Evidence for Retromutagenesis as a Mechanism for Adaptive Mutation in Escherichia coli.

Adaptive mutation refers to the continuous outgrowth of new mutants from a non-dividing cell population during selection, in apparent violation of the neo-Darwinian principle that mutation precedes selection. One explanation is that of retromutagenesis, in which a DNA lesion causes a transcriptional mutation that yields a mutant protein, allowing escape from selection. This enables a round of DNA replication that establishes heritability. Because the model requires that gene expression precedes DNA replication, it predicts that during selection, new mutants will arise from damage only to the transcribed DNA strand. As a test, we used a lacZ amber mutant of Escherichia coli that can revert by nitrous acid-induced deamination of adenine residues on either strand of the TAG stop codon, each causing different DNA mutations. When stationary-phase, mutagenized cells were grown in rich broth before being plated on lactose-selective media, only non-transcribed strand mutations appeared in the revertants. This result was consistent with the known high sensitivity to deamination of the single-stranded DNA in a transcription bubble, and it provided an important control because it demonstrated that the genetic system we would use to detect transcribed-strand mutations could also detect a bias toward the non-transcribed strand. When residual lacZ transcription was blocked beforehand by catabolite repression, both strands were mutated about equally, but if revertants were selected immediately after nitrous acid exposure, transcribed-strand mutations predominated among the revertants, implicating retromutagenesis as the mechanism. This result was not affected by gene orientation. Retromutagenesis is apt to be a universal method of evolutionary adaptation, which enables the emergence of new mutants from mutations acquired during counterselection rather than beforehand, and it may have roles in processes as diverse as the development of antibiotic resistance and neoplasia.

Gene-based single nucleotide polymorphism discovery in bovine muscle using next-generation transcriptomic sequencing.

BACKGROUND: Genetic information based on molecular markers has increasingly being used in cattle breeding improvement programmes, as a mean to improve conventionally phenotypic selection. Advances in molecular genetics have led to the identification of several genetic markers associated with genes affecting economic traits. Until recently, the identification of the causative genetic variants involved in the phenotypes of interest has remained a difficult task. The advent of novel sequencing technologies now offers a new opportunity for the identification of such variants. Despite sequencing costs plummeting, sequencing whole-genomes or large targeted regions is still too expensive for most laboratories. A transcriptomic-based sequencing approach offers a cheaper alternative to identify a large number of polymorphisms and possibly to discover causative variants. In the present study, we performed a gene-based single nucleotide polymorphism (SNP) discovery analysis in bovine Longissimus thoraci, using RNA-Seq. To our knowledge, this represents the first study done in bovine muscle. RESULTS: Messenger RNAs from Longissimus thoraci from three Limousin bull calves were subjected to high-throughput sequencing. Approximately 36-46 million paired-end reads were obtained per library. A total of 19,752 transcripts were identified and 34,376 different SNPs were detected. Fifty-five percent of the SNPs were found in coding regions and ~22% resulted in an amino acid change. Applying a very stringent SNP quality threshold, we detected 8,407 different high-confidence SNPs, 18% of which are non synonymous coding SNPs. To analyse the accuracy of RNA-Seq technology for SNP detection, 48 SNPs were selected for validation by genotyping. No discrepancies were observed when using the highest SNP probability threshold. To test the usefulness of the identified SNPs, the 48 selected SNPs were assessed by genotyping 93 bovine samples, representing mostly the nine major breeds used in France. Principal component analysis indicates a clear separation between the nine populations. CONCLUSIONS: The RNA-Seq data and the collection of newly discovered coding SNPs improve the genomic resources available for cattle, especially for beef breeds. The large amount of variation present in genes expressed in Limousin Longissimus thoracis, especially the large number of non synonymous coding SNPs, may prove useful to study the mechanisms underlying the genetic variability of meat quality traits.

RNA-driven genetic changes in bacteria and in human cells.

As recently demonstrated in the yeast Saccharomyces cerevisiae model organism using synthetic RNA-containing oligonucleotides (oligos), RNA can serve as a template for DNA synthesis at the chromosomal level during the process of double-strand break (DSB) repair. Herein we show that the phenomenon of RNA-mediated DNA modification and repair is not limited to yeast cells. A tract of six ribonucleotides embedded in single-strand DNA oligos corresponding to either lagging or leading strand sequences could serve as a template to correct a defective lacZ marker gene in the chromosome of the bacterium Escherichia coli. In order to test the capacity of RNA to modify DNA in mammalian cells, we utilized DNA oligos containing an embedded tract of six ribonucleotides, as well as oligos mostly made of RNA. These oligos were designed to repair a chromosomal break generated within a copy of the green fluorescent protein (GFP) gene randomly integrated into the genome of human HEK-293 cells. We show that these RNA-containing oligos can serve as templates to repair a DSB in human cells and can introduce base changes into genomic or plasmid DNA. In both E. coli and human cells, the strand bias of chromosomal gene correction by the single-strand RNA-containing oligos was the same as that obtained for the corresponding DNA molecules. Therefore, the RNA-containing oligos are not converted into a cDNA before annealing with complementary DNA. Overall, we demonstrate that in both bacterial and human cells, as in yeast, RNA sequences can have a direct role in DNA genetic modification and remodeling.

Evaluation of multiple neurotoxic outcomes in cancer chemotherapy.

Although it is now clear that cognitive dysfunction is a common accompaniment of cancer chemotherapy, its implications await further research and direction. Most of the clinical research relies on standard neuropsychological tests that were developed to diagnose stable traits. Cognitive dysfunction in patients undergoing treatment varies with time, however. Its dimensions will vary during the course of treatment, which generally consists of cycles of drug administration followed by recovery periods. To effectively determine the connection between chemotherapy and cognitive function requires neuropsychological tests based on performance, so that they can be administered repeatedly at specified times during the entire course of treatment and beyond. A number of computerized test batteries, many of which have been developed for environmental neurotoxicology, are now available that fit such criteria. Moreover, cognitive impairment is only one aspect of chemotherapy-induced neurotoxicity. A full appreciation of its scope requires assessment of sensory functions such as vision, audition and somatosensory properties and assessment of motor function. A program of research based on animal models is also essential. Only with animal models is it possible to determine dose-response relationships and to couple behavioral with mechanistic indices such as neuroplasticity. Animal behavior models play a vital role in environmental toxicology because, from them, it is possible to derive some index of exposure that limits adverse effects. However, as in human testing, it is critical to choose situations whose properties remain stable over long periods of time so as to trace the time course of neurotoxicity. Schedule-controlled operant behavior offers the most promising source of animal models.

The Seychelles Child Development Study of methyl mercury from fish consumption: analysis of subscales from the Child Behaviour Checklist at age 107 months in the main cohort.

Methyl mercury (MeHg) is neurotoxic and all fish contain at least trace amounts. Consequently, prenatal or fetal exposure occurs when pregnant women consume fish and children are exposed postnatally when they breastfeed or consume fish. However, the level of exposure at which toxicity occurs is presently unknown. Since behavioural endpoints can be sensitive indicators of toxic exposure, we administered the Child Behaviour Checklist (CBCL) to measure behaviour as part of a prospective, longitudinal, double blind study (n=779) of prenatal MeHg exposure, the Seychelles Child Development Study (SCDS). The CBCL Total T score was a primary endpoint at 66 and 107 month evaluations of the cohort and showed no association with prenatal or postnatal MeHg exposure. This paper reports the results of a secondary analysis of the CBCL subscales to see if specific aspects of behaviour might show associations. The SCDS main cohort was enrolled in 1989-90 and evaluated on five occasions through 107 months of age. The child's primary caregiver completed the CBCL at the 107 month evaluation. Prenatal exposure was determined by measuring total mercury (THg) in maternal hair growing during pregnancy and recent postnatal exposure by analysing the child's hair taken at the 107 month evaluation. Analysis included linear and nonlinear multiple regression models. For prenatal MeHg exposure, the Social Problems subscale was significantly associated and the Somatic Complaints subscale was marginally associated. Both were beneficial associations. For postnatal exposure the Thought Problems subscale was associated in an adverse direction. This secondary analysis identified a small number of subtle beneficial and adverse associations with prenatal and postnatal MeHg exposure for specific CBCL subscales. These analyses provide no evidence for an adverse effect of prenatal exposure. The adverse postnatal association is difficult to interpret because we measured only recent (about one month) exposure and no adjustment was made for the multiplicity of endpoints.

Production of 3-nitrosoindole derivatives by Escherichia coli during anaerobic growth.

When Escherichia coli K-12 is grown anaerobically in medium containing tryptophan and sodium nitrate, it produces red compounds. The reaction requires functional genes for trytophanase (tnaA), a tryptophan permease (tnaB), and a nitrate reductase (narG), as well as a natural drop in the pH of the culture. Mass spectrometry revealed that the purified chromophores had mass/charge ratios that closely match those for indole red, indoxyl red, and an indole trimer. These compounds are known products of chemical reactions between indole and nitrous acid. They are derived from an initial reaction of 3-nitrosoindole with indole. Apparently, nitrite that is produced from the metabolic reduction of nitrate is converted in the acid medium to nitrous acid, which leads to the nitrosation of the indole that is generated by tryptophanase. An nfi (endonuclease V) mutant and a recA mutant were selectively killed during the period of chromophore production, and a uvrA strain displayed reduced growth. These effects depended on the addition of nitrate to the medium and on tryptophanase activity in the cells. Unexpectedly, the killing of a tnaA(+) nfi mutant was not accompanied by marked increases in mutation frequencies for several traits tested. The vulnerability of three DNA repair mutants indicates that a nitrosoindole or a derivative of a nitrosoindole produces lethal DNA damage.

Removal of deoxyinosine from the Escherichia coli chromosome as studied by oligonucleotide transformation.

Deoxyinosine (dI) is produced in DNA by the hydrolytic or nitrosative deamination of deoxyadenosine. It is excised in a repair pathway that is initiated by endonuclease V, the product of the nfi gene. The repair was studied in vivo using high-efficiency oligonucleotide transformation mediated by the Beta protein of bacteriophage lambda in a mismatch repair-deficient host. Escherichia coli was transformed with oligonucleotides containing a selectable A-G base substitution mutation. When the mutagenic dG was replaced by a dI in the oligonucleotide, it lost 93-99% of its transforming ability in an nfi(+) cell, but it remained fully functional in an nfi mutant. Therefore, endonuclease V is responsible for most of the removal of deoxyinosine from DNA. New nfi mutants were isolated based on the strong selection provided by their tolerance for transformation by dI-containing DNA. The repair patch for dI was then measured by determining how close to the transforming dG residue a dI could be placed in the oligonucleotide before it interferes with transformation. At the endonuclease V cleavage site, three nucleotides were preferentially removed from the 3' end and two nucleotides were removed from the 5' end. dI:dT and dI:dC base pairs gave the same results. Caveats include possible interference by Beta protein and by mispaired bases. Thus, oligonucleotide transformation can be used to determine the relative importance of redundant repair pathways, to isolate new DNA repair mutants, and to determine with high precision the sizes of repair tracts in intact cells.

Escherichia coli HU protein has a role in the repair of abasic sites in DNA.

HU is one of the most abundant DNA binding proteins in Escherichia coli. We find that it binds strongly to DNA containing an abasic (AP) site or tetrahydrofuran (THF) (apparent K(d) approximately 50 nM). It also possesses an AP lyase activity that cleaves at a deoxyribose but not at a THF residue. The binding and cleavage of an AP site was observed only with the HUalphabeta heterodimer. Site-specific mutations at K3 and R61 residues led to a change in substrate binding and cleavage. Both K3A(alpha)K3A(beta) and R61A(alpha)R61A(beta) mutant HU showed significant reduction in binding to DNA containing AP site; however, only R61A(alpha)R61A(beta) mutant protein exhibited significant loss in AP lyase activity. Both K3A(alpha)K3A(beta) and R61K(alpha)R61K(beta) showed slight reduction in AP lyase activities. The function of HU protein as an AP lyase was confirmed by the ability of hupA or hupB mutations to further reduce the viability of an E. coli dut(Ts) xth mutant, which generates lethal AP sites at 37 degrees C; the hupA and hupB derivatives, respectively, had a 6-fold and a 150-fold lower survival at 37 degrees C than did the parental strain. These data suggest, therefore, that HU protein plays a significant role in the repair of AP sites in E. coli.

Expression of Mycobacterium tuberculosis Ku and Ligase D in Escherichia coli results in RecA and RecB-independent DNA end-joining at regions of microhomology.

Unlike Escherichia coli, Mycobacterium tuberculosis (Mt) expresses a Ku-like protein and an ATP-dependent DNA ligase that can perform non-homologous end-joining (NHEJ). We have expressed the Mt-Ku and Mt-Ligase D in E. coli using an arabinose-inducible promoter and expression vectors that integrate into specific sites in the E. coli chromosome. E. coli strains have been generated that express the Mt-Ku and Mt-Ligase D on a genetic background that is wild-type for repair, or deficient in either the RecA or RecB protein. Transformation of these strains with linearized plasmid DNA containing a 2bp overhang has demonstrated that expression of both the Mt-Ku and Mt-Ligase D is required for DNA end-joining and that loss of RecA does not prevent this double-strand break repair. Analysis of the re-joined plasmid has shown that repair is predominantly inaccurate and results in the deletion of sequences. Loss of RecB did not prevent the formation of large deletions, but did increase the amount of end-joining. Sequencing the junctions has revealed that the majority of the ligations occurred at regions of microhomology (1-4bps), eliminating one copy of the homologous sequence at the junction. The Mt-Ku and Mt-Ligase D can therefore function in E. coli to re-circularize linear plasmid.

Chemobrain: a translational challenge for neurotoxicology.

Neurotoxicity is a frequent accompaniment of cancer chemotherapy, and held by many oncologists to be the major dose-limiting side effect. It appears in many forms, but attracted attention during the past decade primarily because of complaints by patients of impaired cognitive function they have labeled as "chemobrain". Neuropsychological testing confirmed the validity of these complaints and has generated a substantial literature examining different aspects of cognitive impairment in various clinical populations undergoing a variety of treatments. Cognitive impairment is far from the only manifestation of neurotoxicity induced by chemotherapy, however. It alters sensory function and motor function as well. A critical need for patients is a suite of methods that will enable clinicians to trace the onset and progression of neurotoxicity so as to guide and balance decisions about the course of chemotherapy. This commentary describes some of the potential methods and encourages neurotoxicologists to enlist their unique skills in the service of these needs.

The new tapestry of risk assessment.

Neurotoxicology is entering a new phase in how it views and practices risk assessment. Perhaps more than any of the other disciplines that comprise the science of toxicology, it has been compelled to consider a daunting array of factors other than those directly coupled to chemical and dose, and the age and sex of the subject population. In epidemiological investigations, researchers are increasingly cognizant of the problems introduced by allegedly controlling for variables classified as confounders or covariates. In essence, they reason, the consequence is blurring or even concealing interactions of exposure with modifiers such as the individual's social ecology. Other researchers question the traditional practice of relying on values such as NOAELs when they are abstracted from a biological entity that in reality represents a multiplicity of intertwined systems. Although neurotoxicologists have come to recognize the complexities of assessing risk in all its dimensions, they still face the challenge of communicating this view to the health professions at large.

Can endocrine disruptors influence neuroplasticity in the aging brain?

Only within the last two decades has the adult mammalian brain been recognized for its ability to generate new nerve cells and other neural structures and in essence to rewire itself. Although hippocampal structures have received the greatest scrutiny, other sites, including the cerebral cortex, also display this potential. Such processes remain active in the aging brain, although to a lesser degree. Two of the factors known to induce neurogenesis are environmental enrichment and physical activity. Gonadal hormones, however, also play crucial roles. Androgens and estrogens are both required for the preservation of cognitive function during aging and apparently help counteract the risk of Alzheimer's disease. One overlooked threat to hormonal adequacy that requires close examination is the abundance of environmental endocrine-disrupting chemicals that interfere with gonadal function. They come in the form of estrogenic mimics, androgen mimics, anti-estrogens, anti-androgens, and in a variety of other guises. Because our brains are in continuous transition throughout the lifespan, responding both to environmental circumstances and to changing levels of gonadal steroids, endocrine-disrupting chemicals possess the potential to impair neurogenesis, and represent a hazard for the preservation of cognitive function during the later stages of the life cycle.

Toxicoanthropology: Phthalate exposure in relation to market access in a remote forager-horticulturalist population.

Phthalates are a class of plasticizing chemicals produced in high volume and widely found in consumer products. Evidence suggests that phthalates may have non-monotonic effects on reproductive hormone activity. With exposure to phthalates virtually ubiquitous among industrialized populations, identifying unexposed and/or minimally exposed human populations is essential for understanding the effects of low level exposures. Our primary objective was to quantify urinary phthalate metabolite concentrations in the Tsimane', a remote population of Bolivian forager-horticulturalists. Our secondary objectives were to determine if phthalate metabolite concentrations vary in relation to access to market goods; and to explore relationships between phthalate and reproductive hormone metabolite concentrations. Given that phthalate exposure is of particular concern during fetal development, we focused on reproductive age women in the current analyses. Phthalate metabolites were assayed in urine samples from 59 naturally cycling, reproductive age Tsimane' women. Market access was assessed as: (1) distance from residence to the largest nearby town (San Borja, Bolivia) and (2) Spanish fluency. Urinary reproductive hormone metabolite concentrations were quantified using enzyme immunoassays. We fit linear models to examine: (1) predictors of phthalate exposure; and (2) relationships between urinary phthalate and reproductive hormone metabolite concentrations. Eight phthalate metabolites were detectable in at least 75% of samples. Median concentrations were up to an order of magnitude lower than industrialized populations. Proximity to San Borja and Spanish fluency were strong predictors of exposure. In exploratory analyses, the sum of the di-2-ethylhexyl phthalate metabolites (∑DEHP) and Mono-isobutyl phthalate (MiBP) were significantly associated with altered concentrations of urinary reproductive hormone metabolites. Remote, subsistence populations, like the Tsimane', offer a unique window into the health effects of endocrine active compounds because: (1) exposures are low and likely to be first generation; (2) a natural fertility lifestyle allows for exploration of reproductive effects; and (3) ever-increasing globalization will result in increasing exposure in the next decade.

Social ecology of children's vulnerability to environmental pollutants.

BACKGROUND: The outcomes of exposure to neurotoxic chemicals early in life depend on the properties of both the chemical and the host's environment. When our questions focus on the toxicant, the environmental properties tend to be regarded as marginal and designated as covariates or confounders. Such approaches blur the reality of how the early environment establishes enduring biologic substrates. OBJECTIVES: In this commentary, we describe another perspective, based on decades of biopsychological research on animals, that shows how the early, even prenatal, environment creates permanent changes in brain structure and chemistry and behavior. Aspects of the early environment-encompassing enrichment, deprivation, and maternal and neonatal stress-all help determine the functional responses later in life that derive from the biologic substrate imparted by that environment. Their effects then become biologically embedded. Human data, particularly those connected to economically disadvantaged populations, yield equivalent conclusions. DISCUSSION: In this commentary, we argue that treating such environmental conditions as confounders is equivalent to defining genetic differences as confounders, a tactic that laboratory research, such as that based on transgenic manipulations, clearly rejects. The implications extend from laboratory experiments that, implicitly, assume that the early environment can be standardized to risk assessments based on epidemiologic investigations. CONCLUSIONS: The biologic properties implanted by the early social environment should be regarded as crucial elements of the translation from laboratory research to human health and, in fact, should be incorporated into human health research. The methods for doing so are not clearly defined and present many challenges to investigators.

Economic implications of manganese neurotoxicity.

Manganese neurotoxicity is linked primarily to inhalation exposure, and its clinical features are almost totally based on high doses, such as those experienced by miners. Manifestations of lower level exposures can take two forms. One is the appearance of neurobehavioral deficits. A second, equally subtle, form is as a promoter, borrowing the term used in carcinogenesis, of neurodegenerative disease. Such low-level environmental exposures may be more potent than expected if they occur as ultrafine particles able to penetrate directly into the brain. The neurological disorder linked most closely to manganese is Parkinson's disease (PD). Although most observers recognize that the features of manganese-induced parkinsonism differ from those of idiopathic PD, they overlap considerably. The overlaps should be expected because the underlying lesions, although distinguishable, are closely linked because they belong to structures with complex interdependent circuitry. Such interdependence makes it feasible to undertake an analysis of how manganese neurotoxicity might elevate the risks of PD. A relatively small increment in risk, expressed as a leftward shift in the age prevalence of PD, incurs significant economic costs.

A rationale for lowering the blood lead action level from 10 to 2 microg/dL.

Fifteen years ago, in 1991, the U.S. Centers for Disease Control and Prevention (CDC) established 10 microg/dL as the lowest level of concern for children's blood lead levels. This value is extremely important because, historically, policy makers and public health officials generally have acted to remove sources of lead exposure only after the CDC's level of concern had been exceeded. A growing body of evidence, however, reveals that blood lead levels below 10 microg/dL may impair neurobehavioral development. There is now sufficient and compelling scientific evidence for the CDC to lower the blood lead action level in children. This review argues that a level of 2 microg/dL is a useful and feasible replacement. Although it can be argued, in turn, that no threshold for the health effects of lead is demonstrable, analytically a blood level of 2 microg/dL is readily and accurately measured and provides a benchmark for successful prevention. Lowering the level of concern would encourage and accelerate the investments needed to ensure that children are protected from lead exposure in their homes, schools, and play settings. Such a program would also offer economic advantages because of the coupling between lead, educational attainment, earnings and anti-social conduct. By lowering the blood action level, CDC will promote policies and initiatives designed to further reduce children's exposure to this potent developmental neurotoxicant.

Prenatal methyl mercury exposure from fish consumption and child development: a review of evidence and perspectives from the Seychelles Child Development Study.

Evidence from an outbreak of methyl mercury (MeHg) poisoning in Iraq suggested that adverse effects of prenatal exposure on child development begin to appear at or above 10ppm measured in maternal hair. To test this hypothesis in a fish-eating population, we enrolled a cohort of 779 children (the main cohort) in the Seychelles Child Development Study (SCDS). The cohort was prenatally exposed to MeHg from maternal fish consumption, and the children started consuming fish products at about 1 year of age. Prenatal exposure was measured in maternal hair and recent postnatal exposure in the child's hair. The cohort has been examined six times over 11 years using extensive batteries of age-appropriate developmental tests. Analyses of a large number of developmental outcomes have identified frequent significant associations in the appropriate direction with numerous covariates known to affect child development, but only one adverse association between prenatal MeHg exposure and a developmental endpoint. Because such results could be ascribed to chance, there is no convincing evidence for an association between prenatal exposure and child development in this fish-eating population. Secondary analyses have generally supported the primary analyses, but more recently have suggested that latent or delayed adverse effects might be emerging at exposure above 10-12ppm as the children mature. This suggests that the association between prenatal exposure and child development may be more complex than originally believed. This paper reviews the SCDS main cohort study results and presents our current interpretations.

Development and validation of a test battery to assess subtle neurodevelopmental differences in children.

There is increasing concern over the impact of low-dose exposures to environmental chemicals on children's neurobehavioral function. To determine subtle alterations in children's function, it is necessary to move beyond global measures such as IQ and employ tests that can detect small, subtle neurodevelopmental effects across a broad array of behavioral domains. We investigated the sensitivity and specificity of a battery of 63 neurodevelopmental tests or tasks designed to detect outcomes representing the type of subtle neurodevelopmental deficits caused by exposure to neurotoxicants in school-aged children. We studied Neonatal Intensive Care Unit (NICU) graduates, a population known to be at risk for both major and mild anomalies in perception, motor functioning, learning, memory and cognition. This population served as a surrogate to evaluate the capacity of these tests and tasks to predict such deficits. The subjects' histories of previous exposures to any environmental neurotoxicants was not ascertainable, but exposures to elevated levels was not suspected. Over one-third of the 63 measures proved capable of detecting pre-diagnosed lower IQ, the presence of a learning disability (LD) or a neonatal risk profile with at least 70% sensitivity and specificity. Some tests were differentially sensitive and specific, depending upon the presence or absence of one or more of several covariates such as gender, age, hearing status, or familiarity with computers. Tests were also eliminated from the battery if they were affected by too many covariates. We propose calling the final battery of tests that are specific and sensitive to subtle neurodevelopmental changes the Rochester test battery (RTB). Further studies are needed to confirm the capability of the RTB to detect subtle changes associated with neurotoxic exposures.

Methylmercury contamination of laboratory animal diets.

In the midst of research focusing on the neurodevelopmental effects of mercury vapor in rats, we detected significant levels of mercury (30-60 ng/g) in the blood of nonexposed control subjects. We determined that the dominant form of the mercury was organic and that the standard laboratory chow we used in our vivarium was the source of the contamination. The dietary levels were deemed of potential biologic significance, even though they might have fallen below the limits of measurement specified by the supplier. All investigators employing animals in research must assess such potential contamination because dietary agents may alter a) conclusions based on intentionally administered doses, b) outcomes by interacting with other agents that are the primary focus of the research, and c) outcomes of research unrelated to the toxic effects of experimentally administered agents.