Thionein Can Serve As a Reducing Agent for the Methionine Sulfoxide Reductases

Thionein can serve as a reducing agent for the methionine sulfoxide reductases

Daphna Sagher*†, David Brunell*†, J. Fielding Hejtmancik‡, Marc Kantorow*, Nathan Brot§, and Herbert Weissbach*¶

*Center for Molecular Biology and Biotechnology, Florida Atlantic University, Boca Raton, FL 33431; ‡Ophthalmic Genetic and Visual Function Branch, National Eye Institute, National Institutes of Health, Bethesda, MD 20892; and §Department of Microbiology and Immunology, Hospital for Special Surgery, Cornell University Medical Center, New York, NY 10021

Contributed by Herbert Weissbach, April 7, 2006 It has been generally accepted, primarily from studies on methio- not been reported. Most in vitro studies have used DTT as the nine sulfoxide reductase (Msr) A, that the biological reducing agent reducing agent for both MsrA and MsrB, because this agent for the members of the Msr family is reduced thioredoxin (Trx), works very well with the Msr family of proteins. although high levels of DTT can be used as the reductant in vitro. Recently, the human MsrB genes have been of interest to us as Preliminary experiments using both human recombinant MsrB2 part of our studies on the role of the Msr system in protecting lens (hMsrB2) and MsrB3 (hMsrB3) showed that although DTT can and retinal cells against oxidative damage (18–20). Using a color- function in vitro as the reducing agent, Trx works very poorly, imetric assay for Msr activity, based on the reduction of the prompting a more careful comparison of the ability of DTT and Trx individual epimers of 4-N,N-dimethylaminoazobenzene-4-sulfonyl to function as reducing agents with the various members of the (DABS)-Met(o), we were surprised to find that Trx serves very Msr family. Escherichia coli MsrA and MsrB and bovine MsrA poorly as the reducing system for human recombinant MsrB2 efﬁciently use either Trx or DTT as reducing agents. In contrast, (hMsrB2) and MsrB3 (hMsrB3). This result prompted us to exam- hMsrB2 and hMsrB3 show <10% of the activity with Trx as ine in more detail the reducing requirement for different members compared with DTT, raising the possibility that, in animal cells, Trx of the Msr family and to search for a reducing system in mammalian may not be the direct hydrogen donor or that there may be a cells that could function in place of Trx. A heat-stable factor in Trx-independent reducing system required for MsrB2 and MsrB3 bovine liver was detected that, in the presence of EDTA, could activity. A heat-stable protein has been detected in bovine liver serve as the reducing system for several of the Msr proteins. This that, in the presence of EDTA, can support the Msr reaction in the factor was purified and identified as a zinc-containing metal- absence of either Trx or DTT. This protein has been identiﬁed as a lothionein (Zn-MT). The results indicate that loss of zinc from zinc-containing metallothionein (Zn-MT). The results indicate that Zn-MT after treatment with EDTA yields cysteine-rich thionein thionein (T), which is formed when the zinc is removed from Zn-MT, (T), which can provide the reducing power for the Msr enzymes. can function as a reducing system for the Msr proteins because of The possible significance of this finding relative to the redox state its high content of cysteine residues and that Trx can reduce of the cell, oxidative damage, and the role of the Msr system will be oxidized T. discussed.

metallothioneins ͉ reducing systems Results Reduced Trx Is Not an Efficient Reducing Agent for hMsrB2 and he methionine sulfoxide reductases (Msrs) are a family of hMsrB3. In the course of developing the DABS colorimetric assay Tproteins that reduce methionine sulfoxide [Met(o)] to me- for Msr activity (see Materials and Methods), it was confirmed thionine (Met) (1). The oxidation of Met to Met(o) results in the that eMsrA and eMsrB could use either DTT or Trx to supply formation of two epimers of Met(o), called Met-S-(o) and the reducing power, with similar or more activity observed with Met-R-(o). In Escherichia coli, there are at least six members of Trx in vitro. However, although both hMsrB2 and hMsrB3 could this family that differ in their substrate specificity and location use DTT as the reducing agent, these proteins showed very little within the bacterial cell (2, 3). The protein that was first isolated activity with Trx. Table 1 compares the activity of several and studied in greatest detail was E. coli MsrA (eMsrA) (4–8), recombinant Msr proteins using either DTT or Trx. It can be which specifically reduces Met-S-(o), whether in peptide linkage seen that eMsrA, bovine MsrA (bMsrA), and eMsrB are active or as a free amino acid (8, 9). MsrB proteins specifically reduce with either DTT or Trx as the reducing system. In fact, eMsrB the Met-R-(o) in proteins but have very weak activity with free was much more active with Trx than with DTT. In contrast, Met-R-(o) (10, 11). In addition, in E. coli, there are specific Msr hMsrB2 and hMsrB3 work very poorly with Trx, having Ͻ10% reductases that reduce free Met-S-(o) and Met-R-(o) (2) as well of the activity seen with DTT. One possibility was that the as membrane-associated activities that can reduce both epimers hMsrB proteins specifically required mammalian Trx and not the of Met(o), whether free or peptide-linked (3). bacterial Trx that was used in these experiments. We therefore The situation in animal cells is somewhat different. It appears tested hMsrB3, as well as eMsrA and eMsrB, with mammalian that there is one gene that codes for MsrA (12, 13) but three Trx and mammalian Trx reductase. Reduced mammalian Trx, separate genes that code for MsrB proteins (14–16). MsrA like the bacterial Trx, gave very poor activity with hMsrB3, but localizes primarily in the mitochondria or cytoplasm (12, 13), whereas the three MsrB proteins have different subcellular localizations (16). MsrB1 (originally called Sel-x) is a seleno- Conﬂict of interest statement: No conﬂicts declared. protein that is primarily found in the cytoplasm and nucleus, Abbreviations: Met(o), Met sulfoxide; Msr, Met sulfoxide reductase; Trx, thioredoxin; hMsr, MsrB2 (originally called CBS-1) is present mainly in the mito- human recombinant Msr; eMsr, Escherichia coli Msr; bMsr, bovine Msr; MT, metallothionein; Zn-MT, zinc-containing MT; DABS, 4-N,N-dimethylaminoazobenzene-4-sulfo- chondria, and MsrB3 is localized primarily in the endoplasmic nyl; PAR, 4-(2-pyridylazo)resorcinol; T, thionein; T(o), oxidized T. reticulum and mitochondria. The biological reducing system for †D.S. and D.B. contributed equally to this work. MsrA appears to be reduced thioredoxin (Trx) (4, 17), and Trx ¶To whom correspondence should be addressed at: Center for Molecular Biology and has also been assumed to be the biological reductant for the Biotechnology, Florida Atlantic University, 777 Glades Road, Boca Raton, FL 33431. E-mail: MsrB proteins. However, detailed studies comparing the reduc- [email protected]. ing system requirement for the MsrA and MsrB proteins have © 2006 by The National Academy of Sciences of the USA

8656–8661 ͉ PNAS ͉ June 6, 2006 ͉ vol. 103 ͉ no. 23 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0602826103 Downloaded by guest on September 30, 2021 Table 1. Trx and DTT as reducing agents with various Msr proteins DABS-Met, nmol

Msr proteins Trx DTT Trx͞DTT ratio eMsrA 46.8 46.2 1.0 eMsrB 60.3 11.1 5.4 bMsrA 15.8 30.0 0.53 hMsrB2 0.8 31.3 0.03 hMsrB3 1.7 53.8 0.03

DABS-Met-S-(o) was used as substrate with MsrA proteins, and DABS-Met- R-(o) was used as a substrate with MsrB proteins. The incubation conditions and assay are described in Materials and Methods. The amounts of Msr protein used were as follows: eMsrA, 1.6 ␮g; eMsrB, 2.7 ␮g; bMsrA, 2 ␮g; hMsrB2, 2.7 ␮g; hMsrB3, 2.3 ␮g.

both eMsrA and eMsrB efficiently used Trx from either source (data not shown). It should be noted that we were not able to test mammalian MsrB1, which differs from MsrB2 and MsrB3 in being a selenoprotein, because all attempts to obtain sufficient amounts of this recombinant protein were unsuccessful. The weak activity of hMsrB2 and hMsrB3 with Trx suggested that there may be another reducing system for these proteins in mammalian cells that either functions in place of Trx or is an intermediate hydrogen carrier between Trx and the human MsrB BIOCHEMISTRY proteins.

Zn-MT in the Presence of EDTA Can Serve as a Reducing Agent for Msr. In our attempts to search for a biological factor that was more Fig. 2. Purification and properties of the active factor. (A) Elution profile efficient than Trx in supplying the reducing system for hMsrB2 from a DE-52 column of the factor showing Msr activity and zinc content. Two and hMsrB3, we initially tested an S-100 from bovine liver. Using peaks of reducing activity with hMsrB3 could be separated, and they are hMsrB3, we were able to detect significant reducing activity in labeled MT-1 and MT-2. Activity (F) is expressed as total nanomoles of the liver S-100 fraction, but only in the presence of EDTA. The DABS-Met formed per 1-ml fraction in the Msr reaction using hMsrB3. Zinc concentration (␮M) also is shown (E). (B) Spectra of purified factor at pH 7.4 active material was stable to heating at 80°C for 10 min. Fig. 1 (solid line) and pH 2.0 (dashed line). An extinction coefficient of ␧220 ϭ 48,600 shows the effect of protein concentration of the heated S-100 MϪ1⅐cmϪ1 at pH 2.0 was used to calculate the amount of MT in the fractions. extract and the almost complete dependency on EDTA for hMsrB3 activity. Optimal activity was seen with levels of EDTA Ͼ2.5 mM (data not shown). Routinely, 5 mM EDTA has been EDTA, whereas EGTA (5 or 20 mM), deferoxamine (5 mM), used in the experiments. EDTA by itself had no significant effect and zincon (500 ␮M) were inactive. Thus, EDTA was used on hMsrB3 activity, although hMsrB3 contains zinc. Other throughout the present studies. A series of metal salts could not chelating agents were tested in place of EDTA with Zn-MT. replace EDTA or the heated S-100 in the reaction (data not 1,10-Phenanthroline (5 mM) gave Ϸ40% of the activity of shown). The heat stability and EDTA requirement suggested that the active factor might be a metallothionein (MT), and the heat- stable factor was further purified as described in Materials and Methods. Fig. 2A shows the elution profile from a DE-52 cellulose column, the last step in the purification. Two distinct peaks of reducing activity were observed, and the fractions in both peaks were active in the Msr assay using hMsrB3 only in the presence of EDTA. The purification profile suggested that the two peaks correspond to MT-1 and MT-2, based on their elution from the DE-52 column. Because of the requirement for EDTA for the fraction to be active with hMsrB3, metal analyses were initially performed on purified preparations by using inductively coupled plasma MS. Zinc was found in significant amounts [60,795 parts per billion (ppb)], with trace levels of copper and silver (688 and 739 ppb, respectively). Besides using nanopure water, no special precau- tions were taken to remove trace metals, so the source of these trace metals in the protein sample is unknown. As shown in Fig. 2A, the active, highly purified fractions from the DE-52 column Fig. 1. Effect of heated liver S-100 concentration and EDTA on MsrB3 activity. Activity was measured without EDTA (E) or with 5 mM EDTA (F)inthe contained high levels of zinc that coeluted with the fractions reaction mix. MsrB3 (2 ␮g) was incubated with the indicated amount of heated active in the Msr assay. The amount of MT could be determined Ϫ1 Ϫ1 S-100 in the absence of a reducing system (DTT or Trx) as described in Materials spectrophotometrically (␧220 ϭ 48,600 M ⅐cm at pH 2.0), and and Methods. zinc analyses using the 4-(2-pyridylazo)resorcinol (PAR) reagent

Sagher et al. PNAS ͉ June 6, 2006 ͉ vol. 103 ͉ no. 23 ͉ 8657 Downloaded by guest on September 30, 2021 Table 2. Comparison of the activity of Msr proteins in the presence of Zn-MT or DTT DABS-Met, nmol

Msr proteins Zn-MT DTT eMsrA 33.9 45.7 eMsrB 8.3 27.8 bMsrA 14.1 38.9 hMsrB2 0.9 27.1 hMsrB3 18.0 53.7

Msr proteins were incubated as described in Materials and Methods with either 20 nmol of puriﬁed Zn-MT or 15 mM DTT. Incubations with Zn-MT routinely contained 5 mM EDTA, and no signiﬁcant activity was detected in the absence of EDTA. The amounts of proteins used were as follows: eMsrA, 1.6 ␮g; eMsrB, 5.4 ␮g; bMsrA, 2.0 ␮g; hMsrB2, 2.7 ␮g; hMsrB3, 2.3 ␮g.

(see Materials and Methods) showed that there were close to seven zinc atoms per mole of MT in each fraction. Although zinc appears to be the major metal associated with the active factor, we cannot eliminate the presence of lower levels of other metals Fig. 3. T can supply the reducing system for hMsrB3 activity in the absence in the sample. Fig. 2B shows the UV spectrum of a fraction from of EDTA. The incubations contained 4.5 ␮g of MsrB3, 20 nmol of T, or 20 nmol peak 2 from the DE-52 column (both peaks displayed similar of Zn-MT. The incubations with T did not contain EDTA, but 5 mM EDTA was spectral characteristics). It can be seen that the active factor has added to the incubations with Zn-MT. At 20 min, Zn-MT in the absence of high absorption in the 200- to 250-nm range but essentially no EDTA formed 1.3 nmol, whereas T in the presence of 5 mM EDTA formed 23.5 F E absorbance at 280 nm, indicating the absence of aromatic amino nmol. ,T; , Zn-MT plus EDTA. acids. Upon acidification, the high UV absorption is markedly ͞ decreased. On SDS PAGE, the purified protein, as well as a Zn-MT is to release the zinc from Zn-MT and form T and that commercial rabbit liver MT preparation, migrated as a diffuse T is able to provide the reducing system for the Msr enzymes. band in the 13- to 16-kDa range (data not shown), double the size Ϸ of Zn-MT, which is 6 kDa. This gel migration pattern could be Trx Can Reduce Oxidized T [T(o)]. The reaction mechanism for both due to the unique shape of the protein or the presence of dimers MsrA and MsrB involves the formation of an oxidized enzyme through intermolecular bond formation. The liver MT obtained intermediate that must be reduced for the Msr protein to act from a commercial source also supported hMsrB3 activity in the catalytically (6, 7, 11, 21, 22). If T is capable of reducing oxidized presence of EDTA (data not shown). The presence of zinc as well Msr, the T would become partly or fully oxidized to T(o), and, as the spectral and other characteristics of the active fractions ideally, there should be an enzymatic system that could regen- indicated that the two peaks off the DE-52 column were erate T and permit it to recycle. T(o) was prepared as described Zn-MT-1 and Zn-MT-2. These peak fractions were further in Materials and Methods. This material had generally lost analyzed by electrospray MS, and the molecular weights matched Ϸ50–60% of its free SH groups but still remained mostly soluble those of bovine MT-1 and MT-2 (5,987 and 6,013, respectively). Ͼ Ͻ (see Materials and Methods). Any insoluble material that was EDTA removed 90% of the zinc from Zn-MT in 10 min, as formed was removed by centrifugation. Trx was considered a measured by the appearance of free SH groups (data not shown). possible candidate to reduce T(o), which could be shown directly We concluded that the purified factor is a Zn-MT, which, in the by measuring NADPH oxidation in the presence of the Trx presence of EDTA, is converted to the metal-free reduced T, and reducing system and T(o). As shown in Fig. 4, the oxidation of that T, because of the high content of cysteine residues, is able NADPH depended on Trx, Trx reductase, and T(o). In addition, to supply the reducing system for the Msr reaction. The results shown below used Zn-MT-2, although similar results were obtained with Zn-MT-1. As seen in Table 2, the purified Zn-MT is not a specific reducing agent for hMsrB3 because it also supports eMsrA, eMsrB, and bMsrA, dependent on EDTA. However, to our surprise, the liver factor showed very little activity with hMsrB2 under the conditions used in Table 2.

T Can Function in the Msr System in the Absence of EDTA. Although it appeared likely that the requirement for EDTA was to release zinc from Zn-MT to form T, this system was obviously artificial, and it was important to demonstrate directly that T could serve as the reducing agent for the Msr system. T was prepared as described in Materials and Methods and tested with hMsrB3 as shown in Fig. 3. It can be seen that hMsrB3 activity was supported by both T and Zn-MT, although T was active in the absence of EDTA, whereas Zn-MT required EDTA for activity. Shorter incubations were used for these experiments to minimize Fig. 4. Reduction of T(o) by Trx. The preparation of T(o) and the incubation the oxidation of T that occurred at neutral pH. T also was active conditions are described in Materials and Methods. The oxidation of NADPH with MsrA in the absence of EDTA (data not shown). These was followed at 340 nm. ■, complete system; ᮀ, minus Trx; E, minus Trx results support the view that the requirement for EDTA with reductase; F, minus T(o).

8658 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0602826103 Sagher et al. Downloaded by guest on September 30, 2021 Table 3. Trx stimulates the activity of MsrB3 in the presence T, the apoform of MT, which can function as a reducing agent of T(o) because of its high content of cysteines. In support of this Trx conclusion, it was shown that T, prepared by acid treatment (see No. T(o) Trx NADPH reductase DABS-Met, nmol Materials and Methods), could function as a reducing agent in the Msr system without EDTA. It is known that T is a small protein 1 ϩϩ ϩ ϩ 23.4 having Ϸ60 amino acids and a molecular mass in the range of 6–7 2 Ϫϩ ϩ ϩ 2.4 kDa. Of the total amino acids, approximately one-third are 3 ϩϪ ϩ ϩ 0.5 cysteines, which could make this protein an important cellular 4 ϩϩ Ϫ ϩ 3.4 source of sulfhydryl groups. For many years, it was felt that MT’s 5 ϩϩ ϩ Ϫ 3.9 primary function was to scavenge free radicals and͞or detoxify The incubations contained hMsrB3 (2.3 ␮g) as described in Materials and metals. However, in 1998, Maret and Vallee, in a seminal study Methods. Where indicated by plus sign, 8.3 nmol of T(o), 10 ␮g of Trx, 2.4 ␮g (24), postulated that the zinc-sulfur clusters in MT also acted as of Trx reductase, and 100 nmol of NADPH were added. In this experiment, with a sensor for the redox state of the cell. Oxidation of Zn-MT the amount of hMsrB3 used, 52.7 nmol of DABS-Met was formed in the resulted in release of the zinc so it could be mobilized within the presence of 15 mM DTT. cell, whereas under reducing conditions, T would efficiently bind zinc. Thus, the major role of MT may be to control cellular zinc mobilization as a function of the redox state of the cell. However, as shown in Table 3, T(o) could support hMsrB3 activity in the there does not appear to be much information on other possible presence of the complete Trx reducing system (row 1) but not in functions of T, the reduced apoform of MT, in addition to its the absence of Trx, Trx reductase, or NADPH (rows 3–5). As critical role in binding zinc. In one report similar to our studies, discussed previously, the Trx system alone showed very low it was reported that Zn-MT in the presence of EDTA can activity (row 2). It is also apparent from the results in Table 3 that reactivate the S-sulfonated (inactive) form of ribonuclease (25). the free SH groups remaining in the T(o) cannot support the Msr These researchers also suggested that the thiol groups in T are reaction, indicating that the SH groups in T are not all equivalent part of the pool of cellular thiols that can regulate redox with respect to their ability to function as a reducing agent for reactions in a mechanism that is modulated by zinc chelation. the Msr system. The results in Fig. 4 and Table 3 indicate that Our results on the ability of T to supply the reducing system for disulfide bonds in T(o) can be reduced by the Trx system. Thus, some of the Msr proteins support this conclusion and link the Trx may be one of the cellular agents that can enable T(o) to BIOCHEMISTRY MT proteins to another cellular antioxidant system. recycle and function as a metabolic reducing system. Although the results indicate that T can supply the reducing In contrast to the results with hMsrB3, hMsrB2, which had low activity with either Trx or Zn-MT (see Table 2), was also not system for all of the Msr enzymes tested, with the exception of stimulated when both T(o) and the Trx reducing system were hMsrB2, it is clear that the Trx system is the preferred reducing present (data not shown). system for MsrA and eMsrB. If there is an important reducing role of T, it is with hMsrB3. One of the unexplained findings in Discussion this study was the failure of T (or Zn-MT and EDTA) to Until the present studies, it has been assumed that Trx was the stimulate hMsrB2. As mentioned, all of the other Msr proteins biological reducing system in cells for all of the Msr proteins. that were tested showed significant activity with Zn-MT in the The initial experiments using eMsrA indicated that Trx was the presence of EDTA. Because MsrB2 and MsrB3 are both zinc biological reducing agent (4, 17), in agreement with earlier proteins, they are thought to have similar reaction mechanisms experiments (23). In those experiments, it was shown that Met(o) (16, 22), which makes the lack of activity of T (the active agent) could support growth of a Met-requiring strain but not if the with hMsrB2 puzzling. One possibility is that different sulfhy- organism was also Trx deficient, indicating that Trx is necessary dryls on T react with hMsrB3 and hMsrB2. Thus, the active for the conversion of free Met(o) to Met in E. coli. The present sulfhydryls in T that can interact with hMsrB3 cannot reduce the experiments are in agreement with these earlier results. It oxidized hMsrB2 intermediate. Because, in our hands, hMsrB2 appears that MsrA from both bacterial and mammalian sources shows only weak activity with the Trx system, we are uncertain utilizes Trx very efficiently, as does MsrB from E. coli. However, as to what may be the normal reducing system for this enzyme. the studies reported here show that hMsrB2 and hMsrB3 (and, Because both MT-1 and MT-2 gave similar results in support- presumably, MsrB proteins from other mammalian sources) use ing Msr activity in the presence of EDTA, we have assumed that Trx very poorly. As noted above, hMsrB1 was not tested in the T derived from other MTs, such as MT-3 (found in the brain and present studies because of our inability to overexpress and purify reported to have growth inhibitory activity) and MT-4 (26, 27), the protein from E. coli, but Trx may work well with this protein. would behave in a similar fashion. The electrospray MS analysis Recently, Kim and Gladyshev (22) postulated that, in MsrB1, a did not show the presence of MT-3 in our samples. However, it cysteine was required in addition to selenocysteine for Trx to is possible that slight structural differences in the MTs might be function. In contrast, with MsrB2 and MsrB3, only the active-site important, and it will be necessary to test the individual MT cysteine was required, and Trx was thought to directly reduce the species for their ability to provide a reducing system for the Msr sulfenic acid intermediate on the enzyme (22). It seems clear enzymes. from the low activity using Trx with both MsrB2 and MsrB3 that The in vivo significance of the present results is clearly not this reaction is not efficient, which raises the possibility that Trx known. Certainly, there is no chelating agent, like EDTA, that may not be the direct biological reducing system for MsrB2 and is present in the cell to convert the MT to T. However, Yang et MsrB3. The ability of a heated bovine liver extract to support al. (28) have reported that as much as 50% of the total MT in Msr activity with hMsrB3 in the absence of an exogenous mammalian tissues is present as T. Thus, the high concentration reducing system provides evidence that animal cells contain a of T in tissues is consistent with a possible role of T as a cellular factor that, in the presence of EDTA, can substitute for Trx in reducing agent, especially if there are mechanisms to regenerate this reaction. The identification of Zn-MT as the active factor T from T(o), as shown here with Trx. The heat step should have was based on the heat stability, purification characteristics, destroyed any T in our liver preparations, although, as shown in absorption spectra at neutral and acidic pH values, gel analysis, Fig. 1, there was a slight activity in the heated S-100 in the metal determination, and molecular weight analyses. The role of absence of EDTA that could have been due to T that was not EDTA appears to be to release the zinc from the Zn-MT to form destroyed by the heat step.

Sagher et al. PNAS ͉ June 6, 2006 ͉ vol. 103 ͉ no. 23 ͉ 8659 Downloaded by guest on September 30, 2021 Fig. 5. Postulated role of Trx and MT in supplying the reducing requirement for the Msr enzymes.

Because oxidative stress is believed to release zinc and other bMsrA, eMsrA, eMsrB, and hMsrB2, the latter generously metals from MT, one can postulate a reaction sequence, sum- provided by Todd Lowther (Wake Forest University School of marized in Fig. 5, in which cells, under oxidative stress, mobilize Medicine, Winston–Salem, NC), were overexpressed in E. coli, zinc from Zn-MT for use for the hundreds of zinc-containing and the respective proteins were purified as described in refs. 11, proteins. The loss of the zinc from MT as a result of oxidation 33, and 34. The hMsrB3 cDNA from the human lens was would yield T(o). As postulated in Fig. 5, T(o) can be reduced amplified by PCR, cloned into a pET vector, and overexpressed to T by the Trx system, and evidence for this reaction is shown in BL21 E. coli cells. The harvested cells were suspended in in Table 3 and Fig. 4. T can serve as a cellular reducing agent and 1͞100 volume of original culture by using 50 mM Tris (pH 7.4). reduce the oxidized Msr intermediates, either an enzyme-bound After sonication and centrifugation at 10,000 ϫ g, the superna- disulfide or sulfenic acid (6, 7, 21, 22). At present, we do not tant was fractionated on a Sephadex G-75 column. Active know how many of the cysteines in T can function to reduce the fractions were combined, and protein purity (Ͼ80%) was con- oxidized Msr proteins, but we have evidence that more than one firmed by SDS͞PAGE. cysteine on T is functional. Trx may be only one of the possible cellular reducing systems that can reduce T(o). It is known that Purification of an Active Factor from Bovine Liver. Fresh bovine liver oxidized glutathione can oxidize MT and cause the release of Zn was homogenized in three volumes of 50 mM Tris (pH 7.4) and from Zn-MT and that reduced glutathione can reduce T(o), centrifuged at 10,000 ϫ g for 30 min and then at 100,000 ϫ g for which can bind zinc. Of interest were the findings that certain 16 h (S-100). The S-100 fraction was heated at 80°C for 5 min and selenium compounds, such as selenocystamine, can accelerate centrifuged to remove precipitated proteins (heated S-100). these reactions (29, 30). Preliminary experiments in our system Once the active material was suspected to be a MT, further have indicated that selenocystamine can be reduced to seleno- purification followed an established method for MT (35). Using cysteamine by T and that selenocysteamine can supply the a Bio-Rad DuoFlow HPLC system, the heated S-100 was placed reducing system for both hMsrB2 and hMsrB3. Further studies on a sizing column (Superdex 75 HR 10͞30) followed by DE-52 are required to determine whether the interaction between the anion-exchange chromatography. The fractions were routinely Msr system and T has physiological relevance, especially because monitored at 240 and 280 nm. As reported in ref. 35, two distinct both may play an important role in protecting cells against peaks of activity were eluted from the DE-52 column that oxidative damage. In addition, the possibility should be consid- corresponded to Zn-MT-1 and Zn-MT-2, as described in Results. ered that T may be playing an important role as a cellular reductant for other systems. Preparation of T and T(o) and Assay of T(o) Reduction by Trx. T and T(o) were prepared from Zn-MT by modification of the proce- Materials and Methods dure described in ref. 36. Briefly, purified Zn-MT was dialyzed Met(o), DABS chloride (DABS-Cl), PAR, and other chemicals, against 10 mM HCl (pH 2.0) containing 150 mM NaCl for 12 h including rabbit liver Zn-MT, were purchased from Sigma- at 4°C. The protein after dialysis is reduced, metal-free T and Aldrich, unless specified otherwise. DABS-Met-S-(o) and appears stable when left at pH 2.0. To study the activity of T in DABS-Met-R-(o) were prepared by derivatizing the amino the Msr system, T was neutralized and added to the reaction group of the Met-R-(o) or Met-S-(o) epimers (31) with mixtures immediately before the incubations were initiated. To DABS-Cl (32). Trx and Trx reductase (E. coli) were overex- oxidize T, 0.75 volumes of 50 mM Tris base were added to the pressed and purified from E. coli, and human Trx was purchased T sample to bring the pH to 8.5. Under these conditions, Ϸ50% from Sigma. Rat Trx reductase (TR3) was a generous gift from of the sulfhydryls will become oxidized after4hatroom Vadim Gladyshev (University of Nebraska, Lincoln). Clones for temperature or2hat37°C. With longer incubations, the T(o)

8660 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0602826103 Sagher et al. Downloaded by guest on September 30, 2021 started to precipitate. The assay for free sulfhydryl groups used otherwise. In experiments using T(o), the incubations did not 5,5Ј-dithiobis(2-nitrobenzoic acid) as described in ref. 37. contain EDTA. The quantitation of DABS-Met formed used a To study the reduction of T(o) by Trx, the reaction mixtures slight modification of an extraction procedure previously de- contained (in a total volume of 1 ml) 100 ␮M NADPH, 26 ␮g scribed by Etienne et al. (40) for the reduction of sulindac to of Trx, 6 ␮g of TrxB, and 28 ␮g of partially oxidized T (see sulindac sulfide by MsrA. The reactions were stopped by the above). The oxidation of NADPH was followed at 340 nm at addition of 200 ␮l of 1 M sodium acetate (pH 6.0), followed by room temperature. the addition of 100 ␮l of acetonitrile and 1 ml of benzene. After thorough shaking and centrifugation, the optical density of the Analysis of Zinc Content. Zinc was quantitatively determined in the benzene layer was read at 436 nm. One hundred nanomoles of MT preparations by using the PAR reagent (38, 39). The samples the product, DABS-L-Met, carried through this procedure gave (100 ␮l) were incubated with 10 mM N-ethylmaleimide for 1 h an optical density reading of 1.7, whereas 100 nmol of the at room temperature. PAR (100 nmol) was then added, and the substrate, either the R or S epimer of DABS-Met(o), read Ͻ0.04. sample was diluted with water to 1 ml. The Zn–PAR complex Under these conditions, the reaction was proportional to Msr was measured at 500 nm. Ten nanomoles of zinc gave a reading concentration until Ͼ75% of the substrate was reduced. Unless of 0.720 at 500 nm. A complete metals analysis of the purified indicated otherwise, the results are presented as nmol of DABS- MT preparation was performed by Joseph Caruso (University of Met formed in 60 min. As little as 2 nmol of product could be Cincinnati, Cincinnati) by using an Agilent 7500 inductively measured. This assay could be used with purified preparations of coupled plasma mass spectrometer (Agilent Technologies, Palo MsrA and MsrB as well as crude extracts of mammalian tissues. Alto, CA). Molecular weight determinations of the purified Bacterial extracts, in the presence of a reducing system, de- protein were performed by Peter Yau (University of Illinois, stroyed the substrate, and further studies are needed to adapt the Urbana–Champaign) by using electrospray MS. assay to bacterial extracts.

Colorimetric Assay for Msr Activity Based on the Reduction of DABS- We thank Dr. Vadim Gladyshev for advice and the gift of the mammalian Met(o). The reaction mixture (200 ␮l) to measure Msr activity Trx reductase and MsrB1 clones, Dr. Todd Lowther for the hMsrB2 contained 100 mM Tris⅐Cl (pH 7.4), 100 nmol of the indicated clone and helpful discussions, Dr. Joseph Caruso for the metal analyses, DABS-Met(o) epimer, either 15 mM DTT or the Trx regener- Dr. Peter Yau for helpful discussions and MS analysis, Dr. Bert Vallee ␮ ͞ ␮ ͞ ␮ for helpful advice and guidance, and Frantz Etienne, who was instru- ating system (10 gofTrx2.4 g of Trx reductase 500 M mental in the development of the colorimetric assay. This article is NADPH), and Msr enzyme as indicated. When the liver frac- contribution no. P200524 from the Center of Excellence in Biomedical BIOCHEMISTRY tions [Zn-MT, T, or T(o)] were tested, DTT was omitted, and the and Marine Biotechnology of Florida Atlantic University, which partially Trx reducing system and 5 mM EDTA were added where funded this work. This work was also supported by National Institutes of indicated. Incubations were for 60 min at 37°C unless noted Health Grant EY13022MK (to M.K.).

1. Weissbach, H., Resnick, L. & Brot, N. (2005) Biochim. Biophys. Acta 1703, 20. Lee, J. W., Gordiyenko, N. V., Marchetti, M., Tserentsoodol, N., Sagher, D., 203–212. Alam, S., Weissbach, H., Kantorow, M. & Rodriguez, I. R. (2006) Exp. Eye Res. 2. Etienne, F., Spector, D., Brot, N. & Weissbach, H. (2003) Biochem. Biophys. 82, 816–827. Res. Commun. 300, 378–382. 21. Lowther, W. T., Brot, N., Weissbach, H., Honek, J. F. & Matthews, B. W. 3. Spector, D., Etienne, F., Brot, N. & Weissbach, H. (2003) Biochem. Biophys. (2000) Proc. Natl. Acad. Sci. USA 97, 6463–6468. Res. Commun. 302, 284–289. 22. Kim, H. Y. & Gladyshev, V. N. (2005) PLoS Biol. 3, e375. 4. Brot, N., Weissbach, L., Werth, J. & Weissbach, H. (1981) Proc. Natl. Acad. Sci. 23. Russel, M. & Model, P. (1986) J. Biol. Chem. 261, 14997–15005. 78, USA 2155–2158. 24. Maret, W. & Vallee, B. L. (1998) Proc. Natl. Acad. Sci. USA 95, 3478– 5. Weissbach, H., Etienne, F., Hoshi, T., Heinemann, S. H., Lowther, W. T., 3482. Matthews, B., St. John, G., Nathan, C. & Brot, N. (2002) Arch. Biochem. 25. Zhang, S., Li, J., Wang, C. C. & Tsou, C. L. (1999) FEBS Lett. 462, 383–386. Biophys. 397, 172–178. 26. Uchida, Y., Takio, K., Titani, K., Ihara, Y. & Tomonaga, M. (1991) Neuron 7, 6. Lowther, W. T., Brot, N., Weissbach, H. & Matthews, B. W. (2000) Biochem- istry 39, 13307–13312. 337–347. 7. Boschi-Muller, S., Azza, S., Sanglier-Cianferani, S., Talfournier, F., Van 27. Hidalgo, J., Aschner, M., Zatta, P. & Vasak, M. (2001) Brain Res. Bull. 55, Dorsselear, A. & Branlant, G. (2000) J. Biol. Chem. 275, 35908–35913. 133–145. 8. Moskovitz, J., Poston, J. M., Berlett, B. S., Nosworthy, N. J., Szczepanowski, 28. Yang, Y., Maret, W. & Vallee, B. L. (2001) Proc. Natl. Acad. Sci. USA 98, R. & Stadtman, E. R. (2000) J. Biol. Chem. 275, 14167–14172. 5556–5559. 9. Sharov, V. S., Ferrington, D. A., Squier, T. C. & Schoneich, C. (1999) FEBS 29. Jacob, C., Maret, W. & Vallee, B. L. (1999) Proc. Natl. Acad. Sci. USA 96, Lett. 455, 247–250. 1910–1914. 10. Grimaud, R., Ezraty, B., Mitchell, J. K., Lafitte, D., Briand, C., Derrick, P. J. 30. Chen, Y. & Maret, W. (2001) Eur. J. Biochem. 268, 3346–3353. & Barras, F. (2001) J. Biol. Chem. 276, 48915–48920. 31. Lavine, T. (1947) J. Biol. Chem. 169, 477–491. 11. Lowther, W. T., Weissbach, H., Etienne, F., Brot, N. & Matthews, B. W. (2002) 32. Minetti, G., Balduini, C. & Brovelli, A. (1994) Ital. J. Biochem. 43, Nat. Struct. Biol. 9, 348–352. 273–283. 12. Hansel, A., Kuschel, L., Hehl, S., Lemke, C., Agricola, H. J., Hoshi, T. & 33. Rahman, M. A., Nelson, H., Weissbach, H. & Brot, N. (1992) J. Biol. Chem. 267, Heinemann, S. H. (2002) FASEB J. 16, 911–913. 15549–15551. 13. Vougier, S., Mary, J. & Friguet, B. (2003) Biochem. J. 373, 531–537. 34. Moskovitz, J., Weissbach, H. & Brot, N. (1996) Proc. Natl. Acad. Sci. USA 93, 14. Kryukov, G. V., Kumar, R. A., Koc, A., Sun, Z. & Gladyshev, V. N. (2002) Proc. 2095–2099. Natl. Acad. Sci. USA 99, 4245–4250. 35. Vasak, M. (1991) Methods Enzymol. 205, 41–44. 15. Huang, W., Escribano, J., Sarfarazi, M. & Coca-Prados, M. (1999) Gene 233, 36. Klein, D., Sato, S. & Summer, K. H. (1994) Anal. Biochem. 221, 405–409. 233–240. 16. Kim, H. Y. & Gladyshev, V. N. (2004) Mol. Biol. Cell 15, 1055–1064. 37. Li, T. Y., Minkel, D. T., Shaw, C. F., III, & Petering, D. H. (1981) Biochem. 17. Brot, N. & Weissbach, H. (1983) Arch. Biochem. Biophys. 223, 271–281. J. 193, 441–446. 18. Kantorow, M., Hawse, J. R., Cowell, T. L., Benhamed, S., Pizarro, G. O., Reddy, 38. Hunt, J. B., Neece, S. H. & Ginsburg, A. (1985) Anal. Biochem. 146, 150–157. V. N. & Hejtmancik, J. F. (2004) Proc. Natl. Acad. Sci. USA 101, 9654–9659. 39. Shaw, C. F., III, Savas, M. M. & Petering, D. H. (1991) Methods Enzymol. 205, 19. Marchetti, M. A., Pizarro, G. O., Sagher, D., Deamicis, C., Brot, N., Hejtman- 401–414. cik, J. F., Weissbach, H. & Kantorow, M. (2005) Invest. Ophthalmol. Vis. Sci. 40. Etienne, F., Resnick, L., Sagher, D., Brot, N. & Weissbach, H. (2003) Biochem. 46, 2107–2112. Biophys. Res. Commun. 312, 1005–1010.

Sagher et al. PNAS ͉ June 6, 2006 ͉ vol. 103 ͉ no. 23 ͉ 8661 Downloaded by guest on September 30, 2021