Silk—A new substrate for UDP-d-xylose: Proteoglycan core protein β-d-xylosyltransferase

The formation of most connective tissue polysaccharides is initiated by transfer of d-xylose from UDP-d-xylose to specific serine residues in the core proteins of the putative proteoglycans. The substrate specificity of the xylosyltransferase catalyzing this reaction has not yet been examined in detail, but it appears that a -Ser-Gly- pair is an essential part of the substrate structure. Since the preparation of the known acceptors (e.g., Smith-degraded or HF-treated cartilage proteoglycan) involves a substantial effort, we have searched for readily available proteins with the -Ser-Gly-sequence, which might serve as alternative substrates. In the present work, it was found that silk fibroin from Bombyx mori, which consists, in large part, of the repeating hexapeptide, Ser-Gly-Ala-Gly-Ala-Gly, is an excellent substrate for the xylosyltransferase from embryonic chick cartilage. Pieces of silk were used directly in the reaction mixtures, and [¹⁴C]xylose transferred from UDP-d-[¹⁴C]xylose was measured by liquid scintillation spectrometry after rinsing the silk in 1 m NaCl and water. Substantially greater incorporation was observed with preparations of silk or fibroin which had been dissolved in 60% LiSCN and subsequently dialyzed exhaustively or diluted appropriately. Under standard reaction conditions, the V_max for fibroin was 531 pmol/h/mg enzyme protein, as compared to 223 pmol/h/mg for Smith-degraded proteoglycan. K_m values were 182 mg/liter (fibroin) and 143 mg/liter (Smith-degraded proteoglycan). The product of [¹⁴C]xylose transfer to silk was alkali labile, and [¹⁴C]xylitol was formed when [¹⁴C]xylosylsilk was treated with borohydride in alkali. Proteolytic digestion with papain, Pronase, leucine aminopeptidase, and carboxypeptidase A yielded a radioactive product which was identified as [¹⁴C]xylosylserine by electrophoresis and chromatography. The identity of the isolated [¹⁴C]xylosylserine was further supported by its resistance to treatment with alkali (0.5 m KOH: 100°C; 8h) and by acid hydrolysis which yielded [¹⁴C]xylose. Tryptic and chymotryptic fragments from fibroin were also good xylose acceptors and had V_max values 60–70% of that observed for the intact protein. Substantial acceptor activity was displayed also by the sericin fraction of silk and by the silk sequence hexapeptide, Ser-Gly-Ala-Gly-Ala-Gly; the latter had a V_max value close to 20% of that of intact fibroin.     

Determination of the multiplicity of the silk fibroin gene and detection of fibroin gene-related DNA in the genome of Bombyx mori.

The number of silk fibroin genes per genome in the silkworm Bombyx mori has been determined by hybridization using fibroin [¹²⁵I]mRNA. The purified [¹²⁵I]mRNA had an oligonucleotide pattern after RNAase T₁ digestion which was characteristic of fibroin mRNA (Suzuki &; Brown, 1972) and it hybridized specifically to DNA with a G + C content expected for a fibroin gene. Thermal denaturations indicated that these hybrids were mismatched by about 3%, which probably indicates some variation among the sequences encoding the internal repetitions of the fibroin protein.The concentration of fibroin gene sequences in B. mori DNA was measured by saturation hybridization of [¹²⁵I]mRNA to filter bound DNA. The same saturation level of 1.8 × 10^?5 μg mRNA per μg DNA was calculated from data obtained with unfractionated DNA and with fibroin gene sequences which had been separated from bulk B. mori DNA by actinomycin $\text{D}\text{CsCl}$ centrifugation. Scatchard plots of the subsaturation data extrapolated to an identical saturation value. Internal reiteration of the fibroin mRNA molecule was apparent from the high association constant of hybridization. An exhaustive hybridization experiment showed that such repetitions comprise at least 90% of each mRNA molecule. The saturation value, in conjunction with the genome DNA content and the mRNA size, indicated the presence of only one fibroin gene per haploid B. mori genome.Hybridization of actinomycin $\text{D}\text{CsCl}$ fractionated DNA indicated that fibroin mRNA can form hybrids with DNA that bands with bulk B. mori DNA. These hybrids appear to involve DNA which is related to, but distinguishable from, true fibroin gene sequences. The fibroin gene-related sequences form mismatched hybrids with the mRNA, are much shorter than the fibroin gene and are dispersed in B. mori DNA of much lower G + C content, and there are many copies of these sequences per B. mori genome.     

Syntheses of glycine and L-serine by their interconversion in the posterior silkgland of the silkworm, Bombyx mori

Summary. It was observed by solution-state ¹³C NMR spectroscopy that a great portion of the ¹³C of [1-¹³C]L-serine fed to the 5th instar larvae of the silkworm, Bombyx mori was incorporated into C1 of glycine in silk fibroin. [1-¹³C]Glycine was detected along with [1-¹³C]serine in fibroin of the posterior silkgland cultured in a medium containing [1-¹³C]serine. This formation of [1-¹³C]glycine was inhibited by addition of aminopterin to the culture medium. These findings suggest that an active conversion from serine to glycine, which needs tetrahydrofolate, occurs in the posterior silkgland for fibroin synthesis. Moreover, the solid-state ¹³C CP/MAS spectrum of the fibroin prepared from cocoons spun by larvae fed with [¹³C]formate revealed that serine C3 was labelled specifically with ¹³C, suggesting that the reverse conversion from glycine to serine took place in the silkworm. The posterior silkgland has the ability to synthesize not only fibroin but also its major materials, glycine and serine. Received May 4, 1999 Accepted December 10, 1999     

The biosynthesis of fibroin. I. The intracellular form of silk fibroin of Bombyx mori L

After in vivo labeling with [³H]glycine the synthesis and transport of fibroin has been studied by radioautography and cell fractionation.Radioactivity appearing in the cytoplasm is rapidly transferred to the lumen where it accumulates in the so-called silk layer before reaching the central core of secreted fibroin. By sucrose density gradients it was demonstrated that the radioactivity appears immediately in the fibroin fraction, no precursors being observed.A simple fractionation procedure, based on the utilization of detergent, gives three fractions tentatively interpreted as synthesis, transport, and accumulation compartments in accordance with their kinetics of labeling.     

Localization of the genes for silk fibroin in silk gland cells of Bombyx mori.

Radioactive iodinated silk fibroin messenger RNA and ribosomal RNA have been used as probes to localize their genes in tissue sections of Bombyx mori by in situ hybridization. From filter hybridization experiments it is inferred that the majority of the grains produced by in situ hybridization with fibroin mRNA represents specific hybridization to fibroin genes. Sections of the posterior silk gland where silk is synthesized have been compared with those of the middle gland which does not synthesize fibroin. Glands have been analyzed from the second through the fifth (last) larval instar during feeding and moulting periods. During later stages when the gland cells increase their DNA content by polyploidization, serial sections were required to follow the distribution of grains through entire nuclei. At all stages, both ribosomal DNA and fibroin genes are distributed randomly throughout the nuclei without a preferred relationship to any nuclear structure.     

Isoaccepting tRNAS of posterior silkgland of mutant Bombyxmori

The amino acid acceptor activities and the electrophoretic patterns of tRNAs from the normal posterior silkgland producing fibroin and from the posterior silkgland of mutant (symbolized as Nd-s) which secretes little fibroin were compared. The tRNA from the normal silkgland incorporated much [¹⁴C]glycine and [¹⁴C]alanine, and less [¹⁴C]leucine and [¹⁴C]lysine, which reflects the amino acid composition of fibroin. However, this was not observed with the silkgland tRNA of the Nd-s mutant of the silkworm. In the case of two-dimensional electrophoresis of the silkgland tRNA of the Nd-s mutant, fewer isoacceptors were recognized compared with tRNA from the normal silkgland, especially regarding isoacceptors of glycyl-, and alanyl-tRNAs.     

The function of the pentose phosphate pathway in Phaseolus mungo hypocotyls

The metabolism of d-gluconate-[1-¹⁴C] and -[6-¹⁴C] by segments from etiolated hypocotyls of Phaseolus mungo has been studied. The release of ¹⁴CO₂ from gluconate-[1-¹⁴C] was greater than that from gluconate-[6-¹⁴C] in all parts of hypocotyls examined. Incorporation of the radioactivity from gluconate-[6-¹⁴C] into RNA, lignin and aromatic amino acid fractions was greater in the upper (younger) part of the hypocotyls. Incorporation into sugars was greater in the lower (more mature) parts.     

Quantitative measurements of fibroin messenger RNA synthesis in the posterior silk gland of normal and mutant Bombyx mori

The amount of newly synthesized and accumulated fibroin messenger RNA has been measured quantitatively at various stages of posterior silk gland development in Bombyx mori. The two-step method involves fractionation on a Bio-Gel column which excludes the large mRNA, followed by RNAase T₁ digestion, and fractionation of the oligonucleotides on DEAE-Sephadex. Larvae in the feeding stages of the third and fourth instar synthesize and accumulate fibroin mRNA to about 2% of cellular RNA; this corresponds to 0.2 and 2 μg per pair of posterior glands in the third and fourth instars, respectively. More than 70% of this mRNA is degraded in vivo during the third and fourth moulting stages. Fibroin mRNA synthesis resumes again within the first 24 hours of the fifth instar; the mRNA accumulates and predominates over other DNA-like RNAs as the stage proceeds until finally it comprises about 3.5% of cellular RNA in a mature larva (170 μg per pair of posterior glands). These results indicate that more than 99% of the fibroin mRNA detected in the fifth instar is synthesized during this stage.Four spontaneous mutants of B. mori which synthesize very low levels of fibroin have been analyzed for their RNA content in the middle fifth instar. The total cellular RNA of the posterior gland is reduced to 4 to 7% of normal. Fibroin mRNA is more severely reduced to 1% of normal. In three heterozygotes, which have mutant phenotypes with respect to fibroin production, only slight increases of total cellular RNA and fibroin mRNA were observed. Thus, the primary biochemical lesion in these mutants is still unknown.The presumed ancestor to B. mori, the wild silkworm B. mandarina, was also analyzed for its fibroin mRNA. The mRNA isolated from fifth instar larvae of B. mandarina is indistinguishable from that of B. mori with respect to its nucleotide sequence, molecular weight and fraction of total cellular RNA.     

Studies on the posterior silk gland of the silkworm Bombix mori. IV. Ultracentrifugal analyses of native silk proteins, especially fibroin extracted from the middle silk gland of the mature silkworm

Ultracentrifugal analyses of the native silk proteins extracted from the various parts of the middle silk gland of the mature silkworm have revealed that there exist four components with S°_20,w values of 10S, 9–10S, 9S, and 4S in the extract. It is suggested that the fastest 10S component is the native fibroin synthesized in the posterior silk gland and transferred to the middle silk gland to be stored there, while the slower three components probably correspond to inner, middle, and outer sericins which were synthesized in the posterior, middle, and anterior portion of the middle silk gland, respectively. Native fibroin solution was prepared from the most posterior part of the middle silk gland. Ultracentrifugal analyses have shown that the solution contains considerable amounts of aggregates in addition to the main 10S component. Treatment with lithium bromide (LiBr), urea, or guanidine hydrochloride solution up to 6 M all have failed to dissociate the 10S component. From the sedimentation equilibrium analyses and partial specific volume of 0.71₆, the molecular weight of the 10S component of the native fibroin solution was found to be between 3.2 – 4.2 x 10⁵, with a tendency to lie fairly close to 3.7 x 10⁵.     

LIM-homeodomain transcription factor Awh is a key component activating all three fibroin genes,fibH, fibL and fhx,in the silk gland of the silkworm,Bombyx mori

In the silkworm Bombyx mori, three fibroin genes, fibroin-heavy-chain (fibH), fibroin-light-chain (fibL) and fibrohexamerin (fhx), are coexpressed only in the posterior silk gland (PSG) cells, while the sericin genes encoding silk glue proteins are expressed in the middle silk gland (MSG) cells. Silk gland factor-2 (SGF-2) is a PSG-specific activator complex of fibH, composed of a LIM-homeodomain protein, Awh, and its cofactors, Ldb and Lcaf. We investigated whether SGF-2 can activate other fibroin genes using transgenic silkworms. The genes for Ldb and Lcaf were expressed ubiquitously in various tissues, while the gene for Awh was expressed strictly specific in PSG of the wild type silkworms. Misexpression of Awh in transgenic silkworms induced ectopic expression of fibL and fhx as well as fibH in MSG. Coincidently with the induction of fibL and fhx by Awh, binding of SGF-2 to the promoter of fibL and fhx was detected in vitro, and SGF-2 binds directly to the fhx core promoter. Ectopic expression of the fibroin genes was observed at high levels in the middle part of MSG. Moreover, fibL and fhx were induced in the anterior silk gland (ASG) of the transgenic silkworms, but fibH was not. These results indicate that Awh is a key activator of all three fibroin genes, and the activity is probably regulated in conjunction with additional factors.     

Biosynthesis of alkane-2,3-diols: enzymatic reduction of 3-hydroxyoctadecane-2-one to octadecane-2,3-diol by a NADPH (B-side) specific microsomal reductase from the uropygial glands of ring-necked pheasants (Phasianus colchicus).

Cell-free preparations from the uropygial gland of ring-necked pheasant catalyzed the reduction of a synthetic R,S-mixture of 3-hydroxyl[3-¹⁴C]octadecane-2-one (acyloin) to a mixture of threo- and erythro-[3-¹⁴C]octadecane-2,3-diol, the final step in the postulated pathway for the biosynthesis of alkane-2,3-diols. The product of enzymatic reduction was identified by Chromatographic techniques and chemical degradation studies. The acyloin reductase showed a pH optimum near 4.0 and specificity for NADPH. With stereospecifically labeled [³H]NADPH, it was shown that acyloin reductase preferentially transferred hydride from the B-side of the nicotinamide ring to the acyloin. A typical Michaelis-Menten substrate saturation was observed for the acyloin and an apparent K_m of 70 μm was calculated from linear double reciprocal plots. Acyloin reductase was inhibited by thioldirected reagents such as p-chloromercuribenzoate and N-ethylmaleimide. Subcellular fractionation of the gland homogenates using sucrose density gradient centrifugation showed that acyloin reductase activity coincided with NADPH:cytochrome c reductase activity, strongly suggesting that acyloin reductase is localized in the microsomal membranes.     

Integrated 13C-metabolic flux analysis of 14 parallel labeling experiments in Escherichia coli

The use of parallel labeling experiments for ¹³C metabolic flux analysis (¹³C-MFA) has emerged in recent years as the new gold standard in fluxomics. The methodology has been termed COMPLETE-MFA, short for complementary parallel labeling experiments technique for metabolic flux analysis. In this contribution, we have tested the limits of COMPLETE-MFA by demonstrating integrated analysis of 14 parallel labeling experiments with Escherichia coli. An effort on such a massive scale has never been attempted before. In addition to several widely used isotopic tracers such as [1,2-¹³C]glucose and mixtures of [1-¹³C]glucose and [U-¹³C]glucose, four novel tracers were applied in this study: [2,3-¹³C]glucose, [4,5,6-¹³C]glucose, [2,3,4,5,6-¹³C]glucose and a mixture of [1-¹³C]glucose and [4,5,6-¹³C]glucose. This allowed us for the first time to compare the performance of a large number of isotopic tracers. Overall, there was no single best tracer for the entire E. coli metabolic network model. Tracers that produced well-resolved fluxes in the upper part of metabolism (glycolysis and pentose phosphate pathways) showed poor performance for fluxes in the lower part of metabolism (TCA cycle and anaplerotic reactions), and vice versa. The best tracer for upper metabolism was 80% [1-¹³C]glucose+20% [U-¹³C]glucose, while [4,5,6-¹³C]glucose and [5-¹³C]glucose both produced optimal flux resolution in the lower part of metabolism. COMPLETE-MFA improved both flux precision and flux observability, i.e. more independent fluxes were resolved with smaller confidence intervals, especially exchange fluxes. Overall, this study demonstrates that COMPLETE-MFA is a powerful approach for improving flux measurements and that this methodology should be considered in future studies that require very high flux resolution.     

Involvement of glycine transfer ribonucleic acids in development of the posterior silk gland of Bombyx mori

Glycine transfer RNAs from the two physiological phases, V-2, the stage of maximum growth, and V-5, the stage of maximum fibroin production, during the development of the posterior silk gland of Bombyx mori were examined. The tRNAs from both phases could be fractionated into two major isoaccepting species on a benzoylated DEAE-cellulose column. No significant qualitative differences were observed among the tRNAs, but the total amount of the isoaccepting species of tRNA^Gly in each gland of V-5 stage was 6-fold higher than the amount of tRNA^Gly in the V-2 gland. The codon recognition properties of the tRNA^Gly species were examined. It was found that tRNA^Gly₁ responded to the copolymer (G:U) preferentially while tRNA^Gly_II recognized the copolymer (A:G). The ratio between the extent of incorporation of labeled glycine from glycyl-tRNA^Gly₁ and glycyl-tRNA^Gly_II into protein in a cell-free system utilizing polysomes from the V-5 glands was similar to the relative abundance of the isoaccepting species present in the glands at that time. It also reflected the ratio between the corresponding codons assigned for glycine based on the sequence analysis of fibroin-mRNA [Suzuki, Y., and Brown, D. D. (1972) J. Mol. Biol.63: 409]. These results suggest that the abundance of tRNA^Gly in the posterior silk gland and the changes in the relative amounts of the isoaccepting species are quite specific for the development of the gland.     

Biosynthesis of alkane-2, 3-diols: chemical synthesis of 3-hydroxy-[3-14C]octadecane-2-one and its reduction to [3-14C]octadecane-2, 3-diol in the uropygial glands of ring-necked pheasants (Phasianus colchicus).

Acyloin has been proposed to be an intermediate in the biosynthesis of long chain alkane-2,3-diols. In order to test this possibility, specifically labeled 3-hydroxyoctadecane-2-one (acyloin) was synthesized by coupling 2-methyl-1,3-dithiane with [1-¹⁴C]hexadecanal followed by cleaving of the thioketal. Injection of the synthetic 3-hydroxy [3-¹⁴C]octadecane-2-one into the uropygial gland of the ring-necked pheasant resulted in the formation of labeled octadecane-2,3-diol. Chemical degradation of this diol showed that all of the ¹⁴C was contained in C-3 of the diol showing direct conversion of acyloin to the diol. These observations support the hypothesis that alkane-2,3-diols might be biosynthesized by reduction of the acyloin derived from a condensation between hydroxyethyl thiamine pyrophosphate and fatty aldehyde. Gas-liquid chromatographic analysis of the alkane-2,3-diols, as their isopropylidene derivatives, of the pheasant strongly suggests that they are of the erythro-configuration; however, alkane-2,3-diol enzymatically formed from the racemic acyloin injected into the gland contained 59.5% erythro- and 40.5% threo-diastereoisomers. This distribution was identical to that produced by chemical reduction of the synthetic racemic acyloin. These results clearly show that the reduction step does not show a preference for either of the enantiomers of the acyloin and that the stereospecificity in diol biosynthesis probably resides in the condensation step.     

Microbial growth on C1 compounds. Synthesis of cell constituents by methane- and methanol-grown Pseudomonas methanica

1. A study has been made of the incorporation of carbon from [¹⁴C]methane, [¹⁴C]methanol and [¹⁴C]bicarbonate by cultures of Pseudomonas methanica growing on methane, and [¹⁴C]methanol by cultures of the same organism growing on methanol. 2. The distribution of radioactivity within the non-volatile constituents of the ethanol-soluble fractions of the cells, after incubation with labelled compound for periods up to 3min., has been analysed by chromatography and radioautography. 3. Over 90% of the radioactivity fixed from [¹⁴C]methane or [¹⁴C]methanol at the earliest times of sampling appeared in phosphorylated compounds. Glucose phosphate and fructose phosphate together constituted the largest part of the radioactive phosphates (70–90%); phosphoglycerate was a relatively minor component (2–17%). Other compounds becoming labelled during the incubation included glycine, serine, glutamate, aspartate, malate, citrate and alanine. 4. The first stable products of [¹⁴C]bicarbonate fixation were malate and aspartate (containing between them over 90% of the total radioactivity fixed at the earliest times of sampling). 5. The percentage of the total radioactivity fixed that was contained in each of the radioactive compounds has been plotted against time. The slopes of the curves obtained show that hexose phosphates are primary stable products of [¹⁴C]methane and [¹⁴C]methanol incorporation and that aspartate and malate are primary stable products of [¹⁴C]bicarbonate incorporation. 6. No carboxydismutase activity has been found in cell-free extracts of the organism. This fact, together with the other findings, shows that an autotrophic metabolism involving the ribulose diphosphate cycle of carbon dioxide fixation cannot be operating.     

REGULATION OF PROTEIN SYNTHESIS IN THE VENOM GLAND OF VIPERID SNAKES

Morphological changes in the venom gland of V. ammodytes were studied after the removal of the venom from the gland lumina (milking) It was found that the height of the secretory cells was changed during the secretory cycle. The patterns of the rough endoplasmic reticulum and of the Golgi complex were changed as well Milking induced an increased incorporation of [¹⁴C]amino acids into total and venom proteins In V ammodytes, during the first day after milking, 25% of the total counts in protein were precipitable by anti-venom serum, while at 8 days, 80% of the proteins synthesized were venom proteins At this stage, the incorporation was 10- and 20-fold that of unmilked glands for total and venom proteins, respectively. Venom was accumulated (secreted) in the gland lumina of V. ammodytes at a relatively high rate up to 2 wk after milking and leveled off afterwards. Intact glands and gland slices of V ammodytes and V palaestinae, taken from snakes a few days after milking, incorporated [¹⁴C]amino acids into proteins in vitro at a rate higher than that of unmilked glands. The activity of two exportable enzymes (phosphodiesterase and benzoyl arginyl ethyl esterase) was assayed in gland homogenates of V. ammodytes. It was found that 2–3 wk after milking, the intracellular level of these enzymes was up to 2-fold that of unmilked glands.     

SECRETORY PROTEIN SYNTHESIS IN THE STIMULATED RAT PAROTID GLAND : Temporal Dissociation of the Maximal Response from Secretion

Administration of the β-adrenergic drug, isoproterenol (IPR), affects the release of 98% of stored amylase from rat parotid gland acinar cells. A period of 6 h elapses from the onset of secretion to the maximum [¹⁴C]phenylalanine (Phe) incorporation into total protein and amylase. 10 h after IPR administration the rate of [¹⁴C]Phe incorporation into total protein was no longer elevated above that of control. Incorporation into amylase, however, remained elevated above the control by 2.3 times. This latent period may reflect: (a) reduced amounts of available ATP which occurs as a result of the process of secretion as well as (b) the time required for reorganization of cellular organelles and membranes after secretion. The latent period after IPR-induced secretion appears similar to the latent period which has recently been reported to occur after physiologic release of amylase from the parotid gland during the diurnal feeding cycle of the rat. These observations support the existence of a positive feedback system operant in the parotid acinar cell linking the release of secretory proteins with their synthesis. The period of greatest protein synthesis is, however, temporally dissociated from the secretory process.     

Metabolic pools associated with monoterpene biosynthesis in higher plants

Partial degradations of (+)-isothujone biosynthesised in Tanacetum vulgare after feeding IPP-[4-¹⁴C], DMAPP-[4-¹⁴C] or 3,3-dimethylacrylate-[Me-¹⁴C], and of geraniol and (+)-pulegone formed in Pelargonium graveolens and Mentha pulegium respectively after uptake of 3,3-dimethylacrylate-[Me-¹⁴C], indicated that none of these metabolites was a direct source of the part of the monoterpene skeleton derived hypothetically from DMAPP. Uptake of glucose-[U¹⁴C] into P. graveolens led, in contrast, to both IPP and DMAPP-derived moieties of geraniol being extensively labelled. Feeding of l-valine-[U-¹⁴C] and l-leucine-[U-¹⁴C] to all three plants resulted in negligible incorporation of tracer into monoterpenes. A soluble enzyme system prepared from foliage of T. vulgare that had been exposed to CO₂-[¹⁴C] for 20 days converted isotopically-normal IPP into GPP with the DMAPP-derived portion containing essentially all (>98%) of the radioactivity present. These observations and those previously obtained from feeding experiments with other [¹⁴C]-labelled precursors on the same plant species are consistent with the occurrence of two metabolic pools of intermediates for monoterpene biosynthesis, one of which is probably protein-bonded.     

C Nuclear Magnetic Resonance Study of Mannitol Cycle and Trehalose Synthesis during Glucose Utilization by the Ectomycorrhizal Ascomycete Cenococcum graniforme

¹³C nuclear magnetic resonance spectroscopy has been used to follow the utilization of glucose for the synthesis of carbohydrates in the ectomycorrhizal ascomycete Cenococcum graniforme. The fate of ¹³C label was analyzed in vivo and in mycelial extracts. The major carbohydrates produced from [1-¹³C]glucose and [6-¹³C]glucose were mannitol and trehalose. Mannitol was mainly synthesized via a direct route from glucose. Scrambling of the ¹³C label was observed to occur in trehalose during glycolysis. From the analysis of the scrambling patterns, it is concluded that the mannitol cycle was operative and that a large part of the carbon of glucose was used to form trehalose after cycling through the mannitol pool. The activities of NAD-mannitol-l-P dehydrogenase (EC 1.1.1.17) and NADP-mannitol dehydrogenase (EC 1.1.1.138), which participate in the mannitol cycle relative to the activity of glycolytic enzymes, provide evidence that the cycle is important for NADPH production.