Bioaccessibility of total bound fumonisin from corn flakes

Fumonisin B₁ (FB₁) is often found as a natural contaminant of corn and corn-based food. Several publications have demonstrated the presence of fumonisin bound to proteins and to other compounds of the matrix. In spite of the low oral bioavailability of FB₁ in rats, pigs, chickens, cows, and monkeys, FB₁ can cause agriculturally significant disease and possibly human cancer. The aim of this work was to determine the bioaccessibility of total bound FB₁ (TB FB₁) (percentage of TB FB₁, released from corn flakes to the chyme) after in vitro digestion. Two samples of corn flakes washed with solvents were incubated with gastrointestinal tract solutions simulating saliva plus stomach and duodenal juices. After hydrolysis of the chyme with KOH, TB FB₁ was determined as hydrolyzed FB₁ (HFB₁). The bioaccessibility of TB FB₁ in chyme from corn flakes was 37–64%, indicating that these derivatives should be considered in evaluation of exposure to fumonisin.     

In vitro and in vivo studies to assess the effectiveness of cholestyramine as a binding agent for fumonisins

Several adsorbent materials were tested at 1 mg/ml for their in vitrocapacity to adsorb fumonisin B₁ (FB₁) from aqueous solutions. Cholestyramine showed the best adsorption capacity (85% from a solution containing 200 g/ml FB₁) followed by activated carbon (62% FB₁). Bentonite adsorbed only 12% of the toxin from a solution containing 13 g/ml FB₁, while celite was not effective even at the lowest tested FB₁ concentration (3.2 g/ml). Cholestyramine was tested in vivoto evaluate its capacity to reduce the bioavailability of fumonisins (FBs) in rats fed diet contaminated with toxigenic Fusarium verticillioidesculture material. Rats were exposed for one week to FBs-free diet, FBs-contaminated diet containing 6 or 20 g/g FB₁ + FB₂ and the same FBs-contaminated diet added of 20 mg/g cholestyramine. The increase of sphinganine/sphingosine (SA/SO) ratio in urine and kidney of treated rats was used as specific and sensitive biomarker of fumonisin exposure.The addition of cholestyramine to the FBs-contaminated diets consistently reduced the effect of FBs by reducing significantly (P < 0.05) both urinary and renal SA/SO ratios.This revised version was published online in October 2005 with corrections to the Cover Date.     

Effect of<Emphasis Type="Italic">in vitro</Emphasis> digestion on fumonisin B<Subscript>1</Subscript> in corn flakes

Low levels of fumonisins have been found frequently in corn based breakfast cereals and can occur bound to protein and other matrix components.In vitro digestion of two samples of corn flakes was carried out under "fed conditions." Fumonisins were measured as o-phthaldialdehyde/mercaptoethanol derivatives by LC-fluorescence. One sample of corn flakes (FN12) had high concentrations of fumonisin B₁ (FB) (average 125 ng/g) and total bound FB₁, (TB FB₁) (average 92 ng/g) and the other (FN11) had a low level of free FB₁ (average 29 ng/g) and no detectable TB FB₁. After incubation of the samples with gastrointestinal tract solutions simulating saliva plus stomach and duodenal juices, chyme was analysed for FB₁, hydrolyzed FB₁ (HFB₁) and partially hydrolyzed fumonisin B₁ (PHFB₁). The bioaccessibility (percentage of FB₁ released from corn flakes into chyme) was 38-78% for incurred FB₁ in FN12, 8-54% for incurred plus spiked FB₁ in FN12, and 19-66% for incurred plus spiked FB₁ in FN11. HFB₁ and PHFB₁ were not detected. If free FB₁ was first extracted from sample FN12, no FB1 was detected in the chyme, indicating no contribution from TB FB₁. Concentrations were corrected for method recovery of FB₁ or, for bound FB₁, partial method recovery of HFB₁ Presented at the XIIth IUPAC International Symposium on Mycotoxins and Phycotoxins, Istanbul, Turkey, 21–25 May, 2007     

Experimental reproduction of ELEM

An experiment to gain insight into the minimum toxic dose of fumonisins was conducted by feeding ponies rations with known fumonisin concentrations. Naturally contaminated corn screenings (CS) were blended with pellets, corn, and molasses to formulate individual daily diets. One group of 4 ponies was fed a ration with fumonisin B₁ (FB₁) varying from <1 ppm to 22 ppm. A second group of 5 ponies was fed a ration at varying rates containing 8 ppm FB₁ for 180 days. A panel of clinical chemistry parameters was evaluated twice weekly for both groups. One pony in the first group died of equine leukoencephalomalacia (ELEM) after 225 days of which the final 55 days' diet contained 22 ppm FB₁. Approximately 9 days prior to death, this animal experienced elevated liver chemistry values. All 5 ponies in the second group experienced mild, transient, clincal signs; were euthanized at 180 days; and had mild, histopathological brain lesions.     

Mutagenic potentials of fumonisin contaminated corn following ammonia decontamination procedure

Naturally contaminated corn implicated in an outbreak of equine leukoencephalomalacia (ELEM) in southeastern Arizona was analyzed for mutagenic potential using the Salmonella/microsome mutagenicity assay before and after treatment with the ammonia procedure. Crude acetonitrile: water (1+1) extracts of high-pressure/ambient temperature (HP/AT) ammonia decontaminated, HP/AT plus low pressure/high temperature (LP/HT), and non-ammoniated fumonisin contaminated corn were tested for mutagenic potentials. Relatively pure (approx. 90%) fumonisin B₁ standard was also tested for comparison purposes. The results of this experiment indicate that there was no mutagenic potential for the fumonisin B₁ standard at the concentrations tested (100 g/plate). Also, neither the naturally-contaminated corn nor the ammonia decontaminated samples elicited a positive mutagenic response. Fumonisin B₁ levels, as determined by HPLC methods, were reduced by an average of 79% via the ammonia decontamination process. It is encouraging to note that, while further work is necessary to increase the efficacy of the ammonia process to reduce fumonisin levels, the ammonia process did reduce fumonisin levels and no mutagenic potentials were apparent in the treated corn.Abbreviations HP/AT high pressure/ambient temperature - LP/HT low pressure/high temperature - ELEM equine leukoencephalomalacia - FB₁ fumonisin B₁ - FB₂ fumonisin B₂     

Production of fumonisins B1, B2 and B3 byFusarium proliferatum Isolated from rye grains

Using the seed- plate technique, we have isolated a strain ofF. proliferatum from rye grains that produces 3 fumonisins, fumonisin B₁ (FB₁), FB₂ and FB₃ on inoculated rice and corn. Inoculated corn and rice were extracted with an aqueous methanol solution and fumonisin concentrations estimated using high performance liquid chromatography. Production of all 3 fumonisins (FB₁, FB₂ and FB₃) was much higher on rice than corn; ranging from 3816, 1068 and 985 ppm to 1643, 350 and 162 ppm respectively. We conclude that all natural substrates whereF. proliferatum is present as a component of the mycoflora should be monitored for the presence of fumonisins.     

Production of fumonisins byFusarium species of Taiwan

Twenty-nineFusarium species isolated from various sources in different districts of Taiwan were tested for their ability to produce fumonisins in corn cultures. OnlyFusarium moniliforme produced fumonisin B₁ (FB₁) and fumonisin B₂ (FB₂). The finding that the other 28Fusarium species produced neither FB₁ nor FB₂ is preliminary because only one strain per species was studied. The detection of FB₁ and FB₂ in cultures ofF. moniliforme was demonstrated by TLC and HPLC, and FB₁ was further confirmed by mass spectrometry. In a separate experiment, in which 38 strains ofF. moniliforme were tested for fumonisins, approximately 66% (25/38) produced FB₁ and/or FB₂. Of the 25 strains, 14 produced only FB₁ and 11 produced both FB₁ and FB₂, and the amounts of FB₁ and FB₂ produced by different strains varied greatly. This is the first report that fumonisins are found in corn cultures experimentally infected withF. moniliforme strains from Taiwan. It is safe to assume that fumonisin producing strains ofF. moniliforme are widely distributed among the economic crops such as corn, rice, sugarcane, and sorghum throughout the Island.Abbreviations FB₁ Fumonisin B₁ - FB₂ Fumonisin B₂ - OPA o-phthalidialdehyde     

Temporal and dose-response features in swine fed corn screenings contaminated with fumonisin mycotoxins

Fumonisin B₁ (FB₁), a mycotoxin produced byFusarium moniliforme andF. proliferatum, induces liver damage and pulmonary edema in swine. We examined the temporal and dose-response features of FB₁ toxicosis in male weanling crossbred pigs fed nutritionally balanced diets, containing corn screenings naturally contaminated with fumonisins, for 14 days. Total fumonisins (FB₁ and FB₂) in diets 1 through 6 were assayed at 175, 101, 39, 23, 5, and <1 ppm (below detectable concentrations), respectively. Clinical signs, serum biochemical alterations, and morphologic changes were evaluated. Pigs were weighed, and bled for hematologic and clinical chemistry evaluation on days 5 and 14. They were euthanized on day 14, or earlier if respiratory distress was observed. Respiratory distress developed in 3/5 pigs fed diet 1 between days 4 and 6 due to severe pulmonary edema and pleural effusion. Histologic evidence of hepatic injury was present in all pigs fed diets 1 and 2, 3/5 on diet 3, and 1/5 on diet 4. Serum bilirubin and cholesterol concentrations, gamma-glutamyl transferase (GGT), alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and arginase (ARG) activities were elevated in pigs fed diets 1 and 2. Based on liver histopathology, the no observed adverse effect level (NOAEL) for fumonisin toxicity in swine was <23 ppm total fumosins for the 14-day period. Based on regression analyses of the clinical chemistry profiles at 14 days, the NOAEL was <12 ppm, with ALP being the most sensitive parameter. In conclusion, pulmonary edema occurred only at the highest fumonisin concentration (175 ppm), while liver damage occurred at much lower concentrations with a NOAEL of <12 ppm.Abbreviations ALP alkaline phosphatase - ALT alanine aminotransferase - ARG arginase - AST aspartate aminotransferase - ELEM equine leukoencephalomalacia - FB₁ fumonisin B₁ - GGT gamma-glutamyl transferase - NOAEL no observed adverse effect level     

Fumonisin and Beauvericin Induce Apoptosis in Turkey Peripheral Blood Lymphocytes

Fumonisins, a family of mycotoxins produced by Fusarium verticillioides (synonym Fusarium moniliforme Sheldon) and F. proliferatum, have been associated with various deleterious effects in different animal species. Serological, hematological and pathological effects and mortality have previously been observed in broiler chicks fed F. proliferatum culture material containing known concentrations of fumonisin, moniliformin and beauvericin. Turkey peripheral blood lymphocytes were exposed in vitro for 72 hours to fumonisin B₁(FB₁), fumonisin B₂(FB₂), hydrolyzed fumonisin B₁ (HFB₁), moniliformin and tricarballylic acid (TCA) (0.01-25 g/ml). A decrease in cell proliferation, as determined by the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide] bioassay, occurred in the order: FB₂ > FB₁ > HFB₁, with IC₅₀ = 0.6 M, 1 M and 10 M, respectively. Internucleosomal DNA fragmentation and morphological features characteristic of apoptosis were observed following exposure to fumonisin B₁ and beauvericin; cytoplasmic condensation and membrane blebbing were seen by light microscopy. Tricarballylic acid and moniliformin did not interfere with cell proliferation. Results suggested that fumonisin B₁ and beauvericin may affect immune functions by suppressing proliferation and inducing apoptosis of lymphocytes.     

Fumonisins: Isolation,chemical characterization and biological effects

The fumonisin B mycotoxins (FB₁ and FB₂) have been purified and characterized from corn cultures of Fusarium moniliforme strain MRC 826. Fumonisin B₁ (FB₁, the major fumonisin produced in culture, has been shown to be responsible for the major toxicological effects of the fungus in rats, horses and pigs. Recent investigations on the purification of compounds with chromatographic characteristics similar to FB₁ have led to the identification of two new fumonisins, FB₃ and FB₄. Fumonisins A₁ and A₂, the N-acetyl derivatives of FB₁ and FB₂ respectively, were also purified and shown to be secondary metabolites of the fungus. Short-term carcinogenesis studies in a rat liver bioassay indicated that over a period of 15 to 20 days, at dietary levels of 0.05–0.1%, FB₂ and FB₃ closely mimic the toxicological and cancer initiating activity of FB₁ and thus could contribute to the toxicological effects of the fungus in animals. In contrast, no biological activity could be detected for FA₁ under identical experimental conditions. These studies and others have indicated that the fumonisin B mycotoxins, although lacking mutagenicity in the Salmonella test or genotoxicity in the DNA repair assays in primary hepatocytes, appear to induce resistant hepatocytes similar to many known hepatocarcinogens.     

Natural occurrence of fumonisins and their correlation to Fusarium contamination in commercial corn hybrids growth in Argentina

Fifty commercial corn hybrids with different endosperm characteristics, vegetative cycle length and cross class grown in the same geographical area (Cordoba Province, Argentina) were analysed for fumonisin accumulation. All hybrids analysed showed fumonisin B₁ and B₂ contamination ranging from 185 to 27,050 ng/g for FB₁ and from 40 to 9950 ng/g for FB₂. Although most of the hybrids analysed had flint-type endosperm, two hybrids with dent-type endosperm (e.g. Prozea 10 and AX 746) showed the highest level of fumonisin (37,000 ng/g) and more FB₂ than FB₁ (FB₂/FB₁ ratio 2.42), respectively. There was no correlation between fumonisin concentration and length of the vegetative cycle. Among 18 hybrids examined for Fusarium species contamination there was also no correlation between fumonisin contamination and the level of infection with Fusarium species (Section Liseola). Eighteen hybrids showed fumonisin levels lower than 1000 ng/g. This result suggests that there is some possibility of selecting hybrids resistant or less susceptible to fumonisin and Fusarium contamination.     

Natural occurrence of Fusarium species and fumonisin-production by toxigenic strains isolated from poultry feeds in Argentina

Fusarium species and fumonisin production by toxigenic strains were investigated. During 1996–1998, 158 samples of poultry feeds were collected from a factory located in the department of Río Cuarto Córdoba province, Argentina. The most common species of Fusarium were F. moniliforme (60.7%) and F. nygamai (35.4%) followed by F. semitectum, F. subglutinans, F. proliferatum, F. dlamini, F. solani, F. oxysporum and F. napiforme. Fungal counts ranged from 1 × 10³ to 8 × 10⁵ CFU/g with mean values from 1.5 × 10³ to 2.3 × 10⁵ CFU/g. The highest counts were for F. dlamini, F. subglutinans, F. moniliforme and F. nygamai. Strains of F. moniliforme, F. nygamai, and F. proliferatum were screened for their potential to produce fumonisin B₁ (FB₁), fumonisin B₂ (FB₂) and fumonisin B₃ (FB₃) in corn grain. The samples were analysed using a modified high performance liquid chromatography method. The strains assayed, 43 strains, produced three fumonisins. There was a high degree of variability in the quantities of FB₁, FB₂, and FB₃ produced. The toxin produced in highest levels by the majority of the strains was FB₁. The range of concentration varied from 5.4 to 3,991, 1.01 to 189 and 0.4 to 765 ppm per gram of corn for FB₁, FB₂ and FB₃ respectively. The toxigenic pattern of strains was normal, although two strains of F. moniliforme produced exceptionally high concentrations of FB₃ and minor concentrations of FB₂ and FB₁. This is the first report from Argentina on Fusarium species in poultry feeds and fumonisin production by these strains.This revised version was published online in October 2005 with corrections to the Cover Date.     

Occurrence of Fumonisin B<Subscript>1</Subscript> in Corn from the Main Corn-Producing Areas of China

Fumonisin B₁ (FB₁) is the most abundant of the fumonisin mycotoxins, mainly produced in maize by F. verticillioides and F. proliferatum. A total of 282 corn samples harvested in 2005 from six provinces, the main corn-producing areas of China, were analyzed for FB₁ using high-performance liquid chromatography. All samples except one were (99.6%) positive for FB₁ at levels varying from 3 to 71,121 ng/g with mean and median levels for all samples of 6,662 and 1,569 ng/g, respectively. During an analysis of the distribution pattern for FB₁, it became apparent that 43.6% of tested samples had FB₁ concentrations below 1,000 ng/g, while 25.2% contained in excess of 5,000 ng/g. The average exposure to FB₁ (1.1 μg/kg body weight/day) is within the provisional maximum tolerable daily intake of 2 μg/kg body weight/day set by the Joint FAO/WHO Expert Committee on Food Additives.     

Fumonisins: current research trends in developmental toxicology

Fumonisin B₁ (FB₁) is a mycotoxin produced byFusarium verticillioides that is found in maize and maize-based foods. Reproductive studies in CD1 mice, rats and rabbits initially found no evidence that fumonisins are teratogenic. However, more recent findings suggest that they might increase the risk of neural tube defects (NTDs) in populations consuming large amounts of fumonisin-contaminated corn. When ≥15 mg/kg body weight fumonisin B₁ (FB₁) was given to pregnant LM/Bc mice by intraperitoneal (ip) injection, all litters were positive for NTDs. To determine if NTD induction is unique to the inbred LM/Bc mouse strain, NTD induction in LM/Bc and CD1 mice was compared: (a) in a study in whichF. verticillioides culture material providing ≤150 ppm FB₁ was fed to female mice before and during gestation, and (b) in a study in which FB₁ was given by ip injection to CD1 dams on gestation days 7 and 8, the critical time for NTD development. In the feeding study, one of five LM/Bc litters from dams fed the 150 ppm FB₁ diet was positive for NTDs whereas no NTDs were found in the CD1 litters. In the ip injection study, 40% of the litters at the highest dose tested, 45 mg/kg body weight, were positive for NTDs and one of nine low-dose (15 mg/kg body weight) litters was also positive. Thus, FB₁ induced NTDs in both LM/Bc and CD1 mice although the latter strain appears less sensitive. Comparative investigations using these strains will be useful for elucidating the mechanisms underlying fumonisin-induced NTDs in mice and determining the suitability of mouse models for studying the relationships between fumonisins and NTDs in humans.     

Preliminary determination of mycotoxin binding to soil when leaching through soil with water

Mycotoxin-contaminated crops that are left in the field are potential contaminants of groundwater. Aflatoxin B₁ (AFB₁) and fumonisin B₁ (FB₁) distribution in soil-water systems and the comparative response of aflatoxin-contaminated corn and pure aflatoxin when leached through soil were investigated using columns. Each experiment was repeated once. Eluates and soil extracts were analyzed for AFB₁ and its metabolites and FB₁ along with different amounts of pure FB₁ and water mixtures. AFB₁ was detected in water samples from columns containing 10% and 20% silty clay loam soil and aflatoxin-contaminated corn mixtures and in the upper (top) 2.5 cm of soil from the 10 cm soil column. Aflatoxin B₂ (AFB₂) was detected in eluates from the column containing 10% soil and aflatoxin-contaminated corn mixture and from the column containing aflatoxin-contaminated corn alone. No AFB₂ was detected in eluates from the column containing 20% soil and aflatoxin-contaminated corn mixture. No detectable amount of aflatoxin was observed in eluates from the containing 50% silty clay loam soil and aflatoxin contaminated corn. No detectable amount of FB₁ was observed in eluates or soil extracts, but FB₁ was detected in the mixtures of pure FB₁ and water.     

Trends of fumonisin contamination and animal intoxication through monitoring 1991 to 1997 corn crop in the State of Paraná, Brazil

Eleven feed samples associated with six animal (horse and poultry) intoxication outbreaks (1991) in the state of Paraná, Brazil, were evaluated for fungal and fumonisin contamination. In order to estimate the␣trend of livestock intoxication, fumonisin contamination was monitored in corn produced both at the commercial level (1991, 1995 crop), and in an experimental field at a local Agronomy Institute (1997 crop). The total mould count in the feed samples ranged from 2.9 × 10³ to 1.9 × 10⁷ CFU/g, with Fusarium verticillioides as the predominant species, at a high count of 2.4 × 10⁴–6.5 × 10⁵ CFU/g. Fumonisins (FB₁ + FB₂) were detected in all corn-based feed samples at levels ranging from 2.89 to 14.54 μg/g. All 27 Northern corn samples (1991 crop) were contaminated with fumonisins at levels ranging from 2.32 to 16.64 μg/g. Twenty-six (96.3%) out of 27 corn samples from the Central-Southern region (1995 crop) were positive for fumonisins (FB₁+FB₂), with the range of 0.07–3.66 μg/g, while all 37 Northern samples (1995 crop) were contaminated with fumonisins ranging from 0.57 to 9.97 μg/g. Twenty-one out of 37 corn samples from the Northern region (1997 crop) were positive for fumonisins, but at low level (range of 0.05–2.67 μg/g). The results showed a decreasing trend in fumonisin contamination over the years. Nowadays animal intoxication outbreaks rarely occur in this State, as both animal producers and feed industries have become conscious about monitoring of corn and other raw materials at the quality control level.     

Developmental toxicity of fumonisin in Syrian hamsters

The effects of fumonisin on development of Syrian hamster fetuses were studied using fumonisin B₁ and B₂ extracted fromFusarium moniliforme corn-culture and purified fumonisin B₁. A significant increase in litters with fetal deaths occurred with the high doses of purified (18 mg FB₁/kg) and culture-extracted (18 mg FB₁ plus 4.5 mg FB₂) fumonisin. It is concluded that prenatal exposure to fumonisin on days 8 and 9 of gestation is detrimental to fetal hamster survivability but does not induce clinical maternal intoxication at these doses. Equivalent doses of fumonisin B₁, whether from culture-extract or pure solution produced similar results.     

Factors affecting the growth of Fusarium proliferatum and the production of fumonisin B1: oxygen and pH

Fumonisins are mycotoxins produced primarily by Fusarium moniliforme and Fusarium proliferatum in corn. In liquid culture, production of fumonisin B₁ (FB₁), the most common moiety of the family of fumonisins, can be obtained using a defined medium that is nitrogen-limited. Under nitrogen-limited conditions both growth and the production of FB₁ were greatly influenced by pH and aeration. At pH above 5.0, F. proliferatum grew normally but produced little FB₁ (<100 μg m⁻¹). At pH below 5.0, there was less growth but substantially more FB₁. Below a pH of 2.5, both growth and metabolism were slower with very little FB₁ produced. When the optimal pH range of between 3.0 and 4.0 under well-aerated conditions was used, both growth and FB₁ production were high. However, under oxygen-limited conditions, less growth occurred, glucose consumption was increased, and no FB₁ was produced. Received 16 May 1997/ Accepted in revised form 03 September 1997     

In vivo effects of fumonisin B1-producing and fumonisin B1-nonproducing Fusarium moniliforme isolates are similar: Fumonisins B2 and B3 cause hepato- and nephrotoxicity in rats

Fumonisins are mycotoxins produced by Fusarium moniliforme, F. proliferatum, and related Fusarium species found on corn. They occur naturally in corn-based feeds and foods and are suspected human esophageal carcinogens. Fumonisin B₁ (FB₁), the most common homologue, causes the animal diseases associated with F. moniliforme. Hepato- and nephrotoxicities, disrupted sphingolipid metabolism, and liver cancer have been found in rats fed FB₁. To determine the in vivo effects of diets containing fumonisins B₂ (FB₂) or B₃ (FB₃), male rats were fed culture materials (CM) of FB₁ non-producing F. moniliforme isolates to provide low (4.6–6.7 ppm), mid (32–49 ppm) or high (219–295 ppm) dietary levels of either FB₂ (FB₂CM) or FB₃ (FB₃CM). Other groups were fed culture material of an FB₁ producing isolate (FB₁CM) providing 6.9, 53 or 303~ppm total fumonisins (FB₁ : FB₂ : FB₃ = 1.0 : 0.38 : 0.15) and a tenth group was fed a control diet having no detectable fumonisins. One-half (n = 5/group) the animals were killed after three weeks, at which time the toxicological and histopathological effects of the three culture materials were similar, mimicked the effects of FB₁, and included decreased body weight gains, serum chemical indicators of hepatotoxicity, decreased kidney weights, and apoptosis of hepatocytes and kidney tubular epithelium. FB₁CM, FB₂CM, and FB₃CM affected sphingolipids, causing increased sphinganine to sphingosine ratios (Sa/So) in both liver and kidneys. The remaining animals (n = 5/group) were fed a control diet for three additional weeks. All body weight and tissue specific effects, including increased Sa/So, induced by the FB₂CM, FB₃CM and low level FB₁CM diets were absent following the recovery period. Except for mild biliary lesions found in the high dose FB₁CM group and a few apoptotic hepatocytes present in one mid- and two high-dose FB₁CM rats, no evidence of toxicity remained in these groups following the recovery period.This revised version was published online in October 2005 with corrections to the Cover Date.