Iowa State University Capstones, Theses and Retrospective Theses and Dissertations Dissertations

1985 C-10(19) oxidation of vitamin D by anaerobic gastrointestinal bacteria Robert Marshall Gardner Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/rtd Part of the Microbiology Commons

Recommended Citation Gardner, Robert Marshall, "C-10(19) oxidation of vitamin D by anaerobic gastrointestinal bacteria " (1985). Retrospective Theses and Dissertations. 7849. https://lib.dr.iastate.edu/rtd/7849

This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. INFORMATION TO USERS

This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quahty of the reproduction is heavily dependent upon the quality of the material submitted.

The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.

1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.

2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.

3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of "sectioning" the material has been followed. It is customary to begin filming at the upper left hand comer of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.

4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.

5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.

University Microfilms Intemationcil 300 N. Zeeb Road Ann Arbor, Ml 48106

8514401

Gardner, Robert Marshall

C-10(19) OXIDATION OF VITAMIN D BY ANAEROBIC GASTROINTESTINAL BACTERIA

Iowa State University PH.D. 1985

University Microfilms Intsmstionsl 300 N. Zeeb Road, Ann Arbor, Ml 48106

C—10(19) oxidation of vitamin D by anaerobic

gastrointestinal bacteria

by

Robert Marshall Gardner

A Dissertation Submitted to the

Graduate Faculty in Partial Fulfillment of the

Requirements for the Degree of

DOCTOR OF PHILOSOPHY

Departments: Microbiology Animal Science Co-majors: Microbiology Nutritional Physiology

Approved; Members of the Committee:

Signature was redacted for privacy.

Signature was redacted for privacy.

In Charge of Maior Work

Signature was redacted for privacy.

Signature was redacted for privacy.

Iowa State University Ames, Iowa 1985 ii

TABLE OF CONTENTS

Page

GENERAL INTRODUCTION 1

Dissertation Format 2

LITERATURE REVIEW 3

Bacterial Transformations of Steroids in the Gastrointestinal (GI) Tract 3

Importance and implications 3 Bile 6 Cholesterol 23 Other biologically important steroids 28 Vitamin D 33

Vitamin D 35

Importance and implications 35 Metabolism 38 Biological activity 41 Enterohepatic cycling of vitamin D 42 Requirements and toxicity 43

SECTION I. FACTORS AFFECTING THE PRODUCTION OF 5(E) 19-NOR-lO-KETO-VITAMIN D3 BY RUMEN BACTERIA 46

INTRODUCTION 47

MATERIALS AND METHODS 49

General 49

Steroids 49

Enzymes 50

Extraction and Chromatography 50

Preparation of Media 51

Assay Conditions 52 iii

Rumen Fluid Studies 52

Oxygen studies 53 Preincubation treatments 53

Mixed Culture Studies 53

Filtration 54 Time study 54 Glucose inhibition 55

Isolation of Clostridium sp. Strain A264 55

Pure Culture Studies 55

Glucose inhibition studies 56

Cysteine Studies 56

RESULTS 57

Rumen Fluid Studies 57

BHI Broth Culture Studies 58

DISCUSSION 62

LITERATURE CITED 77

SECTION II. THE BIOLOGICAL ASSESSMENT OF VITAMIN Dg METABOLITES PRODUCED BY RUMEN BACTERIA 80

INTRODUCTION 81

MATERIALS AND METHODS 83

General 83

Sterols 83

Isolation of lO-Keto-Dg From Bovine Plasma 84

Radioligand Binding Assays 85

Pharmacokinetics 85 iv

Biological Evaluation 86

Preparation "of Bacterial Culture Medium 87

Isolation of Clostridium sp. Strain A264 87

Rumen Fluid 87

Assay Conditions 88

Extraction and Chromatography 89

RESULTS 91

Isolation and Mass Spectral Detection of lO-Keto-Dg From Calves 91

Pharmacokinetics 91

Biological Activity Evaluation 92

Relative Affinity for Plasma Vitamin D Binding Protein 93

Relative Affinity for Thymus 1,25-(OH)2-Vitamin D3 Receptor Protein 93

Characterization of "Peak A" 93

DISCUSSION 96

LITERATURE CITED 114

SUMMARY AND DISCUSSION 117

LITERATURE CITED 119

ACKNOWLEDGMENTS 133 1

GENERAL INTRODUCTION

The ability of gut bacteria to structurally modify steroids has important ramifications. Slight structural changes can dramatically affect function or biological activity. Previous research on bacterial steroid biotransformations has primarily dealt with the three major classes of steroids which pass through the gastrointestinal tract: cholesterol, steroid hormones, and bile acids. Recently, vitamin D3 (a

9,10 seco-steroid) has been shown to undergo bacterial-catalyzed C-10(19) oxidation, yielding the major product 5(E)-19-nor-10-keto- vitamin D3 (lO-keto-Da). This compound has also been detected in the plasma of cows receiving large oral doses of vitamin D3, thus denoting in vivo production.

Biological assays with lO-keto-Dg revealed that the C-10(19) oxidative pathway represents a deactivation process in vitamin D me­ tabolism. Whether or not the vitamin D status of an animal is affected by the bacterial removal of vitamin D has yet to be determined.

The production of lO-keto-Ds could also be an important process in monogastric animals since vitamin D3, 25-hydroxyvitamin D3, and 1,25- dihydroxyvitamin D3 circulated in the enterohepatic cycle. In addition to this cycle, diet represents the other major route by which vitamin D enters the gut. Results obtained from ruminai systems should also transfer to monogastric animals since the rumen is a model of study for the microbiology of the cecum and large intestine. In light of current trends of providing large oral doses of vitamin D for domestic animals 2

in feeds, or administered clinically, a great deal of substrate may be available for the metabolism of vitamin D by gut bacteria.

Therefore, the first phase of this study involves defining the conditions necessary for bacterial C-10(19) oxidative metabolism to occur. It is hoped that this work will help to clarify the situation in vivo. The second part reports the isolation, purification, and bio­ logical assessment of lO-keto-Dg and "Peak A," a new bacterial metabo­ lite of vitamin Dg.

Dissertation Format

This dissertation is presented in the alternate thesis format, and includes two manuscripts to be submitted for publication in (1) Applied and Environmental Microbiology, and in (2) Biochemical and Biophysical

Research Communications. The format used in this dissertation is that of the American Society for Microbiology. An in-depth literature re­ view precedes the first manuscript. At the end of each section, a sep­ arate literature cited section has been included to facilitate publica­ tion. A general summary and discussion follow the final manuscript.

The doctoral candidate, Robert Marshall Gardner, was the principal investigator in each of these studies, with the exception of the pharmacokinetics and binding assays which were performed by R. L.

Horst. 3

LITERATURE REVIEW

This literature review is divided into two major sections. Part I will deal with bacterial biotransformation of steroids in the gastro­ intestinal tract; Part II contains a summary of selected literature on the vitamin D-endocrine system.

Bacterial Transformations of Steroids in the Gastrointestinal (GI) Tract

Importance and implications

Anaerobic bacteria in the GI tract play an important role in al­ tering the structure of many endogenous and dietary steroids (38, 41,

89). Certain structural modifications of the parent steroid can impart unique biological activity, or inhibit inherent activity, resulting in important physiological changes in the host. Both situations should be considered when steroids are administered orally (2, 14, 28) or if they undergo enterohepatic circulation (8). These biologically altered com­ pounds, in many instances, can be reabsorbed back into the body, either directly through the cell membranes of the intestinal epithelium or by the enterohepatic cycle (30, 38, 89). After absorption, further modi­ fication of the steroid may occur in the liver (89). Research with steroid biotransformations by intestinal bacteria has emphasized the three major classes of steroids which pass through the GI tract; these are the bile acids (43, 78, 92, 94, 103), cholesterol (18, 39, 143,

147), and steroid hormones (14, 16, 154). Although this review will almost exclusively deal with bacteria, it should be noted that protozoa 4

and fungi can also play an important role in steroid biotransformations

(21, 60, 121).

Several epidemiology studies provide evidence for a strong corre­ lation between the levels (in the colon) of certain bacterial metabo­ lites of bile and cholesterol and the incidence of colon cancer (58,

59, 102, 108, 116). These transformation products are thought to act as carcinogens or cocarcinogens. The bacteria responsible for produc­ tion of these suspect compounds increase markedly on diets high in fat content (under these conditions bile production likewise increases), which have been frequently linked with the etiology of colon cancer

(68, 108). A high-risk diet can alter the metabolic activity of the gut microflora and this in turn could play a role in the etiology of cancer of the large bowel (116). The hypothesis supporting bacterial involvement in the etiology of colon cancer is still controversial (37,

108). Unfortunately, for rapid progress in this area, dietary factors are so interwoven with bacterial involvement that interactions between the two result in a complex situation which tends to obscure individual contributions.

Bacterial steroid transformations may also exert subtle influences on host physiology. The modification of bile acids by bacteria can influence, for example, the absorption of other neutral steroids and fatty acids. The intestinal flora influences the types of bile acids produced, their turnover time, and excretion rate, thereby affecting the absorption of other lipophilic material (38). In addition to bac­ teria, the types of bile acids along the GI tract are also influenced 5

by other factors such as diet, enterohepatic circulation, and intesti­ nal motility (19)• Fecal bile acid composition might change markedly with changes in the intestinal flora (30). This is supported by the observation that the small intestines of gnotobiotic rats contain at least twice the amount of bile acids found in conventional rats.

Lithocholic (3a—hydroxy-56-cholan-24-oic) acid, a product of microbial bile metabolism, is a toxic compound and provokes cholangiofibrosis.

In rats given a protein-deficient diet, lithocholic acid induced ductu- lar cell proliferation and formation of gall stones (38). Gastroin­ testinal bacteria influence plasma cholesterol levels by metabolizing cholesterol in the gut (109). Therefore, it is logical that these bac­ teria may also be a contributing factor in atherosclerosis. These ex­ amples illustrate the extent with which the flora can interact with its host.

Bacteria also influence the fate of steroids by methods other than direct metabolism. Mott et al. (109) believe that factors such as bulk of feces, intestinal transit time, and the binding of steroids (all three of which are affected by intestinal bacteria) are important.

Steroid-transforming bacteria are not unique to any particular genus, habitat, or mode of metabolism. They represent a very diverse group of organisms, as might be expected from the diversity of steroid- transforming reactions documented. These are classified into six major reactions categories: (1) oxidation; (2) reduction; (3) esterification, amide formation, and hydrolysis; (4) isomerization; (5) miscellaneous addition, rearrangement, and elimination processes; and (6) resolution 6

of d-1 mixtures (21). In many instances, the bacterial transformation

of a steroid (even in pure culture) is not limited to any one reaction

type. Biotransformations of steroids by bacterial enzymes are, how­

ever, stereo-specific, and usually represent a very small change in

structure — hence, the term "transformation" is more appropriate than

the term "degradation." Industries which produce biologically active

steroid drugs for clinical use commonly take advantage of a particular

microbial transformation to synthesize a desired steroid from rela­

tively common and easily attained starting material (i.e., cholesterol

or bile) (3, 84). Given the wide diversity of reaction types, the pro­

duction of very complex steroids can be greatly aided by bacteria.

Microbiological methods are often easier and less expensive than organ­

ic chemical syntheses.

Bile

Perhaps the most important steroid transformations that take place

in the GI tract involve the bile acids (64, 105). Most bile acids are

derived from cholesterol and qualitatively represent the most important

end products of cholesterol metabolism (38). Conventional animals

(those with a normal flora) can compensate for the increased loss of

bile acids due to bacteria by increasing the rate of oxidative catabo-

lism of cholesterol.

Bile acids are classified as either primary or secondary deriva­

tives, depending upon their origin. Primary bile acids are produced in

the liver and are unaltered by intestinal bacterial enzymes. Usually,

the primary bile acids are differentiated from secondary or microbial 7

metabolites by the use of germ-free animals. In germ-free rats vs. conventional, the major fecal bile acids are 6-muricholic (3a,66,76- trihydroxy-56-cholan-24-oic) acid and cholic (3a,7a,12a-trihydroxy-

5B-cholan-24—oic) acid. Chenodeoxycholic (3a,7a-dihydroxy-56-cholan-

24-oic) acid and a-muricholic (3a,66,7a-trihydroxy-56-cholan-24-oic) acid are also present in minor amounts (118). In humans, the major primary bile acids are cholic and chenodeoxycholic acids. Human gall­ bladder bile contains about 70% primary bile acids, 25% 7-dehydroxy- lated bile acids (mostly deoxycholic acid and to a lesser extent lithocholic acid), and 5% urso-bile acids [almost all ursodeoxycholate

(3a,73-dihydroxy-53-cholan-24-oic)].

Bacterial metabolism of intestinal steroids may be closely tied to the enterohepatic cycle, which allows many endogenous steroids to gain entrance into the gut. The utilization of these steroids as a sub­ strate for biotransformation reactions may affect the circulating lev­ els of various steroids.

Most of the steroids in the enterohepatic cycle are conjugated to some compound which renders them more polar and water-soluble. The secondary bile acids, as well as the primary bile acids, are conjugated with either glycine or taurine during hepatic passage (89). When iso­ lated from feces, however, they are found as free acids due to the ex­ tensive microbial hydrolysis (7), which is considered the major initial reaction in the microbial transformation of bile salts (6). Species of secondary bile acids predominate in feces, of which deoxycholic

(3a,12a-dihydroxy-56-cholan-24-oic) acid accounts for over 50%. 8

Conjugation of a neutral steroid is determined by its chemical struc­ ture. For example, and steroids with 3cx-hydroxy-56-struc- tures are usually conjugated with glucuronic acid, while 3a-hydroxy-5a and 36-hydroxy-5a steroids generally form sulfates (89).

Numerous intestinal bacteria deconjugate taurine or glycine conju­ gates (63, 92). "Clostridia," which deconjugate both bile salts, in­ clude C^. perfringens, _C. bifermentans, and C^. sordelli. Clostridium septicum, however, could only hydrolyze the glycine conjugate (63).

Aries and Hill (7) found that the cholanylglycine hydrolase in several fecal bacteria, including Bacteroides sp., Bifidobacterium sp.,

Clostridium sp., and Streptococcus sp., all had pH optima below 6.0.

The enzyme also required highly anaerobic conditions or the presence of a reducing agent.

The mouse liver sulfoconjugation of cholic acid occurs predomi­ nantly at the 7a-hydroxyl group, and cholic acid 7-sulfate accounts for more than 40% of the total bile acids in mice. Humans and rats sulfo- conjugate bile acids predominantly at the 3-hydroxyl group. Man and the rat contain only trace amounts of sulfated bile acids (74) due to microbial deconjugation. Through desulfation, bile sulfates are con­ verted into less polar and thus more efficiently absorbed substrates.

Huijghebaert et al. (74) demonstrated that the microbial bile salt sulfatase in fecal cultures of man, rat, and mouse is limited to the hydrolysis of the 3-sulfate esters of bile acids. Sulfation at the C-7 or C-12 hydroxy1 group decreases or completely prevents further mi­ crobial transformation in the gut. 9

Several steroids can also be conjugated to glucuronides. Cortico­ steroids and vitamin D are two examples (8, 48, 89). The ability of fecal microbes to hydrolyze glucuronide conjugates is higher in Ameri­ cans than in populations at lower risk of developing colon cancer (102).

These findings suggest that dietary composition is a major factor in determining the bacterial ^-glucuronidase activity.

When the primary bile acids enter the GI tract, bacteria transform them into a large number of secondary metabolites. The major reactions inc3.ude the hydrolysis of bile salt conjugates, dehydroxylation of the hydroxy1 group at C-7, and dehydrogenases that oxidize the hydroxy1 groups at carbons 3, 7, or 12, as well as reducing the 3-keto group to both a- and S-hydroxyls (78). Secondary bile acids can be further metabolized by the liver, which complicates their classification as to origin (38).

The most important bile acid transformations, from a physiological perspective, are the conversion of cholic and chenodeoxycholic acids into deoxycholic and lithocholic acids, respectively, by 7a-dehydroxy- lation (64, 105). This reaction is thought to occur in or on the cytoplasmic membrane of the bacterial cell (133). In humans, 7a-dehy- droxylating bacteria do not become established in the gut before the ages of 12-18 months (38).

A two-step biochemical mechanism for 7a-dehydroxylation was pro­ posed by Samuelsson in 1960 (120). The first step involves the elimi­ nation of the 6g-hydrogen and the 7a-hydroxyl group. This results in a

structure, which is rapidly transhydrogenated at the 6a- and 78- 10

positions. The complete dehydroxylation reaction may involve the same

enzyme or an enzyme complex (148). Hydrogénation probably benefits

from a low redox potential where electron donors predominate; this fact

can account for the additional emphasis placed on the need for strict

anaerobic conditions (6). If the hydrogénation step is rapid and not a

rate-limiting step, the apparent need for strict anaerobiosis is less

obvious (4). This is supported by Masuda and Oda (103), who explored

the dehydroxylation process under different atmospheric conditions with washed cells of a Clostridium-like organism. When oxygen was allowed

to enter the system by diffusion, cholic acid was 7a-dehydroxylated to

deoxycholic acid. The resulting deoxycholate was further 12a-dehydro-

genated under aerobic conditions to 12-keto-lithocholate and the 3,12- diketo derivative of deoxycholic acid. However, when the reaction mix­

ture was aerobically incubated on a shaker, the dehydroxylation process was completely eliminated, and only oxidative dehydrogenation took

place. Under these conditions, cholate was converted to 7-keto-cholate, which was further oxidized to 7,12—diketo-cholate. The dehydroxylase enzyme in Eubacterium lentum also is active under aerobic conditions

(105). Archer et al. (4) investigated specific inhibitors of 7a-dehy- drogenation which would have no effect on dehydroxylation. Both 7a-de- hydrogenation and 7a-dehydroxylation were dramatically increased when washed cells of Clostridium bifermentans were incubated in the presence of molecular oxygen as compared with yields under oxygen-free condi­ tions (4). The stimulatory effect of oxygen on dehydrogenation agrees with the observations of Aries and Hill (6) that oxygen is necessary 11

for dehydrogenation of cholxc acid by cell-free extracts of several anaerobes. There have also been other reports of aerobic 7a-dehydroxy- lation. Hayakawa and Samuelsson (54) discovered 7-deoxy-bile acids among the degradation products when Corynebacterium simplex (not a mem­ ber of normal flora) was grown aerobically with sodium cholate as the sole carbon source. The mechanism for the reaction is probably differ­ ent from that employed by intestinal bacteria.

Several investigators have isolated bacteria from the intestine, cecum, and feces that can catalyze the 7a-dehydroxylation of cholic and chenodeoxycholic acids (6, 51, 64, 103, 105). Aries and Hill (6) iso­ lated this activity from Escherichia coli, Bacteroides sp., Clostridium welchii, and Streptococcus feacalis. Masuda et al. (104) noted that with resting cells of Eubacterium lentum a progressive enhancement of the 7a-dehydroxylase activity was caused by increasing the amount of flavin mononucleotide (FMN). Almost complete dehydroxylation of che­ late and chenodeoxycholate occurred when the concentration of FMN was equivalent to that of the substrate bile acids. Reduced flavins are thought to be required as electron donors in the reductive process of

7a-dehydroxylation. Hirano et al. (64) examined 13 bacteria from feces which could dehydroxylate 7-hydroxy-bile acids. Most of the fecal bac­ teria that could dehydroxylate bile acids could also catalyze 7a-dehy- drogenation, giving rise to the formation of 7-keto derivatives. This process, unlike that of dehydroxylation, was reversible. Archer et al.

(3) noted that Clostridium blfermentans (a bacterium capable of 7a-dehy- droxylation) also produced 7-keto-deoxycholic acid. Cell growth and 12

7-keto-deoxycholate production were largely complete by 12 hours, whereas dehydroxylation was strongest at about 18 hours of incubation.

Hirano et al. (64) contended that 7a-dehydroxylation took place early in incubation and ceased after 24 hours.

It has generally been observed that 7a-dehydroxylase is produced only in a medium containing 7a-hydroxy bile acids (103). The relative activity of dehydroxylating organisms differed with substrates; cheno- deoxycholate acid induced higher rates of 7a-dehydroxylation than did cholate (64). Many of the bacteria isolated from feces that were posi­ tive for 7a-dehydroxylation were also capable of deconjugating glycine and taurine conjugates.

Optimal 7a-dehydroxylation by fecal bacteria (from man) was ob­ served when the pH was 7.0, and was markedly inhibited at the low pH obtained when fermentation was stimulated in man by adding lactulose to the diet (27). Under these conditions, the proportion of deoxycholic acid in the bile also decreases significantly. Masuda and Oda (103) noted that when excessive glucose was added to the growth medium, it appreciably reduced the levels of 7a-dehydroxylase. This may be a pH effect because he reported pH optima of 6.5 to 7.0.

White et al. (148) described the regulation by pyridine nucleo­ tides of a 7a-dehydroxylase in Eubacterium sp. V.P.I. 12708. The pro­ duction of maximal 7-dehydroxylase activity required the addition of

NAD when extracts prepared from cholic acid-induced cells had been dialyzed under anaerobic conditions. NAD"*" is also required for the reduction of the intermediate proposed by Samuelsson (120). 13

Reduction of the intermediate (3a-hydroxy-56-6-cholan-24-oic) to

lithocholic acid significantly increased with the addition of NAD^, and also FADH^ and (added to mixtures containing NAD^) (148). NADH appeared to be inhibitory. White et al. (148) concluded that the

NAD^/NADH ratio may be important in the regulation of 7-dehydroxylase activity within the cell. The exact role of NAD"^ remains unclear, since no net reduction or metabolism of NAD"*" associated with 7-dehy- droxylation was observed. NAD"*" may be acting as an activator. The data also indicated that 7-dehydroxylase required a free C-24 carboxyl group because conjugates were not active substrates. An uninhibited

7a- or 76-hydroxyl group on the S-ring of the steroid nucleus was also essential.

Lipsky and Hylemon (87) reported that cholic acid induced 7a-dehy- droxylase and NADH:flavin oxidoreductase activities in cell extracts of

Eubacterium sp. The oxidoreductase had an absolute specificity for

NADH; this differs from the results of White et al. (148). A relative molecular weight of 260,000 was determined; this is much higher than similar enzymes reported from other workers. The exact physiological relationship of NADH:flavin oxidoreductase and 7a-dehydroxylase in this bacterium is not clear. However, 7a-dehydroxylation activity in an— aerobically dialyzed cell extracts of this organism were highly stimu­ lated by reduced flavins. Furthermore, 7a-dehydroxylase activity and oxidoreductase are both rapidly induced by sodium cholate. Hence,

NADH:flavin oxidoreductase may supply reduced free flavins for 7a-dehy- droxylase activity (87). Masuda et al. (105) observed that the 14

addition of flavin mononucleotide (FMN) to the growth medium signifi­

cantly increased the 7a-dehydroxylation of bile without an increase in

cell growth. The addition of FMN to the growth medium did not affect

the production of 7a-dehydroxylase. When FMN was included in a reac­

tion mixture of resting cells, enzyme activity was inhibited.

Some bacteria capable of 7a-dehydroxylating chenodeoxycholic acid

were also able to oxidize the hydroxy1 groups at positions C-3 and C-7

to keto functions (3, 51, 105). Gustafsson et al. (51) isolated the

following metabolites from pure cultures of fecal bacteria incubated

with chenodeoxycholate: 3-keto-7a-hydroxy-56-cholanoic acid, 3ct-

hydroxy-56-cholanoic acid, and 3-keto-5S-cholanoic acid. Taxo-

nomically, these active bacteria appeared to be members of the genus

Clostridium. However, these strains were unable to deconjugate bile

acids. In mixed cultures of fecal bacteria, extensive reduction of the

3-keto group to 36-hydroxy occurred. No organism was isolated for this

reductive activity, however.

Due to the potential harm that arises from 7-dehydroxylation, nu­

merous efforts to inhibit this reaction i^ vivo have been attempted

(91, 94, 95, 98). Epimerization (involves a hydroxysteroid dehydro­

genase, and involves the conversion of an a-hydroxyl group to the B-po-

sition) is a reaction which competes with 7a-dehydroxylation and can

benefit the host (95). Epimerization appears to go through a 7-keto

intermediate, whereas dehydroxylation by Eubacterium sp. and other

fecal bacteria appears to go through ^ unsaturated intermediates.

The major product of hydroxysteroid dehydrogenase is ursodeoxycholate 15

(the fourth most prevalent bile acid in human bile), the 7g-hydroxy

epimer of chenodeoxycholic acid (62). This acid is in current use as

an efficient cholesterol gallstone-dissolving agent.

Ursodeoxycholic acid is formed vivo from chenodeoxycholic acid,

largely via the intermediate 7-keto-lithocholic acid by a microbial

7a-hydroxysteroid dehydrogenase. The product of this reaction is sub­

sequently reduced by 7 6-hydroxysteroid dehydrogenase to yield ursode­

oxycholic acid which is very resistant to 7a-dehydroxy3ation (96). A

7a-hydroxycholanyl dehydrogenase has been observed in Clostridium welchii, Escherichia coli, and Bacteroides sp. The enzyme from C^. welchii was NADP^-dependent, while that of coli and Bacteroides sp.

were NAD^-dependent. Generally, the hydroxycholanyl dehydrogenase

produced by gram-negative bacteria are NAD^-dependent, while these

enzymes from gram-positive bacteria are NADP^-dependent, and enzymes from both gram-positive and gram-negative required oxygen (6). Conju­

gated bile acids are poor substrates for hydroxycholanyl dehydrogenase and 7-dehydroxylase enzymes; this supports the contention that the major initial reaction in the catabolism of bile salts by intestinal bacteria is hydrolysis of the amide bond to release free bile acids

(6).

Macdonald et al. (98) developed a method to rapidly quantitate

76-hydroxyl groups in bile samples (from humans and bears) by using purified 76-hydroxysteroid dehydrogenase from a preparation of

Clostridium absonum (98). Clostridium absonum formed ursocholate (from cholate), and ursodeoxycholate (from chenodeoxycholate), but could not 16

transform deoxycholate in whole-cell cultures (91, 95). This organism is considered to be a "low population" intestinal anaerobe, and is taxonomically very similar to C^. perfringens and C^. paraperf ring ens, neither of which can transform bile acids.

Epimerases appear to be dependent on either NAD or NADP^ (93).

Two different types of hydroxysteroid dehydrogenase have been docu­ mented. The form specific for a 7a-hydroxyl group uses chenodeoxy- cholic acid as a substrate. The other requires a 76-hydroxyl group and uses ursodeoxycholate as a substrate. Optimal pH for the 7c(-enzyme ranged from 9.5 to 11, the 76-enzyme's optimal range was from 9 to 10.

These may represent two entirely different enzjTnes (93). It was ob­ served that when C_. absonum was added to cultures of Eubacterium sp.

(active in 7a-dehydroxylating bile), 7a-dehydroxylation of chenodeoxy- cholate to lithocholate was totally prevented (90). Deoxycholate for­ mation was reduced by 50% when the substrate was cholate. In the pres­ ence of C^. absonum, as expected, the formation of 76-hydroxy and

7-keto-bile acids took precedence over the formation of 7a-dehydroxy- lated bile acids. When C^. absonum was added to mixed cultures of fecal bacteria, lithocholic acid production did not decrease, but the forma­ tion of ursodeoxycholic acid was produced at the expense of 7—keto acids. The 7a- and 76-epimerases in C^. absonum are induced by primary bile salts or deoxycholate and repressed by the end product, ursodeoxy­ cholate (95). Macdonald et al. (97) have partially purified these two enzymes. Hylemon and Sherrod (77) demonstrated that multiple forms of the enzyme 7a-hydroxysteroid dehydrogenase were present in Bacteroides fragilis. The enzymes differed with respect to their pyridine nucleo­ tide specificity, thermal stability, divalent metal cation require­ ments, and elution profiles from Sephadex G-200 columns. The NAD"*"- dependent enzyme had a molecular weight of approximately 80,000, and the NADP^-dependent enzyme 120,000. It was also noted that the

7a-hydroxy group was oxidized only after glucose was nearly depleted from the growth medium. Sherrod and Hylemon (124) purified a NAD"""— dependent 7c(—hydroxysteroid dehydrogenase from Bacteroides thetaiomicron which oxidized cholate to 7-keto-deoxycholate. This ability increased when cells were incubated in the presence of poten­ tial electron acceptors such as fumarate, menadione, or molecular oxy­ gen. This suggests that bile acids may serve as potential energy sources in the absence of fermentable carbohydrates (77, 124).

Eubacterium aerofaciens and Peptostreptococcus productus have only a 76-hydroxysteroid dehydrogenase that specifically oxidizes the

76-hydroxyl group in the cholane nucleus to a 7-keto group; NADP"*" is an electron acceptor. This enzyme activity was inhibited when excessive amounts of fermentable substrate were added to the medium. It was also observed that sulfhydryl inhibitors, such as p-chloromercuribenzoate and iodoacetate, greatly repressed the enzymic activity, suggesting the involvement of active sulfhydryl groups in dehydrogenation (62).

Clostridial species have both a 7a- and 76-enzymes, and can perform epimerization by themselves (62). Peptostreptococcus productus 18

and aerofaciens are certainly major species of the intestinal micro­

flora as opposed to the clostridial species, which are rare in the GI

tract. Peptostreptococcus productus and aerofaciens affected only

bile acids or bile salts which have a 7g-hydroxy group with the reduc­

tion to NADP"*" (not NAD^) and the formation of a 7-keto derivative as

the sole product.

Oral administration of chenodeoxycholic-oxazoline in low amounts

effectively blocks 7-dehydroxylation in rats. Macdonald et al. (96) studied the effect of bile acid oxazolines on 7-epimerization. The bile acid oxazolines appear to interfere with the epimerization of pri­ mary bile acids at the 7-position by inhibiting the growth of the or­

ganisms that participate in the reaction (with the notable exception of

_E. coli).

Hylemon and Stellwag (78) studied the conversion of cholate to

7-keto-deoxycholate and found that the peptidcglycan layer of the cell wall acts as a significant permeability barrier to cholate; greater rates of transformation in lysozyme-treated cells were obtained. Gram- negative bacteria may play a more important role in cholate transforma­ tion in the gut than do the gram-positive bacteria. Hylemon and

Stellwag (78) suggest that the transformation reaction occurs in or on the cytoplasmic membrane.

Eubacterium lentum 7a-dehydrogenates cholic and chenodeoxycholic acids to their respective 7-keto acids, while concurrently catalyzing the conversion of the 3a-hydroxyl group to 3—keto and 36-hydroxy deriv­ atives (61). An interesting relationship between products obtained and 19

the incubating atmosphere was observed. When compared, anaerobically growing and resting cells possessed essentially the same metabolism.

However, aeration of resting cells completely inhibited the dehydro- genation of the 7a-hydroxyl group of cholic and chenodeoxycholic acid, whereas dehydrogenation at positions C-3 or C-12, or both, were signifi­ cantly enhanced. Under anaerobic conditions, cholic and chenodeoxy­ cholic acids were 7a-dehydrogenated, yielding a large quantity of 7- keto-lithocholic acid. The 3—position was epimerized as well as oxi­ dized under aerobic conditions. There seems to exist the following incompatibility between the dehydrogenation at C-7 and at C-12 in cholic acid: extensive 7a-dehydrogenation under anaerobic incubation and only 12a—dehydrogenation in an aerated system. Eubacterium lentum seems to elaborate the whole series of Sa-, 7a-, and 12a-dehydrogenases, and the 3a—dehydrogenase activity of this organism is closely associ­ ated with the epimerization of the concerned hydroxyl group (61), As a rule, dehydrogenating keto formation is stimulated by aeration, presum­ ably because of the use of oxygen as an effective terminal electron acceptor, or by raising the Eh, but this organism appears to be an ex­ ception (61). The epimerizing conversion of the 3a-hydroxyl group by

lentum, which presumably involves a reductive process from a 3-keto to a 36-hydroxyl group, was suppressed by aerobic conditions. Beher et al. (9) have developed a rapid photometric method to determine free and esterified 3a-hydroxy fecal bile acids (9).

Another important transformation is the epimerization of the

3a-hydroxy group to a 3g-position. This reaction involves those bile 20

acids with 3ct-hydroxyl groups; cholate, chenodeoxycholate and deoxy- cholate= The reaction mechanism presumably precedes through a 3-keto intermediate (63, 92). Macdonald et al. (92) demonstrated that with C^. perfringens the 3-keto compound accumulated late in the growth curve when Eh values were above -120 mV. These findings were consistent with those of Hirano et al. (63), who showed that the 3—keto derivative could be obtained with aerated washed cells (of C^. perfringens), but not with whole cells maintained under anaerobic conditions. A broad pH optimum between 7.0 and 9.0 was noted for 3a-epimerization. The activ­ ity was significantly reduced at pH 6.0 and barely detectable at pH

5.0. These data strongly suggested that ketone production was inhib­ ited in media containing a fermentable substrate. Oxygen may serve as a terminal electron acceptor for the dehydrogenation reaction, or it may be involved in elevating the Eh or redox potential of the medium.

Macdonald et al. (92) offered an alternate explanation for the role of oxygen which involves the NADP'^/NADPH ratios in the cell. This ratio increased as the Eh of the medium increased. Interestingly, although the 3—keto compound was formed by different strains of C^. perfringens, not all of these strains produced the 36-epimers, suggesting the need for two distinct enzymes.

3B,12a-dihydroxy—58-cholanic acid has been recently isolated from rabbit feces and identified (29, 30). It was found to be a metabolite of deoxycholic acid, and could not be found in the bile nor could it be reabsorbed due to this transformation (30, 102). 21

There have been several reports of lithocholic acid metabolism by bacteria, but few which involve extensive degradation of the side-chain

(114). Owen and Bilton (114) demonstrated that Pseudomonas sp. strain

NCIB 10590 could metabolize the side-chains of these steroids. Metabo­ lism differed markedly with the type of incubating atmosphere used.

Under aerobic conditions, neutral metabolites were the major products, whereas under anaerobic conditions significant accumulation of acidic metabolites were observed. Leppik et al. (84) reported that optimal bile acid metabolism by a Pseudomonas sp. occurred when low aeration rates were used (84). The major metabolites detected were 3,17-keto-4 ene or 17,3-keto-l,4 diene compounds. This pseudomonad grows in the presence of 4% bile and has potential as an agent for industrial bio- conversion. Hayakawa et al. (53) suggested that the introduction of

A4 -3—keto structure for the microbial breakdown of bile an indispens­ able process for the shortening of the side-chain of cholic acid by

Streptomyces gelaticus.

Humans are unable to transform 6-muricholic acid (118), but rats and mice transform it into w-muricholic (3a,6a,76—trihydroxy-53-cholan-

24-oic) acid (40). Sacquet et al. (118) explored whether bacteria from the human digestive tracts were capable of initiating the transforma­ tion of B-muricholic acid (which does not exist in rats) in gnotobiotic rats by inoculating them with human feces. Their results support the finding that 6-muricholic acid-transforming bacteria are absent in the human microflora. However, 6-muricholic acid-transforming bacteria 22

must have been present in the human donors, but they did not thrive in

the rat.

Secondary bile acids derived from S-muricholic acid include

(D-muricholic and hyodeoxycholic (3a,6a-dehydroxy-56-cholanoic acid)

acid. Eyssen et al. (40) studied this conversion and found that it

proceeded in a two-step process. The first reaction was catalyzed by

an lentum strain and involved the oxidation of the 66-hydroxyl group

of 6-muricholic acid to a 6-keto group. The 6-keto group could be sub­ sequently reduced by a Fusobacterium sp. to a 5a-hydroxy group, yield­

ing w-muricholic acid. Thus, through the concerted action of two dif­ ferent bacteria an epimerization reaction occurred. This illustrates a

commonly observed phenomena; different end products are often detected in pure and in mixed cultures. Often, what appears as a stable end product in pure culture is further metabolized to other compounds in a mixed culture.

Not all biotransformations of bile acids are carried out by intes­ tinal bacteria. Sawada et al. (121) demonstrated that the soil fungus,

Fusarium equiseti M41, could transform lithocholic acid to ursodeoxy­ cholic acid very efficiently.

Steroid transformation is now being proposed as a stable taxonomic method to differentiate different strains of bacteria. Bokkenheuser et al. (15) support the use of this in the differentiation for Eubacterium sp. (15, 43).

The physiological effects exerted by transformed steroids are of­ ten difficult or impossible to separate from closely interrelated 23

factors, such as diet, intestinal motility, or enterohepatic circula­

tion. The effects of dietary fiber on steroid metabolism are mainly

physical and not of microbial origin. Diets high in fiber did not

change the total bile acids produced, but did affect their distribution

and qualitative pattern in the small intestine and colon.

Trans-fatty acids also affect the fecal excretion of bile acids;

this property is of interest in light of the postulated relationship

between cancer of the colon and dietary fat (139). It has been hypoth­

esized that a high fat diet not only changes the composition of bile

acids but also the activity of the gut microflora. The results of var­

ious studies strongly suggest that the gut microflora is modified by

dietary hydrogenated fats.

Cholesterol

The biotransformation of cholesterol (5-cholesten-36—ol) is proba­ bly the most quantitatively important in the gut (26). Cholesterol is characterized by a trans-junction of the A/B rings and a 5,6-double bond in the steroid nucleus. Many plant sterols such as B-sitosterol and campesterol share these structural characteristics, differing only in side-chain modifications. These steroids are often metabolized by the gut flora in the same manner as cholesterol. Roughly 400-1000 mg of cholesterol pass through the colon daily in a healthy, nonpregnant human on a "mixed western diet;" thus, a situation exists where bacte­ rial impact could be great.

Bacterial metabolism of cholesterol is largely of a reductive na­ ture. The major by-product of microbial cholesterol metabolism is 24

coprostanol (56-cholestan-3B-ol). A minor product, coprostanone (56-

cholestan-3-one) can also be detected. Depending upon the diet, more

than 50% of the total fecal sterols can be present in the form of

coprostanol (89). Germ-free animals excrete only unmodified cho­

lesterol and plant sterols (38). However, gut microbes in conventional

animals reduce (5B-hydrogenate) the 5,6-double bond to yield the less

well absorbed coprostanol. Many investigators believe that bacterial

metabolism of cholesterol can have a dramatic impact on the circulating

levels in plasma and its pool size (39, 109), even though absolute

proof of this is lacking.

The biohydrogenation of steroids is of particular interest be­

cause this process yields the 56-hydrogen saturated derivative, whereas

catalytic hydrogénation (in vitro) with hydrogen almost exclusively

yields the 5a-hydrogen saturated derivative. Two major pathways have

been proposed for the formation of coprostanol (10, 11, 89, 115). The

first involves the direct reduction of the 5,6-double bond. An alter­

nate pathway for the reduction of cholesterol involves a two-step mech­

anism. The initial step is the oxidation and isomerization of choles­

terol to 4-cholesten-3-one (10, 11, 89, 115). Bjorkhem and Gustafsson

(10) demonstrated that the oxidation of cholesterol to 4-cholesten-3-one

probably involved the removal of the 3a-hydrogen as a rate-limiting

step. This step is then followed by isomerization of the A^-double 4 bond to a A -double bond by a mechanism using partial transfer of a

hydrogen from the 46-position to the 6-position (10, 11). Mammalian

enzymes can also catalyze the conversion of 4-cholesten-3-one into 25

coprostanone (11) by the enzyme 3-oxo-A -steroid 5B-reductase. The

bacterial form of the enzyme requires NADH, whereas the mammalian form

requires NADPH. Both probably act by the same mechanism since both are

highly sensitive to sulfhydryl reagents (11). The second major step in

the biohydrogenation of cholesterol is the successive reduction to

coprostanone and coprostanol (41). Logically, the indirect hypothesis

for the formation of coprostanol implies a two-enzyme system (11, 115).

Surprisingly, the direct reduction of cholesterol using no intermediate

is thought to be a relatively minor pathway (115).

The reduction of cholesterol to coprostanol occurs predominantly

in the ceca in cholesterol-fed rats (79), and a number of intestinal

bacteria obtained from cholesterol-fed rats can metabolize cholesterol in vitro. Similar bacteria are also in humans (119). Significant amounts of cholesterol were metabolized when mice were fed a diet high in cholesterol. When, however, the intestinal flora of mice was modi­ fied by adding sulfasuxidine or streptomycin to the diet, practically all the cholesterol fed was recovered in the excreta and carcass. Bac­

terial metabolism of cholesterol may represent a natural way to dispose of cholesterol ûi vivo (144). Similar conversions occur when plant sterols pass through the digestive tracts of animals.

Many intestinal bacteria possess one or the other, but not both, of the enzymes involved in the reduction of cholesterol (18). Recent­ ly, Eyssen and Paramentier (41, 115) isolated a Eubacterium sp. from the cecum that possessed both, and is therefore quite unique. Bjorkhem et al. (11) suggested that this Eubacterium sp. has a 3—oxo—A^-steroid 26

56-reductase because it used a two-step mechanism to reduce the 5,6-

double bond in cholesterol. This activity depended upon the addition of cholesterol or similar steroids, such as campesterol, 6-sitosterol, and stigmasterol to the growth medium. Cholesterol was thought to be acting as a hydrogen acceptor. Certain structural requirements had to be met for hydrogénation to occur, however. When the 3-hydroxyl group was absent or in the a-position, or was substituted as in cholesteryl esters, no reduction of the double bond was observed (42). This

Eubacterium sp. could also reduce the 3-keto groups of cholestanone

(5a-cholestan-3-one) and coprostanone (56-cholestan-3-one) to the cor­ responding 56-hydroxyl derivative. When gnotobiotic rats which had been inoculated with this strain of Eubacterium sp. were cecectomized, this hydrogénation ceased within 2 days.

The biohydrogenation of steroids was suppressed by diets contain­ ing high levels of lactose (41). Lactose-supplemented diets dramati­ cally increased cholesterol absorption in the rat when measured as the difference between cholesterol intake and excretion, and by measure­ ments of elevated liver—cholesterol concentrations. This effect may be due to pH, since biohydrogenation of cholesterol is inhibited at low pH. Wells et al. (147) offered an alternative explanation, which re­ lates to bile production. Bile excretion in rats that had been main­ tained on a lactose, cholesterol—free diet was significantly higher than in sucrose-fed controls. Lactose may inhibit bile acid-metabo­ lizing bacteria, which thus allows for higher reabsorption of bile acids. The net result would be an increased circulating bile acid pool 27

size. Higher bile acid levels would also directly favor the absorption

of cholesterol from the intestine (147).

Tanaka et al. (143) found that fecal bacteria from rats fed soy­

bean protein diets demonstrated an increased ability to produce copro- stanol when compared to rats receiving casein as a protein source. The observed increase in biohydrogenation was attributed to soybean pro­

tein's ability to maintain higher levels of coprostanol-generating bac­ teria or bacteria with higher metabolic rats. Rats fed a liquid diet containing no fiber, and virtually no fat, demonstrated that the neu­ tral steroid fraction was virtually unmetabolized by the gut flora (26).

Kobashi et al. (81) examined the effects of oral administration of

Clostridium butyricum on the catabolism and excretion of ^^C-labelled cholesterol in rats. Bacterial treatments accelerated the oxidative cleavage of the cholesterol side-chain in normal rats (using 14C- labelled cholesterol). Clostridium butyricum was also able to sig­ nificantly affect plasma cholesterol levels in rats fed high cholester­ ol diets. Rats treated with _C. butyricum had significantly lower plas­ ma cholesterol levels compared to control rats on the same diet. Mice and rats on an ordinary diet (normal cholesterol) did not show a reduc­ tion in plasma cholesterol levels. Also, these bacteria were unable to cleave the side-chain of cholesterol vitro, suggesting that the ef­ fect is of an indirect nature.

Although fungi and various soil microorganisms cleave the side- chain of cholesterol, there is little evidence that bacteria isolated from humans share this ability (89, 101). 28

Other biologically important steroids

Many other biologically active steroids undergo microbial trans­

formations in the intestinal tract. Among the most important are the

steroid hormones.

Enzymes originating from both the intestinal mucosa and from the

intestinal flora are important in the metabolic transformation of

estrogens during absorption (28). Such gut metabolism plays a signifi­

cant role in reducing the bioavailability of these orally administered

compounds. The therapeutic use of drugs can affect this relationship

and alter absorption characteristics and the nature of the flora. In

this regard, the use of antibiotics is particularly important. Both

natural and synthetic estrogens and progestagens are extensively ex­

creted in bile, principally as glucuronide conjugates. Normally, these

conjugates are hydrolyzed in the gut and subsequently can undergo

enterohepatic circulation (1, 86). The gut microflora is a major

source of hydrolytic enzymes, especially B-glucuronidase, and treatment

with antibiotics may interfere with the hydrolysis of steroid conju­

gates and hence reduce or eliminate steroid reabsorption from the gut.

Estrone sulfate is a naturally occurring estrogen which is being

widely used for the relief of menopausal and post-menopausal symptoms

associated with estrogen deficiency. The ^ vivo importance of bacte­

rial hydrolysis of estrogen conjugates is not known, especially since

unequivocal evidence for the intact absorption of estrone sulfate

(without the aid of bacterial enzymes) from the proximal and distal

small intestine has been obtained (126). In contrast, almost 29

negligible absorption of estrone sulfate from the cecum of germ-free

rats indicates that in this region of the tract, hydrolysis of the

sulfate by bacterial enzymes is essential for quantitative absorption

(126).

Contraceptive failures often occur after the administration of

ampicillin to women (2). Most contraceptive drugs are estrogen-based,

and as such represent an important potential for bacterial steroid me­

tabolism to influence physiology. Reductive pathways dominate the nor­

mal estrogen metabolism by bacteria in the gut. It has been observed

that antibiotics reduce the mass of intestinal microflora, causing a

reduction in the transformation of estrone to estradiol and an increase

in estrone/estradiol and estrone + estradiol/estriol ratios in feces

(2). Simultaneously, the hydrolysis of the biliary estrogen conjugates

by bacterial g-glucuronidase is reduced causing a shift in the excre­

tion of estrogen from urine to feces, because conjugated estrogens are

poorly absorbed from the intestine. These results thus strongly sup­

port the concept that intestinal bacteria play a significant role in

overall metabolism of estrogens (2).

Bokkenheuser et al. (13) demonstrated biotransformations of corti­

costeroids which involved 21-dehydroxylation. In humans, the bacterial

metabolism of corticosteroids appears to be restricted to those corti—

coids which undergo biliary excretion. It is believed that in the

human gut 11-deoxycorticosterone is metabolized to corticosterone and

11-dehydrocorticosterone by a bacterial-mediated 21-dehydroxylation.

The presence of 21-dehydroxylated metabolites of corticoids in human 30

urine can be explained by postulating that after hepatic metabolism, conjugation, and biliary excretion, the biliary steroids undergo decon- jugation and bacterial 21-dehydroxylation in the intestine, followed by reabsorption, and further metabolism such as reduction of the 20-ketone, conjugation and finally urinary excretion. Bokkenheuser et al. (13) state that aldosterone may also be affected in this manner. It is note­ worthy that cysteine hydrochloride was a component of most media that supported the growth of 21-dehydroxylating organisms.

20-Hydroxylated metabolites of corticosteroids are found in normal human urine. Originally, they were thought to be formed exclusively by hepatic reduction of 20-keto structures during the enterohepatic circu­ lation. Winter et al. (154), however, isolated a strain of

Bifidobacterium adolescentis from human and rat feces which has 20- hydroxysteroid dehydrogenase activity (154).

Some investigators have demonstrated that the side-chains of corti- coids can be removed by intestinal microorganisms (12). Bokkenheuser et al. (12) isolated a Clostridium sp. from human feces that has a con­ stitutive desmolase that cleaves the side-chain of Cortisol (a C-21 steroid) to form llB-hydroxy-4-androstene—3,17 dione (a C-19 steroid).

A similar reaction requires desmolase (of animal origin) to convert

C-21 steroids into the C-19 sex hormones in mammals.

Fecal bacteria from rats and humans convert 16a-hydroxypro- gesterone to 17a-progesterone by dehydroxylating at the 16-position

(16). The incidence of these bacteria in humans is low and variable; however, in the rat, these bacteria are typically present at about 10^ 31

per gram of feces. Bacterial involvement in 16o(-dehydroxylation reac­ tions is supported by the fact that the 17a-steroid is lacking in the feces of germ-free rats (16). The reaction mechanism is thought to be similar to formation of urinary pregnanolones by the body. Calvin and

Lieberman (20) suggested that the conversion of 15a-hydroxyprogesterone to 17a-pregnanolone takes place via an intermediate C-21 steroid.

The latter steroid is metabolized both vitro and i£i vivo to 17a- pregnanolone (20). Bokkenheuser et al. (16) believe that the same mechanism occurs with isolates obtained from fecal material (rat), but they could not detect the C-21 intermediate, possibly due to its transient nature.

Glass et al. (46) demonstrated that the 16a-dehydroxylation of

16a-hydroxyprogesterone by Eubacterium sp. 144 resulted in the produc­ tion of 17a-progesterone (46). This reaction proceeded through a A^^ progesterone intermediate which initially accumulated in the culture medium and then was reduced to 17a-progesterone. The A^^- progesterone reacted chemically with L-cysteine to form a water—soluble derivative.

The appearance of 17a-progesterone was dependent on the method of media preparation — whether cysteine was added before or after autoclaving the media. When cysteine was added separately after sterilization, it reacted chemically with A^^-progesterone to form a water-soluble deriv­ ative. The net effect was to remove this intermediate from further metabolism to 17a-progesterone.

With the rapid increase in the number of synthetic steroids now available and administered, a number of investigators determined if 32

these would serve as suitable substrates for bacterial transformation.

Bokkenheuser et al. (14) isolated three intestinal strains of

Clostridium that possessed reductases which were active upon the A-ring of natural steroids and synthetic progestins. Substrates metabolized include: progesterone, Cortisol, deoxycortisterone, and two synthetic steroids - norgestrel and dimethisterone. Activities in mixed and pure culture were found to differ markedly. Three isolates of Clostridium were obtained and identified as C^. paraputrificum, C^. innocuum, and

Clostridium sp. J-1. Both mixed cultures and C^. paraputrif icum reduced the A-ring of natural steroids to 3a,5 6-compounds. In contrast, pure cultures of jC. innocuum, and Clostridium sp. J-1 converted the sub­ strates to 33,56- and 36,5a-compounds, respectively. When synthetic progestins were used, the mixed culture reduced them to a mixture of

3a,56- and 36,56-derivatives; furthermore, the natural steroids were metabolized at much faster rates than the synthetic steroids. The slower metabolism of synthetic steroids and the observation of minimal loss of biological activity of synthetic progestins in the intestinal environment is attributed to the presence of an ethynyl group which enhances the resistance of ring-A to reducing enzymes (14).

Many techniques have emerged to study biotransformations of steroids by bacteria. Most involve preliminary separations by either gas chromatography, thin-layer chromatography, or high-pressure liquid chromatography. Recently, it has been discovered that steroid trans­ formation can be monitored by a powder diffraction technique for those steroids which form crystals (130). Subsamples from a fermenter can be 33

analyzed for crystal growth, thus providing information on the extent of the transformation.

Vitamin D

Recently it was discovered that, in the presence of rumen bacte­ ria, vitamin D, a 9,10 seco-steroid, undergoes C-10(19) oxidation (110,

127, 128, 129). Napoli et al. demonstrated that mixed populations of rumen bacteria produce 5(E)-19-nor-10-keto-vitamin Dg (lO-keto-Ds),

5(E)-19-nor-10-keto-vitamin Dg (lO-keto-Da), and 5(E)-19-nor-10-keto-

25-hydroxyvitamin D3 (10-keto-25-0HD3) when incubated with vitamin D3, vitamin D2, and 25-hydroxyvitamin D3, respectively (110). This also represents the first documentation of bacteria-metabolizing vitamin D

(a eucaryotic vitamin). In addition, the bacterial metabolism of vita­ min D to a more polar compound parallels its metabolism in higher ani­ mals.

Although most of the work by Sommerfeldt and colleagues (110, 127,

128, 129) was done by using mixed cultures that originated from the bovine rumen, it is likely that bacteria isolated from the cecum and colon of monogastric animals can also metabolize vitamin D — since rumen bacteria are similar to many bacteria in the cecum and colon of monogastric animals (5). Some support for this contention is provided by experiments which used isolates of C^. innocuum and C^. paraputrificum from human feces (documented as bile acid transformer); these bacteria were also able to produce lO-keto-Dg (unpublished results). The discovery of new microbial pathways for vitamin D metabolism is 34

exciting because these bacterial products could have biological activity in the host animal.

In vitro bioassays have shown that lO-keto-Dg is significantly more active at stimulating bone calcium resorption than the parent vi­ tamin D3 (134). However, results obtained from in vivo biological as­ says (everted gut sac technique) with lO-keto-Dg demonstrate some re­ sidual biological activity, but 40- to 80-fold less than that obtained

with vitamin D3. Similar vitro and vivo differences in biologi­ cal activity have been shown with 10-keto-25-0H-D3. The in vivo data suggest that the C-10(19) oxidation is a deactivation process for vita­ min D. The discrepancy between i^ vivo and vitro results probably involves differences in half-life or turnover times. More research relating to the biological activity of these metabolites is needed to clarify their physiological importance.

Pharmacologically, this new C-10(19) oxidative pathway for vitamin

D may have particular relevance. With the current trend of supplement­ ing foods with excessive amounts of vitamin D, more is reaching the intestine than is physiologically required or safe (67). This condi­ tion would provide large amounts of substrate for bacteria capable of biotransforming vitamin D. It may be possible that, in cases of vita­ min D excess, the natural flora reduce the risks of toxicity by oxidiz­ ing vitamin D to its less active forms. Related findings from vitro experiments, which incubated bacteria in the presence of high levels of vitamin D (which may occur i^ vivo under large doses of vitamin D in­ take), detected several other uncharacterized bacterial metabolites of 35

vitamin D. Unfortunately, the biological role of these uncharacterized compounds is unknown, but they should be considered seriously because

1,25—dihydroxyvitamin D3 is active in picomolar concentrations.

The compound lO-keto-ZS-OH-Dg has also been produced in cultures of kidney cells (125) and phagocytic cells (25, 85). C-10(19) oxida­ tion of the vitamin D nucleus does not, therefore, appear to be limited to procaryotic organisms. It is not known at this time if production of the 10-keto compounds in these eucaryotic systems is similar to the bacterial system.

The assumption that the intestinal flora cannot significantly af­ fect the host's physiology by the transformation of intestinal steroids is rapidly losing support. Frequently, the flora and its host are treated as mutually exclusive subjects, with no acknowledgment of any interaction. However, the flora and its respective host evolved to­ gether in a host-symbiont relationship and, as such, should not be treated as disjointed units.

Vitamin D

Importance and implications

This part of the literature review is not meant to be a complete review of the vitamin D area. It is hoped, however, that this summary of selected citations will provide the reader, especially those not intimately involved in vitamin D research, a working knowledge and an appreciation of this unique compound.

Along with the two peptide hormones, calcitonin and parathyroid hormone (PTH), vitamin D is one of the most important biological 36

regulators of calcium and phosphorus metabolism (32, 35, 82, 83, 111).

This field of research has undergone a staggering metamorphism in the last two decades (33). Vitamin D should not be regarded as a single compound with a single biological activity, but as a whole family of compounds. Approximately 27 known metabolites of vitamin D are now documented, but the biological activities of only a few of these are known.

The designation of cholecalciferol as vitamin D is actually a mis­ nomer, since it is now accepted that vitamin D actually functions as a steroid hormone (32, 35, 111). In this regard, 1,25-(OH)2D, the active metabolite of vitamin D, fits the classical definition of a steroid hormone — that is a compound produced in one organ that exerts its biological activity in another part of the body. Vitamin D should only be considered as a vitamin under conditions of insufficient light since only then is vitamin D required in the diet.

The recognition of vitamin D as the antirachitic factor in foods is attributed to Sir Edward Mellanby (106) as early as 1919. The dis­ tribution of rickets over the world and its correlation with the inten­ sity of sunlight was recognized by Hess (57) in 1929. These two obser­ vations, that rickets was induced by a lack of sunlight or that rickets was induced by a lack or deficiency of a nutritional factor, as demon­ strated by Mellanby in 1921, appeared to be contradictory when first examined together. Mellanby (107) wrote, "The action of fats in rickets is due to a vitamin or accessory food factor which they con­ tain, probably identical with the fat-soluble vitamin." Furthermore, 37

to prove his hypothesis he established that cod-liver oil was an excel­ lent antirachitic agent. Huldschinsky (75) demonstrated that it was the ultraviolet (U.V.) spectrum in sunlight that was effective in increasing the degree of calcification in the epiphysis of rachitic infants. No connection was made between the curative powers of U.V. light and the existence of a dietary factor in cod-liver oil until the work of Goldblatt and Soames (47) and Steenbock and Black (131). Goldblatt and Soames (47) found that irradiated livers from rachitic rats became antirachitic when fed to other rachitic rats. Steenbock et al. (132) found the same effect in certain other foods which had been irradiated with U.V. light. What was unknown to these investigators was that two major forms of vitamin D exist — vitamin D2 and D3. The metabolism of vitamin D2 and vitamin D3 were once thought to be identical, except in the chick (22); however, many investigators are finding that many spe­ cies discriminate between these two major forms of vitamin D (72, 76,

129).

In 1932, Windaus et al. (152) identified the structure of vitamin

D2 and showed that it was a seco-steroid. In 1936, the structure of vitamin D3 was chemically characterized from studies involving the U.V. irradiation of 7-dehydrocholesterol (153). Natural vitamin D from cod-liver oil was determined to be identical to vitamin D3, and

7-dehydrocholesterol, not ergosterol, is the true provitamin of the

"natural vitamin." Ergosterol, found in plants and yeast, is the provitamin for vitamin D2 (111). 38

Metabolism

Vitamin D3 is produced by a photochemical activation of 7-hydro-

cholesterol in the skin of most higher animals (65, 66). Exogenous

vitamin D also enters the body through the diet as either vitamin D2

(associated with plants and yeast) or vitamin D3 (associated with ani­

mal products). Absorption of vitamin D occurs through the lymphatics

and requires bile salts (36, 50) similar to other intestinal steroids.

Both the vitamin D3 synthesized in the skin and the vitamin D ab­

sorbed by the gut are transported via a plasma transport protein or

6-lipoprotein (65) to the liver where vitamin D is rapidly cleared from

the plasma. In the liver, vitamin D is hydroxylated at the 25-posi-

tion; this step represents the major initial reaction in the activation

process. 25-Hydroxylation is required for all subsequent metabolism of

vitamin D (34). 25-Hydroxyvitamin D (25-OHD) is also the major circu­

lating form of the vitamin. Normal levels in humans range from 20 to

40 ng/ml (52, 71). The mechanism for the 25-hydroxylation of vitamin D is thought to involve a cytochrome P-450, molecular oxygen, and NADPH

(34, 44, 99, 155).

Although reports of 25-hydroxylation of vitamin D in the intestine

and kidney exist, the liver is considered to be the major organ in­ volved. After hydroxylation, the product 25-OHD is rapidly removed from the liver (bound to a transport protein) and is carried to the kidney where it subsequently undergoes further hydroxylation.

The most important reaction occurring in the kidney (with regard

to vitamin D) is the la-hydroxylation of 25-hydroxyvitamin D3 to 39

produce la,25-dihydroxyvitainin Dg [l,25-(OH)2D] (35, 111). The discov­ ery of la,25-dihydroxyvitamin D (the most biologically active form of the vitamin known) is contested between DeLuca and Norman (31, 111).

The hydroxylase system in the kidney is similar to that found in the liver, as both are mixed function oxidases that require a cyto­ chrome P-450 (44, 45). The product 1,25-(0H)2D is transported to the target tissues in bone, intestine, and kidney. In regard to bone and intestinal tissue, the kidney can be considered an endocrine gland.

1,25-Dihydroxyvitamin D can also be produced in the placenta (145, 146) and bone (73). However, the in vivo significance of non-renal

1,25-(OH)2D production is unknown.

Although considerations of the regulation of vitamin D metabolism are beyond the scope of this review, it should be noted that the regu­ lation and metabolism of 1,25-(0H)2D synthesis is exceedingly complex.

1,25-Dihydroxyvitamin D is closely regulated by plasma calcium and phosphorus levels and the peptide hormone PTH. Interactions have also been documented with estrogen, calcitonin, and growth hormone (111).

1,25-Dihydroxyvitamin D is thought to enter the cell and exert its biological activity by mechanisms traditionally associated with steroid hormones. The localization of 1,25-(0H)2D3 in target tissues is one of the most active areas of investigation in vitamin D metabolism (35).

Biochemical evidence, using subcellular fractionation and [^H]-labelled

1,25-(0H)2D3, suggested that the 1,25-(0H)2D localized in the nuclear- chromatin fraction before initiation of a target organ response (23).

Specific nuclear localization of 1,25-(0H)2D3 has been reported for 40

intestinal villus cells (156), osteoblasts (136), distal renal tubular cells of the kidney (137), islet cells of the pancreas (24), brain

(135), skin (136), and pituitary gland (136).

The appearance of 1,25-(0H)2D3 in these different tissues suggests that 1,25-(0H)2D may have multiple functions. The numerous interac­ tions of vitamin D with hormones that traditionally have not been asso­ ciated with calcium homeostasis confirm this.

Vitamin D can undergo numerous other reactions to produce some 27 different metabolites. Many of these compounds probably represent in­ termediates in the degradation of vitamin D. The biological activities of many of these compounds have yet to be determined.

Quantitatively, the second most important reaction that occurs in the kidney concerning vitamin D is the production of 24R,25-dihydroxy- vitamin D [24,25-(OH)2D] by the action of 24R-hydroxylase (35, 80,

141). A great deal of controversy surrounds this compound since many investigators believe it has an important biological role, while others think it has none. Norman supports the contention that 24,25-(OH)2D and 1,25-(0H)2D are both required for normal egg hatchability (112).

The 24R-hydroxylase is also considered to be a mixed-function monooxygenase (100) like the la-hydroxylase and the 25-hydroxylase enzymes. Production of 1,25-(OH)2D and of 24,25-(OH)2D are diametrically opposed. In animals with normal concentrations of plasma calcium, 24-hydroxylation predominates, whereas under conditions of calcium demand (low plasma calcium), the la-hydroxylation becomes the major reaction. 41

Another dihydroxylated metabolite, 25,26-dihydroxyvitamin D

[25,26-(0H)2D]> was discovered in pigs by Suda et al. (138). It has also been noted in other species (71). The exact function is not known

(111).

One of the most recent pathways of vitamin D metabolism to be dis­ covered involves the production of 25-0HD3-26,23 lactone (lactone)

(149). Lactone has been found in the plasma of vitamin Dg-toxic pigs and cows (59, 70), rats (70, 123), and man (88). Although 25,26-(0H)2D was originally thought to be the precursor to lactone, later work clearly demonstrated that 23S,25-(0H)2D3 was the most efficient precursor (142). 23,25,26-Trihydroxyvitamin D represents the next intermediate in sequence of producing this unique compound (111).

Horst (69) has demonstrated that the lactone binds plasma proteins with higher affinity than any other known vitamin D compound.

There are several other minor metabolites of vitamin D (most appear in the case of excessive vitamin D intake); however, their pres­ ence should not be considered unimportant since the biological activity of many of these compounds is unknown. Some of the other minor metabolites are la,25S,26-(0H)3D3 (117), 24-oxo-25-0HD3 (140),

23,24,25-(0H)3D3 (150), 24,25,26-(0H)3D3 (150), 24-OHD3 (150), and

23-dehydro-25-0HD3 (150). The biological activity of these compounds is also unknown.

Biological activity

It is generally accepted that 1,25-(OH)2D represents the most bio­ logically active form of vitamin D. It also represents the most 42

studied form. The classical biological functions of 1,25-(OH)2D are:

(a) In conjunction with PTE, it stimulates the mobilization of bone calcium and phosphorus. Parathyroid hormone induces a phosphaturia, so the net effect is an increased serum calcium concentration. (b) Al­ though still controversial, 125-(0H)2D is thought to be involved in the conservation of calcium by stimulating renal resorption. (c) 1,25-Di- hydroxyvitamin D can also stimulate the transport of calcium and phos­ phorus from the intestine. (d) 1,25-Dihydroxyvitamin D is thought to be involved in normal egg shell development in birds (112).

24,25-Dihydroxyvitamin D has also been extensively studied, but its biological activities are still the subject of controversy.

24,25-Dihydroxyvitamin D may be involved in the following, but more proof is required: (a) it may be required for normal embryo develop­ ment in birds (54, 55, 111); (b) it also may play a role in bone forma­ tion (113) and maintenance (17); (c) 24,25-(0H)2D may suppress PTH levels and thereby influence 1,25-(0H)2D3 metabolism (56).

Enterohepatic cycling of vitamin D

The C-10(19) pathway for production of lO-keto-Dg may have an im­ portant physiological role, especially when one considers that mi­ crobial metabolism of vitamin D may be closely tied to enterohepatic recycling. As much as 30 to 40% of the major metabolites of vitamin D generated by host enzyme systems have been documented to travel in an enterohepatic cycle (8, 49, 151). When [^H]-labelled 25-OHD was in­ jected into humans, approximately one-third of the label appeared in the duodenum within 24 hours - via the bile (8, 48). However, fecal 43

excretion could only account for about 3% of the total radioactivity during the same time period, suggesting that the majority of the label was reabsorbed from the small intestine. This implies that the entero- hepatic cycle is highly conservative and that the gut contains a sig­ nificant vitamin D pool. Similar results have been reported using

1,25-(0H)2D3 (49, 152).

Vitamin D seems to mimic bile with regard to its enterohepatic recycling, possibly due to structural similarities (48). Like bile, 80 to 90% of vitamin D and its metabolites are resorbed in the jejunum.

With the evidence for bacterial transformation of steroids in the gut, substantial support also exists for the biotransformation and absorp­ tions of vitamin D. It has been shown that lO-keto-Dg can be detected in both the plasma and rumen of a cow receiving large internal doses of vitamin D (127). The enterohepatic cycle may act as a regulator which controls the circulating levels of the various metabolites, as noted by the rapid removal of 1,25-(0H)2B3 shortly after injection (49, 152).

Naturally, any gastrointestinal disease (8) or bacteria which can in­ terrupt the recycling of 25-OHD should be considered as a factor which would limit the availability of vitamin D.

Requirements and toxicity

The unique nature of vitamin D that allows both endogenous synthe­ sis and dietary supplementation to occur simultaneously can create a potentially harmful situation. Substantial evidence has accumulated to suggest that there is a narrow tolerance between the amounts needed and that which may produce deleterious results (67). 44

Most animals, including humans, that are not totally confined re­

ceive adequate sunlight which will provide enough vitamin D to meet

their requirements. There is also an increasing tendency to supplement

both animal and human foods with higher levels of vitamin D. This, in

conjunction with the "fad" of taking mega doses of vitamins, can result

in dramatic consequences. Many investigators believe that intakes of

vitamin D often far exceed the official recommendations and greatly

exceed the physiological requirement of the body (122). Emphasis

should be placed on how best to protect against rickets without devel­

oping associated risks of accentuating or causing cardiovascular and

renal disease (122). Practices that result in supplying excessive

amounts of vitamin D (either directly from supplemented foods for human

consumption or foods derived from animals that consume supplemented feeds) can result in vitamin D toxicity to a proportion of the popula­

tion. These practices require reexamination (44, 67, 122).

Like the other fat-soluble vitamins, vitamin D can accumulate in fatty tissues; it is then slowly released back into the blood stream regardless of the physiological needs of the animal. Excessive levels of vitamin D are toxic. The major symptom associated with vitamin D toxicity is primarily soft tissue calcification.

In humans, the actual requirement for vitamin D is not known, but based on information that a supplement of 2.5 yg of vitamin D3 (0.025 pg of vitamin D3 equals 1 international unit [I.U.]) prevents rickets and ensures adequate calcium absorption (67), the current recommended levels in humans are 400 I.U. per day. The National Academy of 45

Sciences has established the National Research Council which prescribes the recommended dietary nutrient requirements (including vitamin lev­ els) for all the major forms of livestock and pets. The amount of vi­ tamin D recommended varies with species, age, type of housing, and phase of reproduction. 46

SECTION I.

FACTORS AFFECTING THE PRODUCTION OF 5(E) 19-NOR-lO-KETO-VITAMIN D3

BY RUMEN BACTERIA 47

INTRODUCTION

Bacteria in the gastrointestinal (GI) tract are known to play an active role in altering the structure and function of many endogenous and dietary steroids (11, 12, 19). Previous research with bacterial steroid transformations has focused on the three major classes of steroids passing through the GI tract; these include cholesterol (8,

13, 31, 32), steroid hormones (6, 7, 33), and the bile acids (1, 19,

21, 22). Similarly, groups in Ames, lA, and Dallas, TX, collaborated in demonstrating that cholecalciferol (vitamin D3), a 9,10 seco-steroid, undergoes C-10(19) oxidation when incubated in vitro with strained bo­ vine rumen fluid (24, 29, 30). They isolated and identified the com­ pounds 5(E)-10-keto-19-nor-vitamin D3 (lO-keto-Dg), and 5(E)-lO-keto-

19-nor-25-hydroxyvitamin D3 (10-keto-25-0H-D3) from incubations with vitamin D3 and 25-hydroxyvitamin D3, respectively (24). In this sys­ tem, up to 75% of the [^H]-vitamin D3 was converted into polar metabo­ lites; 10-keto-[^H]-D3 was the major product (24). These compounds could not be detected in incubations with sterile rumen fluid, thus implying that bacterial metabolism was involved in their production.

Although the vivo significance of 10—keto-D3 production has yet to be determined, the discovery of its vitro production by microbes may have special relevance because several vitamin D sterols, including vitamin D3, 25-hydroxyvitamin D3, and 1,25-dihydroxyvitamin D3 [the latter two are major products of host metabolism of vitamin D3] (25), circulate in the enterohepatic cycle of mammals and therefore may be 48

available for oxidation by gut bacteria (3, 16). Also, the recent finding that lO-keto-ZS-OH-Dg is produced by kidney cell lines (28) and phagocytic cells (10, 17) established that eucaryotic systems can affect this oxidative pathway as well.

Little is known about the pathway leading to the production of lO-keto-Dg by bacteria. However, given the potential vivo impor­ tance of this oxidative pathway, the present vitro studies were con­ ducted to identify conditions required for the optimum production of lO-keto-Dg, and to characterize the processes involved. This report establishes that oxygen, pH, and reducing agents are of primary impor­ tance in the expression of the C-10(19) oxidative pathway for vitamin

D. 49

MATERIALS AND METHODS

General

High-pressure liquid chromatography (HPLC) was performed with a

Waters Model LC-204 chromatograph fitted with a Model 6000-A pump and

Model 440 fixed-wavelength (313 nm) detector (Waters Associates,

Milford, MA). HPLC columns (0.46 x 25 cm) were purchased from DuPont

Instruments (Wilmington, DE), unless otherwise noted. Normal phase refers to microparticulate silica gel columns. Reverse phase refers to microparticulate silica gel columns derivatized with octa decylsilane

(ODS). Spectral-grade solvents (Burdick and Jackson, Inc., Muskegon,

MI) were used in all chromatography systems. The U.V. spectra were obtained in ethanol by using a recording spectrophotometer (Model 25,

Beckman Instruments Co., Lincolnwood, IL). A molar extinction coeffi­ cient of 27,000 was used for lO-keto-Dg.

Steroids

Vitamin D3 was purchased from Sigma Chemical Co. (St. Louis, MO).

[3a-^H]-Vitamin D3 (1.2 Ci/mmole) was prepared by the sodium boro-[^H]- hydride reduction of the a-tricarboxyliron complex of 3—keto-vitamin D3

(4). Radiochemical purity was confirmed by normal-phase HPLC by using hexane/isopropanol (99/1) at a flow rate of 2 ml/min. The lO-keto-Dg used for standards was a gift from Hoffmann-La Roche, Inc. (Nutley,

NJ). All lO-keto—[^H]-vitamin D3 (1.2 Ci/mmole) used as internal stan­ dards was prepared from [^H]-vitamin D3, as described by Napoli et al.

(24). 50

Enzymes

Superoxide dismutase (SOD) and catalase were purchased from Sigma

Chemical Co.; both enzymes originated from bovine liver extracts. The specific activity of the SOD and catalase preparations were 2500 units/mg protein and 11,000 units/mg protein, respectively.

Extraction and Chromatography

Two extraction and chromatographic techniques were used. The first (System I) was employed to define the conditions required for maximum production of lO-keto-Dg. Samples were extracted for total lipids (5) and applied to a Lipidex (Packard Instrument Co., Inc.,

Downers Grove, IL) 5000 column (0.9 x 20 cm). The column was developed in a two-step elution sequence (Fig. 1). The lO-keto-Dg eluted follow­ ing the change of solvent from 10% chloroform in hexane to 50% chloro­ form in hexane.

A second chromatographic system (System II) was used for quantita­ tive analyses. This method involved a solid-phase extraction technique using C^g "Sep-Paks" (Waters Associates). Following an incubation, 30 ml of acetonitrile were added to 20 ml of either strained rumen fluid, a culture of rumen bacteria in Brain Heart Infusion (BHI) (BBL Labora­ tory Systems, Cockeysville, MD), or a cysteine'HCl solution. To moni­ tor recovery, 10,000 dpm of 10-keto-[^H]-D3 was added to each serum bottle. The mixture was placed on a shaker for 20 min and then trans­ ferred to a 50-ml conical centrifuge tube. After centrifugation (2000

X g for 10 min), the supernatant was decanted into a clean serum bottle 51

(100 ml) containing 20 ml of water. The entire 70 ml was passed

through a C^g Sep-Pak, followed by 3 ml of (70/30) methanol/water. The lO-keto-Ds fraction eluted with 8 ml of (90/10) acetonitrile/water, and was collected and dried under nitrogen. The residue was dissolved in

50 ul of (98/2) hexane/isopropanol, and applied to a Waters silica

Sep-Pak. The eluate, 8.5 ml of (98/2) hexane/isopropanol, contained the vitamin D3 fraction. The addition of 8 ml of (90/10) hexane/iso­ propanol was required to elute lO-keto-Dg. This fraction was col­ lected, dried under nitrogen, and prepared for reverse-phase HPLC by redissolving in 200 pi (90/10) methanol/water. Reverse-phase HPLC was accomplished in (90/10) methanol/water at a flow rate of 2 ml/min. lO-Keto-Ds eluted at 32 ml in this system. The lO-keto-Ds peak was collected, dried and analyzed for radioactivity with a Beckman LS-9800 liquid scintillation counter. All reported values were corrected for quench and recovery of the 10-keto-[^H]-D3.

Preparation of Media

All manipulations were done anaerobically, as described by Hungate

(18), except that cysteine'HCl was added to the media, and the media were dispensed (20 ml) into 100-ml serum bottles prior to steriliza­ tion. The basal medium consisted of BHI broth supplemented with

(wt/vol) 0.05% cysteine'HCl, 0.1% resazurin, 0.3% arginine, and 0.4%

NazCOg. Basal medium supplemented with 1.0% (wt/vol) glucose was used to determine if glucose inhibited the production of lO-keto-Dg. All media were adjusted to pH 6.8, and anaerobically dispensed under 52

02-free COg. Cultures were transferred between serum bottles by using

syringes which were pre-gassed with anaerobic CO2.

Assay Conditions

All cultures in BHI broth were grown overnight before being

assayed for their ability to produce 10—keto-Dg. Incubations were

conducted in 100-ml serum bottles to accommodate rapid atmospheric

changes. This was accomplished by inserting gas lines with needle

attachments through the butyl-rubber stoppers. Rumen fluid

incubations, and mixed cultures derived from rumen fluid, were

performed under an atmosphere of 10% O2 in CO2. Cultures of the

ruminai isolate of Clostridium sp. strain A264 required a lower O2

concentration (4%), since higher levels were inhibitory. The O2/CO2

mixture was obtained by using a dual flow meter (Matheson Co., Inc.,

Rutherford, NJ). The concentrations of O2 and CO2 were measured by

using a Loenco respiration and blood gas analyzer (Altadena, CA). All

incubations were conducted at 39°C.

Rumen Fluid Studies

Samples of rumen fluid were obtained from a nonlactating, nonpreg­

nant, fistulated Jersey cow fed a maintenance diet of 90% alfalfa hay

cubes and 10% cracked corn. The samples were taken 4 to 5 h after feeding, strained through 4 layers of cheesecloth, and used within 30 min. 53

Oxygen studies

The effects of varying O2 concentration on lO-keto-Dg production were performed exclusively in strained rumen fluid containing 1 x 10^ dpm of [^H]-vitamin Dg. The impact of different atmospheres was as­ sessed by overlaying the rumen fluid with the desired concentration of

O2 (0-10% in CO2 or room air). One set of bottles was sealed with butyl-rubber stoppers (closed system), and another set of bottles was closed with cotton stoppers (open system). The effects of shaking were studied at a speed of 132 oscillations/min. Bubbling experiments using gas mixtures were performed in Pyrex test tubes (2.5 x 21 cm). All bottles and tubes were incubated 20 h at 39°C. System I was used for extraction and chromatography, as previously described.

Preincubation treatments

To ascertain the importance of viable cells in the transformation of vitamin D3 to lO-keto-Dg, both sterilized (121°C for 15 min) and cell-free rumen fluid samples (10,000 x g for 20 min followed by fil­ tration through a 0.45 ym Millipore filter) were prepared. These two treatments, and an unmanipulated control, were incubated under static

(no shaking) conditions with an atmosphere of 10% O2 in CO2. The reac­ tion was initiated with the addition of 1 x 10^ dpm of [^H]-vitamin D3.

System I was used for extraction and analysis.

Mixed Culture Studies

Cultures of mixed rumen bacteria were maintained in BHI broth and transferred (0.5% inocula) every 3 days. To conduct an assay, a 54

culture was incubated overnight under COg, the atmosphere was then changed to 10% 0% in CO2, and the reaction was initiated by the addi­ tion of 400 ug of vitamin D^. After a 4-h incubation, the presence of lO-keto-Ds was determined by using System II.

Filtration

A mixed culture was grown overnight in 500 ml of BHI broth and then heat-killed (121°C for 15 min). Three 20-ml portions were taken and anaerobically dispensed into serum bottles to serve as controls. A

300-ml sample was clarified by centrifugation at 10,000 x g for 20 min.

Two hundred ml of supernatant were filtered through an Amicon PM-10

62-mm filter (10,000 M.W. cutoff). Three 20-ml portions of the fil­ trate were removed for analysis. The Amicon "retentate" was resus- pended in 100 ml of sterile BHI broth, and divided into two 50-ml aliquots. One aliquot was combined with 50 ml of sterile BHI broth; the other was combined with 50 ml of filtrate. Three 20-ml subaliquots were taken out of each mixture and anaerobically dispensed into each of

3 serum bottles. Four hundred pg of vitamin D3 in 50 ul ethanol were added to each of the 15 bottles. The bottles were incubated under an atmosphere of 10% O2 in CO2 for 20 h, and then the presence of

10—keto-Ds was determined by using System II.

Time study

The production of lO-keto-Dg was plotted against time (0 and 7 h); analysis was by System II. Production peaked at 4-5 h; thus, a 4-h incubation period was used for subsequent studies. 55

Glucose inhibition

Mixed cultures were grown under CO2 for 20 h in basal medium and

basal medium plus 1.0% glucose. At the end of this incubation period,

the atmosphere was changed to 10% O2 in CO2, and the reaction was ini­

tiated by the addition of 400 pg of vitamin D3. The bottles were incu­

bated for 4 h, after which the pH was measured. Extraction and analy­

sis was by System II.

Isolation of Clostridium sp. Strain A264

Mixed cultures grown in BHI were heat-shocked at 80°C for 15 min,

and transferred into fresh BHI broth. The resulting suspension was

then streaked for isolation on BHI roll tubes (2% agar). After incuba­

tion, colonies were picked and assayed for their ability to produce

lO-keto-Dg. All isolates were morphologically similar; one strain,

A264, was selected for further studies. The culture was maintained on

BHI slants (BHI broth + 1.5% agar) and transferred weekly. Identifica­

tion was as described by the Anaerobic Laboratory Manual, Virginia

Polytechnic University (Blacksburg, VA).

Pure Culture Studies

These studies used identical media and growth conditions as de­ scribed in the "Mixed Culture" section. Cultures were grown overnight in BHI broth under CO2, the atmosphere was then changed to 4% O2 in

CO2, and the reaction was initiated by the addition of 400 ug of vita­

min D3. The atmosphere used in these pure culture studies had a lower

O2 content because higher levels were inhibitory to the production of 56

lO-keto-Ds by this bacterium. System II was used for extraction and chromatography.

Glucose inhibition studies

These were done as previously described for the mixed cultures.

Cysteine Studies

Solutions of cysteine'HCl (0.1% wt/vol) were made with spectral grade water, and the pH was adjusted to 7.0. Prior to dispensing (20 ml/serum bottle) and sterilization, the solutions were bubbled with

Oz-free Ng gas to remove traces of Og. For assay, a serum bottle was injected with 500 yg of vitamin D3 in 50 pi ethanol followed by 2 ml of

O2. Incubations were under static conditions. At the end of 4 h, the samples were extracted and chromatographed by using System II, except that no centrifugation step was used. The effect of catalase (1000 units/ml) and SOD (500 units/ml) were studied by using the same cysteine'HCl solution. The enzymes were introduced into the serum bot­ tles in 150 lil of anaerobic water carrier. As controls, the enzymes were denatured by heating them at 100°C for 5 min prior to injection. 57

RESULTS

Rumen Fluid Studies

In initial studies with strained rumen fluid, an elaborate chro­ matographic system (System I) was used because rumen fluid contains a variety of lipids and particles. Recoveries of 10-keto—D3 were low, and the sensitivity of System I was poor compared to System II.

We extended the earlier work of Sommerfeldt et al. (29) by demon­ strating that production of lO-keto-Dg by mixed microbial populations

(strained rumen fluid) required molecular O2. As shown in Tables 1 and

2, the ability to produce lO-keto-Dg was affected by the atmosphere and incubation conditions. The production of 10—keto-Dg was negligible under strictly anaerobic conditions; however, atmospheres which con­ tained low levels of O2 were stimulatory. When 0% concentrations were too high (open system), production was diminished, although to a lesser degree, than with total anaerobiosis (Table 1).

Variations in O2 concentration, and the method used to introduce

O2» demonstrate that in the 100-ml serum bottle system which relied on diffusion (static), the optimal atmospheric O2 concentration was approximately 10% (Table 2). However, since these results were obtained under conditions where an O2 gradient was formed, the active region for lO-keto-Dg production within this gradient (or the effective oxygen concentration) is not known. Rapid O2 introduction by shaking, dramatically reduced the amount of lO-keto-Ds formed. Production of

10—keto-Dg in the shaking system peaked at 4% O2, but the yield never approached that of the static system using 10% Og. When gas mixtures 58

containing between 1.22 and 0.10% Oz in COz were bubbled through rumen

fluid, total inhibition of 10—keto-Dg formation was observed (data not

shown).

The extent of anaerobiosis (static vs. shaking) was estimated by

monitoring methane production after incubation in closed systems (Fig.

2). Under static conditions, even with a 10% oxygen head space, parts

of the serum bottle (presumably at the bottom) were still highly an­

aerobic, as indicated by the high concentrations of Oz required to in­

hibit methane production. The lowest level of O2 used (2%) reduced

methane production in the shaking system.

Heat-sterilized rumen fluid did not support the production of

lO-keto-Dg (Table 3). When rumen fluid was clarified by centrifugation

at 10,000 X g, 10-keto—D3 formation was greatly reduced, indicating

that viable cells were needed for maximal conversion of vitamin D3 to

lO-keto-Dg. However, slow speed centrifugation (1000 x g) to remove

large particles and protozoa did not significantly reduce the amount of

10-keto—D3 in the supernatant (data not shown), suggesting that bacte­

ria play the main role in the formation of lO-keto-Dg. Resuspension of

the pellet in the supernatant restored 10—keto-Dg production to levels nearly equivalent to those obtained in unmanipulated samples.

BHI Broth Culture Studies

The extraction and chromatography of System I was simplified after

it was discovered that the microorganisms responsible for lO-keto-Dg

production could be cultured in BHI broth. Time-course studies (Fig. 59

3) revealed that mixed cultures in BHI broth produced lO-keto-D^ rapidly. Production reached a maximum level between 4 and 5 h, com­ pared with rumen fluid which required 18 to 20 h incubation (J. L.

Sommerfeldt, Ph.D. Dissertation, Iowa State University, Ames, lA).

During attempts to isolate a bacterium which could catalyze the production of 10-keto—D3, we observed that the ability to produce lO-keto-Ds was not destroyed upon heat treatment (Table 4). This sug­ gested that a spore-forming organism was responsible. From heat- shocked cultures we isolated a gram-positive rod which produced lO-keto-Dg, and upon further examination, strain A264 was determined to be a member of the genus Clostridium, possibly C_. hastiforme. When compared to mixed cultures from the bovine rumen, this organism was less efficient at producing lO-keto-Dg (Table 5), possibly because the optimal concentrations of O2 required for lO-keto-Dg production by

Clostridium sp. strain A264 were not known. Both mixed and pure cul­ tures produced significantly more (P < .05) lO-keto-Dg (20-fold) than sterile, uninoculated BHI broth.

It is not known how unique C-10(19) oxidation is among bacteria.

Experiments with other Clostridia demonstrated that several other spe­ cies could also catalyze production of this compound (data not shown).

Highly active species included C^. bif ermentans, C^. sporogenes, C^. pasteuranium, C^. hastif orme, and C_. tetanomorphum. Those species dem­ onstrating little ability to produce lO-keto-Dg include C. paraputrificum, _C. innocuum, C^. butyricum, C^. septicum, and C. perfringens. 60

The addition of glucose to medium inhibited the production of lO-keto-Dg. This effect is probably attributable to acid production which lowered the pH of the media, rather than to a lack of 0% (Table

5). Support for this contention is provided by results obtained with

Clostridium sp. strain A264. Inhibition (P < .05) was noted, in the production of lO-keto-Dg obtained in pure or mixed cultures grown in basal medium plus 1.0% (wt/vol) glucose, when compared to cultures grown in basal medium without glucose. Clostridium sp. strain A264 can utilize glucose; however, it is also highly proteolytic, which dimin­ ishes its tendency to lower the pH. Being a strict anaerobe argues against the stimulation of oxygen uptake by the addition of increased substrate. The mixed culture lowered the medium pH from 6.8 to 5.2 when BHI was supplemented with 1.0% (wt/vol) glucose. The amounts of lO-keto-Dg obtained with mixed cultures grown in the presence of 1.0%

(wt/vol) glucose were not significantly different from those produced in sterile, uninoculated BHI broth (Table 5).

When cultures were grown overnight in BHI broth, and then heat killed, the ability to catalyze lO-keto-Dg production increased dramat­ ically (Table 4). Levels of lO-keto-Dg were 10 to 12 times higher than those obtained with viable cells, and approximately 230 times greater than those obtained by nonspecific oxidation in sterile, uninoculated

BHI broth. Attempts to fractionate the vitamin D3 oxidase activity from heat-killed samples revealed that at least two factors of bac­ terial origin were responsible for the observed increase in oxidative potential (Table 4). The first factor is smaller than 10,000 61

daltons, the second was larger than 10,000 daltons. Neither the

filtrate nor the "retentate" alone could approach the values obtained

in an unmodified system. Values obtained with supernatant from

heat-killed cultures were slightly lower than when a complete system

was used with cells.

Production of 10-keto—D3 in a chemically-defined, nonbiological

system revealed that cysteine was probably responsible for the produc­

tion of lO-keto-Da in sterile medium (Table 6). The effect of 0.1%

cysteine was more pronounced in water than in a complex medium. The

product obtained by this method had an identical E and E . as au- max mxn thentic lO-keto-Dg and the biologically-produced compound (Fig. 4); it

also comigrated on three separate HPLC systems with authentic lO-keto-

D3. The addition of a solution containing both catalase (1000

units/ml) and SOD (500 units/ml) markedly reduced (P < .05) the ability

of this system to oxidize vitamin D3 (Table 6). T'Then heat-inactivated

enzyme was used, about 60% of the ability to produce lO-keto-Dg was restored. 62

DISCUSSION

In light of the anaerobic environment in the rumen, the participa­

tion of O2 in the production of lO-keto-Dg was a surprising result; however, several investigators have noted a stimulatory effect of O2 on bile acid oxidation by anaerobic bacteria (2, 20, 21). As in our stud­

ies, the mechanism of oxidation and the concentration of O2 needed were ill defined. Macdonald et al. (20), however, did propose a possible mechanism, and they suggested that formation of 3-keto derivatives of bile acids may simply relate to an increased ratio of intracellular

NADP'*'/NADPH which would result from increased Eh values. Although the stimulatory effect of 0% was most dramatic vitro (with rumen fluid and cultures derived from rumen inocula grown in BHI), the availability and effects of O2 vivo is not known. The dissolved O2 concentration required for lO-keto-Dg formation must be very small because optimal yields were obtained when O2 entered by diffusion. Experiments in which low levels of oxygen were bubbled through the medium demonstrated that lO-keto-Da production was inhibited by low levels of O2, not by the lack of O2 due to facultative removal. The use of methane produc­ tion as an indicator of anaerobiosis supported the concept that the diffusion system still contained anaerobic areas. It appeared that O2 could also be inhibitory to this reaction, as noted by the experiments with shaking and those using a pure culture of Clostridium sp. strain

A264. Experiments in which O2 was bubbled into strained rumen fluid also demonstrated that rapid O2 introduction was inhibitory to lO-keto-

D3 formation. As low as 0.1% O2 was inhibitory in strained rumen fluid 63

incubations. If O2 is obligately linked to the ^ vivo formation of lO-keto-Dg, it could be argued that this reaction might occur at the epithelial surface in the gut where O2 is available (27).

During the course of studying lO-keto-Dg production in mixed cul­ tures, we observed that 10-keto—D3 could be produced spontaneously in sterile uninoculated BHI broth (Table 5). This phenomenon was unde­ tected in rumen fluid incubations, presumably because of the relative insensitivity of chromatographic System I and the trace amounts of sub­ strate converted. lO-Keto-Dg formation, therefore, does not appear to be enzymatic in nature, but results from indirect effects of bacterial metabolism. This contention is supported by our studies showing sig­ nificant lO-keto-Ds production by heat-killed mixed cultures in BHI broth. Not many enzymes can survive autoclaving. Results obtained with the chemical production of lO-keto-Dg, using cysteine'HCl (0.1% wt/vol), also provide support for this assertion.

The formation of lO-keto-Dg by anaerobic bacteria may involve the generation of superoxide anions, hydrogen peroxide, or other oxidative radicals. These compounds are among the strongest oxidizing agents known (14, 23). A role for hydrogen peroxide or superoxide anion in lO-keto-Dg production was supported by experiments in which the addi­ tion of catalase and SOD inhibited vitamin D3 transformation. Carlsson et al. (9) proposed a mechanism for the appearance of these agents.

Their work showed that anaerobically prepared BHI broth can undergo autoxidation to produce hydrogen peroxide and superoxide anions. The active components in media which produce these radicals are thought to 64

involve glucose and phosphate. After they had been autoclaved together in media at a neutral pH, they rapidly autoxidized and ultimately formed superoxide radicals and hydrogen peroxide (9). Hydrogen peroxide is the major radical produced. The potential also exists for hydrogen peroxide to react with superoxide anions to form hydroxyl radicals (23). Autoxidation of cysteine is known to result in the formation of hydrogen peroxide (26). Frolander and Carlsson (15) demonstrated an increased toxicity of anaerobic media exposed to oxygen. In their studies, the addition of SOD to anaerobic media exposed to oxygen significantly decreased the toxicity of the media, confirming the presence of oxidative radicals. Nyberg et al. (26) stated that the rate of hydrogen peroxide formation from the autoxidation of cysteine could not be higher than 1 nmol/ml per min in a reaction mixture containing 260 uM cysteine. However, this level of production was more than sufficient to be toxic to Peptostreptococcus anaerobicius. This low rate of peroxide release could explain why our nonbiological system required incubation for 4 h or longer to effect maximal formation of lO-keto-Dg from vitamin D3.

The results of high-pressure filtration studies suggested that at least two heat-stable factors that affect lO-keto-Dg production are released upon disintegration of the bacterial cell. Together, these factors apparently are capable of generating oxygen radicals which re­ sult in the formation of lO-keto-Dg. This mechanism is similar to that of the nonbiological system (cysteine'HCl), except that the bacteria probably produce some factor(s) more potent than cysteine. Production 65

of lO-keto-Ds in the nonbiological system was higher than those pro­

duced in sterile medium which contained a similar concentration of

cysteine. One possible explanation for this observation is that in

sterile media many other components can compete for the oxidative radi­

cals produced. In the nonbiological system, vitamin Da experiences

less competition for oxidation, which is ultimately expressed in the

form of enhanced lO-keto-Dg formation. On this basis, the biological

system must be more active than the chemical system because a complex

medium such as BHI broth contains large amounts of components which can

compete for free radicals.

In our experiments, glucose was inhibitory to lO-keto-Dg formation

(Table 5). Several other investigators describe inhibitory effects of

carbohydrates on the biotransformation of steroids by gut bacteria (12,

21). They concluded that the observed inhibition was caused by a low­

ering of the pE. The same mechanism is probably responsible for the

inhibition of lO-keto-Dg production in basal medium supplemented with

glucose. Clostridium sp. strain A264 is highly proteolytic and tends

to keep the medium around neutrality. Since this species is a strict

anaerobe, it would not use O2 for glucose metabolism, negating the idea

that insufficient O2 was available for vitamin D3 oxidation in medium

containing 1% glucose. The mixed culture, on the other hand, contained

organisms which produced acidic end products; the lowering of the pH

was paralleled by inhibition of lO-keto-Dg production.

Although 10-keto—D3 formation may be a nonenzymatic process, the

production of lO-keto-Dg is unquestionably potentiated by the presence 66

of bacterial cells. As shown in Table 4, a 100-fold increase in the formation of lO-keto-Ds took place when cell products were present.

This nonenzymatic metabolism of vitamin D3 by bacteria may be important in vivo. lO-Keto-Dg has been detected in the plasma of calves receiving oral doses of [^H]-vitamin D3 (30). The 10-keto-derivative of 25-OHD3 (10-keto-19-nor-25-0H-D3) has also been produced by cultured kidney and phagocytic cells (28). The mechanism by which these eucaryotic cells produce 10-keto derivatives of vitamin D is probably similar to that used by bacteria. Interestingly, both phagocytic cells and the P-450 cytochrome produce significant amounts of oxidative radi­ cals as a result of normal metabolism. 10-Keto-25-0H-D3 is also of interest because HPLC using silica gel columns with hexane/iropropanol solvents do not discriminate it from 1,25-dihydroxyvitamin D3, the most biologically active form of the vitamin (25).

If the production of lO-keto-Ds by bacterial-mediated oxidation is important i^ vivo, it would probably be so only under circumstances where the animal is challenged by large quantities of vitamin D. In this situation, based on the biological assays of Napoli et al. (24),

C-10(19) oxidative metabolism of vitamin D may benefit the host by re­ sulting in the removal of excess vitamin D from the circulation. 67

5000-1 VITANIN-D: PEAK Q (LO-TOTO-DJL 4000 PEAK

>- 3000-

0, Ë 2000- STRAINED ' RUMEN FLUID

1000- PEAK-R (HOT YET IDOITIFIEDT

0- 100 120 140 160 ELUTION VOLUME (ml)

Figure 1. An elution profile from a Lipidex 5000 column of the total lipid extract taken from rumen fluid incubated with [^H]- labelled vitamin D3 for 18 h. A 4% oxygen atmosphere was used. The Lipidex column was eluted with 50 ml of hexane/chloroform (9/1), followed by 100 ml of hexane/ chloroform (1/1) 68

6 — 'STATIC SYSTEM' (WATER BATH)

•••D. 3- SHAKING SYSTEM ( 132 OSCILLATIONS /min.)

2-

2 4 6 8 10 INITIAL Og CONCENTRATION (%)

Figure 2. The inhibition of methane production by oxygen in the static and shaking incubation systems. Methane production was used as an indicator of anaerobiosis 69

g 3000

2000

1000 o

0 0 2 3 4 5 6 7 8 TIME (hrs)

Figure 3. Quantities of 10—keto-Dg produced by a mixed culture of rumen bacteria in anaerobic BHI broth. Each point repre­ sents the mean of three replications 70

0.8-1

I I I I I I I I I 200 220 240 260 280 300 320 340 360 WAVELENGTH IN nm

Figure 4, U.V. spectra of authentic lO-keto-Dg and that produced by the oxidation of D3 catalyzed by bacteria and the auto- oxidation of cysteine ( ). Vitamin D3 absorbs maximally at 264 nm, whereas the 10-keto derivatives of D have an E at 312 nm, indicating disruption of the triene double bonS^ system in D 71

Table 1. The effects of different atmospheres on the production of lO-keto-Dg in rumen fluid (in vitro)

Atmosphere % of total radioactivity in peak Q (lO-keto-Dg)

Nitrogen 2

Argon 1

Hydrogen 1

Room air (open system) 18

Room air (closed system) 34

10% O2 + 90% CO2 46 72

Table 2. A comparison of two incubation systems: The relationship of oxygen and lO-keto-Dg production in strained rumen fluid

Initial oxygen Ratio of radioactivity % oxygen remaining^ concentration recovered in peak Q after incubation of gas phase vitamin D3 peak (18-20 h) (%) Water bath Shaker Water bath Shaker

0 0.19 0.12 0 0

2 0.63 0.21 0.24 0

4 0.97 0.32 0.90 0

6 2.52 0.27 1.80 0

10 6.37- 0.18 3.80 0

20.9 1.48 0.10 Not determined (Room air)

^In gas phase. 73

Table 3. Comparison of preincubation treatments of rumen fluid on lO-keto-Dg production

Treatment^ % of total radioactivity in peak Q (lO-keto-Ds)

None (control) 54.0

Heat sterilized 0

Supernatant obtained by centrifugation at 10,000 x g 8.0

Supernatant obtained by centrifugation at 10,000 x g and filtered through a 0.45 vm Millipore filter 16.0

Resuspension of pellet from centrifugation step back into supernatant 47.0

^Incubations were under an atmosphere that was 10% O2 in CO2 at the start of incubation.

^All values are corrected for background. Extraction and chroma­ tography were by System I. 74

Table 4. Fractionation of vitamin Dg oxidase activity from heat-killed mixed cultures in BHI broth^

Amount of lO-keto-Dg produced in Treatment 20 h (ng)

Prior to centrifugation

Uninoculated sterile BHI broth 93 ± 32^

Mixed culture grown for 18 h in BHI broth and heat killed 22,060 ± 343

After centrifugation^

Filtrate <10,000 M.W. (supernatant passed through a 10,000 M.W. filter) 9,512 ± 121^

Resuspension of "retentate" (>10,000 M.W.) into sterile uninoculated BHI broth 764 ± 76

Resuspension of retentate (>10,000 M.W.) with the filtrate (<10,000 M.W.) 18,206 ± 2501

^All incubations used 400 yg D3 and 10% 0%.

^Values with the same letter are not statistically different (P < 0.05).

^The next 3 treatments all use supernatant obtained from a heat- killed culture grown in BHI broth. These treatments were performed prior to the addition of vitamin D3. 75

Table 5. Comparison of C-10(19) oxidase activity in mixed and pure cultures - grown in basal BHI and basal plus 1% glucose

Amount of lO-ketogOg Final pH of medium Treatment^ produced in 4 h (after incubation) (ng)

Sterile uninoculated basal BHI 103 43A 6.80

Mixed culture grown in basal BHI 2638 236^ 6.90

Mixed culture grown in basal + 1% glucose 94 ± 22^ 5.20

Clostridium sp. strain A264 grown in basal BHI 1906 i 142® 6,85

Clostridium sp. strain A264 grown in basal + 1% glucose 1557 + 124° 6.50

^All cultures were grown overnight. The atmosphere was then changed from COg to 96/4 (CO2/O2), and incubated another 4 h with

400 ug D3. ^Values with the same letter are not statistically different (P < 0.05). 76

Table 6. Nonbiological production of lO-keto-Da by the autoxidation of cysteine

a b Treatment Amount of lO-keto-D^ produced in 4 h (ng)

Anaerobic water 138 ± 53'^

Cysteine'HCl solution (0.1% wt/vol) 1945 ± 251

Cysteine'HCl solution (0.1% wt/vol) + super­ oxide dismutase (500 units/ml) and catalase (1000 units/ml) 134 ± 24^

Cysteine'HCl solution (0.1% wt/vol) + heat- inactivated superoxide dismutase (500 units/ml) and catalase (1000 units/ml) 1200 ± 56

^All treatments were anaerobically prepared and heat-sterilized prior to the addition of O2.

^Values with the same letter are not statistically different (P < 0.05). 77

LITERATURE CITED

1. Archer, R. H., I. S. Maddox, and R. Chong. 1982. Transformation of cholic acid by Clostridium bifermentens. J. Appl. Bacteriol. 52:49-56.

2. Aries, V., and M. J. Hill. 1970. Degradation of steroids by- intestinal bacteria. II. Enzymes catalysing the oxidoreduction of the 3a,7a- and 12a-hydroxyl groups in cholic acid, and the dehydroxylation of the 7-hydroxyl group. Biochim. Biophys. Acta 202:535-543.

3. Arnaud, S. B., R. S. Goldsmith, P. W, Lambert, and V. L. W. Go. 1975. 25-Hydroxyvitamin D^: Evidence of an enterohepatic circulation in man. Proc. Soc. Exp- Biol. Med. 149:570-572.

4. Barton, D. H. R., and H. Patin. 1976. Chemistry of the tricar- boxyliron complexes of calciferol and ergosterol. J. Chem. Soc. Perkin Trans. 1:829-831.

5. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37:911-917.

6. Bokkenheuser, V. D., J. Winter, B. I. Cohen, S. O'Rourke, and E. H. Mosbach. 1983. Inactivation of contraceptive steroid hormones by human intestinal Clostridia. J. Clin. Microbiol. 18:500-504.

7. Bokkenheuser, V. D., J. Winter, S. O'Rourke, and A. E. Ritchie. 1980. Isolation and characterization of fecal bacteria capable of 16a-dehydroxylating corticoids. Appl. Environ. Microbiol. 40:803-808.

8. Brinkley, A. W., A. R. Gottesman, and G. E. Mott. 1982. Isolation and characterization of new strains of cholesterol reducing bacteria from baboons. Appl. Environ. Microbiol. 43:86-89.

9. Carlsson, J., G. Nyberg, and J. Wrethen. 1978. Hydrogen peroxide and superoxide radial formation in anaerobic broth media exposed to atmospheric oxygen. Appl. Environ. Microbiol. 36:223-229.

10. Cohen, M. S-, and T. K. Gray. 1984, Phagocytic cells metabolize 25-hydroxyvitamin D3 in vitro. Proc. Natl. Acad. Sci. USA 81:931-934.

11. Eyssen, H. 1973. Role of the gut microflora in metabolism of lipids and sterols. Proc. Nutr. Soc. 32:59-63. 78

12. Eyssen, H., G. DePauw, and G. Paramentier. 1974. Effect of lactose on 5—steroid-reducing activity of intestinal bacteria in gnotobiotic rats. J. Nutr. 104:605-612.

13. Eyssen, H., G. Parmentier, F. C. Compernolle, G. DePauw, and M. Piessens-Denef. 1973. Biohydrogenation of sterols by Eubacterium ATCC 21,408—nova species. Eur. J. Biochem. 36:411-421.

14. Fridovich, I. 1978. The biology of oxygen radicals. Science 8:875-879.

15. Frolander, F., and J. Carlsson. 1977. Bactericidal effect of anaerobic broth exposed to atmospheric oxygen tested of Peptostreptococcus anaerobius. J. Clin. Microbiol. 6:117-123.

16. Goldsmith, R. S. 1982. Enterohepatic cycling of vitamin D and its metabolites. Min. Electrolyte Metab. 8:289-292.

17. Gray, T. K., F. W. Maddux, G. E. Lester, and M. E. Williams. 1982. Rodent macrophages metabolize 25-hydroxyvitamin D3 in vitro. Biochem. Biophys. Res. Commun. 109:723-729.

18. Hungate, R. E. 1969. A roll tube method for cultivation of strict anaerobes. Pages 117-132 in J. R. Norris and D. W. Ribbons, eds. Methods in Microbiology, vol. 3B. Academic Press, Inc., New York.

19. Macdonald, I- A., V- D. Bokkenheuser, J. Winter, A. M. McLeron, and E. H. Mosbach. 1983. Degradation of steroids in the human gut. J. Lipid Res. 24:675-700.

20. Macdonald, I. A., D. M. Hutchison, T. P. Forrest, V. D. Bokken­ heuser, J. Winter, and L. V. Holdeman. 1983. Metabolism of primary bile acids by Clostridium perfringens. J. Steroid Biochem. 18:97-104.

21. Masuda, N., and H. Oda. 1983. 7a-Dehydroxylation of bile acids by resting cells of an unidentified, gram-positive, nonspore- forming anaerobic bacterium. Appl. Environ. Microbiol. 45:456-462.

22. Masuda, N., H. Oda, S. Hirano, M. Masuda, and H. Tanaka. 1984. 7a-Dehydroxylation of bile acids by resting cells of a Eubacterium lentum-like intestinal anaerobe, strain c-25. Appl. Environ. Microbiol. 47:735-739.

23. Morris, J. G. 1976. Fifth Stenhouse-Williams memorial lecture. Oxygen and the obligate anaerobe. J. Appl. Bacterid. 40:229:244. 79

24. Napoli, J. L., J. L. Sommerfeldt, B. C. Pramanik, R. Gardner, A. D. Sherry, J. J. Partridge, M. R. Uskokovic, and R. L. Horst. 1983. 19-Nor-lO-keto—vitamin D derivatives: Unique metabolites of vitamin D3, vitamin Dg, and 25-hydroxyvitamin D3. Biochemistry 22:3636-3640.

25. Norman, A. W. 1979. Metabolism of vitamin D. Pages 134-197 in W. J. Darby, ed. The Calcium Homeostatic Steroid Hormone. Academic Press, Inc., New York.

26. Nyberg, G. K., G. P. D. Granberg, and J. Carlsson. 1979. Bovine superoxide dismutase and copper ions potentiate the bactericidal effect of autoxidizing cysteine. Appl. Environ, Microbiol. 38:29-34.

27. Scott, R. I., N. Yarlett, K. Hillman, T. N. Williams, A. G. Williams, and D. Lloyd. 1983. The presence of oxygen in rumen liquor and its effects on methanogenesis. J. Appl. Bacteriol. 55:143-149.

28. Simpson, R. U., J. K. Wichmann, H. E. Paaren, H. K. Schnoes, and H. F. DeLuca. 1984. Metabolism of 25-hydroxyvitamin D3 by rat kidney cells in culture: Isolation and identification of cis- and

trans-19-nor-10-oxo-25-hvdroxyvitamin D3. Arch. Biochem. Biophys. 230:21-29.

29. Sommerfeldt, J. L., R. L. Horst, J. L. Napoli, E. T. Littledike, D. C. Beitz, and R. L. Horst. 1983. Evidence for in vitro production of vitamin D- and vitamin D3 metabolites by rumen microbes. J. Dairy Sci. Suppl. 1:88-89.

30. Sommerfeldt, J. L., J. L. Napoli, E. T. Littledike, D. C. Beitz, and R. L. Horst. 1983. Metabolism of orally administered [^H]-ergocalciferol and [^H]—cholecalciferol by dairy cows. J. Nutr. 113:2595-2600.

31. Wainfan, E., G. Henkin, S. C. Rittenberg, and W. Marx. 1953. Metabolism of cholesterol by intestinal bacteria in vitro. J. Biol. Chem. 207:843-849.

32. Wells, W., S. C. Anderson, and R. Quan Ma. 1960. Lactose diets and cholesterol metabolism. Cholesterol absorption, coprostanol formation and bile acid excretion in the rat. J. Nutr. 71:405-410.

33. Winter, J., A. C. McLeron, S. O'Rourke, L. Ponticorvo, and V. D. Bokkenheuser. 1982. Formation of 206—dihydroxysteroids by anaerobic bacteria. J. Steroid Biochem. 17:661-667. 80

SECTION II.

THE BIOLOGICAL ASSESSMENT OF VITAMIN D3 METABOLITES PRODUCED

BY RUMEN BACTERIA 81

INTRODUCTION

When incubated in the presence of rumen bacteria, cholecalciferol

(vitamin D3) and 25-hydroxycholecalciferol (25-OH-D3) undergo oxidative

metabolism to 5 (E)-19-nor-10-keto-vitamin D3 (lO-keto-Dg) and 5(E)-19-

nor—lO-keto—25—OH-D3 (10-keto—25-OH—D3), respectively (17). These

metabolites were first isolated and characterized from vitro bacte­

rial incubations, and Sommerfeldt et al. (22) recently presented pre­

liminary evidence that 10-keto-[^Hj-Ds is present in the plasma of

calves receiving oral [^H]-vitamin D3. However, the in vivo signifi­

cance of bacterial metabolism of vitamin D3 is not known.

The rumen, its bacteria, and its physiological functions are simi­

lar to those found in the cecum and colon of nonruminant animals (1);

thus, any significant rumen bacterial impact on vitamin D metabolism

may have important ramifications in monogastric species such as humans.

Supplementing foods (both livestock and human) with normal to excessive levels of vitamin D (10, 20) exposes gut bacteria to a significant pro­

portion of the vitamin D that animals receive. Thus, gut bacteria may influence the vitamin D status of the host. In addition to diet, the enterohepatic cycle represents the other major route for vitamin D to

enter the intestine. Most of the major vitamin D sterols [vitamin D3,

25-OH-D3, and 1,25-dihydroxyvitamin D3 (1,25-(0H)2D3] circulate in this cycle (3, 8). The relative contributions of diet and the enterohepatic cycle in providing vitamin D substrate for bacterial oxidation is not known. 82

The present study addresses the vivo significance, as measured by biological activity, entry into the circulation and pharmacokinetics of the recently described bacterial metabolites of vitamin D (17). We also briefly describe the production, purification, and initial charac­ terization of a new bacterial metabolite of vitamin D3, designated

"Peak A." This compound is produced when vitamin D3 is incubated with

Clostridium sp. strain A264, a rumen isolate known to affect the pro­ duction of lO-keto-Dg (7). 83

MATERIALS AND METHODS

General

High-pressure liquid chromatography (HPLC) was performed with a

Waters Model LC-204 chromatograph fitted with a Model 6000-A pump and a

Model 440 fixed-wavelength (254 and 313 nm) dual channel detector

(Waters Associates, Milford, MA). HPLC columns (0.46 x 25 cm) were

purchased from DuPont Instruments (Wilmington, DE), unless otherwise

noted. Normal phase refers to microparticulate silica gel columns.

Reverse phase refers to microparticulate silica gel columns derivatized

with octa decylsilane (CDS). Spectral-grade solvents (Burdick and

Jackson, Inc., Muskegon, MI) were used in all chromatography systems.

Low resolution mass spectra were taken with a Finnigan Model 4021 qua­

druple mass spectrometer, using a solid probe. The U.V. spectra were

obtained in ethanol with a Beckman Model 25 recording spectro­

photometer.

Sterols

Vitamin D3 was purchased from Sigma Chemical Co. (St. Louis, MO).

[3a,^H]-Vitamin D3 was prepared as described by Barton et al. (4).

Radiochemical purity of [^H]-vitamin D3 was confirmed by HPLC on a

Zorbax Sil silicic acid column (0.46 x 25 cm) developed in hexane/iso- propanol (99/1) at a flow rate of 2 ml/min. lO-Keto-Dg, used for stan­ dards, was a gift from Hoffmann-LaRoche, Inc. (Nutley, NJ). The 10- keto-[^H]-D3, for internal standards, was prepared from [^H]-vitamin 84

Dg, as described by Napoli et al. (17). lO-Keto-25—OH-D3, for biologi­

cal assays, was also prepared as described by Napoli et al. (17).

Isolation of lO-Keto-Dg From Bovine Plasma

Two 150-kg Jersey steers were fed vitamin D3 (125 mg/day) for 30

days. Twenty-four hours after administration of the last dose, the

steers were euthanized with sodium pentobarbital in succinyl choline,

and blood was collected in heparinized containers. All subsequent work was done under incandescent lights. Plasma was separated from whole

cells by centrifugation. Eight hundred fifty ml of plasma were

"spiked" with 100,000 dpm each of 5(E)- and 5(Z)-10-keto-[^H]-D3 (1.2

Ci/mMol) and the plasma was then extracted (5). The extract was dried under reduced pressure and then was redissolved in ethanol and cooled in a dry ice/acetone bath. Insoluble lipids were removed by centrifu­ gation (2000 X g for 15 min) at -20°C. The supernatant was dried under reduced pressure to a volume of approximately 20 ml. This concentrated extract was mixed with 400 ml of acetonitrile and allowed to stand at room temperature for 30 min. Insoluble lipids were removed by centri­ fugation (2000 X g for 15 min) at 20°C. The supernatant which con­ tained lO-keto-Dg was dried and dissolved in 500 pi of hexane/isopro- panol (9/1). The sample was applied in toto to a normal-phase HPLC prep column (Zorbax Sil, 2.2 x 25 cm), and the column was developed by using the same solvent at a flow rate of 5 ml/min. One-ml fractions were collected and the region containing radioactivity (30-40 ml) was collected for further purification, using successive normal-phase HPLC 85

analytical columns (Zorbax Sil, 0.45 x 25 cm) developed in hexane/iso-

propanol (99/1) at a flow rate of 2.5 ml/min, followed by hexane/iso-

propanol (97/3). The metabolite was recycled 2 times, for a total of 3

passes through the final HPLC column. The fraction containing 10- keto-Dg was then collected, dried under nitrogen, and redissolved in absolute ethanol for further characterization.

Radioligand Binding Assays

The comparison of relative binding affinities for lO-keto-Dg,

10-keto-25-0H-D3, and "Peak A" to other vitamin D3 metabolites was ac­ complished with two sources of vitamin D-binding proteins: (1) the rat plasma vitamin D-binding protein diluted 1/5000 in 0.05 M potassium phosphate buffer containing 0.01% gelatin (11); (2) the specific l,25-(0H)2-vitamin D receptor isolated from bovine thymus, and assayed as described by Reinhardt et al. (19).

Pharmacokinetics

A plasma turnover rate for lO-keto-f^Hj-Dg was obtained in a

150-kg castrated, male Jersey. Due to a limited supply of lO-keto-

[^Hj-Ds, only one animal could be used. The animal received 9.87 x 10^ dpm of 5(E)-10-keto-[^H]-D3 tracer. Blood samples (5 ml) were taken frequently beginning at 1 min for a period of 720 min. The radioactiv­ ity present as 10-keto-[^H]-D3 was assayed by adding 100 ng of radio- inert lO-keto-Dg as an internal standard and extracting by the method of Bligh and Dyer (5). The extracts were chromatographed by using

Sephadex LH-20 in hexane/chloroform/methanol (79/20/1) (17) and the 86

region corresponding to the elution of lO-keto-Ds was collected for

HPLC. The samples were applied to a HPLC column and the column was developed in isopropanol/hexane (3/97). Recovery was estimated by us­ ing the internal standard peak heights. The data were analyzed by Dr.

Charles Ramberg at the University of Pennsylvania by using CONSAM, a conversational version of the SAAM computer program (CONSAM User Guide,

U. S. Department of Health and Human Services, Washington, DC).

Biological Evaluation

Weanling male rats were housed individually in overhanging plastic cages. They were fed vitamin D-deficient diets containing low calcium

(0.005%) and normal phosphorus (0.3%) for 3 weeks prior to experimental use. At the end of 3 weeks, the rats were anesthetized with halothane and the test compounds given intrajugularly in 50 vl of a carrier solu­ tion containing ethanol/propylene glycol (3/7). Controls received the carrier solution alone. Twenty-four hours after injection, the rats were sacrificed by decapitation and their duodena were removed and used to measure intestinal calcium transport by the everted intestinal sac procedure (15). In addition, blood was collected from the animals in heparinized tubes; the blood was clarified by centrifugation (500 x g for 15 min at 25°C). Plasma calcium was measured by atomic absorption spectroscopy (24). The rats were on a diet essentially devoid of cal­ cium; therefore, serum calcium increases reflect mobilization of calcium from bone. 87

Preparation of Bacterial Culture Medium

All manipulations were done anaerobically, as described by Hungate

(12), except that cysteine'HCl was added to the media, and the media

(20 ml) were dispensed into 100-ml serum bottles prior to steriliza­

tion. The basal medium consisted of Brain Heart Infusion (BHI) broth

(BBL Microbiology Systems, Cockeysville, MD) supplemented with (wt/vol)

0.05% cysteine'HCl, 0.1% resazurin, 0.3% arginine, and 0.04% Na2C03.

All media was adjusted to pH 6.8, and anaerobically dispensed under

Og-free COg. Cultures were transferred between serum bottles by using

syringes which were pregassed with anaerobic COg.

Isolation of Clostridium sp. Strain A264

A mixed culture of rumen bacteria grown in BHI was heat-shocked at

80°C for 15 min, and then transferred into fresh BHI broth. The re­

sulting suspension was streaked for isolation on BHI roll tubes (BHI

broth + 2% agar). After incubation, colonies were picked and assayed

for their ability to produce 10—keto—Dg. Active colonies were main­

tained on BHI slants (BHI broth + 1.5% agar); one strain, A264, was

chosen for further study.

Rumen Fluid

Samples of rumen fluid were obtained from a fistulated Jersey cow

(dry), fed a maintenance diet of alfalfa hay cubes and cracked corn

(high roughage). The samples were taken 4 to 5 h after feeding. 88

strained through 4 layers of cheesecloth, and used within 30 min. Rumen

fluid was used to determine if "Peak A" could be produced when mixed

bacterial populations, similar to those found in vivo, were incubated

with vitamin D3.

Assay Conditions

All clostridial cultures were grown overnight anaerobically in BHI

broth before being assayed for the ability to metabolize vitamin D3.

At the start of an experiment, the atmosphere was changed from CO2 to

4% O2 in CO2. This was accomplished by inserting gas lines through the

rubber stoppers. The reaction was then initiated by the addition of

400 lig vitamin D3 and allowed to incubate 18 h at 39°C. Rumen fluid

incubations were overlayed with 10% 0% in CO2 before the addition of

either 400 yg of vitamin D3 or 2 yg of "Peak A." This concentration

was previously shown to be optimal for lO-keto-Ds production by mixed

cultures of rumen bacteria (7). Pure cultures of Clostridium sp. strain

A264 required a low O2 concentration (4%); higher levels of O2 were

inhibitory. The O2/CO2 mixture was obtained by using a dual flow meter

(Matheson Co., Inc., East Rutherford, NJ). The concentration of O2 and

CO2 were measured by using a Loenco Respiration and Blood Gas Analyzer

(Altadena, CA). To insure that "Peak A" was derived from vitamin D3,

overnight cultures of Clostridium sp. strain A264 were incubated with

300 lig [^H]—vitamin D3 (1 pCi/300 yg)- These were analyzed as previ­ ously described. 89

Extraction and Chromatography

A solid-phase extraction technique, using C^g Sep-Paks (Waters

Associates) was developed. Following an incubation, 30 ml of aceto- nitrile was added to 20 ml of a culture of Clostridium sp. strain A264 or a strained rumen fluid culture. Then, 10,000 dpm of 10-keto-[^H]-D3 were added to each 100-ml serum bottle as an internal standard (no labelled "Peak A" was available). This mixture was placed on a shaker

(132 oscillations/min) for 20 min and then transferred to a 50-ml coni­ cal centrifuge tube. After centrifugation at 2000 x g for 10 min at

25°C, the supernatant was decanted into a 100-ml serum bottle contain­ ing 20 ml of glass-distilled water. The entire 70 ml were passed through a C^g Sep-Pak, followed by 3 ml of methanol/water (70/30).

Both lO-keto-Dg and "Peak A" were eluted with 8 ml of acetonitrile/water

(90/10), collected and dried under nitrogen. The residue was dissolved in 500 ul of hexane/isopropanol (98/2), applied to a Waters Silica

Sep-Pak and eluted with hexane/isopropanol (98/2). The first 8.5 ml of eluate contained the vitamin D3 fraction. The addition of 8 ml of hexane/isopropanol (90/10) was required to elute both 10-keto—D3 and

"Peak A." This fraction was collected, dried under nitrogen and pre­ pared for reverse-phase HPLC by redissolving in 200 ul methanol/water

(90/10). Reverse—phase HPLC was accomplished in methanol/water (90/10) at a flow rate of 2 ml/min. In this system, "Peak A" eluted between

19-21 ml and 10—keto—D3 eluted between 31-33 ml. "Peak A" was col­ lected, dried under nitrogen, and stored in absolute ethanol at 0°C.

The 10—keto—D3 peak was collected, dried and analyzed for radioactivitv 90

with a Beckman LS-9800 liquid scintillation counter. All reported val­ ues were corrected for quench and recovery of lO-keto-f^Hj-Dg.

For mass spectral, or U.V. analysis, "Peak A" was subjected to more rigorous purification. After collecting "Peak A" from the re- verse-phase system, it was dried under nitrogen and prepared for normal- phase HPLC. The residue was dissolved in 97/3 hexane/isopropanol and chromatographed on a Whatman 10-ym Partisil column (4.6 x 25 cm). In this system, "Peak A" eluted between 19-21 ml. The final column was a

Zorbax-NH2 column (4.6 x 25 cm) developed in 96/4 hexane/isopropanol;

"Peak A" eluted between 31-33 ml. The fraction containing "Peak A" was collected and stored under nitrogen until further characterization could be performed. 91

RESULTS

Isolation and Mass Spectral Detection of lO-Keto-Dg From Calves

Approximately 600 ng of lO-keto-Dg were isolated from plasma lipid extracts of vitamin Da-intoxicated calves. The final HPLC column, which separated the 5(Z) and 5(E) stereoisomers of lO-keto-Dg, recon­ firmed the findings of Napoli et al. (17) that the 5(E) isomer is the natural isomeric form because no 5(Z) could be detected in plasma. The

U.V. spectrum of the isolated compound displayed an absorbance maximum at 312 nm (Fig. 1) which is consistent with a 5(E), 7(Z), 10(19)-triene chromophore (17). Low resolution mass spectra (Fig. 2) attained for the putative plasma lO-keto-Dg was consistent with observed spectra for crystalline material (17). The spectrum had a molecular ion at m/z 386 and a large peak at m/z 355 (m^-43). Loss of the side-chain was indi­ cated by a peak at m/z 273. Major peaks were also observed at m/z 179,

177, 133, and 135, which are characteristic of lO-keto-Dg.

Pharmacokinetics

Figure 3 illustrates the rapid turnover rate observed for

10-keto-[^H]-D3 in the bovine plasma. The half-life of 5(E)-10- keto—D3 was approximately 6 min, indicating that this compound is rapidly removed or cleared from circulation and most probably the body.

The mass turnover rate was estimated to be 18 ug/kg/day. 92

Biological Activity Evaluation

The 5(E) (the biologically produced isomer) and 5(Z) (the isomer

produced by U.V. irradiation) of 5(E)-10-keto derivatives of vitamin D3

and "Peak A" were compared with vitamin D3 for biological activity.

The intestinal calcium transport response of rats that received a sin­

gle dose of the test compound 24 h prior to sampling are shown in

Tables 1 and 2. 5(Z)-10-Keto-D3 and 5(E)-10-keto-D3 were 40 and 80

times less potent than vitamin D3, respectively, in stimulating intes­

tinal calcium transport (Table 1). The 5(Z) metabolite, but not the

5(E) metabolite, displayed some bone-resorbing activity at the doses

tested. "Peak A" stimulated calcium absorption from the gut (P < .05); however, it was approximately 20-fold less active than the vitamin D3

(Table 2). "Peak A" did not mobilize detectable quantities of bone calcium at the doses tested. Co-dosing vitamin D3 and "Peak A" did not differ significantly from treatments with vitamin D3 alone, indicating that "Peak A" had no antagonistic effects on vitamin D3 activity. The

10—keto derivatives of 25-OH-D3 (10-keto-25-0H-D3) neither stimulated calcium absorption from the gut nor mobilization of bone calcium at the

doses tested (Table 3). The administration of 10-keto-25-0H-D3 daily over a 4-day period did not appear to enhance its biological activity

(Table 4). It was not possible to test both the 5(E) and 5(Z) isomers of 10-keto-25-0H-D3 because of their limited supply, hence a racemic mixture was used in this study. 93

Relative Affinity for Plasma Vitamin D Binding Protein

Table 5 illustrates the activities of the bacterial metabolites of

vitamin D3 and 25-OH-D3 in the diluted rat plasma radioassay used to quantitate vitamin D and vitamin D metabolites (11). In addition, other vitamin D3 metabolites are included as reference points. About

6000 times more 5(E)-IO—keto-Da than 25-OH—D3 was needed to displace

50% of the 25-OH-the 5(Z) isomer required about 300 times more mass. The rat plasma binding protein did not appear to possess any affinity for "Peak A." Although both the 5(E) and 5(Z) isomers of

10-keto-25-0H-D3 possessed greater displacement activity than any of the other bacterial metabolites tested, they were still 30-fold less

active than 25-OH-D3.

Relative Affinity for Thymus 1,25-(0H)2-Vitamin D3 Receptor Protein

Table 6 illustrates the relative affinity of the bacterial metabo­

lites of vitamin D3 and 25-OH-D3 for the 1,25-receptor from thymus (19).

Of these, only the 5(E) and 5(Z) isomers of 10-keto-25-0H-D3 displayed any significant ability to displace 1,25-(0H)2D3 from this protein.

However, when compared to 1,25-(0H)2D3, both were approximately 140-fold less active.

Characterization of "Peak A"

Under incubation conditions which were optimal for lO-keto-Ds pro­ duction by mixed cultures of rumen bacteria (7), Clostridium sp. strain 94

A264 produced several new metabolites which had absorption maxima at

254 nm (Fig. 4). The major product, designated "Peak A," was isolated and purified. When -vitamin D3 was used as the substrate, a major peak in radioactivity comigrated with "Peak A," indicating that this product originated from the parent [^H]-vitamin D3 (Fig. 5). Extensive metabolism to other compounds by Clostridium sp. strain A264 was noted by the authors (Fig. 4).

In contrast to the results of Gardner et al. (7) for lO-keto-Dg, no "Peak A" could be detected in media without cells (Table 7). How­ ever, similar to results obtained with lO-keto-Dg production (7), "Peak

A" production was also stimulated by the addition of small amounts of molecular Og. The catalytic factor(s) responsible for "Peak A" produc­ tion are somewhat heat-stable because "Peak A" was produced, at a re­ duced rate, following heat treatment (Table 7). Only trace amounts of

"Peak A" could be detected in rumen fluid incubations with vitamin D3

(data not shown). Interestingly, when 2 ug of "Peak A" was added to rumen fluid and incubated under identical conditions (10% O2 in CO2, at

39°C for 18 h), only 10 to 25% could be recovered (data not shown).

Thus, "Peak A" may be produced by rumen bacteria, but is not stable, and is probably metabolized further.

The U.V. spectrum of purified "Peak A" revealed an E at 254 nm max and an at 220 nm (Fig. 6). This absorption maximum is close to that of vitamin D3 (264 nm) and may indicate that the triene double bond system is still intact in "Peak A." Preliminary mass spectral data indicated that the parent ion is at m/z 416, which represents an 95

increase of 32 mass units from the original vitamin D3 substrate.

"Peak A" is located in between 25-OH-D3 and lO-keto-Dg in polarity on

ODS or silica columns (Fig. 4). 5(E)-lO-Keto-Da was isolated from the plasma of calves receiving

large internal doses of vitamin D3. Structural confirmation was

achieved by mass spectroscopy, comigration with synthetic standards and

U.V. analysis. The presence of this compound, and not the 5(Z) isomer,

in bovine plasma is consistent with the findings of Napoli et al. (17),

who demonstrated that only the 5(E) isomer is produced by bacterial

incubations in the absence of light. The source of this compound in

plasma is of considerable interest in view of recent evidence that an

analogous compound, 10-keto-25-0H-D3, is produced by kidney (13, 21)

and phagocytic cells (6, 9). Two lines of evidence suggest that the

oxidative cleavage at position 10 on the A—ring was achieved prior to

entry into systemic circulation: (1) The compound 10-keto—25-OH-D3 was not present in the plasma of calves that received large oral doses of

vitamin D3, albeit circulating concentrations of 25-OH-D3 were greater

than 1000 ng/ml (normal range is 40-50 ng/ml). In view of this result,

the C-10(19) oxidative pathway in the kidney, at least for 25-OH-D3, appears to be nonfunctional vivo and therefore not physiologically significant. (2) C-10(19) oxidation by rumen microbes has been demon­ strated vitro (17). Since vitamin D3 was given orally, it was the only compound to be subjected to microbial metabolism. Hence, 10- keto-D3, shown by us to be present in blood, is probably a direct re­ sult of bacterial metabolism of vitamin D in the rumen and absorption of this compound in the intestine. 97

When tested in vivo in rats, the natural 5(E) of lO-keto-Dg isomer

was approximately 100-fold less potent than vitamin D3 at stimulating

intestinal calcium absorption. This isomer did not cause bone calcium

resorption at the doses tested. The nonnatural 5(Z) isomer, however,

showed slightly greater calcium absorptive activity and minor bone

resorptive activity. Both the 5(E) and 5(Z) isomers of 10-keto-

25-OH—D3 did not appear to have any significant effect on either the

stimulation of calcium absorption from the gut or the mobilization of

bone calcium. Interestingly, "Peak A" could stimulate calcium absorp­

tion (although 20-fold less than vitamin Dg), but could not affect bone

calcium mobilization at the doses tested. The potential for such a

metabolite in the treatment of osteoporosis and other bone diseases, in which normal calcium absorption from the gut is needed, should be ex­

plored further.

Circulating, lO-keto-Dg is rapidly cleared. The plasma T^y^ was approximately 6 min. In addition, the metabolism of lO-keto-Dg to other metabolites vivo, especially to IS-nor-lO-OH-Dg or 19-nor-

10-keto-25-0H-D3, could not be detected (data not shown). The rapid plasma turnover can be partially explained by the poor binding of lO-keto-Ds to the plasma vitamin D binding protein (Table 5). Gen­ erally, compounds with higher affinity for this protein (e.g.,

25-OH-D3) have longer plasma half-lives than compounds with relatively low affinities [e.g., vitamin D3 and 1,25-(0H)2D3] (18). The produc­ tion of this compound, therefore, represents a terminal event - result­ ing in detoxification of large internal doses of vitamin Dy. 98

This hypothesis is also supported by the biological activity data.

The naturally-occurring 5(E) isomer of lO-keto-Ds was biologically inert, in comparison to vitamin D3, at stimulating ^ vivo intestinal calcium absorption and bone calcium resorption. This also applies to

10-keto-25-0H-D3 and "Peak A." The lack of iji vivo biologic activity can be partially explained by the lack of affinity for the 1,25-(0H)2D3 receptor. Both the 5(E) and the 5(Z) isomers of lO-keto-Dg were ap­ proximately 5 X 10^ times less competitive than 1,25-(OH)203 for bind­ ing to the 1,25-(0H)2D3 receptor protein. "Peak A" had no measurable affinity and 10-keto-25-0H-D3 was approximately 150-fold less active than 1,25-(0H)2D3 in receptor-binding affinity. However, when tested 45 in vitro for bone Ca resorptive activity in rat fetal long bone, the lO-keto—D3 metabolites were at least as active as 25-OH-D3 and markedly more active than vitamin D3 (23). The vivo and vitro biological activities of vitamin D compounds have generally been highly correlated with affinities of the metabolites for the vitamin D receptor (18).

Functionally, the metabolite—receptor complex is thought to interact with nuclear components, ultimately leading to transcriptional activa­ tion. The present observation of a dissociation between receptor- binding and in vitro bone-resorptive activities may suggest that: (1) bone resorption is not totally a receptor-mediated event; (2) the bone has a different population of receptors compared to the intestine or thymus, with different affinities for vitamin D metabolites; or (3) the

10—keto-Ds is being metabolized to a compound(s) that binds more avidly to the vitamin D receptor in the in vitro bone culture assays used. 99

No "Peak A" could be detected in the plasma of vitamin Dg-toxic calves, nor could it be detected after in vitro incubations in strained rumen fluid that produced lO-keto-Dg. It will be recalled that "Peak

A" is a product obtained in a pure culture system and in an environment such as rumen fluid (a mixed bacterial population), it might not be stable and may be metabolized by other bacteria. Preliminary ^ vitro experiments with rumen fluid incubated with "Peak A" demonstrated that

75 to 85% of the original starting material could not be recovered; and yet the chromatography system is highly efficient in recovery of both

10-keto—Da and "Peak A" (approximately 76% recovery for "Peak A").

These results support the concept that "Peak A" would not be stable in vivo and, therefore, would be difficult to detect.

In light of the anaerobic environment in the rumen, the participa­ tion of O2 in the production of both lO-keto-Dg and "Peak A" was a sur­ prising result; however, several investigators have noted a stimulatory effect of O2 on bile acid oxidation by anaerobic bacteria (2, 14, 16).

As in our studies, the mechanism of oxidation and the concentration needed were ill-defined. Macdonald et al. (14) proposed a mechanism to explain the formation of 3-keto derivatives of bile, which involves intracellular ratios of NADP^/NADPH resulting from a higher Eh under O2 atmospheres. Not enough is known about the C-10(19) oxidation of vita­ min D to extrapolate from the work of Macdonald et al. (14) at this time.

In summary, the compound lO-keto-Dg was established as an in vivo

metabolite of vitamin D3 in animals intoxicated with vitamin D3. These 100

observations are an extension of in vitro studies which demonstrated that lO-keto-Dg was a metabolite of vitamin D3 produced by rumen bacte­ ria. The low plasma values and rapid plasma turnover in conjunction with its lack of biological activity suggests that the C-10(19) oxida­ tion of vitamin D3 is a deactivation pathway which may play a role in allowing ruminants to tolerate relatively large oral doses of vitamin

D. The quantitative importance of this newly described pathway is cur­ rently under investigation.

In addition to any vivo role these metabolites may play, these transformation products and the responsible bacteria may have industri­ al importance in the production of new metabolites of vitamin D and other biologically active steroids. These reactions may provide desir­ able starting material or precursors for organic chemical syntheses, of new biologically-active metabolites. 101

240 260 280 300 320 340 WAVELENGTH (nm)

Figure 1. U.V. spectrum of 19-nor-lO-keto-vitamin D3 isolated from plasma of vitamin D-toxic calves 102

135 100-1

107

50 — 177 159 119 147 343

Z73 386

191 237 259 222 301 35637' 0-J lllm llJll T I I I I 1 i 1 50 100 150 200 250 300 350 400 M/E

Figure 2. Low-resolution electron impact mass spectrum of 19-nor-lO- keto-vitamin D3 extracted from plasma of vitamin D-toxic calves 103

10.0

CE: LU H- _J LU

0 100 200 300 400 500 600 700 800 MINUTES

Figure 3. The disappearance of 10-keto-19-nor-[^H]-vitamin D3 from plasma 104

0.5 n 254 nm

0.4-

0.3-

0.2-

02-1 B

0.1-1 C 25-HYDfiOXYVrraMIN 0, 5(E)-19-ror-IO-lieto-vrraMIN 0,

0 4 8 12 16 20 24 28 32 36 40 44 48 VOLUME (mis)

Figure 4. Bacterial metabolites of vitamin D3. (A) An elution profile of the lipid-soluble extract taken from an 18-h incubation of an overnight culture of Clostridium sp. strain A264 with 400 pg vitamin D3 and a 4% 0% in CO2 atmosphere. This chromatograph was developed by reverse—phase HPLC in methanol/water (90/10). (B) Same as treatment "A," except no vitamin D3 was added. (C) The elution profile of two standards, 25-hydroxyvitamin D3 and 5(E)-19-nor-10-keto- vitamin D3, developed on the same reverse-phase system 105

21,000- k 12 Q 18,000-

-g 15,000- >- 12,000-^ .8 S3

9000 H

6000 H

3000 H t. 0

VOLUME (mis)

Figure 5. An elution profile of the lipid-soluble extract taken from an 18-h incubation of an overnight culture of Clostridium sp. strain A264 with 1 x 10® cpm of []-vitamin D3 and 300 ug of vitamin Da. A 4% O2 in CO2 atmosphere was used 106

0.8-1

0.7-

0.6 -

0.5-

0.3-

0.2-

T 360 340 WAVELENGTH IN nm

Figure 6. An ultraviolet spectrum of "Peak A" that was purified from in vitro incubations with Clostridium sp. strain A264. "Peak A" had an E of 254 nm and an E . of 220 nm max mxn 107

Table 1. Bone calcium mobilization and intestinal calcium transport in rats administered 19-nor—10-keto-vitamin D3 and vitamin D3

^^Ca serosal/^^Ca Plasma Treatment mucosal calcium

% Control Control (17) 100 ± 3.5 100 ± 1.9

Vitamin D3

50 ng (14) 170 ±7.3 117 ±2.6

5(Z)-19-Nor-10-keto-vitamin D3

2000 ng (5) 150 ± 15 107 ±2.7*

4000 ng (10) 201 ± 12 102 ± 2.2

5(E)-19-Nor-10-keto-vitamin D3

2000 ng (5) 118 ±5.1 101 ± 4.3

4000 ng (5) 161 ± 13.4 99 ± 2.8

Values significantly different from control (P < .05) are indi­ cated by the letter "A".

The numbers in parentheses represent the number of rats/treatment group. 108

Table 2. Bone calcium mobilization and intestinal calcium transport in rats administered "Peak A" and vitamin D3

^^Ca serosal/^^Ca Plasma Treatment mucosal calcium

% Control B,C Control (5) 100 ± 14 100 ± 10

Vitamin D3

50 ng (6) 163 ± 23 119 ± 4

"Peak A'

100 ng (6) 120 23*'= 96 SC.D

500 ng (6) 137 14® 99 gE.D

1000 ng (5) 164 + 15^ 108 9®

"Peak A" + vitamin D3

500 ng + 50 ng (6) 175 ± 3r 125 ± 8

Numbers in parentheses represent the number of rats/treatment group.

^All values with the same letter are not statistically different (P < .05). 109

Table 3. Bone calcium mobilization and intestinal calcium transport in rats administered 19-nor—lO-keto-25-OH-vitamin D3 and 25-OH-vitamin D3

^^Ca serosal/^^Ca Plasma Treatment mucosal calcium

% Control Control (6) 100 ± 17 B 100 ± 9 B

25-OH-vitamin D3

25 ng (6) 179 ± 4f 114 ± 5^

19-Nor—10—keto-25-OH-vitamin Dg

100 ng (6) 84 10® 100 9®

250 ng (6) 95 23® 92 -r 7®

500 ng (6) 97 20® 94 7®

^Number in parentheses represent the number of rats/treatment group.

All values with the same letter are not statistically different (P < .05). 110

Table 4. Bone calcium mobilization and intestinal calcium transport in rats administered 19-nor-10-keto-25-0H—vitamin D3 and 25-OH-vitamin D3 daily^

^^Ca serosal/^^Ca Plasma Treatment mucosal calcium

% Control B Control (6) 100 ± 20 100 ± 10

25-OH-vitamin D3

25 ng (6) 419 ± 83 133 ± 13'

19—Nor-lO-keto-25-OH—vitamin D3

100 ng (6) 141 ± 41 99 ± 7 B

Rats were injected daily for 4 days.

^The numbers in parentheses represent the number of observations/ treatment.

All values with the same letter are not statistically different (P < .05). Ill

Table 5. Ability of vitamin D3 metabolites to displace 25-OH-[^H]- vitamin D3 from the 4.2 S rat plasma vitamin D binding protein

Amount that produces 50% Competitive index displacement of relative to Metabolite 25-OH-[^H]-vitamin D3 25-OH-vitamin D3 (ng) (%)

25-OH-vitamin D3 0.60 100.0

5(Z)-19-nor-lO-keto- 25-OH-vitamin D3 20.0 3.0

5(E)-19-nor-lO-keto- 25-OH-vitamin D3 20.0 3.0

1,25-(0H)2-vitamin D3 30.0 2.0

Vitamin D3 32.0 1.9

5(Z)-19-nor-lO-keto- vitamin D3 200.0 0.3

5(E)-19-nor-lO-keto- vitamin D3 4000.0 0.015

'Peak A" No activity 0.0

^The amount of 25-OH-vitamin D3 that will displace 50% of the 25-0H-[^H]-vitamin D3 divided by the amount of metabolite that will displace 50% of the 25-0H-[^H]-vitamin D3 times 100. 112

Table 6. Ability of vitamin D3 metabolites to displace 1,25-(OH)2- r^H]-vitamin D3 from the 3.75 S calf thymus receptor

Amount that produces 50% Competitive index displacement of 1,25- relative to 1,25- Metabolite (QH)2[^H]-vitamin D3 (OH)2-vitamin D3 (ng) (%)

1,25-(OH)2-vitamin D3 0.044 100.0

5(Z)-19-not-10-keto- 25-OH-vitamin D3 6.31 0.69

5(E)-19-nor-10-keto- 25-OH-vitamin D3 7.41 0.59

25-OH-vitamin D3 7.59 0.58

Vitamin D3 4600.0 0.00094

5(Z)-19-nor-lO-keto- vitamin D3 21200.0 0.00021

5(E)-19-nor-10-keto- vitamin D3 23500.0 0.00019

"Peak A" No activity 0.0

^The amount of 1,25-(OH)2-vitamin D3 that will displace 50% of the l,25-(0H)2-[^H]-vitamin D3 divided by the amount of metabolite that will displace 50% of the 1,25-(OH)2-[^H]-vitamin D3 times 100. 113

Table 7. A comparison of different preincubation treatments on "Peak A" production by Clostridium sp. strain A264

Treatment Amount of "Peak A" produced (ng)

Sterile, uninoculated BHI broth under Not detected 4% OZ

Overnight culture of Clostridium sp. strain A264 incubated under 4% O2 6307 ± 610'

Overnight culture of Clostridium sp. strain A264 incubated under anaerobic conditions 1933 ± 212 B

Overnight culture of Clostridium sp. strain A264 heat-killed and incu­ bated under 4% O2 883 ± 152'

^All samples were incubated 18 h in the presence of vitamin D3.

^All values with the same letter are not statistically different (P < .05). 114

LITERATURE CITED

1. Argenzio, R. A., and C. E. Stevens. 1984. The large bowel—a supplementary rumen. Proc. Nutr. Soc. 43:13-23.

2. Aries, V., and M. J. Hill. 1970. Degradation of steroids by- intestinal bacteria. Enzymes catalyzing the oxidoreduction of the 3a, 7a, and 12a-hydroxyl groups in cholic acid, the dehydroxylation of the 7-hydroxyl group. Biochim. Bxophys. Acta 202:535-543.

3. Arnaud, S. B., R. S. Goldsmith, P. W. Lambert, and V. L. W. Go. 1975. 25-Hydroxyvitamin D3: Evidence of an enterohepatic circulation in man. Proc. Soc. Exp. Biol. Med. 149:570-572.

4. Barton, D. H. R., and H. Patin. 1976. Chemistry of the tricar- boxyliron complexes of calciferol and ergosterol. J. Chem. Soc. Perkin Trans. 1:829-831.

5. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37:911-917.

6. Cohen, M. S., and T. K. Gray. 1984. Phagocytic cells metabolize 25-hydroxyvitamin D3 in vitro. Proc. Nat. Acad. Sci. 81:931-934.

7. Gardner, R. M., M. J. Allison, P. A. Hartman, T. A. Reinhardt, I. M. Robinson, and R. L. Horst. 1985. Factors affecting the production of 5(E) 19-nor-lO-keto-vitamin D3 by rumen bacteria, Appl. Environ. Microbiol. (Submitted for publication)

8. Goldsmith, R. S. 1982. Enterohepatic cycling of vitamin D and its metabolites. Min. Electrolyte Metab. 8:289-292.

9. Gray, T. K., F. W. Maddux, G. E. Lester, and M. E. Williams. 1982. Rodent macrophages metabolize 25-hydroxyvitamin D3 in vitro. Biochem. Biophys. Res. Commun. 109:723-729.

10. Homes, R. P., and F. A. Kummerow. 1983. The relationship of adequate and excessive intake of vitamin D to health and disease. J. Am. Coll. Nutr. 2:173-199.

11. Horst, R. L., E. T. Littledike, J. L. Riley, and J. L. Napoli. 1981. Quantitation of vitamin D and its metabolites and their plasma concentration in five species of animals. Anal. Biochem. 116:189-203. 115

12. Hungate, R. E. 1969. A roll tube method for cultivation of strict anaerobes. Pages 117-132 in J. R. Norris and D. W. Ribbons eds. Methods in Microbiology, vol. 3B. Academic Press, Inc., New York.

13. Lester, G. E., R. L. Horst, and J. L. Napoli. 1984. Formation of 19-nor-10-keto-25-hydroxyvitamin D3 in cultured mammalian cells. Biochem. Biophys. Res. Commun. 120:919-925.

14. Macdonald, I. A., D. M. Hutchison, T. P. Forrest, V. D. Bokken- heuser, J. Winter, and L. V. Holdeman. 1983. Metabolism of primary bile acids by Clostridium perfringens. J. Steroid Biochem. 18:97-104.

15. Martin, D. L., and H. F. DeLuca. 1969. Influence of sodium on calcium transport by the rat small intestine. Am. J. Physiol. 216:1351-1359.

16. Masuda, N., and H. Oda. 1983. 7a-Dehydroxylation of bile acids by resting cells of an unidentified, gram-positive, nonspore- forming anaerobic bacterium. Appl. Environ. Microbiol. 45:456-462.

17. Napoli, J. L., J. L. Sommerfeldt, B. C. Pramanik, R. Gardner, A. D. Sherry, J. J. Partridge, M. R. Uskokovic, and R. L. Horst. 1983. 19-Nor-lO-keto-vitamin D derivatives: Unique metabolites

of vitamin D3, vitamin , and 25-hydroxyvitamin D3. Biochemistry - 22:3636-3640.

18. Norman, A. W. 1979. Metabolism of vitamin D. Pages 134-197 in W. J. Darby, ed. The Calcium Homeostatic Steroid Hormone. Academic Press, Inc., New York.

19. Reinhardt, T. A., R. L. Horst, E. T. Littledike, and D. C. Beitz. 1982. 1,25-Dihydroxyvitamin D3 receptor in bovine thymus gland. Biochem. Biophys. Res. Commun. 106:1012-1018.

20. Seelig, S. M. 1983. Vitamin D—risk vs. benefit. J. Am. Coll. Nutr. 2:109-110.

21. Simpson, R. U., J. K. Wichmann, H. E. Paaren, H. K. Schnoes, and H. F. DeLuca. 1984. Metabolism of 25-hydroxyvitamin D3 by rat kidney cells in culture: Isolation and identification of cis- and trans-19-nor-10-oxo-25-hydroxyvitamin D3. Arch. Biochem. Biophys. 230:21-29.

22. Sommerfeldt, J. L., J. L. Napoli, E. T. Littledike, D. C. Beitz, and R. L. Horst. 1983. Metabolism of orally administered [^H]- ergocalciferol and [^H]-cholecalciferol by dairy calves. J. Nutr. 113:2595-2600. 116

23. Stern, P. H., R. L. Horst, R. Gardner, and J. L. Napoli. 1984. Formation of 19-nor-10-keto-25-hydroxyvitamin D3 in cultured mammalian cells. Biochem. Biophys. Res. Commun. (In press)

24. Willis, J. B. 1960. The determination of metals in blood serum by atomic absorption spectroscopy. I. Calcium. Spectrochima. Acta 16:259-272. 117

SUMMARY AND DISCUSSION

Recently, it was discovered that in the presence of rumen bacteria vitamin D3 (a 9,10 seco-steroid) undergoes C-10(19) oxidation to 10— keto-Dg. This compound was the first documented bacterial metabolite of vitamin D (a eucaryotic vitamin), but relatively little was known about factors influencing its production or its biological signifi­ cance. Our in vitro studies in Section I clearly demonstrate that the

C-10(19) oxidative pathway for vitamin D3 is unquestionably potentiated by bacteria. However, the discovery that lO-keto-Ds could also be pro­ duced by the autoxidation of cysteine or in heat-treated cultures tended to support a nonenzymatic mechanism. We propose that some an­ aerobic bacteria are capable of producing at least two heat-stable fac­ tors (one, <10,000 daltons; the other, >10,000 daltons) which catalyze theTproduction of oxygen radicals. These radicals, in turn, are di­

rectly responsible for the observed oxidation of vitamin D3. The in vitro requirement for molecular O2 by anaerobic bacteria tends to sup­ port this mechanism. If O2 is also an obligate iji vivo requirement, lO-keto-Dg production could occur at the epithelial wall of the gut where 0% is available.

Our studies also resulted in the isolation of Clostridium sp. strain A264, a bacterium capable of generating lO-keto—D3. With this system, we were able to greatly optimize the production of these com­ pounds needed for biological assays.

Section II demonstrated that lO-keto-Ds can be detected in the plasma of vitamin Ds-toxic calves. The short half-life observed for 118

lO-keto-Ds, coupled with its biological activity in vivo and ^ vitro, support that this oxidation is a deactivation pathway in vitamin D3 metabolism.

The metabolite "Peak A" is of particular interest because this compound could not be detected in the absence of cells or by the autoxidation of cysteine, thereby denoting a more direct relationship with bacteria. Interestingly, the 0% requirement demonstrated for lO-keto-Dg production also appeared for "Peak A." Although the role of

O2 in lO-keto-Dg production can be explained by the presence of oxi­ dative radicals, its role in producing "Peak A" is not clear.

The extent to which the bacterial flora interact with the vitamin

D endocrine system under physiological levels of vitamin D is not known, but given the inert biological nature of these metabolites, they are probably significant only under conditions where the animal is challenged by toxic levels of vitamin D. In this situation, the bacte­ rial transformation of vitamin D3 to lO-keto-Ds and "Peak A" serve to deactivate vitamin Dg's biological activity, thus reducing the severity of the symptoms associated with toxicity.

Future studies in this area should explore the metabolite "Peak

A," which may have some use in the treatment of bone diseases marked by excessive calcium resorption. The potential interaction of bacteria with vitamin D in the enterohepatic cycle should also be examined in gnotobiotic animals and in animals on antibiotic regimes. In addition, the several new bacterial metabolites recently isolated should be char­ acterized and tested for biological activity. 119

LITERATURE CITED

1. Adlercreutz, H., R. Martin, M. Pulkkinen, H. Denker, U. Rinner, N-0. Sjoberg, and M. J. Tikkanen. 1976. Intestinal metabolism of estrogens. J. Clin. Endocrinol. Metab. 43:497-505.

2. Adlercreutz, H., M. 0. Pulkkinen, E. K. Hamalainen, and J. T. Korpela. 1984. Studies on the role of intestinal bacteria in metabolism of synthetic and natural steroid hormones. J. Steroid Biochem. 20:217-229.

3. Archer, R. H., I. S. Maddox, and R. Chong. 1981. 7a-Dehydroxy- lation of cholic acid by Clostridium bifermentans. Eur. J. Appl. Microbiol. Biotechnol. 12:46-52.

4. Archer, R. H., I. S. Maddox, and R. Chong. 1982. Transformation of cholic acid by Clostridium bifermentans. J. Appl. Bacteriol. 52:49-56.

5. Argenzio, R. A., and C. E. Stevens. 1984. The large bowel—a supplementary rumen. Proc. Nutr. Soc. 43:13-23.

6- Aries, V., and M. J. Hill. 1970. Degradation of steroids by intestinal bacteria. Enzymes catalysing the oxidoreduction of the 3a,7a-, and 12a—hydroxy groups in cholic acid, and the dehydroxy- lation of the 7-hvdroxyl group. Biochim. Biophys. Acta ' 202:535-543.

7. Aries, V., and M. J. Hill. 1970. Degradation of steroids by intestinal bacteria. Deconjugation of bile salts. Biochim. Biophys. Acta 202:526-534.

8. Arnaud, S. B., R. S. Goldsmith, P. W. Lambert, and V. L. W. Go. 1975. 25-Hydroxyvitamin D^: Evidence of an enterohepatic circu­ lation in man. Proc. Soc. Exp. Biol. Med. 149:570-572.

9. Beher, W. T., S. Stradnieks, and G. J. Lin. 1983. Rapid photo­ metric determination of free and esterified 3a—hydroxy fecal bile acids. Steroids 41:729-740.

10. Bjorklem, I., and J. A. Gustafsson. 1971. Mechanism of microbial transformation of cholesterol into coprostanol. Eur. J. Biochem. 21:428-432.

11. Bjorklem, I., J. Gustafsson, and 0. Wrange. 1973. Microbial transformation of cholesterol into coprostanol. Properties of a 3-oxo-A'^-steroid-g-reductase. Eur. J. Biochem. 37:143-147. 120

12. Bokkenheuser, V. D., G. N. Morris, A. E. Ritchie, L. V. Holdeman, and J. Winter. 1984. Biosynthesis of androgen from Cortisol by a species of Clostridium recovered from fecal flora. J. Infect. Dis. 149:489-494.

13. Bokkenheuser, V. D., J. B. Suzuki, S. B. Stuart, B. Polovsky, J. Winter, and W. G. Kelly. 1975. Metabolism of deoxycorticosterone by human fecal flora. Appl. Microbiol. 30:82—90.

14. Bokkenheuser, V. D., J. Winter, B. I. Cohen, S. O'Rourke, and E. H. Mosbach. 1983. Inactivation of contraceptive steroid hormones by human intestinal Clostridia, J. Clin. Microbiol. 18:500-504.

15. Bokkenheuser, V. D., J. Winter, S. M. Finegold, V. L. Sutter, A. E. Ritchie, W. E. C. Moore, and L. V. Holdeman. 1979. New markers for Eubacterium lentum. Appl. Environ. Microbiol. 37:1001-1006.

16. Bokkenheuser, V. D., J. Winter, S. O'Rourke, and A. E. Ritchie. 1980. Isolation and characterization of fecal bacteria capable of 16a-dehydroxylating corticoids. Appl. Environ. Microbiol. 40:803-808.

17. Bordier, P., H. Rasmussen, P. Marie, L. Miravet, J. Gueris, and A. Ryckwaert. 1978. Vitamin D metabolites and bone mineralization in man. J. Clin. Endocrinol. Metab. 46:284-294.

18. Brinkley, A. W., A. R. Gottesman, and G. E. Mott. 1982. Isola­ tion and characterization of new strains of cholesterol reducing bacteria from baboons. Appl. Environ. Microbiol. 43:86—89.

19. Brydon, W. G., K. Tadesse, M. A. Eastwood, and M. E. Lawson. 1980. The effect of dietary fibre of bile acid metabolism in rats. Br. J. Nutr. 43:101-106.

20. Calvin, H. I., and S. Lieberman. 1962. Studies on the metabo­ lism of 16a-hydroxyprogesterone in humans. Conversion to urinary 17-isopregnanolone. Biochemistry 1:639-645.

21. Charney, W., and H. L. Herzog. 1967. Chemical classification of microbial transformations of steroids. Pages 14-73 in Microbial Transformations of Steroids. Academic Press, New York.

22. Chen, P. S., and H. B. Bosmann. 1964. Effect of vitamin Dg and D3 on serum calcium and phosphorus in rachitic chicks. J. Nutr. 83:133-139. 121

23. Chen, T. C., J. C. Weber, and H. ?. DeLuca. 1970. On the sub­ cellular location of vitamin d metabolites in intestine. J. Biol. Chem. 245:3776-3780.

24. Clark, S. A., W. E. Stumpfy, M. Sar, H. F. DeLuca, and Y. Tanaka. 1980. Target cells for 1,25-dihydroxyvitamin Dg in the pancreas. Cell Tiss. Res. 209:515-520.

25. Cohen, M. S., and T. K. Gray. 1984. Phagocytic cells metabolize 25-hydroxyvitamin Dg in vitro. Proc. Nat. Acad. Sci. USA 81:931-934.

26. Crowther, J. S., B. S. Drasar, P. Goddard, M. J. Hill, and K. Johnson. 1973. The effect of a chemically defined diet on the faecal flora and faecal steroid concentration. Gut 14:790—793.

27. Cummings, J. H. 1983. Fermentation in the human large intestine: Evidence and implications for health. Lancet 1:1206-1209.

28. Dada, 0. A., and 0. 0. Martins. 1983. Drug effects on the intestinal absorption of estrogens. J. Steroid Biochem. 19:821-825.

29. Danielsson, H., P. Eneroth, K. Hellstrom, and J. Sjovall. 1962. Synthesis of some 36-hydroxylated bile acids and the isolation of 3S,12a-dihydroxy-56-cholanic acid from feces. J. Biol. Chem. 237:3657-3659.

30. Danielsson, H., A. Kallner, and J. Sjovall. 1963. On the compo­ sition of the bile acid fraction of rabbit feces and the isolation of a new bile acid: 3ct,12a-dihydroxy-5ct-cholanic acid. J. Biol. Chem. 238:3846-3852.

31. DeLuca, H. F. 1976. Recent advances in our understanding of the vitamin D endocrine system. J. Lab. Clin. Med. 87:7-22.

32. DeLuca, H. F. 1979. The vitamin D system in the regulation of calcium and phosphorus metabolism. Nutr. Rev. 37:161-193.

33. DeLuca, H. F. 1980. Some new concepts emanating from a study of the metabolism and function of vitamin D. Nutr. Rev. 38:169-182.

34. DeLuca, H. F. 1983. Metabolism and mechanism of action of vita­ min D—1982. Pages 7-73 in W. A. Peck, ed. Bone and Mineral Research. Excerpta Medica, Amsterdam.

35. DeLuca, H. F., and H. K. Schnoes. 1983. Vitamin D: Recent advances. Ann- Rev. Biochem. 52:411-439. 122

36. Deuland, S., J. I. Pederson, P. Helgerud, and C. A. Drevon. 1982. Transport of vitamin from rat intestine. Evidence for transfer of vitamin Dg from chylomicrons to a-globulins. J. Biol. Chem. 257:146-152.

37. Edenharder, R., and H. Deser. 1981. The significance of the bacterial steroid degradation for the etiology of large bowel cancer. VIII. Transformation of cholic-, chenodeoxycholic-, and deoxycholic acid by lecithinase-lipase-negative Clostridia. Zbl. Bakt. Hyg. 174:91-104.

38. Eyssen, H. 1973. Role of the gut microflora in metabolism of lipids and sterols. Proc. Nutr. Soc. 32:59-63.

39. Eyssen, H., G. DePauw, and G. Paramentier. 1974. Effect of lac­ tose on A^-steroid-reducing activity of intestinal bacteria in gnotobiotic rats. J. Nutr. 104:605-612.

40. Eyssen, H., G. DePauw, J. Stragier, and A. Verhulst. 1983. Cooperative formation of co-muricholic acid by intestinal micro­ organisms. Appl. Environ. Microbiol. 45:141-147.

41. Eyssen, H., and G. Paramentier. 1974. Biohydrogenation of sterols and fatty acids by the intestinal microflora. Am. J. Clin. Nutr. 27:1329-1340.

42. Eyssen, H., G. Parmentier, F. C. Compernolle, G. DePauw, and K. Piessens-Denef. 1973. Biohydrogenation of sterols by Eubacterium ATCC 21,408—nova species. Eur. J. Biochem. 36:411-421.

43. Eyssen, H., and A. Verhulst. 1984. Biotransformation of linoleic acid and bile acids by Eubacterium lentum. Appl. Environ. Microbiol. 47:39-43.

44. Fraser, D. R. 1980. Regulation of the metabolism of vitamin D. Physiol. Rev. 60:551-613.

45. Ghazarian, J. G., C. R. Jefcoate, J. C. Knutson, W. H. Orme- Johnson, and H. F. DeLuca. 1974. Mitochondrial cytochrome P-450: A component of chick kidney 25-hydroxycholecalciferol-la—hydroxy­ lase. J. Biol. Chem. 249:3026-3033.

46. Glass, T. L., J. Winter, V. D. Bokkenheuser, and P. B. Hylemon. 1982. Biotransformation of 16a—hydroxyprogesterone by Eubacterium sp. 144: Non-enzymatic addition of L-cysteine to A^®-pro- gesterone. J. Lipid Res. 23:352-356.

47. Goldblatt, H., and K. N. Soames. 1923. A study of rats on a normal diet irradiated by the mercury vapor quartz lamp or kept in darkness. Biochem. J. 17:294. 123

48. Goldsmith, R. S. 1982. Enterohepatic cycling of vitamin D and its metabolites. Min. Electrolyte Metab. 8:289-292.

49. Gray, R. W., A. E. Caldas, D. R. Wilz, J. Lemann, G. A. Smith, and H. F. DeLuca. 1978. Metabolism and excretion of [^H]-1,25-(0H)2 vitamin D3 in healthy adults. J. Clin. Endocrinol. Metab. 46:756-765.

50. Greaves, J. D., and C. L. A. Schmidt. 1933. The role played by bile in the absorption of vitamin D in the rat. J. Biol. Chem. 102:101-112.

51. Gustafsson, B. E., T. Midtvedt, and A. Norman. 1965. Isolated fecal microorganisms capable of 7a-dehydroxylating bile acids. J. Exp. Med. 123:413-432.

52. Haddad, J. C., and T. C. B. Stamp. 1974. Circulating 25-hydroxy- vitamin D in man. Am. J. Med. 57:57-62.

53. Hayakawa, S., T. Fujii, Y. Saburi, and T. Eguchi. 1957. Micro­ biological degradation of cholic acid. Nature 179:537-538.

54. Hayakawa, S., and B. Samuelsson. 1964. Transformation of cholic acid in vitro by Corynebacterium simplex. J. Biol. Chem. 239:94-97.

55. Henry, u. L., and A. W. Norman. 1978. Vitamin D: Tt-70 dihydroxy- lated metabolites are required for normal chicken egg hatchabil- ity. Science 201:835-837.

56. Henry, H. L., A. N. Taylor, and A. W. Norman. 1977. Response of chick parathyroid glands to the vitamin D metabolites 1,25-dihy- droxyvitamin D3. J. Nutr. 107:1918-1926.

57. Hess, A. F. 1929. Rickets, Osteomalacia and Tetany. Lea and Febiger, Philadelphia.

58. Hill, M. J. 1974. Steroid nuclear dehydrogenation and colon cancer. Am. J. Clin. Nutr. 27:1475-1480.

59. Hill, M. J., B. S. Drasar, R. E. D. Williams, T. W. Meade, A. G. Cox, J. E. P. Simpson, and B. C. Morson. 1975. Fecal bile acids, Clostridia, and the etiology of cancer of the large bowel. Lancet 1:535-538.

60. Hino, T., and M. Kametaka. 1978. Metabolic origin of sterol in rumen protozoa. J. Gen. Appl. Microbiol. 24:173-184.

61. Hirano, S., and N. Masuda. 1981. Transformation of bile acids by Eubacterium lentum. Appl. Environ. Microbiol. 42:912-915. 124

62. Hirano, S., and N. Masuda. 1982. Characterization of NADP - dependent 7 6-hydroxysteroid dehydrogenase from Peptostreptococcus productus and Eubacterium aerofaciens. Appl. Environ. Microbiol. 43:1057-1063.

63. Hirano, S., N. Masuda, H. Oda, and H. Mukai. 1981. Transforma­ tion of bile acids by Clostridium perfringens. Appl. Environ. Microbiol. 42:394—399.

54. Hirano, S., R. Nakama, M. Tamaki, N. Masuda, and H. Oda. 1981. Isolation and characterization of thirteen intestinal micro­ organisms capable of 7a-dehydroxylating bile acids. Appl. Environ. Microbiol 41:737-745.

65. Holick, M. F., and M. B. Clark. 1978. The photobiogenesis and metabolism of vitamin D. Fed. Proc. 37:2567-2574.

66. Holick, M. F., J. A. Maclaughlin, and S. H. Doppelt. 1981. Regu­ lation of cutaneous previtamin Dg photosynthesis in man. Skin pigment is not an essential regulator. Science 211:590-593.

67. Homes, R. P., and F. A. Kummerow. 1983. The relationship of adequate and excessive intake of vitamin D to health and disease. J. Am. Coll. Nutr. 2:173-199.

68. Hopkins, G. J., and C. E. West. 1976. Possible roles of dietary fats in carcinogenesis. Life Sci. 19:1103-1116.

69. Horst, R. L. 1979. 25—0HD3-26,23-Lactone: A metabolite of

vitamin D3 that is 5 times more potent than 25-OHD3 in the rat plasma competitive protein binding radioassay. Biochem. Biophys. Res. Commun. 89:286-293.

70. Horst, R. L., and E. T. Littledike. 1980. 25-0HD3-26,23-Lactone: Demonstration of kidney—dependent synthesis in the pig and rat. Biochem. Biophys. Res. Commun. 93:149-154.

71. Horst, R. L., E. T. Littledike, J. L. Riley, and J. L. Napoli. 1981. Quantitation of vitamin D and its metabolites and their plasma concentration in five species of animals. Anal. Biochem. 116:189-203.

72. Horst, R. L., J. Napoli, and E. T. Littledike. 1982. Discrimina­ tion in the metabolism of ergocalciferol and cholecalciferol by the pig, rat, and chick. Biochem. J. 204:185-189.

73. Howard, G. A., R. T. Turner, D. J. Sherrard, and D. J. Baylink. 1981. Human bone cells in culture metabolize 25-hydroxyvitamin D3 to 1,25—dihydroxyvitamin D3 and 24,25-dihydroxyvitamin D3. J. Biol. Chem. 256:7738-7740. 125

74. Huijghebaert, S., G. Parmentier, and H. Eyssen. 1984. Speci­ ficity of bile salt sulfatase activity in man, mouse, and rat intestinal microflora. J. Steroid Biochem. 20:907—912.

75. Huldschinsky, K. 1919. Heilung von rachitis durch kuenstlicheho- hensonne. Dtsch. Med. Wochenschr. 45:712.

76. Hunt, R. D., F. G. Garcia, D. M. Hegsted, and N. Kaplinsky. 1967. Vitamin D2 and D3 in New World primates. I. Production and regression of osteodystrophia fibrosa. Lab. Anim. Care 17:222-234.

77. Hylemon, P. B., and J. A. Sherrod. 1975. Multiple forms of 7a- hydroxysteroid dehydrogenase in selected strains of Bacteroides fragilis. J. Bacteriol. 122:418-424.

78. Hylemon, P. B., and E. J. Stellwag. 1976. Bile acid biotrans­ formation rates of selected gram-positive and gram-negative intestinal anaerobic bacteria. Biochem. Biophys. Res. Commun. 69:1088-1094.

79. Kellogg, T. F. 1973. On the site of the microbiological reduc­ tion of cholesterol to coprostanol in the rat. Lipids 8:658-659.

80. Knutson, J. C., and H. F. DeLuca. 1974. 25-Hydroxyvitamin D3-24- hydroxylase: Subcellular location and properties. Biochemistry 13^1543-1548.

81. Kobashi, K., Y. Takeda, H. Itoh, and J. Hase. 1983. Cholesterol- lowering effect of Clostridium butyricum in cholesterol-fed rats. Digestion 26:173-178.

82. Lawson, D. E. M. 1978. Biochemical responses of the intestine to vitamin D. Pages 167-196 in D. E. M. Lawson, ed. Vitamin D. Academic Press, London.

83. Lawson, D. E. M., P. W. Wilson, and E. Kodicek. 1969. Metabolism of vitamin D: A new cholecalciferol metabolite, involving loss of hydrogen at C-1, in chick intestinal nuclei. Biochem. J. 115:269-277.

84. Leppik, R. A., R. J. Park, and M. G. Smith. 1982. Aerobic catabolism of bile acids. Appl. Environ. Microbiol. 44:771-776.

85. Lester, G. E., R. L. Horst, and J. L. Napoli. 1984. Formation of 19-nor-10-keto-25-hydroxyvitamin D3 in cultured mammalian cells. Biochem. Biophys. Res. Commun. (In press). 126

86. Levitz, M., and J. Katz. 1968. Enterohepatic metabolism of estriol-3-sulfate-16-glucosiduronate in women, J. Clin. Endocrinol. Metab. 28:862—868,

87. Lipsky, R. H., and P. B. Hylemon. 1980. Characterization of a NADH:flavin oxidoreductase induced by cholic acid in a' 7a-dehy- droxylating intestinal Eubacterium species. Biochim. Biophys. Acta 612:328-336.

88. Littledike, E. T., R. L. Horst, R. W. Gray, and J. L. Napoli. 1980. Impaired 24,25-(OH)2D and 25-0HD3-26,23 lactone (lactone) production in anephric man and pig. Fed. Proc. 39:662. (Abstr.)

89. Macdonald, I. A., V. D. Bokkenheuser, J. Winter, A. M. McLeron, and E. H, Mosbach. 1983. Degradation of steroids in the human gut. J. Lipid Res. 24:675-700.

90. Macdonald, I. A., and D. M. Hutchison. 1982. Epimerization versus dehydroxylation of the 7a-hydroxyl group of primary bile acids: Competitive studies with Clostridium absonum and 7a- dehydroxylating bacteria (Eubacterium sp.). J. Steroid Biochem. 17:295-303.

91. Macdonald, I. A., D. M. Hutchison, and T. P. Forrest. 1981. Formation of urso- and ursodeoxy-cholic acids from primary bile acids by Clostridium absonum. J. Lipid Res. 22:458-466.

92." Macdonald, I. A., D. M. Hutchison, T. P. Forrest, V. D. Bokken­ heuser, J. Winter, and L. V. Holdeman. 1983. Metabolism of primary bile acids by Clostridium perfringens. J. Steroid Biochem. 18:97-104.

93. Macdonald, I. A., and P. D. Roach. 1981. Bile salt induction of 7a- and 76-hydroxysteroid dehydrogenase in Clostridium absonum. Biochim. Biophys. Acta 665:262-269.

94. Macdonald, I. A., G. Singh, D. E. Mahony, and C. E. Meier. 1978. Effect of pH on bile salt degradation by mixed fecal cultures. Steroids 32:245-256.

95. Macdonald, I. A., and J. D. Sutherland. 1983. Further studies on the bile salt induction of 7a- and 73-hydroxysteroid dehydrogenase in Clostridium absonum. Biochim. Biophys. Acta 750:397—403.

96. Macdonald, I. A., J. D. Sutherland, B. I. Cohen, and E. H. Mosbach- 1983. Effect of bile acid oxazoline derivatives on microorganisms participating in 7a-hydroxyl epimerization of primary bile acids. J. Lipid Res. 24:1550-1559. 127

97. Macdonald, I. A., B. A. White, and P. B. Hylemon. 1983. Separa­ tion of 7a— and 76-hydroxysteroid dehydrogenase activities from Clostridium absonum ATCC #27555 and cellular response of this organism to bile acid inducers. J. Lipid Res. 24:1119-1126.

98. Macdonald, I. A., C. N. Williams, J. D. Sutherland, and A. C. Macdonald. 1983. Estimation of ursodeoxycholic acid in human and bear biles using Clostridium absonum 76-hydroxysteroid dehydro­ genase. Anal. Biochem. 135:349-354.

99. Madhok, T. C., and H. F. DeLuca. 1979. Characteristics of the rat liver microsomal enzyme system converting cholecalciferol into 25-hydroxycholecalciferol. Biochem. J. 184:491-499.

100. Madhok, T. C., H. K. Schnoes, and H. F. DeLuca. 1977. Mechanism of 25-hydroxyvitamin D3 24-hydroxylation: Incorporation of oxygen-18 into the 24 position of 25-hydroxyvitamin D3. Biochemistry 16:2142-2145.

101. Martin, C. K. 1977. Cleavage of sterol side chains. Adv. Appl. Microbiol. 22:29-58.

102. Mastromarino, A., B. S. Reddy, and E. L. Wynder. 1976. Metabolic epidemiology of colon cancer: Enzymic activity of fecal flora. Am. J. Clin. Nutr. 29:1455-1460.

103. Masuda, N., and H. Oda. 1983. 7a-Dehydroxylation of bile acids by resting cells of an unidentified gram-positive, nonspore- forming anaerobic bacterium. Appl. Environ. Microbiol. 45:456—462.

104. Masuda, N., H. Oda, S. Hirano, M. Masuda, and H. Tanaka. 1984. 7a-Dehydroxylation of bile acids by resting cells of a Eubacterium lentum-like intestinal anaerobe, strain C-25. Appl. Environ. Microbiol. 47:735-739.

105. Masuda, N., H. Oda, S. Hirano, and H. Tanaka. 1983. Enhancement of the 7a-dehydroxylase activity of a gram-positive intestinal anaerobe by flavins. Appl. Environ. Microbiol. 45:308-309.

106. Mellanby, E. 1919, A further determination of the part played by accessory food factors in the etiology of rickets. J. Physiol. 52:liii-liv.

107. Mellanby, E. 1921. Experimental rick-ts. Med. Res. Counc. (G.B.) Spec. Rep. Ser. SRS—61.

108. Moore, W. E. C., E. P. Cato, and L. V, Holdeman. 1978. Some current concepts in intestinal bacteriology. Am. J. Clin. Nutr. 31:S33-S42. 128

109. Mott, G. E., R. W. Moore, H. E. Redmond, and R. Reiser. 1973. Lowering of serum cholesterol by intestinal bacteria in cholesterol-fed piglets. Lipids 8:428-431.

110. Napoli, J. L., J. L. Sommerfeldt, B. C. Pramanik, R. Gardner, A. D. Sherry, J. J. Partridge, M. R. Uskokovic, and R. L. Horst. 1983. 19-Nor-lO-ketovitamin D derivatives: Unique metabolites of vitamin D3, vitamin Dg, and 25-hydroxyvitamin D3. Biochemistry 22:3636-3640.

111. Norman, A. W. 1979. Metabolism of vitamin D. Pages 134-197 in W. J. Darby, ed. The Calcium Homeostatic Hormone. Academic Press, New York.

112. Norman, A. W., V. Leather, and J. E. Bishop. 1983. Normal egg hatachability requires the simultaneous administration to the hen of la,25-dihydroxycholecalciferol and 24R,25-dihydroxycholecalci- ferol. J. Nutr. 113:2505-2515.

113. Ornoy, A., D. Goodwin, D. Noff, and S. Edelstein. 1978. 24,25— Dihydroxyvitamin D is a metabolite of vitamin D essential for bone formation. Nature 276:517—519.

114. Owen, R. W., and R. F. Bilton. 1984. Bioconversion of litho- cholic acid under anaerobic conditions by Pseudomonas sp. strain NCIB 10590. Appl. Environ. Microbiol. 48:606-609.

115." Parmentier, G., and H. Eyssen. 1974. Mechanism of biohydrcgena- tion of cholesterol to coprostanol by Eubacterium ATCC 21498. Biochim. Biophys. Acta 348:279—284.

116. Reddy, B. S. 1981. Dietary fat and its relationship to large bowel cancer. Cancer Res. 41:3700-3705.

117. Reinhardt, T. A., J. L. Napoli, D. C. Beitz, E. T. Littledike, and R. L. Horst. 1981. A new in vivo metabolite of vitamin D3:

1,25,25-Trihydroxyvitamin D3. Biochem. Biophys. Res. Commun. 99:302-307.

118. Sacquet, E. C., D. P. Gadelle, M. J. Riottot, and P. M. Raibaud. 1984. Absence of transformation of 8—muricholic acid by human microflora implanted in the digestive tracts of germ-free male rats. Appl. Environ. Microbiol. 47:1167-1168.

119. Sadzikowski, M. R., J. F. Sperry, and T. D. Wilkins. 1977. Cholesterol-reducing bacterium from human feces. Appl. Environ. Microbiol. 34:355-362. 129

120. Samuelsson, B. 1960. Bile acids and steroids. On the mechanism of the biological formation of deoxycholic acid from cholic acid. J. Biol. Chem. 235:361-366.

121. Sawada, H., S. Kulprecha, N. Nilubol, T. Yoshida, S. Kinoshita, and H. Taguchi. 1982. Microbial production of ursodeoxycholic acid from lithocholic acid by Fusarium equiseti m41. Appl. Environ. Microbiol. 44:1249-1252.

122. Seelxg, M. S. 1983. Vitamin D—risk vs. benefit. J. Am. Coll. Nutr. 2:109-110.

123. Shepard, R. M., and H. F. DeLuca. 1980. Plasma concentrations of vitamin D3 and its metabolites in the rat as influenced by vitamin D3 or 25-hydroxyvitamin D3 intakes. Arch. Biochem. Biophys. 202:45-53.

124. Sherrod, J. A., and P. B. H^lemon. 1977. Partial purification and characterization of NAD -dependent 7a-hydroxysteroid dehydrogenase from Bacteroides thetaiotaomicron. Biochim. Biophys. Acta 486:351-358.

125. Simpson, R. TJ., J. K. Wichmann, H. E. Paaren, H. K. Schnoes, and H. F. DeLuca. 1984. Metabolism of 25-hydroxyvitamin D3 by rat kidney cells in culture. Isolation and identification of cis and

trans 19-nor-10-oxo-25-hydroxyvitamin D3. Arch. Biochem. Biophys. 230:21-29.

126. Sim, S. M., S. Huijghebaert, D. J. Back, and H. J. Eyssen. 1983. Gastrointestinal absorption of estrone sulfate in germ-free and conventional rats. J. Steroid Biochem. 18:499-503.

127. Sommerfeldt, J. L., R. L. Horst, E. T. Littledike, and D. C. Beitz. 1979. ^n vitro degradation of cholecalciferol in rumen fluid. J. Dairy Sci. 62:192. (Abstr.)

128. Sommerfeldt, J. L., J. L. Napoli, E. T. Littledike, D. C. Beitz, and R. L. Horst. 1983. Metabolism of orally administered [^H]-ergocalciferol and [^H]-cholecalciferol by dairy calves. J. Nutr. 113:2595-2600.

129. Sommerfeldt, J. L., R. L. Horst, E. T. Littledike, D. C. Beitz, and J. L. Napoli. 1982. Metabolism of orally administered [^H]-vitamin Dg and [^H]-vitamin D3 by dairy calves. Pages 23-26 in A. W. Norman and K. Schaefer, eds. Vitamin D: Endocrinological Aspects and Their Clinical Applications. Walter deGruyter, New York. 130

130. Spassov, G., R. Krutzfeldt, W. S. Sheldrick, W. Wania, R. Vlahov, and G. Snatzke. 1983. Crystallographic monitoring of micro­ biological steroid transformations. Eur. J. Appl. Microbiol. Biotechnol. 17:80-84.

131. Steenbock, H., and A. Black. 1924. Fat soluble vitamins. XXIII. The induction of growth promoting and calcifying properties of a ration by exposure to ultraviolet light. J. Biol. Chem. 61:405-422.

132. Steenbock, H. S., W. R. Kletzein, and J. G. Halpin. 1932. The reaction of the chicken to irradiated ergosterol and irradiated as contrasted with the. natural vitamin D of fish liver oils. J. Biol. Chem. 97:249-264.

133. Stellwag, E. J., and P. B. Hylemon. 1979. 7ct-Dehydroxylation of cholic acid and chenodeoxycholic acid by Clostridium lentum. J, Lipid Res. 20:325-333.

134. Stern, P. H., R. L. Horst, R. Gardner, and J. L. Napoli. 1984. 10-Keto or 25-hydroxy substitution confer equivalent ^ vitro bone-resorbing activity to vitamin D3. Arch. Biochem. Biophys. (In press).

135. Stumpf, W. E., M. Sar, S. A. Clark, and H. F. DeLuca. 1982.

Brain target sites for 1,25-dihydroxyvitamin D3. Science 215.: 1403-1405.

136. Stumpf, W. E., M. Sar, and H. F. DeLuca. 1981. Sites of action of 1,25—(OH)2 vitamin D3 identified by thaw—mount autoradiography. Pages 222-229 in D. V. Cohen, R. V. Talmage, T. L. Matthews, eds. Hormonal Control of Calcium Metabolism. Excerpta Medica, Amsterdam.

137. Stumpf, W. E., M. Sar, R. Narbaitz, F. A. Reid, H. F. DeLuca, and Y. Tanaka. 1980. Cellular and subcellular localization of l,25-(OH)2-vitamin D3 in rat kidney: Comparison with localization of parathyroid hormone and estradiol. Proc. Natl. Acad. Sci. USA 77:1149-1153.

138. Suda, T., H. F. DeLuca, H. K. Schnoes, Y. Tanaka, and M. F. Holick. 1970. 25,26-Dihydroxycholecalciferol, a metabolite of vitamin D3 with intestinal calcium transport activity. Biochemistry 9:4776-4780.

139. Sugano, M., T. Ide, M. Kohno, M. Watanabe, Y. Cho, and Y. Nagata. 1983. Biliary and fecal steroid excretion in rats fed partially hydrogenated soybean oil. Lipids 18:186-192. 131

140. Takasaki, Y., T. Suda, S. Yamada, H. Takayama, and Y. Nishii. 1984. Isolation, identification, and biological activity of 25- hydroxy-24-oxo-vitamin D3: A new metabolite of vitamin D3 generated by vitro incubations with kidney homogenates. Biochemistry 20:1681-1686.

141. Tanaka, Y., and H. F. DeLuca. 1974. Stimulation of 24,25-dihy- droxyvitamin D3 production by 1,25-dihydroxyvitamin D3. Science 183:1198-1200.

142. Tanaka, Y., H. F. DeLuca, H. K. Schnoes, N. Ikekawa, and T. Eguchi. 1981. 23,25-Dihydroxyvitamin D3: A natural precursor in the biosynthesis of 25-hydroxyvitamin 03-26,23 lactone. Proc. Natl. Acad. Sci. USA 78:4805-4808.

143. Tanaka, C., M. Watanoki, and Y. Nozaki. 1983. Effect of soybean protein on coprostanol production and cholesterol metabolism, in cholesterol-fed rats. J. Nutr. Sci. Vitaminol. 29:447-454.

144. Wainfan, E., G. Henkin, S. C. Rittenberg, and W. Marx. 1953. Metabolism of cholesterol by intestinal bacteria in vitro. J. Biol. Chem. 207:843-849.

145. Weisman, Y. A., A. Harrell, S. Edelstein, M. David, Z. Spirer, and A. Colander. 1979. 1,25-Dihydroxyvitamin D3 and 24,25-dihydroxy- vitamin D3 vitro synthesis by human decidua and placenta. Nature 281:317-319.

146. Weisman, Y., A. Vargas, G- Duckett, E. Reiter, and A. Rout. 1978. Synthesis of 1,25-dihydroxyvitamin D in the nephrectomized pregnant rat. Endocrinology 103:1992-1998.

147. Wells, W., S. C. Anderson, and R. Quan Ma. 1970. Lactose diets and cholesterol metabolism. Cholesterol absorption, coprostanol formation and bile acid excretion in the rat. J. Nutr. 71:405-410.

148. White, B. A., D. A. M. Paone, A. F. Cacciapuoti, R. J. Fricke, E. H. Mosbach, and P. B. Hylemon. ^1983. Regulation of bile acid 7-dehydroxylase activity by NAD and NADH in cell extracts of Eubacterium species V.P.I. 12708. J. Lipid Res. 24:20-27.

149. Wichmann, J. K., H. F. DeLuca, H. K. Schnoes, R. L. Horst, R. M. Shepard, and N. A. Jorgensen. 1979. 25-Hydroxyvitamin D3-26,23- lactone: A new in vivo metabolite of vitamin D. Biochemistry 18:4775-4780. 132

150. Wichmann, J. K., H. K. Schnoes, and H. F. DeLuca. 1981. 23,24,25-Trihydroxyvitainin D3, 24,25,26-trihydroxyvitamin D3, 24-keto-25-hydrox}rvitamin D3, and 23-dehydro-25-hydroxyvitamin D3:

New in vivo metabolites of vitamin D3. Biochemistry 20:7385-7391.

151. Wiesner, R. H., R. Kumar, E. Seenan, and V. L. W. Go. 1980. Enterohepatic physiology of 1,25-hydroxyvitamin D3 metabolites in normal man. J. Lab. Clin. Med. 96:1094-1100.

152. Windaus, A., 0. Linsert, A. Luttringhaus, and G. Weidlich. 1932. Crystalline vitamin Hz- Justus Liebigs Ann. Chem. 492:226-241.

153. Windaus, A., F. Schenck, and F. von Werder. 1936. Uber das antirochitish aus 7-dehydrocholesterin. Hoppe-Seyler's Z. Physiol. Chem. 241:100-103.

154. Winter, J., A. C. Mclernon, S. O'Rourke, L. Ponticorvo, and V. D. Bokkenheuser• 1982. Formation of 206-dihydroxysteroids by anaerobic bacteria. J. Steroid Biochem. 17:661-667.

155. Yoon, P. S., and H. F. DeLuca. 1980. Resolution and reconstitu­ tion of soluble components of rat liver microsomal vitamin D-25- hydroxylase. Arch. Biochem. Biophys. 203:529-541.

156. Zile, M., E. C. Bunge, L. Barsness, S. Yamada, H. K. Schnoes, and H. F. DeLuca. 1978. Localization of 1,25-dihydroxyvitamin D3 in intestinal nuclei i^ vivo. Arch. Biochem. Biophys. 186:15-24. 133

ACKNOWLEDGMENTS

The tendency to blame others for our failures, and then claim the

successes for ourselves, is all too common. The fact that I am writing

this thesis at all is because many people had faith in me and provided

the much—needed support. This dissertation is dedicated to my wife,

Lonnie, who, even during her serious illness, managed to be loving,

supportive, and understanding. I wish to also thank my parents and

family because they believed in me and provided the environment which molded what I am.

I would like to thank Dr. P. A. Hartman for his advice and help

throughout my tenure at ISU. To Dr. M. J. Allison, my thanks for your

patience, and the fact that you managed to teach me gut microbiology despite the opposition provided by me. Your timely sense of humor on domestic issues was also greatly appreciated. I would also like to thank Dr. T. A. Reinhardt for the technical advice and experiments he suggested.

To Dr. R. L. Horst, a special thanks is extended. It is hard to describe how much he has contributed to both my personal and scientific development. His enthusiasm for research was contagious, as science was meant to be. Through him, a much needed degree of self-confidence was instilled. The time spent in his laboratory has been some of the most rewarding in my life.

I would also like to thank Drs. J. Young and D. C. Beitz, from the

Nutritional Physiology Department, for their aid and support in 134

obtaining a double major. Their help made a difficult task easier, and their effort on my behalf was greatly appreciated.

I also appreciated the unique opportunity to work at the National

Animal Disease Center, and for the financial aid received from the USDA and the one-year teaching assistantship from the Department of Micro­ biology.

To the many technicians and fellow graduate students who made me feel like I belonged - thank you. I also thank Annette Bates for deci­ phering this and turning it into a thesis. Finally, as a tradition started by a former student, I would also like to thank my "wonder dog," Legion, who gave up many walks so I could be in the lab.