Intestinal Absorption of Water-Soluble Vitamins: Cellular and Molecular Mechanisms

54.2.1

Metabolic Role and Effect of Deficiency

Thiamin ( Fig. 54.1 ) was the first member of the family of water-soluble vitamins to be described, with reference to the effects of its deficiency (beriberi) being suggested some 4000 years ago in old Chinese medical literatures. Thiamin (mainly in its pyrophosphate forms) acts as an enzymatic cofactor for transketolase, pyruvate dehydrogenase, α-ketoglutarate dehydrogenase, and branched-chain ketoacid dehydrogenase, enzymes involved in many critical metabolic reactions that relate to energy metabolism. Because thiamin bridges the glycolytic and the pentose phosphate metabolic pathways (which generate chemical-reducing power in cells), it also plays an important role in reducing cellular oxidative stress. Thus, low intracellular levels of thiamin will lead to impairment in energy metabolism (acute energy failure) and to a propensity for oxidative stress. While most of thiamin’s biological activity is attributed to its pyrophosphate form (i.e., thiamin pyrophosphate, TPP, which is also the most abundant form of this vitamin in cells and accumulates mainly in mitochondria), recent studies have shown that other forms of the vitamin, namely, thiamin triphosphate, also possess biological activity in nerve cells where it regulates the function of membrane chloride channels and acts as a phosphate group donor to other membrane proteins (Ref. and references therein). Thiamin deficiency (and suboptimal levels) represents a significant nutritional problem worldwide and leads to several clinical abnormalities including neurological (neuropathy and/or Wernicke-Korsakoff syndrome) and cardiovascular (peripheral vasodilatation, biventricular myocardial failure, edema, and potentially acute fulminant cardiovascular collapse) disorders. On the other hand, optimization of thiamin levels appears to have the potential of preventing diabetic retinopathy and tissue damage caused by the hyperglycemia of diabetes and alleviating fatigue in patients with inflammatory bowel diseases (IBDs). Systemic thiamin deficiency and suboptimal levels occur in a high percentage of chronic alcoholics, diabetics, patients with celiac sprue and renal diseases, post bariatric surgery, patients with sepsis, and in subjects on long-term furosemide diuretic therapy. In addition, thiamin deficiency and suboptimal levels occur in the elderly population despite an average daily intake of the vitamin that exceeds their recommended requirement. Also recognized is the occurrence of thiamin deficiency disorders that occur despite normal plasma level of this vitamin. Examples of the latter are thiamin-responsive megaloblastic anemia (TRMA) and thiamin-responsive/Wernicke’s-like encephalopathy. TRMA is an autosomal recessive disorder characterized by megaloblastic anemia, sensorineural deafness, and non-type 1 diabetes mellitus, which is caused by mutations in the human thiamin transporter-1 (hTHTR-1; product of the SLC19A2 gene). Such mutations in hTHTR-1 gene led to an impairment in thiamin uptake (by the affected tissues) and the development of localized thiamin deficiency. Thiamin-responsive/Wernicke’s-like encephalopathy is characterized by seizures, ophthalmoplegia, nystagmus, and ataxia and is believed to be due to mutations in the human thiamin-transporter-2 (hTHTR-2; product of the SLC19A3 gene). In both of the latter disorders, oral administration of pharmacological doses of thiamin brings about significant improvements in many of the clinical symptoms of the affected individuals.

54.2.2

Physiology of the Intestinal Thiamin Absorption Process

The human intestine encounters thiamin from two sources: a dietary source and from microbiota. Dietary thiamin exists mainly in the phosphorylated form, which must be hydrolyzed to free thiamin prior to absorption (the small intestine does not have a transport system for intact TPP, that is, it is unlike the large intestine that has such a system; see below), a process that is achieved by the action of the abundant intestinal phosphatases. Absorption of the liberated free thiamin (which exists in the monocationic form at gut luminal pH) then takes place mainly in the proximal part of the small intestine. The mechanism of thiamin absorption in the small intestine has been investigated using a variety of intestinal preparations from different species including human. Collectively, these studies have shown the involvement of a specific carrier-mediated mechanism that is inhibited by thiamin structural analogues, but not by unrelated organic cations. Studies using purified brush-border membrane vesicles (BBMV) isolated from human and animal small intestine have shown the involvement of a pH (but not Na ⁺ )-dependent, electroneutral, and amiloride-sensitive carrier-mediated mechanism for thiamin uptake. The mechanism by which thiamin exits out of enterocytes, that is, transport across the basolateral membrane (BLM), has also been studied using purified human and animal intestinal basolateral membrane vesicles (BLMV) and has been found to involve a specific, pH-dependent, electroneutral, carrier-mediated process. With regard to potential metabolism during absorption, studies have shown that some of the absorbed free thiamin is rephosphorylated (mainly to TPP) in intestinal epithelial cells; however, only free thiamin exits the intestinal absorptive cells into the serosal side. There is recent emerging evidence that indicates that the intracellularly generated TPP accumulates in mitochondria via a carrier-mediated process and this important process has major relevance to intestinal energy metabolism in health and disease.

As to the microflora-generated thiamin, this source provides considerable amounts of free and phosphorylated (TPP) forms of the vitamin. While gut microbiota of all human generate thiamin, recent studies that classified the human gut microbiota into three functionally different enterotypes (enterotypes 1, 2, and 3) have reported that enterotype 2 (which is enriched in Prevotella and Desulfovibrio ) is especially overrepresented in enzymes that are involved in TPP biosynthesis, implying a larger level of thiamin generation in the group that harbors this enterotype. As to absorbability of the microbiota-generated forms of thiamin in the large intestine, recent investigations have shown that the human colonocytes do indeed possess an efficient and specific carrier-mediated mechanism for uptake of free thiamin and a separate and specific carrier-mediated mechanism for uptake of intact TPP. The latter system was shown to have high affinity (apparent K _m of 0.157 μM) and specificity for TPP (it does not transport free thiamin or thiamin monophosphate), is pH- and Na ⁺ -independent, and energy-dependent. It is also interesting to note that the gene that encodes the colonic TPP transporter (TPPT) (i.e., SLC44A4; see below) has recently been identified as an ulcerative colitis susceptibility gene. Based on the mentioned evidence, it is clear that the microbiota-generated thiamin is absorbable and contributes to overall thiamin nutrition, especially when it come to cellular nutrition and health of the local colonocytes.

54.2.3

Molecular Biology of the Intestinal Thiamin Uptake Systems

Mammalian cells appear to utilize two members of the solute carrier SLC19A gene family (i.e., THTR-1 and THTR-2) to transport free thiamin and one to transport TPP. The SLC19A2 encodes a protein of 497 amino acids (THTR-1), while the SLC19A3 encodes a protein of 496 amino acids (THTR-2). Both of the human THTR-1 and -2 are predicted to have 12 putative transmembrane domains (TMDs) with their N- and C-terminal tails extending into cell interior. The hTHTR-1 and hTHTR-2 share 48% identity and 64% similarity with one another. They also share 40% and 39% identity at the amino acid level, respectively, with the human reduced folate carrier (hRFC). However, neither hTHTR-1 nor hTHTR-2 transport folate and the hRFCs do not transport free thiamin.

The hTHTR-1 appears to function in the micromolar range, while the hTHTR-2 appears to function in the nanomolar range. Both the hTHTR-1 and hTHTR-2 are expressed in the human small and large intestine, with expression of the former being markedly higher than that of the latter. The membrane domain of the polarized human enterocyte at which hTHTR-1 and hTHTR-2 proteins are expressed have also been investigated using immunological and confocal imaging approaches. Expression of the hTHTR-1 protein was shown to be at both the apical and BLM domains of the polarized enterocytes, with a slightly higher expression at the latter compared to the apical membrane domain. On the other hand, expression of the hTHTR-2 protein was shown to be restricted to only the apical membrane domain of the polarized absorptive epithelia.

As to the molecular identity of the colonic TPPT, this has been recently delineated in studies by Nabokina et al., which showed that the uptake system is the product of the SLC44A4 gene. Expression of the human TPPT along the intestinal tract was found to be restricted to the colon ( Fig. 54.4 ). Also, cell surface biotinylation assay and live-cell confocal imaging studies have shown that the human TPPT is exclusively expressed at the apical membrane domain of polarized epithelial cells ( Fig. 54.5 ). The tissue-specific expression of the TPPT in the colon appears to be determined by epigenetics as well as microribonucleic acid (miRNA)-mediated mechanisms. The latter was suggested by findings that in the colon, histone H3 in the 5′-regulatory region of Slc44a4 is trimethylated at lysine 4 and acetylated at lysine 9, whereas the trimethylation at lysine 27 was negligible. In the jejunum, on the other hand, histone H3 was found to be hypertrimethylated at lysine 27 (repressor mark). Involvement of miRNA(s) in the tissue-specific expression of TPPT was suggested by findings showing the 3′-UTR of SLC44A4 to be the target of specific miRNAs/RNA-binding proteins in noncolonic, but not in colonic, epithelial cells.

54.2.4

Cell Biology of the Intestinal Thiamin Absorption Process

Insight into the mechanisms involved in intracellular trafficking and membrane targeting of the human thiamin transporters in epithelial cells have also been emerging from studies utilizing the live-cell confocal imaging approach. Using a series of truncated hTHTR-1 constructs fused with the enhanced green fluorescent protein (EGFP) (i.e., hTHTR1-EGFP), expression of the full-length fusion protein was shown to be at both the apical and the BLM domains of intestinal epithelial cells. Analysis of the localization pattern of truncated mutants of the hTHTR-1 has shown that while the C-terminus region of the polypeptide does not appear to have a role in membrane targeting, whereas an essential role for the N-terminal and the backbone was found. Also, truncation of the hTHTR-1 polypeptide within a region where several TMRA truncations are clustered results in intracellular retention of the mutant protein. As to intracellular trafficking, the hTHTR-1 protein was found to reside inside numerous trafficking vesicles whose movement is temperature dependent and requires an intact microtubule network for their mobility (Ref. ; real-time movies can be viewed at website: http://www.jcb.org/cgi/content/full/278/6/3976/DC1 ).

With regard to the hTHTR-2 protein, this carrier is expressed exclusively at the apical membrane domain of the polarized intestinal epithelial cells. Live-cell imaging investigations utilizing a series of hTHTR-2 truncations have shown that while the cytoplasmic C-terminal is not essential for membrane targeting of the protein, an essential role for the transmembrane backbone is evident. In addition, trafficking vesicles were found to be involved in the intracellular movement of hTHTR-2 to the cell surface, and that mobility of these vesicles depends on an intact microtubule network that was disrupted by overexpression of the dynactin subunit dynamitin (p50) (dynactin is involved in trafficking of vesicles via conventional cytoplasmic dynein, a minus end-directed motor protein).

Other studies have identified accessory proteins that influence the functionality and cell biology of the hTHTR-1 and -2 in intestinal epithelial cells ( Fig. 54.6 ). Using a bacterial two-hybrid system to screen a human intestinal cDNA library with the complete coding sequence of hTHTR-1 as bait, studies have identified a member of the tetraspanin family of proteins, Tspan-1, as an interacting partner with hTHTR-1. Coimmunoprecipitation assay, glutathione S -transferase (GST)-pulldown assay, and live-cell confocal imaging have confirmed the existence of such an interaction between hTspan-1 and hTHTR-1 in human intestinal epithelial Caco-2 cells. This interaction appears to have biological and functional consequences as it decreased the rate of degradation of the hTHTR-1 protein, and thus, influenced its functionality. More recent investigations have utilized a yeast split-ubiquitin two-hybrid approach to identify the human transmembrane 4 superfamily 4 (hTM4SF4) as an interacting partner of hTHTR-2 in human intestinal epithelial cells ( Fig. 54.6 ). This interaction was again confirmed by GST-pull-down assay in vitro , by live-cell confocal imaging in intestinal epithelial cells, and shown to have functional consequences on the hTHTR-2 activity.

54.2.5

Structure-Activity Aspects of the Thiamin Transporters

Knowledge about the structure-function activity of the hTHTR-1 and hTHTR-2 proteins has been emerging in recent years from clinical and experimental findings. With regard to the hTHTR-1, over 18 different clinical mutations have been identified in patients with TRMA. These mutations are either mis-sense or nonsense in nature with the latter type leading to an early truncation of the carrier protein. The effect of a number of these mutations (e.g., P51L and T158R, W358X, and Delta383fs) on functionality of the hTHTR-1 has been investigated experimentally using mutants generated by site-directed mutagenesis followed by expression in intestinal and other appropriate cell types. The results revealed a spectrum of mutant phenotypes with all exhibiting impaired thiamin uptake function due to either a change in hTHTR-1 level of expression/stability, mis-targeting, and/or altered transport activity. Other studies have examined the role of the only conserved anionic amino acid residue (located at position 138) in the predicted TMDs of hTHTR-1 in the transport of the cationic thiamin, as well as the role of the two putative N-glycosylation sites (predicted to be at positions 63 and 314). The result showed a critical role for the residue at position 138 in the function of hTHTR-1. On the other hand, neither of the putative glycosylation sites was found to play a role in the function or membrane targeting of the hTHTR-1 protein.

As to the hTHTR-2 protein, again data from clinical and experimental studies have shed light into important aspects of its structure-activity relationship. Two clinical mutations (K44E and E320Q) in hTHTR-2 have been described in patients with thiamin-responsive Wernicke’s-like encephalopathy. Experimental testing of these two mutants showed impaired functionality of hTHTR-2 in both cases. The K44E mutant caused thiamin transport dysfunction by impairing hTHTR-2 trafficking to the cell membrane. Another two mutations in the hTHTR-2 protein (G23V and T422A) have been reported in patients with biotin-responsive basal ganglia disease (although it is unclear at present as to why mutations in a specific thiamin transporter precipitates this condition despite the fact that hTHTR-2 does not transport biotin; Ref. ). Both of these mutations in hTHTR-2 were found to cause functional impairment, with no defect in the expression of this transporter at cell membrane. It is interesting to mention here that both of the affected residues (i.e., at positions 23 and 422) are conserved in the hTHTR-2 sequences of all species cloned so far, as well as in other members of the SLC19A gene family, that is, the hTHTR-1 and the (hRFC; product of the SLC19A1 gene), thus, supporting an essential role of these residues in transporter functionality. Hydropathy analyses suggested that these residues (G23V in TM1, T422A in TM11) lie within the TMDs of the hTHTR-2 protein. Crystallographic data from other major facilitator superfamily transporters showed that both TM1 and TM11 contribute to the central hydrophilic cavity of the transporters, underscoring the likelihood that mutation of residues within these regions could impair transporter function (reviewed in Ref. ). Other studies have examined the role of the three negatively charged conserved glutamic acid residues within the TMDs of hTHTR-1 and hTHTR-2 (i.e., E120, E320, and E346) as well as the role of the two putative N-glycosylation sites in the hTHTR-2 polypeptide (i.e., N45, N166) in the function and cell biology of the hTHTR-2 protein. Results of these investigations showed that mutating any of the conserved glutamic acid residues lead to a significant impairment in transport of the positively charged thiamin. Furthermore, while mutating E120 and E320 led to a decrease in the level of expression of hTHTR-2 at the apical membrane domain of intestinal epithelial cells, mutating E346 did not affect this parameter. Finally, the two putative glycosylation sites of the hTHTR-2 polypeptide (i.e., N45Q, N166Q) do not appear to affect the functionality or level of expression of the hTHTR-2 in intestinal epithelial cells.

With regard to the colonic TPPT, recent studies have shown that the protein is glycosylated at positions N69, N155, N197, N393, and N416. However, only glycosylation at N69, N155, and N393 appears to be important for function, most likely through an effect on protein conformation and/or interaction with the substrate.

54.2.6

Regulation of the Intestinal Thiamin Uptake Process

The intestinal thiamin uptake process appears to be under the regulation of a variety of extracellular/environmental conditions and intracellular factors. Since a number of these conditions/factors exert their regulatory effect(s) at the transcription level, it is relevant to first describe the basal transcriptional regulation of the SLC19A2, SLC19A3 , and SLC44A4 genes.

54.2.7

Conditions/Factors That Negatively Impact the Intestinal Thiamin Uptake Process

54.3.1

Metabolic Role and Effect of Deficiency

Riboflavin (RF; Fig. 54.1 ) is required for normal cellular metabolism, proliferation, and growth. In its biologically active forms [flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD)], the vitamin plays a key metabolic role in the transfer of electrons in biological oxidation-reduction reactions involving carbohydrate, lipid, amino acid, and certain water-soluble vitamins (pyridoxine and folate). Studies have also shown a role for riboflavin in protein folding in the endoplasmic reticulum and in the regulation of cellular energy expenditure. More recent investigations have demonstrated antioxidant and antiinflammatory properties for riboflavin, as well as a role for this vitamin in normal immune function. Riboflavin deficiency leads to a variety of clinical abnormalities, including degenerative changes of the nervous system, anemia, skin lesions, cataract, and growth retardation. It also leads to an increase in the susceptibility to cancer. Deficiency/suboptimal levels of riboflavin occur in chronic alcoholism, diabetes mellitus, IBD, Brown-Vialetto Van Laere, and Fazio Londe syndromes [the latter are rare neurological disorders caused by mutations in riboflavin transporters-2 and -3 (RFVT-2 and RFVT-3)]. In contrast to the deleterious effect of riboflavin deficiency, optimizing riboflavin status has been reported to reduce the risk of esophageal squamous cell carcinoma (Ref. and references therein).

54.3.2

Physiology of the Intestinal Riboflavin Absorption Process

The human intestine encounters two sources of RF: a dietary source and a microbiota source. Dietary riboflavin exists mainly in the form of FMN and FAD and is bound (noncovalently) to proteins. These coenzymes are first released from proteins via the combined action of gastric acid and hydrolases and then hydrolyzed further (prior to absorption) by intestinal phosphatases to free riboflavin. The mechanism of absorption of free riboflavin has been studied using a variety of human and animal intestinal preparations. Collectively, these studies have shown that transport of free riboflavin occurs mainly in the proximal part of the small intestine and involves an efficient and specific Na ⁺ -independent, carrier-mediated mechanism. This mechanism is inhibited by riboflavin structural analogues and by the Na ⁺ /H ⁺ exchange inhibitor amiloride ( K _i for amiloride ≈ 0.48 mM). Other studies have shown that the intestinal riboflavin uptake process is sensitive to the effect of the tricyclic phenothiazine drug chlorpromazine, a compound that shares structural similarities with riboflavin. Once internalized, some of the riboflavin is phosphorylated and used by enterocytes, and the rest is transported out of the cell across the BLM via a specific carrier-mediated process.

With regard to the microbiota-generated riboflavin in the large intestine, the amount of the vitamin produced depends on the type of the ingested diet and is higher following ingestion of a vegetable-based diet compared to a meat-based diet. Also, a considerable amount of the microbiota produced riboflavin was found to exist in the large intestinal lumen in the form of free riboflavin, and thus, available for absorption. Other investigations have shown that the large intestine is capable of absorbing luminally introduced free riboflavin. Insight into the mechanism involved in colonic uptake of riboflavin came from studies using the human colonic epithelial NCM460 cells, which showed the involvement of an efficient and specific carrier-mediated mechanism that is similar to the one described in the small intestine. Subsequent studies in rats have confirmed the existence of a specialized carrier-mediated mechanism for riboflavin uptake in the colon. With the demonstration of existence of an efficient carrier-mediated system for riboflavin uptake in the large intestine, the belief that this source of riboflavin contributes to the overall host nutrition of the vitamin, and especially toward the cellular nutrition of the localized colonocytes, is further strengthened.

54.3.3

Molecular Biology of the Intestinal Riboflavin Uptake Process

Insight into the molecular identity of the transport systems involved in intestinal riboflavin uptake has recently emerged with the identification of three putative human riboflavin transporters-1, -2, and -3 (RFVT-1, RFVT-2, and RFVT-3) that are encoded by three different but homologous genes: the SLC52A1 , SLC52A2 , and SLC52A3 genes, respectively. These transporters exhibit no similarity with the yeast or bacterial riboflavin transporters, nor do they show similarity with any other mammalian transporter. The hRFVT-1 and hRFVT-2 share > 87% sequence similarity at the protein level, while hRFVT-3 shares around 43% similarity with hRFVT-1 and hRFVT-2. The hRFVT-1, -2, and -3 are all expressed in the gut, with expression of the hRFVT-3 being significantly higher than that of hRFVT-1 and hRFVT-2 ; hRFVT-3 is also more efficient in transporting riboflavin than the other hRFVTs. Furthermore, live-cell confocal imaging of polarized intestinal (and renal) epithelial cells showed the hRFVT-3 to be predominantly expressed at the apical membrane domain, while expression of hRFVT-1 is mostly at the BLM domain of these cells; expression of hRFVT-2 was localized inside intracellular vesicles (with some expression at the BLM domain).

The relative contribution of the predominantly expressed riboflavin transporter in the intestine (i.e., RFVT-3) toward total carrier-mediated riboflavin uptake has also been recently addressed using an in vitro SLC52A3 gene-specific silencing siRNA approach and human intestinal epithelial Caco-2 cells with the results showing a major role for these transporters in intestinal riboflavin uptake. This was confirmed in vivo in studies utilizing an intestinal-specific (conditional) RFVT-3-knockout (cKO) mouse model developed by the Cre/Lox approach ( Fig. 54.7 ). In the latter study, all the RFVT-3 cKO mice developed systemic riboflavin deficiency indicating that RFVT-3 is involved in determining the overall riboflavin body homeostasis. Also observed was a significant increase in the level of expression of oxidative stress-responsive genes in the intestine of the cKO mice, which suggest a role for riboflavin in the maintenance of normal intestinal physiology.

54.3.4

Cell biology of the Intestinal Riboflavin Absorption Process

As mentioned earlier, the RFVT-3 is expressed exclusively at the apical membrane domain of intestinal epithelial cells. What dictates cell surface expression of the protein appears to be a sequence in the C-terminal portion of the polypeptide that includes two conserved cysteine residues (one at position 463 and the other at position 467). Mutating these cysteine residues was shown to lead to severe impairment in riboflavin uptake due to retention of the protein in the endoplasmic reticulum. A potential disulfide bridge between C463 and C386 (also predicted by modeling analysis and mutating the C386 residue) was found to lead to intracellular retention of RFVT-3. Intracellular trafficking of RFVT-3 was shown to involve distinct vesicular structures whose mobility depends on existence of an intact microtubule network.

54.3.5

Structure-Activity Aspects of the Riboflavin Transporters

Early studies using Caco-2 cells exposed to specific sulfhydryl group reagents have reported a role for sulfhydryl groups in the function of the intestinal riboflavin uptake carriers. Since the sulfhydryl group reagents used in these studies were membrane impermeable, it was assumed that the affected sulfhydryl group(s) were located at the exofacial surface of the cell membrane. More recent investigations examined the effect of clinical mutations in SLC52A3 found in patients with the Brown-Vialetto-Van Laere syndrome (BVVLS; a rare neurological disease characterized by ponto-bulbar palsy, bilateral sensorineural deafness, and respiratory insufficiency) on cell biology of RFVT-3. Mutants P28T, E36K, E71K, and R132W were all found to be functionally impaired and that the impairment was due to retention of the mutated RFVT-3 within the endoplasmic reticulum. These findings suggest a role for the affected residues in normal cell biology of RFVT-3.

54.3.6

Regulation of the Intestinal Riboflavin Uptake Process

The intestinal riboflavin uptake process is under the regulation of extracellular/environmental conditions and intracellular factors. Since a number of these conditions exert their effect(s) at the transcription level, it is relevant to first describe the basal transcriptional regulation of the intestinally relevant SLC52A1 and SLC52A3 genes.