Ding et al. 2015 — Stability constants of Hg²⁺, Cd²⁺, and Pb²⁺ complexes with thiol-containing peptides isolated from soy glycinin hydrolysates

Ding and colleagues at Jiangnan University (Wuxi, China) hydrolysed soy glycinin (the 11S storage protein of soybean) with three commercial proteases — alcalase, papain, and pepsin — to three degrees of hydrolysis (DH 5%, 15%, 25%), enriched the thiol-containing peptide (TCP) sub-population from each hydrolysate by Thiopropyl-Sepharose 6B covalent chromatography, and measured the stability constants (lgβ₁ and lgβ₂) of the resulting TCPs against Hg²⁺, Cd²⁺, and Pb²⁺ by potentiometric pH titration. The TCPs from the 25% DH papain digest reached lgβ₂ values of 33.4 (Hg), 15.2 (Cd), and 21.1 (Pb), comparable to or exceeding the corresponding glutathione values measured in the same study. The paper frames TCPs as candidate functional-food ingredients for “natural detoxication of heavy metals,” not as a measurement of heavy metals in soy or in any food product.

Why this matters

• This is one of a small set of papers in the HMI peptide-metal-binding corpus that measures stability constants for food-protein-derived peptides against Pb²⁺, Cd²⁺, and Hg²⁺ on a common (potentiometric, 0.1 M KCl, 25 °C) platform, with glutathione as an in-study comparator. The published data make soy-glycinin TCP affinities directly comparable to glutathione without cross-paper buffer/temperature/instrument correction. Companion papers in the same Kimi folder include urbina2018-biomining-peptide-metal-recovery (ITC characterisation of natural and rationally-designed Cu/Zn/Ni-binding peptides), grill1989-phytochelatins-heavy-metal-binding-peptides-plants (the discovery paper on plant phytochelatins), and cobbett2002-phytochelatins-metallothioneins-review (review of PC/MT biology).
• The paper provides no contamination data on any food matrix. Soy glycinin here is a purified protein starting material, not a measured soy product. The wiki should not record Pb/Cd/Hg occurrence values for soy on the basis of this paper.
• The functional-food framing in the paper’s discussion (“TCPs could be used as ingredients in the formulation of functional foods to control and manage diseases associated with heavy metal accumulation”) is a hypothesis the authors put forward, not a demonstrated in-vivo or in-clinical effect. No animal-model or human-clinical chelation data are presented. The chemistry establishes that the peptides bind the metals in vitro; whether oral intake of such peptides would reduce body-burden of Pb/Cd/Hg in humans is not tested here.
• The paper’s Hg²⁺-glutathione comparators are internally inconsistent with one literature source (Cardiano et al. 2011, reference [32] in the paper) but consistent with another (Oram et al. 1996, reference [44]). The authors note this and report their own GSH-Hg lgβ values as 19.2 (β₁) and 30.5 (β₂) in 0.1 M KCl at 25 °C. A future synthesis page that uses Hg-glutathione lgβ as a benchmark should treat this disagreement explicitly rather than averaging across the GSH literature.

Key numbers

Stability constants of TCP-metal complexes at 25 °C, I = 0.1 M KCl (Table 4, p. 8048; mean ± SE, triplicate measurement). Sample codes: Alc/Pap/Pep = alcalase/papain/pepsin; 05/15/25 = degree of hydrolysis in % (the hydrolysate code therefore identifies the protease and the DH level).

Glutathione comparators measured in this study (text, pp. 8048-8049): lgβ₁ = 19.2, lgβ₂ = 30.5 for Hg²⁺-GSH; lgβ₁ = 10.1, lgβ₂ = 15.4 for Cd²⁺-GSH (consistent with Leverrier et al. 2007 values of 8.5 and 12.4 for Cd-GSH and Cd-GSH₂ respectively, paper text).

Acid dissociation constants of glutathione measured in this study (Table 2, p. 8047, “Experimental Values” row; I = 0.1 M KCl, 25 °C): pK₁ = 2.3, pK₂ = 3.6, pK₃ = 8.9, pK₄ = 9.8 (literature values [35]-[37] also tabulated, ~0.2 pK units lower across the range).

Acid dissociation constants of TCPs by hydrolysate (Table 3, p. 8047). For DH 5% TCPs: alcalase gave six titratable pK values (pK₁ = 11.6, pK₂ = 9.9, pK₃ = 7.1, pK₄ = 4.5, pK₅ = 3.5, pK₆ = 2.4); papain gave four (pK₁ = 10.7, pK₂ = 7.5, pK₃ = 4.4, pK₄ = 3.1); pepsin gave six. For DH 25% TCPs: alcalase gave three (pK₁ = 11.4, pK₂ = 8.0, pK₃ = 3.1); papain gave four; pepsin gave four. The number of titratable protons decreased with increasing DH for alcalase and pepsin TCPs but remained constant at four for papain TCPs across all DH levels. Lowest pK values (range 1.8–3.6 across all TCPs) are assigned to the C-terminal carboxylic-acid; values 7.4–9.1 to cysteine sulfhydryl and N-terminal amine groups (which cannot be distinguished in pK alone); values >10 to lysine and arginine side chains.

Total sulfhydryl content of SGHs and enriched TCPs (Figure 2, p. 8044, qualitative reading of axes). Total sulfhydryl content of SGHs was approximately 100–120 µmol/g protein across all enzymes and DH levels (i.e., proteolysis itself caused little loss of -SH). After Thiopropyl-Sepharose enrichment, sulfhydryl content of TCPs ranged from ~400 µmol/g (DH 5%) to ~950 µmol/g (Alc25), with enrichment factors of 4.2–8.2× (alcalase), 4.0–4.8× (papain), and 3.5–4.8× (pepsin) over the unenriched hydrolysate.

Order of TCP-metal stability among the three metals (text, p. 8049):

• Alcalase TCPs and pepsin TCPs: Hg²⁺ > Cd²⁺ > Pb²⁺
• Papain TCPs: Hg²⁺ > Pb²⁺ > Cd²⁺
• The authors note this order does not follow the inverse-ionic-radius rule (Cd²⁺ < Hg²⁺ < Pb²⁺ in ionic radius), so the relative-affinity order is “dependent upon the experimental conditions used” rather than a simple ionic-radius prediction.

Correlations between sulfhydryl group content and lgβ (Figure 5, pp. 8049-8050; Pearson r):

The authors interpret the systematically larger r-values for lgβ₂ than for lgβ₁ as evidence that the sulfhydryl group is more important in stabilising the ML₂ (two-ligand) complex than the ML (one-ligand) complex, consistent with a linear two-S-coordinated Cys–M–Cys binding geometry being a preferred low-strain mode for soft-acceptor metals.

Correlations between TCP molecular-weight-fraction content and lgβ (Table 5, p. 8051; Pearson r):

The authors interpret the much stronger Mw-dependent signal for Pb²⁺ than for Hg²⁺ or Cd²⁺ as evidence that steric accessibility of the binding groups matters more for Pb²⁺ than for the softer metals. No fraction-III (1–3 kDa) correlation reached significance for any metal.

Methods (brief)

Starting material. Defatted soy flakes (Shandong Gushen Industrial, China; 53.13% protein, 87% nitrogen solubility index). Soy glycinin (11S) prepared by the method of Wolf (1993) without further purification: flakes suspended in 30 mM Tris-HCl, pH 8.0, with 10 mM 2-mercaptoethanol; centrifuged; supernatant brought to pH 6.4 with 2 M HCl to precipitate proteins; precipitate redissolved at pH 8.0, dialysed against distilled water, and freeze-dried. Final 11S protein content: 95.3 ± 1.6% (w/w) by micro-Kjeldahl (N × 6.25); purity 90 ± 3.1% by SDS-PAGE.

Hydrolysis. Alcalase (Novozymes 2.4 L FG, 180,000 U/mL), papain, and pepsin (Sigma) used at E/S ratio 5:100 (w/w). Conditions: alcalase 50 °C pH 8.0; papain 50 °C pH 7.0; pepsin 37 °C pH 2.0. pH-stat method with 0.5 M NaOH (alcalase, papain) or 1 M HCl (pepsin). Hydrolysis stopped at 5%, 15%, and 25% DH by boiling 5 min. Supernatant lyophilised.

TCP enrichment. Hydrolysate at 50 g/L reduced with 30 mM DTT, pH 8.0, 50 °C, 30 min. pH adjusted to 3.0; loaded on Sephadex LH-20 to remove excess DTT (eluted with 1 mM HCl). Eluate pH-adjusted to 7.5, brought to 0.5% SDS, and applied to Thiopropyl-Sepharose 6B beads. Bound peptides liberated with 25 mM DTT in 10 mM Tris-HCl pH 7.5. Excess DTT and 2-thiopyridine removed on a C-18 column.

Sulfhydryl content. 4,4’-Dithiodipyridine (4-DPS) method (Riener et al. 2002): 0.3 mL TCP solution (~25 µmol/L sulfhydryl) + 2.7 mL 0.1 M citrate-Na₂HPO₄ buffer pH 4.5 with 1% SDS + 125 µL 4-DPS (4 mM); A₃₂₄ read after 30 min ambient; ε = 21,400 M⁻¹ cm⁻¹. Total sulfhydryl after NaBH₄ reduction of disulfides (Cavallini et al. 1966 modified).

Molecular-weight distribution. SEC on TSK gel 2000 SW_XL (Tosoh, 300 × 7.8 mm) at 0.5 mL/min in acetonitrile/water/TFA = 45/55/0.1 (v/v/v); detection at 214 nm. Calibration: cytochrome C (12,500 Da), bacitracin (1450 Da), tetrapeptide GGYR (451 Da), tripeptide GGG (189 Da).

Potentiometric titration. Lab-assembled cell: 50 mL thermostatted, Teflon-stoppered reactor with high-precision pH probe (Mettler-Toledo FE 20) and inert-gas (N₂) atmosphere. Buffers pH 4.00, 6.86, 9.18 used for calibration. Titrations at 25 ± 0.1 °C, pH range 2.0–12.0, with 0.1 M KOH as titrant. Sample volume 25 mL; TCP concentration 4 mM (assumed equal to sulfhydryl concentration on the basis of one –SH per TCP molecule); metal-ion (Hg²⁺, Cd²⁺, or Pb²⁺) concentration 1 mM where present.

Stability-constant calculation. Bjerrum method. Protonation constants β_j^H calculated from titrations in absence of metal ions by non-linear least-squares (equations 1–3 in the paper, p. 8053). Average ligand-per-metal coordination number n̄ obtained by the half-integral method by plotting against pL (=−log[L]); stepwise stability constants Kₙ read from the formation curve at the pL where n̄ = n − ½; overall stability constants βₙ computed as K₁·K₂·…·Kₙ.

Statistics. ANOVA with SPSS; LSD test at 95% confidence; triplicate runs reported as mean ± SD; Pearson correlation coefficients reported in Figures 5 and Table 5.

What this paper does not measure. No Pb, Cd, or Hg occurrence in soy beans, soy flakes, soy protein isolate, soy-based infant formula, or any food product. No in-vivo or in-vitro digestion model; no animal or human chelation outcome. The “functional food” framing in the conclusion is forward-looking; no human or animal data on whether oral TCP intake reduces body-burden of these metals are presented.

Implications

Certification: Not directly applicable. The paper measures no food matrix or supply-chain occurrence and therefore cannot contribute occurrence data to any product-category page or threshold workbench. It is, however, a relevant mechanistic source for mercury-total, cadmium, and lead explainer sections that describe how thiol-rich peptides bind these metals (soft-acceptor behaviour, Cys-S-donor preference, the ML₂ vs ML coordination distinction). It is also a counter-evidence source against any future claim that “soy glycinin is itself a source of these metals” — the paper says nothing about the contaminant content of soy.

Courses: Useful as a worked example in a future advanced module on peptide-based metal chelation chemistry, alongside urbina2018-biomining-peptide-metal-recovery and grill1989-phytochelatins-heavy-metal-binding-peptides-plants. Specifically: how potentiometric titration extracts protonation and metal-binding constants on the same instrument; how a food-protein hydrolysate can be ranked as a candidate chelator against the glutathione benchmark; and what the lgβ₁ vs lgβ₂ split tells you about coordination geometry.

App: Not applicable. No contamination-profile data.

Microbiome: Not applicable. No microbiota or gut-microbial measurements.

Verification notes

• Cite-key. ding2015-soy-glycinin-thiol-peptides-heavy-metal-complexation chosen for specificity; the paper is the only Ding 2015 in the wiki at ingest, but the qualifier protects against later additions.
• Evidence tier B. Peer-reviewed journal article (Int. J. Mol. Sci., MDPI, CC BY 4.0), reference-method instrumentation (Mettler-Toledo pH probe, pH-stat calibration with three standard buffers, triplicate runs). Tier B rather than A because the wiki’s A-tier is reserved for direct-contamination measurement with traceable LOD/LOQ and certified reference materials; this is mechanism/chemistry work outside that scope.
• Frontmatter metals: [tHg, Cd, Pb]. Hg²⁺ here is inorganic Hg(II); HMI conventions use tHg for total/inorganic mercury and MeHg for methylmercury. No MeHg measurements in this paper.
• Frontmatter ingredients: [], products: [], matrices: []. Deliberately empty. The paper measures no food matrix or supply-chain occurrence; soy glycinin appears here as a purified protein starting material for chemistry work, not as a measured food. Adding soy or soy-protein-isolate would mis-signal the routing layer to fan this paper out to soy product/ingredient pages as an occurrence source, which would corrupt those pages’ contamination profiles.
• Frontmatter jurisdictions: []. Chinese authors and Chinese-sourced soy flakes, but no jurisdictional or regulatory framing; the paper is laboratory chemistry with no regulatory connection.
• Glutathione lgβ disagreement noted but not resolved here. The paper’s own GSH-Hg comparators (lgβ₁ = 19.2, lgβ₂ = 30.5 measured in this study) agree with Oram et al. 1996 (reference [44]) but disagree with Cardiano et al. 2011 (reference [32]). The paper does not adjudicate. This wiki page records both the in-study value and the discrepancy without picking a winner; a future Hg-chelation synthesis should re-read both literature sources directly rather than relying on Ding et al.’s comparator alone.
• “Functional food” framing reported, not endorsed. The paper concludes that “soy TCPs could be used as ingredients in the formulation of functional foods to control and manage diseases associated with heavy metal accumulation.” This is repeated above as the paper’s own framing, not as a wiki-endorsed claim. There is no in-vivo, in-vitro digestion, or human clinical data in the paper to support an oral-chelation conclusion.
• Audit subagent (2026-06-08) flagged a Cardiano-vs-Leverrier misattribution — verified against source, false positive. The audit claimed the wiki misattributes the Cd-GSH stability values 10⁸·⁵ and 10¹²·⁴ to Leverrier when (per the audit) they should be Cardiano. Independent re-read of page 8048 confirms the paper’s body text reads “Leverrier et al. [32] reported that stability constants of CdGSH and Cd(GSH)₂ were 10⁸·⁵ and 10¹²·⁴”. The paper’s reference list has [31] = Leverrier (Anal. Biochem. 2007, Cadmium complexes with glutathione revisited) and [32] = Cardiano (J. Chem. Eng. Data 2011, Sequestration of Hg²⁺). The bracket numbers in the body text are swapped (a typo in the source), but the surname and the subject-matter (Cd-GSH binding) both attribute these values to Leverrier’s cadmium-glutathione paper, not to Cardiano’s mercury-thiols paper. The wiki page is correct as written. The audit subagent was misled by the inverted bracket numbers.
• Audit subagent (2026-06-08) flagged a source-internal SD-vs-SE inconsistency — verified, real concern, recorded here. Section 3.10 of the paper (p. 8054) states “Data were expressed as the mean ± SD (n = 3)”; the Table 4 footnote (p. 8048) states “Values expressed as means of triplicate ± standard error.” The source is internally inconsistent about whether the Table 4 stability constants are reported with SD or SE. The wiki Key-numbers Table-4 transcription mirrors the Table 4 footnote (SE); the Methods Statistics paragraph mirrors Section 3.10 (SD). Both wiki transcriptions are faithful to the corresponding source text. A reader using the lgβ uncertainty values for downstream propagation should treat the ± values as approximate and read the original Table 4 directly before making a precision claim.

Wiki pages updated on ingest

• mercury-total
• cadmium
• lead

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.