INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE CURRENT INTERNATIONAL RECOMMENDATIONS for indispensable amino acid requirements of adult humans (12) are derived from nitrogen balance studies in men and women (22, 36). Earlier studies by us and others, using approaches based on ¹³C-labeled amino acid oxidation, suggested that adult lysine requirements are considerably higher than previously thought (23, 44, 45). Careful 24-h tracer studies have provided general support for the minimal physiological requirement as estimated in short-term experiments for phenylalanine, leucine, and lysine, and possibly for other indispensable amino acids, and they have established that the tracer-balance concept is valid (3, 11, 20). However, there remains controversy about the quantitative values for the amino acid requirements in adult humans, as reviewed by Waterlow (42).

Among the reasons suggested for the discrepancy between nitrogen balance and tracer-derived amino acid requirement estimates is the possibility that the metabolic requirement, i.e., the irreversible loss of indispensable amino acids (of which oxidation is the major component), is met not only by the diet but also by amino acids synthesized de novo by the gastrointestinal microflora, which are then absorbed. On the basis of the interpretation of urinary ¹⁵N urea excretion after labeled urea administration, it has been claimed that urea nitrogen can be salvaged by urea hydrolysis in the colon and that this nitrogen can be incorporated by the intestinal microflora into amino acids that are subsequently absorbed by the host (18). Although it has been shown in the pig that colonic absorption of amino acids is possible (30, 31), experimental evidence in nonruminant animals does not indicate a quantitatively important amino acid absorption from the colon (7, 17, 19). Nevertheless, results in the pig demonstrated that amino acids synthesized by the microflora can be absorbed supposedly in the small intestine and used for host protein synthesis (25, 40). Tracer studies in animals and humans have shown a transfer of nonspecific nitrogen into dispensable and indispensable amino acids (1, 25, 32). For a majority of amino acids, this may reflect nitrogen exchange or reversible transamination. However, lysine does not undergo transamination in mammalian tissue (6). Therefore, appearance of ¹⁵N-labeled lysine in body proteins and plasma amino acids after administration of a ¹⁵N-labeled nitrogen source, such as ammonium or urea, must reflect the de novo synthesis of lysine and its subsequent absorption from the gastrointestinal tract. Comparative experiments with germ-free and conventional rats (39) have confirmed that lysine de novo synthesis is due to the activity of the indigenous microflora in the gastrointestinal tract.

Although in uremic patients and human subjects consuming a low-protein diet microbial lysine can be made available to the human host (15, 38), there has been no previous attempt to quantify the significance of this source of lysine in host tissue metabolism. The purpose of the present investigation, therefore, was to quantify the contribution of microbial lysine to whole body lysine turnover in healthy adult human subjects. In a pilot study in pigs (25), we found that the degree of [¹⁵N]lysine labeling is dependent on the ¹⁵N source ([¹⁵N₂]urea or ¹⁵NH₄Cl) used. This article describes the appearance of ¹⁵N-labeled lysine in plasma after oral administration of isonitrogenous amounts of [¹⁵N₂]urea and¹⁵NH₄Cl, respectively, in normal young adults. In addition, lysine turnover, lysine oxidation, and lysine splanchnic uptake were measured during a 24-h oral/intravenous tracer protocol. Because the intestinal site of synthesis and subsequent absorption of microbial lysine is uncertain, we studied, additionally, the microbial synthesis of lysine in otherwise healthy subjects with ileostomies after administration of ¹⁵NH₄Cl by use of the same protocol. This enabled us to compare estimates of the microbial lysine contribution to host tissue lysine homeostasis by using fecal and ileal microbial protein as putative precursor pools for the [¹⁵N]lysine absorption.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Subjects and diets. Six male subjects aged 21.7 ± 1.6 yr (weight 81.7 ± 13.3 kg; height 1.82 ± 0.04 m) were recruited within the student population at the Massachusetts Institute of Technology (MIT) and the community of the Boston-Cambridge area. The study was performed at the Clinical Research Center (CRC) of the MIT. The subjects were healthy according to medical history, physical examination, vital signs, blood parameters, and urinalysis. They had no recent weight loss, unusual dietary practice, or pharmacological treatment. Subjects who smoked and consumed more than five or more alcoholic drinks per week were excluded from the study. Subjects were allowed to engage in their usual daily activities but not to participate in competitive sports and perform strenuous exercise. Body weight was monitored daily and did not change throughout the experimental period. The purpose of the study and the potential risks involved were fully explained to each of the subjects. Written informed consent was obtained, and the subjects received financial compensation for their participation in the study. The study was approved by the MIT Committee on the Use of Humans as Experimental Subjects and the MIT-CRC Advisory Committee.

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

24-h Tracer protocol. At 1600 on day 10, the subjects consumed their last meal in the CRC and were then admitted as inpatients to the MIT Medical Department. Two small catheters were placed into veins of the nondominant arm by use of aseptic sterile procedures. One catheter was introduced into an antecubital vein for tracer infusion, and one catheter was placed into a superficial dorsal vein of the hand for blood collection (10). Beginning at 1800 on day 10 of the diet, oral doses (1/24 of daily dose) of the respective ¹⁵N tracer ([¹⁵N₂]urea or ¹⁵NH₄Cl) were given hourly. Hourly portions of ¹⁵N tracers were weighed individually and dissolved in ~25 ml of tap water immediately before ingestion. Simultaneuosly, a 24-h infusion protocol was performed in which an L-[1-¹³C]lysine · HCl tracer (99 AP; MassTrace, Woburn, MA) was administered intravenously and L-[6,6-²H₂]lysine · HCl (99 AP; MassTrace) was given by mouth in hourly intervals to assess lysine kinetics. Because the analytic data of two of the six subjects were questionable, they were not included in data evaluation. The primed, constant intravenous infusion of [¹³C]lysine was administered at a known rate of 3.5 µmol · kg¹ · h¹ (prime 5.25 µmol · kg¹ · h¹ ). The bicarbonate pool was primed (0.6 µmol/kg) with ¹³C sodium bicarbonate (99 AP; Cambridge Isotope Laboratories, Andover, MA). The oral tracer [²H₂]lysine was given at a rate of 2.5 µmol · kg¹ · h¹ (prime 3.75 µmol/kg). All tracers were tested to be sterile and pyrogen free by an independent laboratory. The intravenous tracers were prepared in physiological saline under sterile conditions and were infused with a screw-driven pump at a rate of 8 ml/h (model 22; Harvard Apparatus, Millis, MA). The oral tracer was dissolved in distilled water and given in hourly doses of 8 ml.

Breath ¹³CO₂ background enrichment. To correct the breath ¹³CO₂ baseline for food-derived changes, a tracer-free study was performed in three additional young MIT subjects (20.7 ± 0.9 yr; 61.6 ± 15.0 kg; 1.65 ± 0.1 m). They received the standard experimental diet plus unlabeled ammonium chloride, as described in Subjects and diets, and 30-min breath samples were taken. To account for the lysine intake by the tracers, unlabeled lysine was added as hourly oral doses. The mean value for each 30-min interval of the 24-h protocol was used to correct the ¹³C enrichment in breath during [¹³C]lysine tracer infusion.

Sample collection. Urine and feces from MIT subjects were collected over 24 h starting on day 3 of the dietary period and continued throughout the remainder of the experimental period (day 3 0800 to day 11 1800). The subjects were instructed to collect feces directly in clean plastic bags and then to freeze the sample promptly at 20°C. During the 24-h tracer protocol, urine was collected at consecutive 3-h intervals. Fasted blood samples were drawn into heparinized tubes from an antecubital arm vein by venipuncture on days 3 or 4 (baseline samples for [¹⁵N]lysine enrichments) and day 8. Throughout the 24-h tracer protocol, hourly blood samples were taken via the catheter in the hand vein. Ten minutes before blood sampling, the hand was placed into a customized warming box (65°C air temperature) to obtain arterialized blood.

Sample analysis. Ileal digesta and feces were pooled daily, and the microbial fraction was obtained by differential centrifugation as recently described (39). The microbial pellet was precipitated, and the protein was hydrolyzed and purified by filtration. Nitrogen content in ileal digesta was determined by a micro-Kjeldahl method. Lysine content in ileal microbial protein was determined after acid hydrolysis by ion exchange chromatography (33). Plasma was separated in a refrigerated centrifuge at 1,000 g and stored frozen at 20°C until analysis. The isolation of free amino acids was performed as described by Metges and Petzke (24).

Evaluation of primary data. Lysine oxidation (Lys Ox) was computed for each half-hour interval during the 24-h tracer protocol, except during the first 2 h of the study, when it was taken to be equal to that measured at the end of the 2-h period to reduce any possible effect of the lysine and bicarbonate prime where ¹³C plasma lysine enrichment is the average of the two enrichments determining the specific half-hour interval. Mean lysine oxidation for the fasted and fed states was calculated during 480-720 min and 1,020-1,320 min of the tracer period, respectively. and # = APE/100
and ## = recovery of ¹³CO₂ [as reported by El-Khoury et al. (9); fasted state: 76.6%; fed state: 85.1%].

1

1

Statistics. All results were expressed as means ± SD. The hypotheses to be tested were 1) that [¹³C]lysine kinetics after intake of isonitrogenous amounts of ¹⁵NH₄Cl and [¹⁵N₂]urea, respectively, would not differ from each other, 2) microbial lysine contribution to host metabolism would not differ when ¹⁵NH₄Cl and [¹⁵N₂]urea, respectively, were given as ¹⁵N tracers to the MIT subjects, and 3) that ¹⁵N enrichment of ileal microbial lysine (in ileostomates) would be lower than fecal microbial lysine (in MIT subjects) after the intake of ¹⁵NH₄Cl. Hence we used a two-sided, paired t-test to evaluate group means for hypotheses 1 and 2 and a one-sided, unpaired t-test to compare group means of ileostomates and MIT subjects (hypothesis 3). To compare the fasted and fed states during the 24-h tracer protocol in the MIT subjects, the mean values were compared using a two-sided, paired t-test. A P value of <0.05 was considered to be significant. Nonsignificantly different means were labeled NS.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

[¹⁵N]lysine labeling. The plasma lysine ¹⁵N enrichments in the MIT subjects on days 3 and 4, before the intake of the two ¹⁵N tracers, did not differ from each other, and they corresponded to the ¹⁵N natural abundance ([¹⁵N₂]urea period: 0.3671 ± 0.0003; ¹⁵NH₄Cl period: 0.3668 ± 0.0004 ¹⁵N AP). Because a baseline plasma sample was not available for the ileostomates, urinary total nitrogen was measured. The result (0.3658 ± 0.0003 ¹⁵N AP) was similar to plasma lysine of the MIT subjects.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Daily pattern of plasma lysine 15N enrichment on days 10-11 throughout 24-h hourly oral intake of 0.14 mg 15N · kg1 · h1, either as 15NH4Cl () or [15N2]urea () after 6 days of intake of the respective 15N tracer. Values are means ± SD for 6 subjects. Results are expressed as difference in enrichment [15N atom percent excess (APE)] from initial value on experimental days 3 and 4 (considered zero enrichment baseline). Feeding began with small meals at 0600 (720 min) and was continued until 1500 (1,260 min), providing in total (tracer included) 45 mg dietary lysine · kg1 · day1.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Microbial fecal (15NH4Cl: ; [15N2]urea: ) and ileal (15NH4Cl: ) lysine enrichments throughout experimental time frame. Daily 15N intake started on day 5. Results are expressed as difference in enrichment (15N APE) from initial value on experimental days 3 and 4 (considered zero enrichment baseline). Samples were not obtained on day 11. Note that on some days (days 5, 7, and 9), fecal samples of only 4 MIT subjects could be obtained and that ileostomates ingested 13% more 15N · kg1 · day1 than MIT subjects because of their lower body weight.

1

1

Lysine kinetics. As expected, the course of ¹³CO₂ background enrichment in three subjects in the tracer-free study after consumption of the experimental diet for 10 days changed only minimally with the experimental diet (typically 0.0008 APE fasted state; 0.0003 APE fed state). The course of ¹³CO₂ enrichment and ¹³CO₂ production throughout the 24-h tracer protocol for the ¹⁵NH₄Cl and [¹⁵N₂]urea intakes was apparently different, although this did not reach statistical significance at any given interval (Fig. 3, A and B).

View larger version (72K): [in this window] [in a new window]   Fig. 3.   Pattern of change during a 24-h primed continuous intravenous infusion of [1-13C]lysine (infusion rate 3.5 µmol · kg1 · h1). Values are 13C enrichment of expired carbon dioxide (A), µmol 13CO2 expired (B), and lysine oxidation (C) for each 30-min interval throughout 24-h tracer protocol (means ± SD of 4 subjects). Feeding began with small meals at 0600 (720 min) and was continued until 1500 (1,260 min), providing in total (tracer included) 45 mg dietary lysine · kg1 · day1. , After 15NH4Cl intake; , after [15N2]urea intake.

1

1

1

1

1

1

1

1

1

1

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Although the present data confirm the less extensive, earlier investigations by others on body [¹⁵N]lysine labeling after administration of ¹⁵N-labeled nitrogen precursors (15, 38), a major question posed at the outset of this investigation was the quantitative extent to which microbially synthesized lysine was made available for metabolism in host tissues and its net contribution to the lysine economy of the individual subject. This question turns out to be far more difficult to answer from the present data than we had anticipated. After methodological and design aspects are addressed, the various reasons for this will be discussed.

We have measured plasma and microbial [¹⁵N]lysine without consideration of its enantiomeric form. A study investigating D-amino acid contents in cultures of several bacteria revealed that D-lysine was not detectable, whereas, for example, D-alanine and D-aspartic acid ranged between 8 and 80% and 2 and 60% of total amino acids, respectively (5). In serum of adult rats ingesting standard food, no D-lysine was detected (37). Also, to the best of our knowledge, D-lysine in human plasma has not been reported in the literature, although other plasma amino acids (serine, alanine, aspartic acid, phenylalanine, tyrosine) have been shown to be present in D form (2, 28). Therefore, we do not consider that a significant contribution of D-lysine is likely here. Nevertheless, if this were not the case, then we would have overestimated the availability of microbially derived lysine for host utilization.

The reason for including ileostomates in this study is that, although microbial activity is associated principally with the large intestine, it is by no means confined solely to that part of the gastrointestinal tract (8, 13), and the flora of the lower ileum is qualitatively similar to that of feces (8). Given that the major sites of amino acid absorption are in the small intestine, it was important to determine the ¹⁵N enrichment in the microbial protein of digesta as close as possible to those sites and where there was no functional large intestine to complicate interpretation of the results. Nevertheless, we recognize that ileal fluid of ileostomates may not be the same as the contents of an intact ileum (4, 16, 29, 41); however, our assessment, for the present purpose, was that conventional ileostomy resembles the environment in the intact ileum more than it does the colon, and this gave us the rationale for comparing and contrasting the data from the two groups of subjects studied here.

To help readers understand our results and the discussion that follows, we present a scheme in Fig. 4 showing that ¹⁵N from either ¹⁵NH₄Cl or [¹⁵N₂]urea, after breakdown by microbial urease, is incorporated either directly or indirectly into microbial lysine (and other microbial amino acids) in the small intestine. This depicts the complexity of the pathways that lysine might take within the intestine and between the intestine and host tissues, which introduces difficulties in readily interpreting the quantitative significance of our findings.

View larger version (47K): [in this window] [in a new window]   Fig. 4.   Schematic outline of metabolism of microbial lysine in gastrointestinal tract.

The ¹⁵N from nonspecific nitrogen sources is returned to the intestinal tract in the ¹⁵N-labeled amino acids in endogenous secretions (pancreatic, biliary, mucosal) and as ¹⁵N urea (Fig. 4). Intravenous infusion of ¹⁵N-labeled amino acids is followed by labeled plasma amino acids appearing rapidly in the jejunal fluids and pancreatic secretions (14, 21). The quantity of endogenous protein that is recycled in the intestine makes it a potentially significant source of nitrogen for microbial growth. In pigs, 90% of all endogenous nitrogen secreted into the gut was found to be reabsorbed (19). Although these experiments do not explain the involvement of the enteric flora in this process, it is probable that microbial proteolytic activity is involved because mucins and some other digestive secretions are resistant to mammalian digestive enzymes.

The differences we observed between ileal and fecal microbial protein in [¹⁵N]lysine enrichment are in accord with investigations in the pig (40) showing that, after the pigs were fed ¹⁵NH₄Cl for 10 days, [¹⁵N]lysine enrichment increased dramatically from the ileum to the caecum. Also, the difference in fecal microbial and plasma lysine ¹⁵N enrichment seen in this study between ¹⁵NH₄Cl and [¹⁵N₂]urea confirms earlier results in the pig (25) and is presumably due, in part, to the greater retention of ¹⁵N label from ammonium chloride (24a). However, because of high interindividual variation the difference did not always reach statistical significance. If allowance is made for the different ¹⁵N retention, then the microbial enrichments would not appear to be different, although an additional reason for the different enrichments might be differences in ammonia nitrogen and urea nitrogen metabolism by the microflora (35, 26).

Using the ¹⁵N enrichment of fecal microbial lysine with ¹⁵NH₄Cl as the label, we estimated an uptake of ~29 mg · kg¹ · day¹; with [¹⁵N₂]urea, the estimate was 12 mg · kg¹ · day¹, although this was not statistically different. It is obvious, however, that the digesta undergo many changes during passage from the duodenum to the rectum, and it is unlikely that the [¹⁵N]lysine in microbial protein from feces is truly representative of that occurring at sites of amino acid absorption in the intestine. It was for that reason that we sampled ileal digesta from ileostomates as being closer to the sites of absorption. The corresponding estimate of microbial lysine absorption for the ileostomates (when the same rate of lysine turnover is assumed as in the MIT subjects) is ~68 mg · kg¹ · day¹. If the ¹⁵N enrichment of lysine in ileal digesta were assumed to represent the ¹⁵N enrichment of absorbed lysine in the MIT subjects given ¹⁵NH₄Cl, their uptake of microbial lysine would be calculated to be ~130 mg · kg¹ · day¹, a value that exceeds by ~10-fold both the apparent dietary requirement and, therefore, the irreversible loss of body lysine (12).

In considering the plausibility of these different estimates of microbial lysine uptake, one must consider that they refer to the gross uptake of amino acid synthesized by the microflora and do not necessarily represent a net increase of that magnitude in the total lysine available for metabolism. To the extent that 1) the growth of the microbes that give rise to the labeled lysine is supported by the degradation of endogenous protein (Fig. 4) and 2) luminal amino acids of endogenous origin are oxidized by intestinal tissues, then the microbially derived lysine in plasma can be seen as part of the normal mechanism by which endogenous nitrogen and amino acids are recycled, rather than microbial amino acids serving as a net source of amino acids, which are additional to those supplied in the diet. Support for this view comes from previous [1-¹³C]leucine oxidation studies (10, 11), where there is an excellent concordance between both predicted and measured protein oxidation when estimated from leucine oxidation and N excretion values, which would not have been the case if there had been a significant net uptake of leucine from the gut microflora. There seems little reason to expect, if there were a significant entry of microbially derived lysine into body tissues, that this would not also apply to the other nutritionally indispensable amino acids, including leucine.

Although the true ¹⁵N enrichment of the labeled lysine that is actually absorbed might be closer to that of the ileal digesta protein than to that of the feces, it could quite conceivably be higher (30). Furthermore, if ¹⁵N-labeled substrates entering the gut lumen were used preferentially by microbes in juxtaposition to the intestinal wall, with turnover and release of their constituent proteins and amino acids in that spatial domain, then the ¹⁵N enrichment of lysine being absorbed from the gastrointestinal tract would not be accurately reflected by either the ileal or fecal microbial protein-bound lysine.

The loss of lysine at the terminal ileum was estimated in the present study to be ~8.5 mg · kg¹ · day¹. Presumably there is also loss of endogenous lysine due to bacterial and tissue intestinal oxidative metabolism (Fig. 4), as indicated by the measured first-pass lysine extraction by splanchnic tissues (and microflora) of ~25% of the dietary lysine (Table 5). It could be, therefore, that any entry of lysine of microbial origin helps to counteract the endogenous lysine losses, rather than adding in a net way to the total exogenous (diet) input of lysine.

The lysine oxidation rates measured in this study are considerably lower than the ~45 mg · kg¹ · day¹ level of lysine intake, indicating that we have significantly underestimated the true whole body lysine oxidation rate. This is in accord with our previous oral/intravenous [1-¹³C]lysine tracer studies (9). Lysine turnover rates were in the same range as reported earlier and were somewhat higher for the orally adminstered lysine tracer (9). It was surprising, however, that there were differences between various measures of lysine kinetics between the two different studies in which either [¹⁵N₂]urea or ¹⁵NH₄Cl was used as a tracer (Table 5). There are a number of possible explanations. First, it could be due to the ¹⁵N tracers per se. However, this seems unlikely, because the amount of nitrogen supplied with the two ¹⁵N tracers was equal. If account is taken of the 1.6-fold higher ¹⁵N retention with ¹⁵NH₄Cl compared with [¹⁵N₂]urea (24a), and it is assumed that all of the additional nitrogen retained was used to produce microbial lysine, this would amount to a difference between the two tracers of only 3 mg · kg¹ · day¹. This does not explain the observed difference in lysine kinetics between the two tracers (Table 5). Second, although administration of NH₄Cl causes acidosis and increases protein breakdown and amino acid oxidation (34), this occurs at doses that are severalfold higher than the total 1 g/day dose used in this study. Third, and on the basis of the foregoing comments, the observed differences would appear to be due to experimental and biological factors that cannot be identified at this time.

In conclusion, to reliably estimate the quantitative contribution made by lysine (or other indispensable amino acids) of bacterial origin to host metabolism, use of the present paradigm seems doubtful. Although our data do suggest a potentially significant nutritional uptake into body tissues of amino acids of microbial origin, the quantification of this uptake, and its net contribution to the amino acid economy, require a better understanding of nitrogen and carbon transactions of amino acids and related compounds between the intestinal tissues and the microflora within the gastrointestinal tract.