Protein-altering variants of PTPN2 in childhood-onset Type 1A diabetes.

AIM: To examine the contribution of PTPN2 coding variants to the risk of childhood-onset Type 1A diabetes. METHODS: PTPN2 mutation analysis was carried out for 169 unrelated Japanese people with childhood-onset Type 1A diabetes. We searched for coding variants that were absent or extremely rare in the general population and were scored as damaging by multiple in silico programs. We performed mRNA analysis and three-dimensional structural prediction of the detected variants, when possible. We also examined possible physical links between these variants and previously reported risk SNPs as well as clinical information from variant-positive children. RESULTS: One frameshift variant (p.Q286Yfs*24) and two probably damaging missense substitutions (p.C232W and p.R350Q) were identified in one child each. Of these, p.Q286Yfs*24 and p.C232W were hitherto unreported, while p.R350Q accounted for 2/121,122 alleles of the exome datasets. The p.Q286Yfs*24 variant did not encode stable mRNA, and p.C232W appeared to affect the structure of the tyrosine-protein phosphatase domain. The three variants were physically unrelated to known risk SNPs. The variant-positive children manifested Type 1A diabetes without additional clinical features and invariably carried risk human leukocyte antigen alleles. CONCLUSIONS: The results provide the first indication that PTPN2 variants contribute to the risk of Type 1A diabetes, independently of known risk SNPs. PTPN2 coding variants possibly induce non-specific Type 1A diabetes phenotypes in individuals with human leukocyte antigen-mediated disease susceptibility. Our findings warrant further validation.

Verification of the Usefulness of an Assessment and Risk Control Sheet that Promotes Management of Cancer Drug Therapy.

Background: Proper management of adverse events is crucial for the safe and effective implementation of anticancer drug treatment. Showa University Hospital uses our interview sheet (assessment and risk control [ARC] sheet) for the accurate evaluation of adverse events. On the day of anticancer drug treatment, a nurse conducts a face-to-face interview. As a feature of the ARC sheet, by separately describing the symptoms the day before treatment and the day of treatment and sharing the information on the medical record, it is possible to clearly determine the status of adverse events. In this study, we hypothesized that the usefulness and points for improvement of the ARC sheet would be clarified by using and evaluating a patient questionnaire. Methods: This study included 174 patients (144 at Showa University Hospital (Hatanodai Hospital) and 30 at Showa University Koto Toyosu Hospital (Toyosu Hospital) who underwent pre-examination interviews by nurses and received cancer chemotherapy at the outpatient center of Hatanodai and Toyosu Hospital. In the questionnaire survey, the ARC sheet's content and quality, respondents' satisfaction, structural strengths, and points for improvement were evaluated on a five-point scale. Results: The patient questionnaire received responses from 160 participants, including the ARC sheet use group (132 people) and the non-use group (28 people). Unlike the ARC sheet non-use group, the ARC sheet use group recognized that the sheet was useful to understand the adverse events of aphthous ulcers (p = 0.017) and dysgeusia (p = 0.006). In the satisfaction survey questionnaire, there was a high sense of security in the pre-examination interviews by nurses using the ARC sheet. Conclusions: The ARC sheet is considered an effective tool for comprehensively evaluating adverse events. Pre-examination interviews by nurses using ARC sheets accurately determined the adverse events experienced by patients with anxiety and tension due to confrontation with physicians.

A case of cerebral aquaporinopathy.

A 35-year-old woman was hospitalized due to impaired consciousness. Magnetic resonance imaging (MRI) revealed multiple parenchymal lesions in supra and infratentorial brain regions, which were considered responsible for her declining consciousness level. She was treated with intravenous methylprednisolone. Neurological symptoms improved and she was discharged. She was readmitted 14 months later due to intractable hiccups. A follow-up brain MRI revealed an abnormal signal near the area postrema in the dorsal medulla. Serum aquaporin-4 antibody levels were positive, but there were no visual manifestations or myelitis. Spinal MRI was negative for longitudinally extended transverse myelitis throughout the clinical course.

Association of single nucleotide polymorphisms in the endothelial differentiation sphingolipid G-protein-coupled receptor 1 gene with marbling in Japanese Black beef cattle.

Marbling defined by the amount and distribution of intramuscular fat, so-called Shimofuri, is an economically important trait of beef cattle in Japan. The endothelial differentiation sphingolipid G-protein-coupled receptor 1 (EDG1) gene, involved in blood vessel formation, has been previously shown to be expressed at different levels in musculus longissimus muscle between low-marbled and high-marbled steer groups. It is located within the genomic region of a quantitative trait locus for marbling, and thus was considered as a positionally functional candidate for the gene responsible for marbling. In this study, two single nucleotide polymorphisms (SNPs) in the 5' untranslated region (UTR) and the 3' UTR of EDG1, referred to as c.-312A>G and c.*446G>A, respectively, were detected between the two steer groups. The two SNPs were associated with the predicted breeding value for beef marbling standard number by analyses using a population of Japanese Black beef cattle. The effect of genotypes at each of the SNPs on the predicted breeding value for subcutaneous fat thickness was not statistically significant (P > 0.05). Reporter gene assays revealed no significant differences in gene expression between alleles at each of the SNPs. These findings suggest that EDG1 SNPs, although they may not be regarded as a causal mutation, may be useful for effective marker-assisted selection to increase the levels of marbling in Japanese Black beef cattle.

Dimethyl sulphoxide relaxes rabbit detrusor muscle by decreasing the Ca2+ sensitivity of the contractile apparatus.

BACKGROUND AND PURPOSE: The intravesical administration of dimethyl sulphoxide (DMSO) is used to alleviate the symptoms of interstitial cystitis. We investigated the relaxant effect of DMSO and its underlying mechanism in the detrusor muscle. EXPERIMENTAL APPROACH: The effects of DMSO on contraction, on Ca2+ sensitivity of myofilaments, and on myosin light chain (MLC) phosphorylation were investigated in both intact and alpha-toxin-permeabilized strips of rabbit detrusor muscle. KEY RESULTS: In fura-PE3-loaded strips, DMSO (>1%) induced a significant relaxation during sustained contractions induced by 60 mM K+-depolarization or 10 microM carbachol, while having no effect on the [Ca2+](i) level. DMSO decreased the level of MLC phosphorylation during the contractions induced by 60 mM K+ and 10 microM carbachol. DMSO also inhibited both the contraction and MLC phosphorylation induced by calyculin-A in intact strips. In the alpha-toxin-permeabilized preparations, DMSO relaxed the Ca2+-induced contraction and also inhibited the tension development induced by a stepwise increment of Ca2+ concentrations. Such a relaxant effect of DMSO was enhanced in the presence of phosphate. CONCLUSIONS AND IMPLICATIONS: DMSO relaxes rabbit detrusor muscle by decreasing the Ca2+ sensitivity of myofilaments. Inhibition of the kinase activities involved in myosin phosphorylation may play a major role in DMSO-induced Ca2+ desensitization. Inhibition of the cross-bridge cycling at the step of phosphate release may also contribute to the relaxant effect of DMSO. Such relaxant effects of DMSO could be linked to the therapeutic effect of DMSO in interstitial cystitis.

The primary structure of rat rig/ribosomal protein S15 gene.

We have isolated rat rig/ribosomal protein S15 gene from a DNA library derived from a rat insulinoma and determined the complete nucleotide sequence. The rat rig/S15 gene is composed of four exons and three introns spanning 2 kbp and exhibits distinctive structural features unique for a ribosomal protein gene.

Phosphorus nuclear magnetic resonance of perfused salivary gland.

Phosphorus nuclear magnetic resonance (31P-NMR) was used for the sequential measurement of phosphorus energy metabolites in perfused canine submandibular gland. Under resting conditions, ATP and creatine phosphate levels were 0.42 +/- 0.11 mM and 0.62 +/- 0.16 mM (mean +/- S.D., in nine glands). When perfusion of the gland was stopped, the tissue contents of ATP and creatine phosphate decreased, that of ADP increased and tissue pH decreased. Restarting perfusion led to recoveries of the tissue content of the phosphorus compounds and tissue pH to normal. Acetylcholine administration induced secretion of saliva, decreased the level of ATP, creatine phosphate and tissue pH, and increased the ADP level.

Sequence of the chicken rig gene encoding ribosomal protein S15.

The nucleotide sequence of the chicken rig gene encoding ribosomal protein S15 was determined. The 1.6-kb gene consists of four exons and three introns. The 5'-flanking region of the gene lacks TATA- or CAAT-box sequences. Several GC-box sequences were found around the transcription start point.

Inhibition of lipid peroxidation and the active oxygen radical scavenging effect of anthocyanin pigments isolated from Phaseolus vulgaris L.

No attention has been paid to anthocyanin pigments from the viewpoint of inhibitors of lipid peroxidation and scavengers of active oxygen radicals; therefore, we investigated the antioxidative, radical scavenging, and inhibitory effects on lipid peroxidation by UV light irradiation of three anthocyanin pigments, pelargonidin 3-O-beta-D-glucoside (P3G), cyanidin 3-O-beta-D-glucoside (C3G), and delphinidin 3-O-beta-D-glucoside (D3G), isolated from the Phaseolus vulgaris L. seed coat, and their aglycons, pelargonidin chloride (Pel), cyanidin chloride (Cy), and delphinidin chloride (Del). All pigments had strong antioxidative activity in a liposomal system and reduced the formation of malondialdehyde by UVB irradiation. On the other hand, the extent of antioxidative activity in a rat liver microsomal system and the scavenging effect of hydroxyl radicals (-OH) and superoxide anion radicals (O2-) were influenced by their own structures.

Preparation of separated alpha and beta subunits of electron-transferring flavoprotein in unfolded forms and their restoration to the native holoprotein form.

Electron-transferring flavoprotein (ETF) isolated from pig kidney is a heterodimer containing one FAD and one AMP [Sato, K. et al. (1993) J. Biochem. 114, 215-222]. This paper presents a method for separation of the alpha and beta subunits on a preparative scale. The subunits were separated by cation-exchange chromatography in the presence of 6 M urea at pH 7.6. The subunits were obtained in unfolded forms with denaturant. The unfolded subunits were restored to the heterodimeric form having FAD and AMP with high yield after incubation with FAD and AMP under non-denaturing conditions. The far-UV CD spectrum, the fluorescence and absorption spectra of the bound FAD, and the enzyme activity of the reconstituted ETF were all identical to those of the ETF isolated from pig kidney. Thus, an experimental system was established for the reconstitution of ETF from the four components, alpha, beta, FAD, and AMP. Analyses of the absorption spectra of the separated subunits suggested that the contents of aromatic residues were one tryptophan and six tyrosines in alpha and one tryptophan and one tyrosine in beta. The possibility that one subunit binds with FAD or AMP in the absence of the other subunit was examined by measuring the flavin and protein fluorescence spectra, but no spectral changes reflecting such binding were detected. This result suggests that the coexistence of both the subunits is necessary for the binding of FAD and AMP with the protein.

The binding of adenine nucleotides to apo-electron-transferring flavoprotein.

Apoprotein of electron-transferring flavoprotein (ETF) reacts with FAD as follows: A*<-->A, A+FAD<-->holoETF. Two different forms of apoETF (A* and A) convert into each other and only one of them, A, can associate with FAD [Sato, K. et al. (1991) J. Biochem. 109, 734-740]. In the present study, the reactions between apoETF and ATP, ADP, AMP, riboflavin, or FMN were investigated. It was revealed that all three adenine nucleotides bind with apoETF with the same kinetic reaction scheme as FAD, and compete with FAD. These results suggest that the nucleotides bind to A with the same location as the ADP part of FAD in holoETF and that the ADP-binding site of apoETF is generated upon conversion from A* to A. Neither riboflavin nor FMN bound to apoETF regardless of the presence or absence of the nucleotides, indicating that the ADP part of the FAD molecule is essential to the incorporation of the isoalloxazine ring into ETF. The binding rate constant of FAD to A was 1/20 of that of ADP while the dissociation rate constant was 1/1,000. This indicates that the riboflavin part of FAD inhibits the binding of FAD by steric hindrance, while after the binding, it stabilizes the complex.

Complex formation between reduced D-amino acid oxidase and pyridine carboxylates.

Picolinate binds to a reduced form of D-amino acid oxidase, and the complex formed has a broad absorption band around 600 nm as in the case of the purple intermediate of the enzyme with a substrate. The dissociation constant at 25 degrees C was 35 microns at pH 7.0. The pH dependence (pH 8.3-pH 6.4) of the dissociation constant indicates that one proton is associated with the complex formation, and picolinate protonated at the N atom binds to the reduced enzyme. Resonance Raman spectra of the complex support that picolinate in the complex is a cationic form protonated at the N atom. Nicotinate also binds to the reduced enzyme, but isonicotinate does not.

Resonance Raman spectra of anionic semiquinoid form of a flavoenzyme, D-amino acid oxidase.

Resonance Raman (RR) spectra of the complex of anionic semiquinoid D-amino acid oxidase (DAO) with picolinate in H2O and D2O were observed in the 300-1,750 cm-1 region. RR spectra were also measured for the complex of the semiquinoid enzyme reconstituted with isotopically labeled FAD's, i.e., [4a-13C]-, [4,10a-13C2]-, [2-13C]-, [5-15N]-, and [1,3-15N2]-FAD. On the basis of the isotope effects, tentative assignments of the observed bands of the anionic semiquinoid flavin were made. The spectra differ from those of oxidized, neutral semiquinoid, and anionic reduced flavins previously reported. The 1,602 cm-1 band was not shifted for any FAD labeled in ring II and/or ring III and was assigned to a ring I mode. The 1,516 cm-1 band underwent an isotopic shift upon [4a-13C]- or [4,10a-13C2]-labeling. The band was assigned to the mode containing C(4a)-C(10a) stretching. The 1,331 and 1,292 cm-1 bands shifted upon [4a-13C]- or [5-15N]-labeling and were assigned to the modes containing C(4a)-N(5) stretching. The 1,217 and 1,188 cm-1 bands were assigned to the skeletal vibrations of ring III coupled with the N(3)-H bending mode. The RR spectrum of the complex of anionic semiquinoid DAO with alpha-iminopropionate or N-methyl-alpha-iminopropionate was essentially identical with that of the complex with picolinate.

Electron-transferring flavoprotein has an AMP-binding site in addition to the FAD-binding site.

Mammalian electron-transferring flavoprotein (ETF) has been reported to consist of two non-identical subunits and one FAD. The present paper shows that ETF purified from pig kidney contains one more molecule, an AMP. ETF was denatured by guanidine hydrochloride and ultrafiltered for the purpose of removing proteins. The filtrate was analyzed by reverse-phase chromatography. Two peaks appeared on the chromatogram: they were identified as FAD and AMP, and their molar amounts were identical, indicating that ETF contains one AMP molecule. ApoETF, which was prepared by KBr treatment of ETF, also contains one AMP molecule. ApoETF, which was prepared by KBr treatment of ETF, also contain one AMP molecule. These results clearly demonstrate that ETF has an AMP-binding site in addition to the FAD-binding site. AMP-free apoETF was prepared by guanidine treatment of ETF. Mixing AMP-free apoETF, FAD, and AMP produced reconstituted ETF, which showed the same properties as native ETF. Mixing AMP-free apoETF and FAD produced AMP-free ETF, regardless of the coexistence of ATP or ADP: the AMP-binding site cannot bind FAD, ADP, or ATP. The enzymatic activity of the AMP-free ETF for electron transfer from substrate-reduced medium-chain acyl-CoA dehydrogenase to 2,6-dichlorophenolindophenol was identical to that of native ETF. This indicates that the AMP contained in holoETF has no apparent influence on this enzymatic activity. A role of AMP recognized in this study is that AMP facilitates the formation of holoETF from AMP-free apoETF, FAD, and AMP.

Anion-induced conformational change of apo-electron-transferring flavoprotein.

Apoprotein of electron-transferring flavoprotein (ETF) exists in an equilibrium between two different forms, only one of which can associate with FAD (Sato, K. et al. (1991) J. Biochem. 109, 734-740), as represented in the following kinetic scheme: A* in equilibrium with A, A+FAD in equilibrium with holoETF, where "A*" and "A" are the different forms of apoETF. In the present study, the effects of various anions on the conversion between the two forms of apoETF were investigated by kinetic analyses on binding of FAD to apoETF. All the anions tested here induced the conversion from "A*" to "A"; the order of the effectiveness was I- approximately Br- greater than Cl- greater than F-. Further, glycerol also induced the conversion from "A*" to "A". The elution pattern of apoETF on molecular sieve chromatography was changed by addition of salts or glycerol; this change was due to the conversion from "A*" to "A" by the added solutes. The "A*" form was eluted more rapidly than the "A" form, indicating that the "A*" form exists in a looser conformation than the "A" form. The far-UV CD spectral change upon addition of salts indicated that a greater part of the secondary structure is retained in the conversion from "A*" to "A," but the "A" form contains a somewhat larger amount of beta-sheet than "A*."

In vitro assembly of FAD, AMP, and the two subunits of electron-transferring flavoprotein: an important role of AMP related with the conformational change of the apoprotein.

Electron-transferring flavoprotein from pig kidney is composed of four non-covalently bound components: alpha and beta subunits, flavin adenine dinucleotide (FAD), and adenosine monophosphate (AMP). This paper reveals the pathway of assembly of the electron-transferring flavoprotein. The holoprotein can be formed by two different pathways. (i) alpha + beta <==> (alpha-beta)*, (alpha-beta)* + AMP <==> (alpha-beta-AMP)*, (alpha-beta-AMP)* <==> alpha-beta-AMP, alpha-beta-AMP + FAD <==> holoprotein. (ii) alpha + beta <==> alpha-beta, alpha-beta + FAD <==> alpha-beta-FAD, alpha-beta-FAD + AMP <==> holoprotein. Here the presence and absence of asterisks distinguish different conformations with the same composition. The monomeric forms of alpha and beta showed no significant binding with FAD and AMP. AMP and FAD associated with different heterodimer forms which were formed as a result of weak binding between alpha and beta. The binding of alpha + beta + AMP was much faster than that of alpha + beta + FAD because the rate of alpha + beta --> (alpha-beta)* was much faster than that of alpha + beta --> alpha-beta. The alpha-beta-AMP complex associated with FAD rapidly. As a result, the binding of FAD with the subunits is promoted by AMP. The alpha-beta-FAD complex associated with AMP much more slowly than the mixture of alpha and beta. Thus the AMP binding with the subunits is inhibited by the preceding FAD binding.

In vitro refolding and unfolding of subunits of electron-transferring flavoprotein: characterization of the folding intermediates and the effects of FAD and AMP on the folding reaction.

Electron-transferring flavoprotein (ETF) from pig kidney is composed of two subunits (alpha and beta, molecular weights of 33,000 and 29,000) and two small molecules, FAD and AMP. In this study, in vitro refolding and unfolding of the subunits of ETF were carried out with urea as the denaturing reagent. The refolding reaction of alpha and beta was revealed to proceed kinetically in two steps: D in equilibrium with I-->N, where D,I, and N denote the denatured, intermediate, and native forms, respectively. The features of the I forms of alpha and beta, described below, are consistent with the concept of the so-called "molten globule state," which is frequently observed in protein refolding. (i) The conversion between D and I was very rapid. (ii) The I form showed as much secondary structure as the N form as judged from the far-UV circular dichroism. (iii) The solvent accessibility of the I form, estimated by the analysis of equilibrium unfolding experiments, was intermediate between those of the D and N forms. (iv) The standard free energy of the I form is almost the same as that of the D form. The refolding reaction progressed more slowly and the environment of the tryptophan chromophore was changed more drastically in beta refolding that in alpha refolding. We previously reported that the reconstitution of holoETF from denatured subunits is speeded up by increasing the AMP concentration. In this study, the effects of AMP, FAD, and the other subunit on the single subunit folding were examined, but no effect was detected. This result suggests that AMP plays a role in a later process, namely, assembly of the four components (refolded alpha and beta, FAD, and AMP).

Isomerization of delta 1-piperideine-2-carboxylate to delta 2-piperideine-2-carboxylate on complexation with flavoprotein D-amino acid oxidase.

The 1,646 cm-1 band in a resonance Raman spectrum obtained with excitation in the charge-transfer band of the complex of oxidized D-amino acid oxidase (DAO) with the oxidation product of D-lysine catalyzed by DAO shifted to 1,617 cm-1 upon 2-13C substitution of lysine. Thus, the band is assigned to a C(2) = C(3) stretching mode of the enamine, delta 2-piperideine-2-carboxylate (En). In the enzyme-free solution, the product is preferentially in the cyclic imine form, delta 1-piperideine-2-carboxylate (Im). Thus, DAO has a higher affinity for the enamine form than for the imine form. The pH effects on the affinity of DAO for the product and on the molar absorption coefficient at 630 nm in the charge-transfer band, suggest that the enzyme-bound product is En in the neutral form at the N atom. As the value of observed rate constant between DAO and the product was constant at high product concentrations, the binding mechanism can be explained as follows; E + Im in equilibrium with EIm in equilibrium with EEN: rapid bimolecular and slow unimolecular processes. The isomerization of the imine form to the enamine form proceeds in the slow process. The low affinity of Im for DAO may be due to a steric repulsion of the hydrogen atoms of Im at C(3) in the active site. The hydrogen atoms of a substrate D-amino acid at C(3), which correspond to the C(3) hydrogens of Im, may act repulsively in the active site and the repulsive energy may induce strain or distortion of the substrate and the enzyme, accelerating the catalytic reaction.

Proton release from flavoprotein D-amino acid oxidase on complexation with the zwitterionic ligand, trigonelline.

Trigonelline, i.e., N-methylnicotinate, which has a zwitterionic structure similar to a substrate D-amino acid, is a useful active site probe for D-amino acid oxidase (DAO). The affinity of trigonelline for DAO in the deprotonated state at the enzyme bound FAD 3-imino group is higher than in the neutral state, contrary to in the case of benzoate, which is a competitive inhibitor and is in a monoanionic form. The time course of the absorbance change was monitored for the binding of DAO with trigonelline by means of a stopped-flow technique. The reaction, on monitoring at 507 nm, was found to be biphasic at pH 8.3, with fast and slow phases. The dissociation of the 3-imino proton of the enzyme bound FAD was observed in the same time course as the slow phase. These results suggest that the positive charge of trigonelline exists near the 3-imino group of the enzyme bound FAD and interacts repulsively with the proton of the 3-imino group. The absorption spectra of the DAO-trigonelline complex at various pHs also support this hypothesis. In the catalysis of DAO, a similar mechanism may be involved, that is, the positive charge of a D-amino acid may interact repulsively with the 3-imino proton of the enzyme bound FAD, and this interaction may be important for the catalysis.

Substrate recognition and activation mechanism of D-amino acid oxidase: a study using substrate analogs.

We investigated the mechanism of recognition and activation of substrate by D-amino acid oxidase (DAO) by thermodynamical and spectrophotometric methods using zwitterionic ligands [N-methylisonicotinate (NMIN), trigonelline, and homarine] and monoanionic ligands as model compounds of the substrate and the product. In terms of the charge within the substrate D-amino acid, monoanionic (e.g., benzoate), zwitterionic (e.g., NMIN), and dianionic (e.g., terephthalate) ligands are thought to be good models for neutral, basic, and acidic amino acids, respectively, because when a substrate binds to DAO, as previously reported, the a-ammonium group (-NH(3)(+)) probably loses a proton to become neutral (-NH(2)) before the oxidation. Zwitterionic ligands can also be good model compounds of product in the purple complex (the complex of reduced DAO with the product imino acid), because the imino nitrogen of the imino acid is in a protonated cationic form. We also discuss electrostatic interaction, steric effect, and charge-transfer interaction as factors which affect the affinity of substrate/ligand for DAO. Monoanionic ligands have high affinity for neutral forms of oxidized and semiquinoid DAO, while zwitterionic ligands have high affinity for anionic forms of oxidized, semiquinoid, and reduced DAO; this difference was explained by the electrostatic interaction in the active site. The low affinity of homarine (N-methylpicolinate) for oxidized DAO, as in the case of o-methylbenzoate, is due to steric hindrance: one of the ortho carbons of benzoate is near the phenol carbons of Tyr228 and the other ortho carbon is near the carbonyl oxygen of Gly313. The correlation of the affinity of meta- and para-substituted benzoates for oxidized DAO with their Hammet's s values are explained by the HOMO-LUMO interaction between the phenol group of Tyr224 and the benzene ring of benzoate derivative. The pK(a) of neutral flavin [N(3)-H of oxidized flavin, N(5)-H of semiquinoid flavin, and N(1)-H of reduced flavin] decreases by its binding to the apoenzyme. The magnitude of the decrement is oxidized flavin < semiquinoid flavin < reduced flavin. The largest factor in the substantially low pK(a) of reduced flavin in DAO is probably the steric hindrance between the hydrogen atom of H-N(1)(flavin) and the hydrogen atom of H-N of Gly315, which becomes significant when a hydrogen is bound to N(1) of flavin.