PAGE per Laemmli, proteinase K at 10 µg/mL with the 3 min / 10 min / 1 h / 2 h / 5 h / 21 h time points, 5 mM DTT for 2 h at 37 °C reduction, 300 nmol Cd²⁺ for 1 h at 37 °C binding, Sephadex G-25 fractionation, flame AAS PerkinElmer AAS3100, Grassetti–Murray 1967 thiol assay, 481-bp lpp-ompA cassette and Richins et al. 1997 source) with no invented analytical methods, LODs, or reference materials; that the only metal measured is Cd and the metals: [Cd] frontmatter is correct without speciation issues; that no consumer-product brand names appear and the scientific-vendor mentions (PerkinElmer, Research Genetics, Promega, New England BioLabs, ICN, Sigma) are method-context per Exception 2; and that the Implications section explicitly disclaims any HMTc threshold-setting contribution. 1 finding flagged, 0 corrections applied (the ⚠️ was a known cross-page taxonomy gap, not a defect on this page), 0 rejected. Audit subagent ID aa8d3586d3cd812a4.

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.