Drug/Xenobiotic Metabolism

https://www.youtube.com/watch?v=Jk5ZuNr55Ww

Contents:

1. What is the problem we worked on and what is the textbook understanding?

2. Here are the arguments why I oppose the textbook belief.

3. What is the alternate hypothesis proposed (for which we have lots of evidence and arguments)?

4. If the arguments against the prevailing understanding were not convincing enough, a head-to-head comparison of the two hypotheses should allow the reader to decide.

5. Still not convinced? Here is conclusive data on binding constants and reaction times!

1. What is the problem we worked on and what is the textbook understanding?

The heme-enzymes- as exemplified by cytochrome P450s (CYPs), are perhaps the most versatile, well-known and well-studied systems. CYPs are identified as the primary enzymes responsible for the first phase of metabolism of a wide variety of xenobiotics (drugs, pesticides etc.) in animals. They are heme-thiolate hydrophobic membrane proteins, with a molecular weight of ~ 45-60 KD. Till date, there are several hundreds of CYPs known and several crystal structures have been solved.

CYPs work in tandem with a membrane-bound diflavoenzyme, cytochrome P450 reductase (CPR, a ~77 KD protein). The overall reaction requires NADPH as the electron donor, molecular oxygen and a substrate. Reaction occurs more efficiently when the two enzymes are stationed on phospholipid membranes. The following is the phenomenological understanding of this system, as evident in all textbooks and research reviews hitherto available.

The overall equation for the productive cycle, as per this hypothesis, is-

RH + O₂ + NADPH + H⁺ → ROH + H₂O + NADP⁺

The overall equation for the non-productive peroxide and water shunts, as per this hypothesis, are-

O₂ + NADPH + H⁺ → H₂O₂ + NADP⁺

O₂ + 2NADPH + 2H⁺ → 2H₂O + 2NADP⁺

The reactive species is a two-electron deficient Compound I, formed at the heme center. This reacts with a substrate, in a single step two-electron reaction, releasing the product.

The 'transition state' shown above is supposedly a summation of the following 'oxygen rebound' process-

This essentially means that in the transition state, all the key atoms, R, H & O are at 'bond-able' distances with respect to each other and within a few Angstroms of the heme iron.

The hypothesis shown above came into prevalence owing to a lot of work on a soluble enzyme from Pseudomonas putida, P450cam and another 'fused' enzyme system (one enzyme contains both heme and flavin groups) from Bacillus megaterium, P450BM3. There is quite a lot of work done that could apparently (mis)lead one to 'believe' in this paradigm's general applicability, but............................

2. Here are the arguments why I oppose the textbook belief-

1. The erstwhile hypothesis flouts-

• Quantitative logic (wrt time):

(a). Considering the diffusion limitation barriers for bulky proteins housed in a phospholipids' micro-environment, this hypothesis cannot account for the high reaction rates (wrt collision frequency). Please see the following data regarding diffusion coefficients (these are not mine but generated experimentally by other scientists and are well accepted by the community; the values are expressed in micrometer squared per second)-

A. Proton (from hydronium to water) = 10⁴

B. Oxygen (free movement in water at RT) = 10³

C. 30 KD protein (in water) = 10²

D. 30 KD protein (in cyto- or nucleo- plasm) = 10¹

E. 70 KD protein (in cytoplasm) = 10⁰

F. 70 KD protein (in plasma membrane) = 10^-1 or 10^-2.

Now, when you take nanomolar levels of proteins and if you get pseudo-first order (in the linear ranges of assay) CPR mediated ET rates at 10¹ to 10² per second and CYP mediated hydroxylations at 10⁰ per second, we cannot have the phenomenon explained with direct interactions of the protein species. (This is because the second order diffusion limitation is ~ 10⁹ per molar per second for small and highly mobile molecules/ions. If you multiply this value with the concentration of the protein, you get its collision frequency.) Even if we afforded any mobility to CPR, the ET rates afforded by collisions of CPR would be several orders lower than the experimentally observed rates. This is an elementary calculation which no wishful thinking can refute! Till date, researchers wrongly assumed that CYPs were slow. We believe that these redox enzymes actually work on the thresholds of limits that can be achieved in semi-liquid phase catalysis.

(b). Overall cycle shows that the residence time (the time that a substrate molecule stays within the heme distal pocket) is in the order of seconds. This is when the breathing time of F & G loops would be in the range of milliseconds and the calculations from K_M gives a residence time range of micro- to milli- seconds!

• Fundamental reasoning and self-consistency (wrt space):

(a). Committed to catalysis: Substrate is bound in the active site to account for “thermodynamic switch” but substrate is freely rotating within the active site to explain large intramolecular KIEs.

(b). Movement of F & G loops: If these moved to incorporate diverse substrates, then why are only some substrates “preferred” by a given CYP?

(c). Topographical recognition by the enzyme: Why don't several large substrates give regio- and stereo- specific reactions, when it is supposed that they are tightly bound at the constrained distal heme pocket?

(d). Water, peroxide and superoxide production in the CYP+CPR system is supposedly owing to a Type I binding of the substrate within the active site of CYPs. But then, why should uncoupling occur in the first place? How or why do the highly reactive Compound I and the substrate sit around in the active site and not react? Why should Compound I wait for another round of two-electron and two proton relays (in the water shunt cycle)? [Please see that the "biologically aesthetic" saving grace of the erstwhile hypothesis is rendered redundant with the uncoupling phenomenon. There exists not even a marginal utility with this concept!]

• Occam's razor:

(a). Termolecular ordered/sequential reactions involve purported high efficiency binding of nM to µM levels of reactants that have little 'topographical binding affinities'.

(b). A unique CPR (which is present at 1:10 to 1:100 ratios wrt CYPs) is supposed to repeatedly bind and pump electrons to diverse CYPs across unfavorable potential/activity gradients.

(c). Diverse CYPs have multiple binding loci for multiple substrates in their distal pockets.

(d). Inhibitions and activations observed by additives (drug-drug interactions) result owing to multiple binding sites of these additives within the distal pocket.

(e). All CYPs have highly efficient proton relays in the distal hydrophobic pocket.

(f). Long range intermolecular electron transfers beyond 15 Angstroms, between two bulky proteins.

2. Further, explanations are lacking why-

(a). CYPs have substrate preference and substrate diversity (why unrelated CYPs metabolize the same substrate, why a CYP cannot metabolize very closely related substrate analogs, etc.

(b). There exist diverse CYPs if the F & G loops of each CYP moved to accommodate diverse substrates.

(c). Idiosyncratic dose-responses, atypical kinetics and bizarre kinetic/equilibrium constants are reported.

(d). Secondary oxidations and reductive reactions are seen.

(e). Substitutions and modifications of surface amino acid residues critically affect reactivity.

(f). Activity differences exist among natural and/or synthetic mutants of a CYP.

.

3. Also, little direct and convincing evidence is available (wrt to functional or physiological conditions in microsomal liver CYPs that have purely hydrophobic distal pockets)

(a) the existence of Compound I or its role in catalysis,

(b) water formation in CYP's active site,

(c) CYP-CPR's complexations' roles in electron transfers, and

(d) oxygen atom transfer occurring solely at the heme center.

[This argument could be negated in the future, but I project that it can be only in highly concentrated systems with micromolar levels of proteins and reactants, as exemplified in synthetic or spectroscopy samples. Further, I state that systems like P450cam and P450BM3 are inappropriate role models for microsomal CYPs, the enzymes of "high significance".]

We have recently shown that catalyses mediated by alternate binding modes and via diffusible species explain a lot many observations of activations and inhibitions in heme peroxidases [please refer the publications listed in the appropriate section]. Therefore, it was only natural to deem that those explanations could hold merit even in the CYP system.

3. What is the alternate hypothesis proposed (for which we have lots of evidence and arguments)?

The basic strings of this work were originally pursued by Aust and Coon groups, more than four decades ago.

The equations involved are-

(1) 2O₂ + NAD(P)H → 2O₂^*- + NAD(P)⁺ + H⁺ (Primarily, CPR's role)

(2) RX + O₂^*- + H⁺ → ROH + OX^* (Primarily, CYP's + milieu role)

(3) O₂^*- / OX^* + O₂^*- / OX^* / RH / OH^- / NAD(P)H / H₂O₂ / Enzymes → Diverse fates

The overall productive equation is-

RH + O₂ + NADPH → ROH + OH^- + NADP⁺

The generic scheme for explaining the CYP-CPR system is- (a) CPR activates molecular oxygen to produce diffusible superoxide/hydroxyl radicals at the phospholipid interface. (b) The distal heme-iron pocket of the CYP (which is co-localized alongside CPR in the microsomal membrane) serves to stabilize the DROS generated by CPR and the apoprotein of CYP transiently presents the xenobiotic molecule (at any favorable loci on the protein). The one-electron equivalents generated by the semiquinone of CPR-flavin(s) can be easily trapped by the solvent accessible thiolate of CYPs. This leads to the reduction of the iron center, to which a molecular oxygen can bind to, via diffusion through the distal pocket. Else, the superoxide generated by CPR can directly bind at the distal heme pocket. In many cases, presence of the substrate can also directly reduce the heme center (and this may have been misinterpreted as a Type I spectral change, because both incur an increase in OD at ~390 nm). Further, there is no requirement for protons (within the highly hydrophobic active site of CYP stationed in membrane) in this catalytic scheme. Hydrophobic and uncharged molecules react to give soluble and charged species at the phospholipid interface, leading to inundation and washing away by water. This phenomenon, together with the 'electron sink' afforded by substrate oxidation, gives the 'thermodynamic pull' for the 'constitutive (but yet inducible)' flow of electrons from CRP+NADPH.

Interaction of CYPs with the substrate(s) is via a novel modality-

The new hypothesis gives a lot of scope for low affinity interactions of various xenobiotic substrates and DROS with different and multiple loci of microsomal cytochrome P450s. The different loci are not brought together for a direct bond formation in the transition state (by induced fit). The substrate, if small enough (like Xenobiotic 4), can also approach the heme center and react there. The substrates 3 and 1 would have better probability of being reacted (than say, a substrate like Xenobiotic 2) by the given CYP. This may be owing to better binding per se or/and because of a more probable presentation to the reactive species as it emerges out from the distal heme pocket. Thus, CYPs serve as the "rendezvous locale" for the reactants. In the absence of CYP within the vicinity, the reactive DROS radicals are scavenged by milieu components or get converted to relatively ineffective DROS, like peroxide. Therefore, the liver cells employ CYP-CPR for an inherently constitutive "mured burning" or "mild unrestricted burning" (or "murburn", in short) in the vicinity of phospholipid membranes of the microsomes. (Larger substrates can also get into the distal pocket of CYPs with the opening of F & G loops. But these would only be possible at very high enzyme and substrate concentrations and/or over much longer periods of time.) For this reason, CYPs and such redox enzymes which show a similar functioning (that extends beyond the conventionally known aspects of "unique binding locus, one reaction mechanism"), could be termed 'murzymes' (enzymes that mediate unrestricted redox catalysis).

The new hypothesis is highly probable because it involves only bimolecular, independent and non-sequential reactions, which also meets the evolutionary/logical mandate of xenobiotic metabolism. To my understanding, this unordered pseudo-pingpong scheme is distinct from the hitherto available enzyme mechanistic models because- (i) There are at least two or more spatially distinct binding sites (which are separated by a dimension unfavorable for a direct bond-formation in the transition state) for the “two substrates" (DROS and various xenobiotic molecules, Figure 1) and (ii) There is an uncertainty regarding the exact locus where the two substrates finally react. The “murburn” hypothesis reconciles with available crystallographic data, affording greater scope for explaining CYPs' broad 'substrate' specificities or preferences, presence or absence of enantio- and/or regio- specificity and multiple products, etc. Further, it explains the relatively high concentrations of diverse CYPs and low concentrations of a unique and promiscuous CPR (and also explains the action of cytochrome b₅). The following enzymatic functionalities of CYPs are also explained now- kinetic isotope effects, atypical kinetics and bizarre values of constants (like K_M, IC₅₀, K_i, K_is, etc.), idiosyncratic dose responses (activations or inhibitions), secondary oxidations, multiple catalytic species (some of which are also capable of reduction), etc. The proposed hypothesis seeks us to go beyond treating kinetic data obtained in CYP systems with the Michaelis-Menten supposition. The ideas derived herein enable us to better design green chemistry reactions with these enzymes and proffers better understanding of in vitro or clinical data like- (i) how large drug molecules serve as substrates for CYPs with very small or occluded distal heme pockets, (ii) reactivities of CYPs across diverse molecules from a class of drugs (like sartans, statins etc.), (iii) substrate preferences and demarcations between related and unrelated CYPs, (iv) differences in activity and metabolic predisposition seen in artificial and natural mutants respectively, (v) drug-drug interactions at the activity level (excluding the effects of enzyme induction), (vi) Why covalent modification or mutation of surface residues of a CYP could mess with its activity, etc.

4. If the arguments against the prevailing understanding were not convincing enough, a head-to-head comparison of the two hypotheses should allow the reader to decide-

5. Still not convinced? Here is conclusive analyses of data on binding constants and reaction times!

Points (a & b) of quantitative logic, when explained in detail would make things evident. Some theory first:

For the reaction-

E + S --> ES -- EP --> E + P

Now, if we take that k₁ and k_-1 are the forward reaction rate constant to form ES and the backward breakdown rate constant of ES, respectively; and k₂ is the forward breakdown rate constant of EP, then, as per Michaelis-Menten supposition-

k₁ = (d[ES] / dt) . (1/ [E][S])

Experimentally, nM levels of CYPs give product formation with pseudo first order rates equaling 1 per second (for efficiently coupled baculosome systems, in vitro). Therefore, the overall reaction's second order rate constant would be 10⁹ M^-1 s^-1. In the steady state conditions, we can have only < nM ranges of [ES] formed; and this formation requires > milliseconds time frame (which is the generally accepted “breathing time” of proteins, the time that would be required for F/G loops or helices to move around). So, the k₁ for this first step may be lower than 10⁷ M^-1 s^-1 or only in the most suitable of cases, approach the theoretical maximum of 10⁹ M^-1 s^-1 (which is also the diffusion limitations for small molecules in water). Therefore, the first "substrate binding" step alone becomes the limitation step. Thereafter, there is little time or "probabilistic scope" for any of the other sequential reactions! Since the erstwhile hypothesis espouses an ordered reaction scheme involving collisions of at least five components (nM levels of CYP & CPR and micromolar levels of oxygen, substrate & NADPH), it cannot explain the low (or higher!) reaction times observed in microsomal CYPs. The erstwhile hypothesis becomes even more improbable considering that it espouses a termolecular complexation scheme of bulky hydrophobic proteins and small molecules on the phospholipid microsomal membrane (and that, when there is little evolutionary or observable/provable affinities for each other). This kinetic’ argument is the most compelling quantitative logic against the erstwhile hypothesis. Yet, let us indulge the erstwhile hypothesis further. If it could be applied, then-

K_M = (k_-1 + k₂) / (k₁) …………..…….. (a)

As per laws of equilibrium-

K_d = (k_-1)/(k₁)………………………... (b)

It is clear that-

K_M = K_d + (k₂/k₁) …………………..... (c)

Therefore,

K_M ~ K_d, when k_-1 >> k₂ ……….…… (d)

and / or

K_M ~ K_d, when (k₂ /k₁) = 0 ………….. (e)

Now, the prevailing hypothesis assumes that once the substrate is hydroxylated, it loses efficiency for the enzyme. Therefore, the forward breakdown of EP may be comparable or efficient with respect to the backward breakdown of ES. In this case, the contribution of k₂ cannot be neglected. Therefore, we might have a higher value of K_M than that of K_d. But under any circumstances, the value of K_M cannot be lower than that of K_d.

As per several leading mechanistic/kinetics papers published in the field, most P450-substrate forward binding is very fast and only diffusion limited (k₁ ~ 10⁹ per molar per sec). If that were true, it is highly unlikely that the physical value of k₂ can approach k₁ and therefore, supposition (e) tends to be more or less the scenario (as is true for many enzymes in normal environments).

Now, please look at the data on microsomal CYPs we collated-