- Email: [email protected]
Human IgG is substrate for the thioredoxin system: Differential cleavage pattern of interchain disulfide bridges in IgG subclasses
Mo/ccw/ur Inwru~/o,q~. Vol. 34. No. IO. ~1’. 70% 717. 1997 , 1997 Elcevier Science Ltd. All nghts reserved
Pergamon PII: SOl61-5890(97)00092-8
Printed 111Great Britain 01 (,I -5X90 97 $17.00 c 0.00
HUMAN IgG IS SUBSTRATE FOR THE THIOREDOXIN SYSTEM: DIFFERENTIAL CLEAVAGE PATTERN OF INTERCHAIN DISULFIDE BRIDGES IN IgG SUBCLASSES CARL
*++ MIKAEL G. M. MAGNUSSON, ARNE H0LMGREN.r
BJdRNSTEDTJ-
and
*Department of Laboratory Medicine, Division of Clinical Immunology, Karolinska Institute. Karolinska Hospital, S-171 76 Stockholm. Sweden: and “rThe Medical Nobel Institute for Biochemistry. Department of Medical Biochemistry and Biophysics. Karolinska Institute. Stockholm. Sweden
Abstract-Thioredoxin. a 12,000 mol. wt protein with two redox-active cysteine residues. togethet with thioredoxin reductase and NADPH. may reduce protein disulfides and thereby act as a molecular probe of their structure and reactivity. Interchain and intrachain disulfides are structural elements in and therefore potential substrates for the reduced thioredoxin, Trx(SH&. It LV;IS all immunoglobulins investigated whether such disulfides are cleaved in human polyclonal IgG and IgG subclass myeloma proteins by both the human and the Eschrri.shitr w/i thioredoxin systems. The reactions wrre montored spectrophotometrically as oxidation of NADPH at 340 nm, and by following the kinetics ol the cleavage patterns with sodium dodecyl sulfate--polyacrylamide gel electrophoresis (SDS-PAGEI). Human IgG was a substrate for both prokaryotic and eukaryotic Trx(SH)?, which directly reduced IgG disulfides in a time and dose-dependent manner. Stoichiometric analyses indicated near-complete reduction of mainly inter-heavy-light chain and inter-heavy chain disulfides. and SDS--PAGE corroborated that the buried intrachain disulfides were left intact. The kinetic studies showed that IgG I. IgG3 and lgG4 were readily reduced into heavy and light chains via the formation of half-molecules with slightly slower kinetics for lgG4. In sharp contrast. IgG2 was not cleaved at all. even with increased thioredoxin concentrations or reduction times. A binall but significant NADPH consumption by IgG2 myeloma proteins suggested reduction of II labile interchain or surface-exposed mixed disulfide. Consistent results were obtained for seceral IgG myeloma proteins within each subclass. The structural and functional importance of interchain disulfides in immunoplobulins suggests physiological implications of the thioredoxin system. t’ 1997 Elsevier Science Ltd. All rights reserved
&I’ ~ort/s: antibodies.
enzyme, human,
IgG subclasses.
INTRODUCTION
+ NADPH
thioredoxin
Trx - (SH)? + Protein - S2 + Trx - S2
Thioredoxin (Trx), a 12.000 mol. wt protein with two redox-active cysteine residues in the conserved active site (-Cys-Gly-Pro-Cys-). operates together with thioredoxin reductase (TR) as an NADPH-dependent protein disulfide reductase system (Holmgren, 1985. 1989) by reactions I and II: Trx-S,
reduction.
+ H’ ZTrx-(SH), + NADP+
(I)
IAuthor to whom correspondence should be addressed. Tel: 46-8-51775938; Fax: 46-8-969475. Ahhrwiatims; Ig. immunoglobulin, 2-ME, 2-mercaptoethanol. Trx. thioredoxin. TR, thioredoxin reductase. E-Trx. E. coli thioredoxin. H-Trx. human thioredoxin, E-TR, E. coli thioredoxin reductase, H-TR, human thioredoxin reductase. TE-buffer. 50 mM Tris+HCI. I mM EDTA, pH 7.5.
+Protein-(SH),.
(II)
Trx has a large and growing number of functions as. for example, hydrogen donor to ribonucleotide reductase at DNA replication (Thelander and Reichard. 1979), regulator of photosynthetic enzymes (Cskke and Buchanan, 1986) and modulator of immunological events (Rosen et al., 1995; Tagaya ct al., 1989: Wakazugi ct ~1.. 1990; Yodoi and Tursz. 1991). Human Trx is secreted by activated human lymphocytes (Ericson et ul., 1992). upregulates IL2 receptors (Tagaya et al.. l989), and promotes growth and proliferation of B-cells (Biguet CI (I/., 1994: RosCn ef ul., 1995: Yodoi and Tursz. 1991). Moreover, Trx is present in human plasma and elevated levels are found in HIV-infected patients (Nakamura rt ~1.. 1996). Interchain
structural 709
and intrachain disultide bonds are common elements in all immunoglobulins and are there-
710
C. G. M. MAGNUSSON et al.
fore potential substrates for the Trx system. IgG is the most common lg in serum and four subclasses exist, i.e. IgGI, IgG2, lgG3 and lgG4, which differ in effector functions such as complement activation and Fey-receptor binding (Burton, 1990; Jefferis, 1990). lgG1 (65%) is the dominating subclass in serum, followed by lgG2 (25%), lgG3 (7%) and lgG4 (3%) (Morel1 et al., 1976). Inter-heavy chain disulfide bridges in the hinge region vary in numbers among the four IgG subclasses (Burton, 1990) and are considered to be involved in both the structural properties and physiological functions of IgG subclasses (Brekke et al., 1993, 1994; Burton, 1990; Dorai et al., 1992; lsenman et al., 1975; Jefferis, 1990; Michaelsen, 1988; Snow and Amzel, 1988). The aim of the present investigation was to study the reactivity of IgG with the Trx system and elucidate whether any differences exist in the cleavage patterns of intrachain and interchain disulfides among the four human IgG subclasses.
Spectrophotometric
measurements
The reducing capacity of the Trx systems was followed in an automatic Zeiss PMQIII spectrophotometer by monitoring the oxidation of NADPH at 340 nm using 1 cm semi-microcuvettes. A change in AXbonrn of 0.0062 between the reaction and reference cuvettes represents the reduction of 1 ,nM disulfide. Following preliminary experiments, the final concentrations of the components were 0.3- 0.6 mg/ml (2-4pM) for the Ig, 3 PM for Trx. 0.1 PM for TR and 300-600 PM for NADPH if not otherwise stated. The reactions were performed in TE-buffer at room temperature for up to 60min in a final volume of 0.5 ml and A340nmwas recorded every 30sec. Equal volumes in the reaction and reference cuvettes were always maintained by the addition of TE-buffer. The NADPH consumption caused by the reduction of Trx itself, initially in mainly the oxidized form, was determined in the absence of lg and subtracted before any calculations. Kinetics and gel electrophoresis
MATERIALS
AND METHODS
Enzyme preparations E. coli thioredoxin reductase (E-TR), E. coli thioredoxin (E-Trx), human recombinant thioredoxin (H-Trx) and human placental thioredoxin reductase (H-TR) were purified preparations (Holmgren and Bjornstedt, 1995). Reduced E-Trx was obtained by treatment with a 20fold molar excess of dithiothreitol (Sigma Chemical Co., U.S.A.) at 37°C for 15 min (Holmgren and Bjbrnstedt, 1995). Excess dithiothreitol was removed by gelchromatography on a NAP-5 column (Pharmacia AB, Uppsala, Sweden) with ice-cold nitrogen-equilibrated 50 mM Tris-HCI, 1 mM EDTA, pH 7.5 (TE-buffer). The reduced product was kept in aliquots at -70°C until used.
Immunoglobulins
IgG myeloma proteins [two lgGl(h-), one lgGl(l), two lgG2(rc), one lgG2(1), four lgG3(k-), one lgG3(1I), three IgG4(rc) and one lgG4(;1)] were purified with standard chromatographic procedures, or have been described previously (Jefferis et al., 1985). Protein concentration was determined by absorbance at 280 nm (EZgOnm,,% = 14.0). Most IgG subclass proteins had a purity exceeding 98% (w/w), calculated on the sum of the other subclasses, respectively, except for two lgG2s and one lgG3 which contained 3-5% IgGl . The degree of contamination of a given IgG subclass protein by the others was determined by isotype-specific competitive enzyme immunoassay based on monoclonal antibodies (mAbs) and WHO-calibrated (67/97) standard (Lilja et al., 1990). Polyclonal IgG for intravenous immunotherapy (165 mg/ml; batch no. 36909-51, Kabi-Pharmacia, Stockholm, Sweden) had the following IgG subclass composition as determined by competitive enzyme immunoassay: IgGl (61%), lgG2 (35%), lgG3 (3.1%) and lgG4 (0.61%).
The reduction of lg by the human or E. coli Trx systems was performed at room temperature in TE-buffer for different times (0, 15 set, 1 min, 4 min, 16 min, 1 hr and 4 hr). Reaction aliquots were treated with one volume of 10.8 mM iodoacetamide (Sigma) dissolved in nitrogensaturated 0.2 M Tris-HCI, pH 8.0, per three volumes of reaction mixture. The final concentrations of lg and all components of the Trx systems were similar to those described for the spectrophotometric experiments above. Control samples consisted of a mixture of all reagents except Trx. Gel electrophoresis under non-reducing and reducing conditions was performed on a PHAST@ system using polyacrylamide gradient gels (4-l 5%; 8-25%) sodium dodecyl sulfate (SDS) buffer strips, sample applicator S/l, and high or low mol. wt standards as outlined by the manufacturer (Pharmacia AB). Samples were prepared by mixing one volume of SDS buffer [lo mM Tris-HCI, 1 mM EDTA, pH 8.0, containing 5% (w/v) SDS and 0.02% (w/v) bromophenol blue] with one volume of test sample and heating the mixture for 5 min at 1OO’C. Reducing sample conditions were obtained after a similar heat-treatment in SDS buffer containing 10% (v/v) 2ME. The gels were stained with Coomassie brilliant blue R250 in the PHAST-system and the mol. wts of the proteins were estimated from a plot of relative migration versus the logarithm of mol. wts of standard proteins. RESULTS Spectrophotometric
analysis
Following preliminary experiments with lgG3, the IgG subclass with most disulfides (Table I), we found it absolutely necessary to add lg, TR and NADPH to both cuvettes and start the reaction by Trx, as the lg concentration interferred at AX40nmand TR itself caused NADPH consumption (0.27 pM/min). To illustrate the
Human Table
IgG is substrate
I. Characteristics
kG
of human
IgGl IgG2 IgG3 IgG4
IgG subclasses study
711
system
used in the calculations
of this
Number of S-S-bridges/lgG molecule ___-__ Inter-H-L chain Total Inter-H-H chain _~~__~_____ ~. _.
Molecular weight
subclass
for the thioredoxin
150,000 150,000 160,000 150,000
2 4 11 2
2 2 2 2
4 6 13 4
Data from Jefferis (1990)
and role of the different components of the Trx system three experiments with IgG3 are shown (Fig. 1). Firstly, addition of E-TR (0.1 PM) gave a minor linear change in the consumption of NADPH (approx. 0.3pM:min) due to oxidation by E-TR as alluded to above. Secondly, addition of E-Trx (3 FM), in the absence of TR, caused no change at all in AX4”“,,,. Thirdly, addition of both E-TR (0.1 PM) and E-Trx (3 PM) to the reaction cuvette gave a rapid and substantial consumption of NADPH that plateaued at 25pM after 5min. As expected. addition of E-Trx (3 PM) to one of two cuvettes containing just 0.1 LIM E-TR and 300 PM NADPH gave a consumption rapidly plateauing at 3pM (data not shown). The NADPH consumption obtained for IgG3 (Fig. 1) showed that about 12 disulfide bridges were reduced, corresponding well with the total number of interchain bridges i.e. 11 inter-heavy chain and two interheavy-light chain disulfides (Table 1). Altogether, these experiments demonstrate the specificity of the detection specificity
system and that Trx is an intermediate electron carrier in the reduction process of IgG3. The reduction of IgG3 by six different concentrations of E-Trx (0.03311 I’M) showed that for concentrations consumption plateaued after 52 1 PM the NADPH 1Omin (Fig. 2) corresponding to the cleavage of about 13 disulfide bridges. It is interesting from a physiological point of view that even 0.03 PM Trx (360 ng/ml) managed to reduce about three disulfide bridges in IgG3 after 1Omin. Reduction of five 2-fold IgG3 concentrations (0.6-9.6pM) with the E-Trx (0.3 FM) system demonstrated that 2 11 disulfides were cleaved after 10 min for all IgG3 concentrations except for 9.6pM. In the latter case only seven disulfides were reduced (data not shown), probably due to NADPH depletion. The reduction of polyclonal IgG and several purified IgG subclass proteins were studied with the H-Trx system and representative reduction patterns are illustrated in Fig. 3. The consumption of NADPH, transformed into
30
-
0
1
2
3
4
5
6
TR O.lpM
7
8
9
10
Time (mid Fig. 1. Disulfide reduction followed by the spectrophotometric procedure using human IgG3 and the E. coli Trx system. The reactions were performed in TE-buffer containing 300lM NADPH. IgG3 was added to both reaction and reference cuvettes, to a final concentration of 2.1 PM. TR ( ?? ), Trx (A) or TR/Trx (0) was added to the reaction cuvettes, to final concentrations of 0.1 PM. 3 ILM or 0.1 PM/~ PM. respectively. Equal volumes of buffer were added to the reference cuvettes and the initial reduction of Trx-S, has been subtracted.
712
C. G. M. MAGNUSSON
c’t N/.
-*
0.03 &I
Trx
-*
0.1 HIM Trx
-
1 jhl Trx
-
3 /A4 Trx
-
11 pM Trx
,#I
7-I
0~:::::::::::::::::;;: 0
1
2
3
4
5
6
7
8
9
10
Time (min) Fig. 2. Reduction of lgG3 by six cuvettes containing IgG3 (2.4pM), final concentrations of 0.03 ,L~M(A), Equal volumes of buffer were added
-2
different concentrations of E. co/i Trx. E-Trx was added to NADPH (300/~M) and E-TR (0.1 /tM) in TE-buffer, to reach 0.1 ,uM (0) 0.3 ,uM (4), I /tM (I). 3 ,LIM (0) and 1I LLM (A). to the reference cuvettes and the reduction of Trx-S, has been subtracted.
I
1
0
5
10
15
20
25
30
35
40
45
Time (min) Fig. 3. Reduction of polyclonal IgG Polyclonal IgG (A). IgGl (A), lgG2 0.1 PM H-TR and 600itM NADPH reactions were started by addition of Equal volumes of buffer were added
and IgG subclass myeloma proteins by the human Trx system. (+). IgG3 (0) and IgG4 (M) were added to cuvettes containing in TE-buffer. to reach final concentrations of 2-3pM. The H-Trx to a final concentration of 3 ;tM to the reaction cuvettes. to the reference cuvettes and the initial reduction of H-Trx has been corrected for.
the number of disulfide bridges reduced per IgG molecule, showed that after 45min about 13 disulfide bridges were reduced in IgG3(x), four in IgGl (x), one in IgG2(k-), three in IgG4(rc) and two or three in polyclonal IgG. An additional set of two IgGl, two IgG2, four IgG3 and three IgG4 subclass proteins was also tested with the E-Trx system for different periods of time. Concordant results with the human system, and no influence from
light chain or G2m23 shown).
Gel electrophoretic
allotype,
were obtained
(data not
anul~wis
To investigate further the kinetics and cleavage pattern of the reduction process, polyclonal IgG and IgG subclass proteins were incubated with the E. coli or human Trx
Human
system for different
periods
(O-l hr). The
1gG is substrate
for the thioredoxin
reactions were
analysed by SDS-poly(PAGE) under non-reducing conditions. The two Trx systems gave near identical results and showed that the cleavage patterns for IgGl, lgG3 and IgG4 were similar but had slightly different kinetics whereas, in sharp contrast, IgG2 was not cleaved at all (Fig. 4). Polyclonal IgG gave a pattern suggesting partial cleavage, as intact Ig persisted after 2 hr of reduction (data not shown). For all IgG subclasses there is mainly one band with a mol. wt of 150,000-162,000, visible in all gels at 0 time (control), representing intact lg. Minor additional bands probably represent disulfide exchange products that may form during heat-treatment of the sample in SDS buffer (Ikeyama, 1987) as all proteins were highly pure and gave two bands after complete reduction with 2-ME (Fig. 4). Trx(SH), reduced IgG3 most efficiently, as intact Ig disappeared after 4min, whereas intact IgGl and IgG4 were still present. Concomitantly. bands with mol. wts of 76,000-83,000 appeared. probably representing half-molecules (HL). Reduction of IgG3 generated many bands smaller than the heavy chain, presumably representing products differently reoxidized in the hinge region. The presence of HL of lgG4 at time 0 (control) is an expected heterogeneity in IgG4 (King et al., 1992). Further Trxmediated reduction of IgGl and IgG3 leads to complete cleavage of HL into 47,000&56,000 heavy (H) chains and 24,000-22,000 light (L) chains, respectively. after 1 hr. IgG4 was also split into heavy and light chains, though not completely. after 1 hr of reduction but well after 2 hi (data not shown). In sharp contrast, IgG2 was cleaved neither after prolonged incubation (4 hr) nor after terminated
acrylamide
by alkylation
and
gel electrophoresis
12
3
4
5
6
7
system
713
increasing the Trx concentration 5-fold (I 5 AIM). despite the latter conditions giving faster kinetics for IgGl and IgG4 (data not shown). Further SDS -PAGE analyses on one IgGl, two IgG2, two IgG3 and two IgG4 of different light chains confirmed the cleavage differences between subclasses and also showed, as expected, that neither the light chain nor the G2m23 allotype had any effect on the cleavage pattern. In control experiments Ig were exposed to all reagents except for Trx and no cleavages uerc observed even after 2 hr (data not shown). confirming the spectrophotometric data. In order to study the reduction of disultides directly (reaction II). E-Trx(SH)? only was added to IgGl. lgG2 and IgG3. and samples were alkylated after different times. Three concentrations of E-Trx(SH): were tested representing molar ratios to the total concentration of interchain disulfide bridges in the Ig of I : IO. I : I and IO: I (Table 1). The results showed that E-Trx(SH), in a IOfold molar excess and at equimolar concentration reduced both IgGl (Fig. 5) and IgG3 (data not shown) but not IgG2 (data not shown). Half-molecules (HL) appeared after 15 set as well as H and L chains at the highest Trx concentration. Although the reduction rate was faster with the highest concentration of Trx the reductions were not complete. as half-molecules (HL) persisted even after 1 hr of reduction. The incomplete reduction may be related to the consumption of limited amounts of reduced E-Trx because of self-oxidation. All the data strongly suggested that mainly interchain disulfides. and no intrachain disulfides. were reduced by the Trx systems. To evaluate this further. IgGl and IgG4 were reduced by the E-Trx system for I hr. alkylated and analysed by SDS-PAGE under both reducing and non-
a
Fig. 4. SDS-PAGE of human monoclonal IgG subclass proteins following incubation for different times with the human Trx system and subsequent alkylation. Lane I, Fully 2-ME-reduced Ig: lane 2. high mol. wt standards; lane 3, control sample with all reagents (lg. H-TR, NADPH and TE-buffer) except for H-Trx; lanes 4-8, Ig treated with the Trx system for I5 set (lane 4) I min (lane 5). 4min (lane 6). 16 min (lane 7) and I hr (lane 8) followed by alkylation with iodoacetamide. (A) IgGl: (B) lgG2: (C) lgG3; (D) IgG4. Samples in lanes 228 were treated with SDS buffer without 2-ME prior to SDS-PAGE on gradient gels (4-15%). Note the differences in size between heavy chains (H) and light chains (L) generated by 2-ME and the Trx system. respectively.
714
C. G. M. MAGNUSSON FlgGt
al
2
3
4
5
6
7
8@3
4
5
et (II. 6
7
8lc.l
4
5
6
7
8
Fig. 5. SDS-PAGE of IgGl (6.7pM) incubated for different times with three different concentrations of E. coli Trx-(SH),, i.e. 270, 27 and 2.7/*M, respectively. The molar ratios of reduced E-Trx(SH)2:IgGl-SS were 1:lO (A), 1:l (B) and 1O:l (C). Lane I, Fully reduced Ig with SDS buffer containing 2-ME; lane 2, low mol. wt standards; lane 3, control sample with all reagents (Ig, NADPH and TE-buffer) except for E-Trx(SH)*; lanes 4-8, Ig alkylated with iodoacetamide following treatment with the E-Trx(SH)2 for 15 set (lane 4), 1min (lane 5) 4 min (lane 6) I6 min (lane 7) and 1hr (lane 8). Samples in lanes 2-8 were treated with SDS buffer without 2-ME prior to SDS-PAGE on gradient gels (4-l 5%). Note the appearance of the E-Trx band (mol. wt 12,000) in panels B and C.
reducing conditions (Fig. 6). The E-Trx-treated alkylated samples of IgGl and IgG4 had, under non-reducing conditions, heavy and light chains with mol. wts of 54,000 and 25,000, respectively. These mol. wts were apparently smaller than those obtained under reducing conditions, i.e. 59,000-62,000 and 28,000, respectively. These latter mol. wts are the same as those observed for controls following complete 2-ME reduction (Fig. 6). The lower mol. wts of lg chains following Trx reduction are interpreted as the result of intact intrachain disulfides giving them a higher electrophoretic mobility that will decrease upon full 2-ME reduction due to linearization of the chains.
1
2
3
4
5
6
7
8
94 kD + 67 kD + 43 kD +
H
30 kD +
L
20 kD + 14kD
j
Fig. 6. SDS-PAGE on gradient gels (S-25%) demonstrating interchain disulfide cleavage of IgGl and IgG4 following reduction for 1 hr by the E. coli Trx system and subsequent alkylation with iodoacetamide. Aliquots of alkylated IgGl and IgG4 were treated (15 min, 1OOC) with SDS buffer containing 2-ME (lanes 2 and 6, respectively) or with SDS buffer without 2-ME (lanes 3 and 7, respectively). Comparison is made to fully 2-ME-reduced IgGl and IgG4 control samples not exposed to the Trx system (lanes 4 and 8, respectively). Lanes 1 and 5, Low mol. wt standards.
DISCUSSION This study demonstrates that human IgG is an efficient substrate for the Trx system with an apparent reaction rate of the same order as for insulin, i.e. 2 x lO”M-‘set-’ (Holmgren, 1985; Holmgren and Bjornstedt, 1995). Our spectrophotometric and gel electrophoretic data show that both the E. coli and the human Trx system, in a doseand time-dependent manner, reduce interchain disulfide bridges in purified human monoclonal IgG proteins of different subclasses and in polyclonal IgG. The reduction was dependent on Trx(SH)z because in control experiments neither E. coli nor mammalian TR alone cleaved any disulfides. This is consistent with the inability of TR to reduce protein disulfides (Holmgren, 1984. 1985), although mammalian TR reduces a wide range of low mol. wt substrates (Holmgren and Bjornstedt, 1995) including the T-cell effector polypeptide NK-lysin (Andersson et al., 1996). Interestingly, the IgG subclasses showed differences in their susceptibility to Trx-mediated reduction, as verified for both the human and the E. coli system. Thus. while IgGl, IgG3 and IgG4 showed similar but kinetically slightly different cleavage patterns with a ranking order of susceptibility IgG3 > IgGl > IgG4, in contrast IgG2 was not cleaved at all. These findings were consistent using several myeloma proteins of each subclass and independent of the Trx system used. The subclass differences could be explained neither by the type of light chain, by the G2m23 allotype, nor as a dose-response effect, as IgG2 was not cleaved, even at a high Trx concentration (15 PM) or after prolonged incubation (4 hr). Thus, the incomplete reduction of polyclonal IgG observed by SDS-PAGE could be explained by its 35% content of IgG2. The cleavage pattern for IgGl, IgG3 and IgG4 observed by SDS-PAGE showed that the inter-heavy chain disulfides first readily reacted, generating half-molecules comprised of one heavy and one light chain. This
Human IgG is substrate for the thioredoxin system was followed by a slower cleavage of the inter-heavylight chain disulfides. This interpretation is supported by the kinetic data from the spectrophotometric experiments indicating the reduction of four bridges in IgG1, 13 in IgG3 and almost four in IgG4 following 45min of reduction. Together, these results indicate not only that Trx reduces the inter-heavy chain disulfides more efficiently than the inter heavy-light chain disulfides but also that the intrachain disulfides are not affected. The SDS-PAGE results showed that further reduction of alkylated Trx-reduced chains with 2-ME gives products of greater size which were interpreted as intrachain disulfide reduction since the compactness of the domains is lost due to linearization (Jack et al., 1993). Trx has previously been used to produce half-molecules of c(?macroglobulin (Larsson et al., 1988) and may be regarded as a sensitive probe for protein structure (Holmgren, 1984). The differences observed among the IgG subclasses to reduction by Trx might be explained from a structural point of view. As IgG2 is a more compact molecule than IgGl and IgG3, sterical constraints may impede access of Trx to the disulfides of this subclass (Burton, 1990; Jefferis, 1990; Snow and Amzel, 1988). The slower reduction rate of IgG4 is in line with this because IgG4 has intermediate structural characteristics. Accessibility problems may also explain the inability of Trx-mediated reduction of buried intrachain disulfides. It is interesting to note that the sensitivity of IgG subclasses to enzymatic cleavage with papain into Fc and Fab fragments follows the same pattern. IgGl and IgG3 are more readily digested by papain than IgG4, whereas IgG2 is quite resistant (Jefferis, 1990). Although we were unable to detect any Trx-mediated reduction in the three IgG2 proteins tested by SDS-PAGE we observed a small but consistent consumption of NADPH in the spectrophotometric experiments corresponding to the reduction of about one disulfide bridge. This cannot be attributed to contamination from other IgG subclasses but rather we consider that one of the four inter-heavy chain disulfides is more labile, or that a surface-exposed thiol is engaged in a mixed disulfide. IgG2 has been proposed to contain such sulfhydryl groups (Schauenstein et al., 1986) as have light chains (Buchwald, 1971). The functional and physiological implications of our findings have yet to be elucidated but its worth mentioning that the concentrations of Trx used in this study are within the micromolar and nanomolar physiological ranges encountered in mammalian tissues (Grankvist rt al., 1982; Holmgren and Luthman, 1978; Padilla et al., 1995; Rozell et al., 1985) and human plasma (Nakamura et al., 1996), respectively. Moreover, activated lymphocytes secrete Trx (Ericson et ul., 1992) potentially giving higher local concentrations. Higher rates of reduction are also to be expected under less aerobic conditions and at 37% Thus, one may envision that Trx exerts stimulatory or down-regulating physiological effects on circulating Abs and their membrane counterparts. Effector functions, which differ among IgG subclasses (Brtiggemann et al.. 1987; Burton, 1990; Jefferis,
715
1990) may be modulated since the hinge region, where the disulfides are, contributes to structural properties (Brekke et al., 1994; Michaelsen, 1988) which are of importance for segmental flexibility (Oi et al., 1984), different effector functions (Aase et al., 1993; Brekke ct al., 1994; Isenman et al., 1975; Michaelsen, 1988) e.g. complement activation and Fey-receptor interaction. as well as Ig catabolism (Kim et al., 1995). Interestingly, inter-heavy-light chain disulfides are also considered to be important for some effector functions and Ag-binding properties (Dorai et al., 1992). Trx may also down-regulate effector functions by rendering Igs susceptible to proteolytic attack analogous to the effects of papain digestion in the presence of cysteine (Jefferis, 1990). In this context the structural compactness of IgG2, the main subclass involved in immune response to polysaccharide antigens (Hammarstrom and Smith, 1986) could be advantageous in maintaining its effector functions in bacterial infections. In a conceivable mechanism bacterial Trx would be unable to exert any effect on this subclass. The Trx system may also play a more general role in the assembly, disassembly and proper folding of Igs in a way similar to that exerted by the two Trx-like domains of protein disulfide isomerase (PDI) on the folding of various proteins (Lundstrom and Holmgren, 1990). In conclusion, we have shown that the thioredoxin system efficiently cleaves interchain disulfides in human IgG subclasses except for IgG2. These in vitro findings suggest that the Trx system may be involved in structural and functional aspects related to Ig or Ig-like receptors. AcknoLt,lr~gements-We wish to thank Mrs Maggy Magnusson for skilful technical assistance. The study was supported by the Swedish Medical Research Council (grants no. 16X-00105. 13X-3529 and 13X1 1213), the Swedish Foundation for Health Care Sciences and Allergy Research, and the Inga-Britt and Arne Lundberg’s Foundation.
REFERENCES Aase A., Sandlie I., Norderhaug
L., Breeke 0. H. and Michaelsen T. E. (1993) The extended hinge region of IgG3 is not required for high phagocytic capacity mediated by Fc receptors but the heavy chains must be disulfide bonded. Europeun Journal cf immunology 23, 1546 1551. Andersson M., Holmgren A. and Spyrou G. (1996) NK-lysin, a disulfide containing effector peptide of T-lymphocytes is reduced and inactivated by thioredoxin reductase. Implication for a protective mechanism against NK-lysin cytotoxicity. Journal of’Biological Chrmi.rtr~~ 271, 10 I 16-l 0 120. Biguet C.. Wakasugi N.. Mishal Z., Holmgren A., Chouaib S.. increases the Tursz T. and Wakasugi H. ( 1994) Thioredoxin proliferation of human B-cell lines through a protein kinase C-dependent mechanism. Journal of’ Biological Chemistr!, 269,28865-28870. Brekke 0. H., Michaelsen T‘. E., Aase A. and Sandlie I. (1994) Structure-function relationships of human IgG. fn?munobgist 2, 125-130. Brekke 0. H., Sandin R., Michaelsen T. E. and Sandlie I. (1993) Activation of complement by an IgG molecule without a genetic hinge. Nuture 363, 628-630. Brtiggemann M., Williams G. T., Bindon C. 1.. Clark M. R..
716
C. G. M. MAGNUSSON
Walker M. R., Jefferis R., Waldmann H. and Neuberger M. S. (1987) Comparison of the effector functions of human immunoglobulins using a matched set of chimeric antibodies. Journal of E.xperitnental Medicine 166, 1351-1361. Buchwald B. M. (1971) An immunoglobulin light chain of subgroup III of 1 type with a free sulfhydryl group. Canadian Journal OfBiochemistry 49,900-902 Burton D. R. (1990) The conformation of antibodies. In Fc Receptors and the Action of' Antibodies (Edited by Metzger H.), pp. 31-54. American Society for Microbiology, Washington DC. Cseke C. and Buchanan B. B. (1986) Regulation of the formation and utilization of photosynthate in leaves. Biochimica et Biophysics Acta 853, 43-63. Dorai H., Wesolowski J. S. and Gillies S. D. (1992) Role of inter-heavy and light chain disulfide bonds in the effector functions of human immunoglobulin IgG 1. Molecular Immunology 29, 1487-1491. Ericson M. L., Hiirling J., Wendel-Hansen V., Holmgren A. and Rosen A. (1992) Secretion of thioredoxin after in ritro activation of human B cells. Lymphokine and Cytokine Reseurch 5, 201-207. Grankvist K., Holmgren A., Luthman M. and TBljedal I.-B. (1982) Thioredoxin and thioredoxin reductase in pancreatic islets may participate in diabetogenic free-radical production. Biochemical und Biophysical Research Communications 107, 1412-1418. Hammarstriim L. and Smith C. I. E. (1986) IgG subclasses in bacterial infections. Monographs in Allergy 19, 122-133. Holmgren A. (1984) Enzymatic reduction-oxidation of protein disulfides by thioredoxin. Methods in Enzymology 107, 295300. Holmgren A. (1985) Thioredoxin. Annuul Review of Biochemistry 54, 237-271. Holmgren A. (1989) Thioredoxin and glutaredoxin system. Journal of Biologic& Chemistry 264, 13963-13966. Holmgren A. and Bjiirnstedt M. (1995) Thioredoxin and thioredoxin reductase. Methods in Enzymology 252, 199-208. Holmgren A. and Luthman M. (1978) Tissue distribution and subcellular localization of bovine thioredoxin determined by radioimmunoassay. Biochemistry, 17,40714077. Ikeyama S. (1987) Thermally induced disulfide bond exchanges in human IgE. Molecuhr Immunology 24,231-237. Isenman D. E., Dorrington K. J., Painter R. H. (1975) The structure and function of immunoglobulin domains. II. The importance of interchain disulfide bonds and the possible role of molecular flexibility in the interaction between immunoglobulin G and complement. Journul qf Immunology 114, 172661729. Jack H.-M., Sloan B., Grisham G., Reason D. and Wabl M. (1993) Large cytoplasmic inclusion body K-chain has unusual intrachain disulfide bonding. Journal of Immunology 150, 49284933. Jefferis R. (1990) Molecular structure of human IgG subclasses. In The Human IgG Subclusses: Molecular Anulysis qf Structure, Function and Regulation (Edited by Shakib F.), pp. 1530. Pergamon Press, Oxford. Jefferis R., Reimer C. B., Skvaril F., de Lange G., Ling N. R., Lowe J.. Walker M. R., Phillips D. J., Aloisio C. H., Wells T. W., Vaerman J. P., Magnusson C. G. M., Kubagawa H., Cooper M., Vartdal F., Vandvik J. P., Haaijamn J. J., Make12 O., Sarnesto A., Lando Z., Gergely J., Rajnaviilgyi E., L&z16 G., Radl J. and Molinaro G. A. (1985) Evaluation of monoclonal antibodies having a specificity for human IgG subclasses: results of an IUISjWHO collaborative study. Immunology~Letters 10, 2233252.
et al.
Kim J.-O., Tsen M.-F., Ghetie V. and Ward E. S. ( 1995) Evidence that the hinge region plays a role in maintaining serum levels of the murine IgGl molecule. Molecular Immurzo/og~~ 32,467475. King D. J., Adair J. R., Angal S., Low D. C., Proudfoot K. A., Lloyd J. R., Bodmer M. W. and Yarranton G. T. (1992) Expression, purification and characterization of a mousehuman chimeric antibody and chimeric Fab’ fragment. Biochemical Journal 281, 317-323. Larsson L.-J., Holmgren A., Smedsrod B., Lindblom T. and Bjork I. (1988) Subunits of human rz-macroglobulin produced by specific reduction of interchain disulfide bonds with thioredoxin. Biochemistry 27, 983-99 1, Lilja G., Magnusson C. G. M., Oman H., Johansson S. G. 0. (1990) Serum levels of IgG subclasses in relation to IgE and atopic disease in early infancy. Clinical and Experimental Allergy 20,40774 13. Lundstrom J. and Holmgren A. (1990) Protein disulfide isomerase is a substrate for thioredoxin reductase and has thioredoxin-like activity. JournalqfBiolo~gical Chemistry 265, 9114-9120. Michaelsen T. J. (1988) Alteration of the conformation of human IgG subclasses by reduction of the hinge SSS bonds. Moleculur Immunolqyy 25, 639-646. Morel1 A., Skvaril F. and Barandun S. (1976) Serum concentrations of IgG subclasses. In Clinical Immunobiolog~ (Edited by Back F. and Good R.). pp. 35-76. Academic Press, New York. Nakamura H. S., DeRosa S., Roederer M., Anderson M. T.. Dubs J. G., Yodoi J., Holmgren A., Herzenberg L. A. and Herzenberg L. A. (1996) Elevation of plasma thioredoxin levels in HIV-infected individuals. Znternationulfmmunology~ 8, 101-109. Oi V. T., Vuong T. M.. Hardy R., Reidler J., Dangl J.. Herzenberg L. A. and Stryer L. A. (1984) Correlation between segmental flexibility and effector function of antibodies. Nature 307, 136-140. Padilla C. A., Martinez-Galisteo E., BBrcena J. A., Spyrou G. and Holmgren A. (1995) Purification from placenta, amino acid sequence structure comparisons and cDNA cloning of Molecular and Cellular Endocrinolo,yy human glutaredoxin. 85, l-12. Rosen A., Lundman P., Carlsson M., Bhavani K., Srinivasa B. R., Kjellstriim G., Nilsson K. and Holmgren A. (1995) A CD4+ T cell line-secreted factor growth promoting for normal and leukemic B cells identified as thioredoxin. Internutional Immunology 7, 6255633. Rozell B., Hansson H.-A., Luthman M. and Holmgren A. (1985) lmmunohistochemical localization of thioredoxin and thioredoxin reductase in adult rats. Europeun Journal of Cell Biology 38, 79-86. Schauenstein E., Dachs F., Reiter M., Gombotz H. and List W. (1986) Labile disulfide bonds and free thiol groups in human IgG. I. Assignment to IgGl and IgG2 subclasses. International Arc~hicesof Allergy and Applied Immunoiogy~80, 174-179. Snow M. E. and Amzel L. M. (1988) A molecular-mechanics study of the conformation of the interchain disulfide of human immunoglobulin G4 (IgG4). Molecular Immunology 25, 1019-1024. Tagaya Y., Maeda Y., Mitsui A., Kondo N., Matsui H., Hamuro J., Brown N., Arai K.-I., Yokota T., Wakasugi H. and Yodoi J. (1989) ATL-derived factor (ADF) an IL2 receptor/Tat inducer homologous to thioredoxin; possible involvement of dithiol-reduction in the IL2 receptor induction. EMBO Journal 8, 757-764.
Human
IgG is substrate
Thelander L. and Reichard P. (1979) Reduction of ribonucleotides. Annual Review qf‘Biochemistr,v 48, 133-l 58. Wakazugi N., Tagaya Y., Wakasugi H.. Mitsu A., Maeda M., Yodoi J.. Tursz T. (1990) Adult T-cell leukemia-derived factor!thioredoxin produced by both human T-lymphotropic virus type I- and Epstein-Barr virus-transformed lymphocytes acts as autocrine growth factor and synergizes with
for the thioredoxin
system
1 and interleukin 2. Procredinys Academy ofSciences U.S.A. 87, 8282-8286.
interleukin
717 of’ the ,litrlionul
Yodoi J. and Tursz T. (1991) ADF a growth-promoting factor derived from adult T cell leukemia and homologous to thioredoxin: involvement in lymphocyte immortalization by HTLV-I and EBV. Adt~unc~c~s in Ctmcw Rc.wtrrch 57, 38 I 411.
Pergamon PII: SOl61-5890(97)00092-8
Printed 111Great Britain 01 (,I -5X90 97 $17.00 c 0.00
HUMAN IgG IS SUBSTRATE FOR THE THIOREDOXIN SYSTEM: DIFFERENTIAL CLEAVAGE PATTERN OF INTERCHAIN DISULFIDE BRIDGES IN IgG SUBCLASSES CARL
*++ MIKAEL G. M. MAGNUSSON, ARNE H0LMGREN.r
BJdRNSTEDTJ-
and
*Department of Laboratory Medicine, Division of Clinical Immunology, Karolinska Institute. Karolinska Hospital, S-171 76 Stockholm. Sweden: and “rThe Medical Nobel Institute for Biochemistry. Department of Medical Biochemistry and Biophysics. Karolinska Institute. Stockholm. Sweden
Abstract-Thioredoxin. a 12,000 mol. wt protein with two redox-active cysteine residues. togethet with thioredoxin reductase and NADPH. may reduce protein disulfides and thereby act as a molecular probe of their structure and reactivity. Interchain and intrachain disulfides are structural elements in and therefore potential substrates for the reduced thioredoxin, Trx(SH&. It LV;IS all immunoglobulins investigated whether such disulfides are cleaved in human polyclonal IgG and IgG subclass myeloma proteins by both the human and the Eschrri.shitr w/i thioredoxin systems. The reactions wrre montored spectrophotometrically as oxidation of NADPH at 340 nm, and by following the kinetics ol the cleavage patterns with sodium dodecyl sulfate--polyacrylamide gel electrophoresis (SDS-PAGEI). Human IgG was a substrate for both prokaryotic and eukaryotic Trx(SH)?, which directly reduced IgG disulfides in a time and dose-dependent manner. Stoichiometric analyses indicated near-complete reduction of mainly inter-heavy-light chain and inter-heavy chain disulfides. and SDS--PAGE corroborated that the buried intrachain disulfides were left intact. The kinetic studies showed that IgG I. IgG3 and lgG4 were readily reduced into heavy and light chains via the formation of half-molecules with slightly slower kinetics for lgG4. In sharp contrast. IgG2 was not cleaved at all. even with increased thioredoxin concentrations or reduction times. A binall but significant NADPH consumption by IgG2 myeloma proteins suggested reduction of II labile interchain or surface-exposed mixed disulfide. Consistent results were obtained for seceral IgG myeloma proteins within each subclass. The structural and functional importance of interchain disulfides in immunoplobulins suggests physiological implications of the thioredoxin system. t’ 1997 Elsevier Science Ltd. All rights reserved
&I’ ~ort/s: antibodies.
enzyme, human,
IgG subclasses.
INTRODUCTION
+ NADPH
thioredoxin
Trx - (SH)? + Protein - S2 + Trx - S2
Thioredoxin (Trx), a 12.000 mol. wt protein with two redox-active cysteine residues in the conserved active site (-Cys-Gly-Pro-Cys-). operates together with thioredoxin reductase (TR) as an NADPH-dependent protein disulfide reductase system (Holmgren, 1985. 1989) by reactions I and II: Trx-S,
reduction.
+ H’ ZTrx-(SH), + NADP+
(I)
IAuthor to whom correspondence should be addressed. Tel: 46-8-51775938; Fax: 46-8-969475. Ahhrwiatims; Ig. immunoglobulin, 2-ME, 2-mercaptoethanol. Trx. thioredoxin. TR, thioredoxin reductase. E-Trx. E. coli thioredoxin. H-Trx. human thioredoxin, E-TR, E. coli thioredoxin reductase, H-TR, human thioredoxin reductase. TE-buffer. 50 mM Tris+HCI. I mM EDTA, pH 7.5.
+Protein-(SH),.
(II)
Trx has a large and growing number of functions as. for example, hydrogen donor to ribonucleotide reductase at DNA replication (Thelander and Reichard. 1979), regulator of photosynthetic enzymes (Cskke and Buchanan, 1986) and modulator of immunological events (Rosen et al., 1995; Tagaya ct al., 1989: Wakazugi ct ~1.. 1990; Yodoi and Tursz. 1991). Human Trx is secreted by activated human lymphocytes (Ericson et ul., 1992). upregulates IL2 receptors (Tagaya et al.. l989), and promotes growth and proliferation of B-cells (Biguet CI (I/., 1994: RosCn ef ul., 1995: Yodoi and Tursz. 1991). Moreover, Trx is present in human plasma and elevated levels are found in HIV-infected patients (Nakamura rt ~1.. 1996). Interchain
structural 709
and intrachain disultide bonds are common elements in all immunoglobulins and are there-
710
C. G. M. MAGNUSSON et al.
fore potential substrates for the Trx system. IgG is the most common lg in serum and four subclasses exist, i.e. IgGI, IgG2, lgG3 and lgG4, which differ in effector functions such as complement activation and Fey-receptor binding (Burton, 1990; Jefferis, 1990). lgG1 (65%) is the dominating subclass in serum, followed by lgG2 (25%), lgG3 (7%) and lgG4 (3%) (Morel1 et al., 1976). Inter-heavy chain disulfide bridges in the hinge region vary in numbers among the four IgG subclasses (Burton, 1990) and are considered to be involved in both the structural properties and physiological functions of IgG subclasses (Brekke et al., 1993, 1994; Burton, 1990; Dorai et al., 1992; lsenman et al., 1975; Jefferis, 1990; Michaelsen, 1988; Snow and Amzel, 1988). The aim of the present investigation was to study the reactivity of IgG with the Trx system and elucidate whether any differences exist in the cleavage patterns of intrachain and interchain disulfides among the four human IgG subclasses.
Spectrophotometric
measurements
The reducing capacity of the Trx systems was followed in an automatic Zeiss PMQIII spectrophotometer by monitoring the oxidation of NADPH at 340 nm using 1 cm semi-microcuvettes. A change in AXbonrn of 0.0062 between the reaction and reference cuvettes represents the reduction of 1 ,nM disulfide. Following preliminary experiments, the final concentrations of the components were 0.3- 0.6 mg/ml (2-4pM) for the Ig, 3 PM for Trx. 0.1 PM for TR and 300-600 PM for NADPH if not otherwise stated. The reactions were performed in TE-buffer at room temperature for up to 60min in a final volume of 0.5 ml and A340nmwas recorded every 30sec. Equal volumes in the reaction and reference cuvettes were always maintained by the addition of TE-buffer. The NADPH consumption caused by the reduction of Trx itself, initially in mainly the oxidized form, was determined in the absence of lg and subtracted before any calculations. Kinetics and gel electrophoresis
MATERIALS
AND METHODS
Enzyme preparations E. coli thioredoxin reductase (E-TR), E. coli thioredoxin (E-Trx), human recombinant thioredoxin (H-Trx) and human placental thioredoxin reductase (H-TR) were purified preparations (Holmgren and Bjornstedt, 1995). Reduced E-Trx was obtained by treatment with a 20fold molar excess of dithiothreitol (Sigma Chemical Co., U.S.A.) at 37°C for 15 min (Holmgren and Bjbrnstedt, 1995). Excess dithiothreitol was removed by gelchromatography on a NAP-5 column (Pharmacia AB, Uppsala, Sweden) with ice-cold nitrogen-equilibrated 50 mM Tris-HCI, 1 mM EDTA, pH 7.5 (TE-buffer). The reduced product was kept in aliquots at -70°C until used.
Immunoglobulins
IgG myeloma proteins [two lgGl(h-), one lgGl(l), two lgG2(rc), one lgG2(1), four lgG3(k-), one lgG3(1I), three IgG4(rc) and one lgG4(;1)] were purified with standard chromatographic procedures, or have been described previously (Jefferis et al., 1985). Protein concentration was determined by absorbance at 280 nm (EZgOnm,,% = 14.0). Most IgG subclass proteins had a purity exceeding 98% (w/w), calculated on the sum of the other subclasses, respectively, except for two lgG2s and one lgG3 which contained 3-5% IgGl . The degree of contamination of a given IgG subclass protein by the others was determined by isotype-specific competitive enzyme immunoassay based on monoclonal antibodies (mAbs) and WHO-calibrated (67/97) standard (Lilja et al., 1990). Polyclonal IgG for intravenous immunotherapy (165 mg/ml; batch no. 36909-51, Kabi-Pharmacia, Stockholm, Sweden) had the following IgG subclass composition as determined by competitive enzyme immunoassay: IgGl (61%), lgG2 (35%), lgG3 (3.1%) and lgG4 (0.61%).
The reduction of lg by the human or E. coli Trx systems was performed at room temperature in TE-buffer for different times (0, 15 set, 1 min, 4 min, 16 min, 1 hr and 4 hr). Reaction aliquots were treated with one volume of 10.8 mM iodoacetamide (Sigma) dissolved in nitrogensaturated 0.2 M Tris-HCI, pH 8.0, per three volumes of reaction mixture. The final concentrations of lg and all components of the Trx systems were similar to those described for the spectrophotometric experiments above. Control samples consisted of a mixture of all reagents except Trx. Gel electrophoresis under non-reducing and reducing conditions was performed on a PHAST@ system using polyacrylamide gradient gels (4-l 5%; 8-25%) sodium dodecyl sulfate (SDS) buffer strips, sample applicator S/l, and high or low mol. wt standards as outlined by the manufacturer (Pharmacia AB). Samples were prepared by mixing one volume of SDS buffer [lo mM Tris-HCI, 1 mM EDTA, pH 8.0, containing 5% (w/v) SDS and 0.02% (w/v) bromophenol blue] with one volume of test sample and heating the mixture for 5 min at 1OO’C. Reducing sample conditions were obtained after a similar heat-treatment in SDS buffer containing 10% (v/v) 2ME. The gels were stained with Coomassie brilliant blue R250 in the PHAST-system and the mol. wts of the proteins were estimated from a plot of relative migration versus the logarithm of mol. wts of standard proteins. RESULTS Spectrophotometric
analysis
Following preliminary experiments with lgG3, the IgG subclass with most disulfides (Table I), we found it absolutely necessary to add lg, TR and NADPH to both cuvettes and start the reaction by Trx, as the lg concentration interferred at AX40nmand TR itself caused NADPH consumption (0.27 pM/min). To illustrate the
Human Table
IgG is substrate
I. Characteristics
kG
of human
IgGl IgG2 IgG3 IgG4
IgG subclasses study
711
system
used in the calculations
of this
Number of S-S-bridges/lgG molecule ___-__ Inter-H-L chain Total Inter-H-H chain _~~__~_____ ~. _.
Molecular weight
subclass
for the thioredoxin
150,000 150,000 160,000 150,000
2 4 11 2
2 2 2 2
4 6 13 4
Data from Jefferis (1990)
and role of the different components of the Trx system three experiments with IgG3 are shown (Fig. 1). Firstly, addition of E-TR (0.1 PM) gave a minor linear change in the consumption of NADPH (approx. 0.3pM:min) due to oxidation by E-TR as alluded to above. Secondly, addition of E-Trx (3 FM), in the absence of TR, caused no change at all in AX4”“,,,. Thirdly, addition of both E-TR (0.1 PM) and E-Trx (3 PM) to the reaction cuvette gave a rapid and substantial consumption of NADPH that plateaued at 25pM after 5min. As expected. addition of E-Trx (3 PM) to one of two cuvettes containing just 0.1 LIM E-TR and 300 PM NADPH gave a consumption rapidly plateauing at 3pM (data not shown). The NADPH consumption obtained for IgG3 (Fig. 1) showed that about 12 disulfide bridges were reduced, corresponding well with the total number of interchain bridges i.e. 11 inter-heavy chain and two interheavy-light chain disulfides (Table 1). Altogether, these experiments demonstrate the specificity of the detection specificity
system and that Trx is an intermediate electron carrier in the reduction process of IgG3. The reduction of IgG3 by six different concentrations of E-Trx (0.03311 I’M) showed that for concentrations consumption plateaued after 52 1 PM the NADPH 1Omin (Fig. 2) corresponding to the cleavage of about 13 disulfide bridges. It is interesting from a physiological point of view that even 0.03 PM Trx (360 ng/ml) managed to reduce about three disulfide bridges in IgG3 after 1Omin. Reduction of five 2-fold IgG3 concentrations (0.6-9.6pM) with the E-Trx (0.3 FM) system demonstrated that 2 11 disulfides were cleaved after 10 min for all IgG3 concentrations except for 9.6pM. In the latter case only seven disulfides were reduced (data not shown), probably due to NADPH depletion. The reduction of polyclonal IgG and several purified IgG subclass proteins were studied with the H-Trx system and representative reduction patterns are illustrated in Fig. 3. The consumption of NADPH, transformed into
30
-
0
1
2
3
4
5
6
TR O.lpM
7
8
9
10
Time (mid Fig. 1. Disulfide reduction followed by the spectrophotometric procedure using human IgG3 and the E. coli Trx system. The reactions were performed in TE-buffer containing 300lM NADPH. IgG3 was added to both reaction and reference cuvettes, to a final concentration of 2.1 PM. TR ( ?? ), Trx (A) or TR/Trx (0) was added to the reaction cuvettes, to final concentrations of 0.1 PM. 3 ILM or 0.1 PM/~ PM. respectively. Equal volumes of buffer were added to the reference cuvettes and the initial reduction of Trx-S, has been subtracted.
712
C. G. M. MAGNUSSON
c’t N/.
-*
0.03 &I
Trx
-*
0.1 HIM Trx
-
1 jhl Trx
-
3 /A4 Trx
-
11 pM Trx
,#I
7-I
0~:::::::::::::::::;;: 0
1
2
3
4
5
6
7
8
9
10
Time (min) Fig. 2. Reduction of lgG3 by six cuvettes containing IgG3 (2.4pM), final concentrations of 0.03 ,L~M(A), Equal volumes of buffer were added
-2
different concentrations of E. co/i Trx. E-Trx was added to NADPH (300/~M) and E-TR (0.1 /tM) in TE-buffer, to reach 0.1 ,uM (0) 0.3 ,uM (4), I /tM (I). 3 ,LIM (0) and 1I LLM (A). to the reference cuvettes and the reduction of Trx-S, has been subtracted.
I
1
0
5
10
15
20
25
30
35
40
45
Time (min) Fig. 3. Reduction of polyclonal IgG Polyclonal IgG (A). IgGl (A), lgG2 0.1 PM H-TR and 600itM NADPH reactions were started by addition of Equal volumes of buffer were added
and IgG subclass myeloma proteins by the human Trx system. (+). IgG3 (0) and IgG4 (M) were added to cuvettes containing in TE-buffer. to reach final concentrations of 2-3pM. The H-Trx to a final concentration of 3 ;tM to the reaction cuvettes. to the reference cuvettes and the initial reduction of H-Trx has been corrected for.
the number of disulfide bridges reduced per IgG molecule, showed that after 45min about 13 disulfide bridges were reduced in IgG3(x), four in IgGl (x), one in IgG2(k-), three in IgG4(rc) and two or three in polyclonal IgG. An additional set of two IgGl, two IgG2, four IgG3 and three IgG4 subclass proteins was also tested with the E-Trx system for different periods of time. Concordant results with the human system, and no influence from
light chain or G2m23 shown).
Gel electrophoretic
allotype,
were obtained
(data not
anul~wis
To investigate further the kinetics and cleavage pattern of the reduction process, polyclonal IgG and IgG subclass proteins were incubated with the E. coli or human Trx
Human
system for different
periods
(O-l hr). The
1gG is substrate
for the thioredoxin
reactions were
analysed by SDS-poly(PAGE) under non-reducing conditions. The two Trx systems gave near identical results and showed that the cleavage patterns for IgGl, lgG3 and IgG4 were similar but had slightly different kinetics whereas, in sharp contrast, IgG2 was not cleaved at all (Fig. 4). Polyclonal IgG gave a pattern suggesting partial cleavage, as intact Ig persisted after 2 hr of reduction (data not shown). For all IgG subclasses there is mainly one band with a mol. wt of 150,000-162,000, visible in all gels at 0 time (control), representing intact lg. Minor additional bands probably represent disulfide exchange products that may form during heat-treatment of the sample in SDS buffer (Ikeyama, 1987) as all proteins were highly pure and gave two bands after complete reduction with 2-ME (Fig. 4). Trx(SH), reduced IgG3 most efficiently, as intact Ig disappeared after 4min, whereas intact IgGl and IgG4 were still present. Concomitantly. bands with mol. wts of 76,000-83,000 appeared. probably representing half-molecules (HL). Reduction of IgG3 generated many bands smaller than the heavy chain, presumably representing products differently reoxidized in the hinge region. The presence of HL of lgG4 at time 0 (control) is an expected heterogeneity in IgG4 (King et al., 1992). Further Trxmediated reduction of IgGl and IgG3 leads to complete cleavage of HL into 47,000&56,000 heavy (H) chains and 24,000-22,000 light (L) chains, respectively. after 1 hr. IgG4 was also split into heavy and light chains, though not completely. after 1 hr of reduction but well after 2 hi (data not shown). In sharp contrast, IgG2 was cleaved neither after prolonged incubation (4 hr) nor after terminated
acrylamide
by alkylation
and
gel electrophoresis
12
3
4
5
6
7
system
713
increasing the Trx concentration 5-fold (I 5 AIM). despite the latter conditions giving faster kinetics for IgGl and IgG4 (data not shown). Further SDS -PAGE analyses on one IgGl, two IgG2, two IgG3 and two IgG4 of different light chains confirmed the cleavage differences between subclasses and also showed, as expected, that neither the light chain nor the G2m23 allotype had any effect on the cleavage pattern. In control experiments Ig were exposed to all reagents except for Trx and no cleavages uerc observed even after 2 hr (data not shown). confirming the spectrophotometric data. In order to study the reduction of disultides directly (reaction II). E-Trx(SH)? only was added to IgGl. lgG2 and IgG3. and samples were alkylated after different times. Three concentrations of E-Trx(SH): were tested representing molar ratios to the total concentration of interchain disulfide bridges in the Ig of I : IO. I : I and IO: I (Table 1). The results showed that E-Trx(SH), in a IOfold molar excess and at equimolar concentration reduced both IgGl (Fig. 5) and IgG3 (data not shown) but not IgG2 (data not shown). Half-molecules (HL) appeared after 15 set as well as H and L chains at the highest Trx concentration. Although the reduction rate was faster with the highest concentration of Trx the reductions were not complete. as half-molecules (HL) persisted even after 1 hr of reduction. The incomplete reduction may be related to the consumption of limited amounts of reduced E-Trx because of self-oxidation. All the data strongly suggested that mainly interchain disulfides. and no intrachain disulfides. were reduced by the Trx systems. To evaluate this further. IgGl and IgG4 were reduced by the E-Trx system for I hr. alkylated and analysed by SDS-PAGE under both reducing and non-
a
Fig. 4. SDS-PAGE of human monoclonal IgG subclass proteins following incubation for different times with the human Trx system and subsequent alkylation. Lane I, Fully 2-ME-reduced Ig: lane 2. high mol. wt standards; lane 3, control sample with all reagents (lg. H-TR, NADPH and TE-buffer) except for H-Trx; lanes 4-8, Ig treated with the Trx system for I5 set (lane 4) I min (lane 5). 4min (lane 6). 16 min (lane 7) and I hr (lane 8) followed by alkylation with iodoacetamide. (A) IgGl: (B) lgG2: (C) lgG3; (D) IgG4. Samples in lanes 228 were treated with SDS buffer without 2-ME prior to SDS-PAGE on gradient gels (4-15%). Note the differences in size between heavy chains (H) and light chains (L) generated by 2-ME and the Trx system. respectively.
714
C. G. M. MAGNUSSON FlgGt
al
2
3
4
5
6
7
8@3
4
5
et (II. 6
7
8lc.l
4
5
6
7
8
Fig. 5. SDS-PAGE of IgGl (6.7pM) incubated for different times with three different concentrations of E. coli Trx-(SH),, i.e. 270, 27 and 2.7/*M, respectively. The molar ratios of reduced E-Trx(SH)2:IgGl-SS were 1:lO (A), 1:l (B) and 1O:l (C). Lane I, Fully reduced Ig with SDS buffer containing 2-ME; lane 2, low mol. wt standards; lane 3, control sample with all reagents (Ig, NADPH and TE-buffer) except for E-Trx(SH)*; lanes 4-8, Ig alkylated with iodoacetamide following treatment with the E-Trx(SH)2 for 15 set (lane 4), 1min (lane 5) 4 min (lane 6) I6 min (lane 7) and 1hr (lane 8). Samples in lanes 2-8 were treated with SDS buffer without 2-ME prior to SDS-PAGE on gradient gels (4-l 5%). Note the appearance of the E-Trx band (mol. wt 12,000) in panels B and C.
reducing conditions (Fig. 6). The E-Trx-treated alkylated samples of IgGl and IgG4 had, under non-reducing conditions, heavy and light chains with mol. wts of 54,000 and 25,000, respectively. These mol. wts were apparently smaller than those obtained under reducing conditions, i.e. 59,000-62,000 and 28,000, respectively. These latter mol. wts are the same as those observed for controls following complete 2-ME reduction (Fig. 6). The lower mol. wts of lg chains following Trx reduction are interpreted as the result of intact intrachain disulfides giving them a higher electrophoretic mobility that will decrease upon full 2-ME reduction due to linearization of the chains.
1
2
3
4
5
6
7
8
94 kD + 67 kD + 43 kD +
H
30 kD +
L
20 kD + 14kD
j
Fig. 6. SDS-PAGE on gradient gels (S-25%) demonstrating interchain disulfide cleavage of IgGl and IgG4 following reduction for 1 hr by the E. coli Trx system and subsequent alkylation with iodoacetamide. Aliquots of alkylated IgGl and IgG4 were treated (15 min, 1OOC) with SDS buffer containing 2-ME (lanes 2 and 6, respectively) or with SDS buffer without 2-ME (lanes 3 and 7, respectively). Comparison is made to fully 2-ME-reduced IgGl and IgG4 control samples not exposed to the Trx system (lanes 4 and 8, respectively). Lanes 1 and 5, Low mol. wt standards.
DISCUSSION This study demonstrates that human IgG is an efficient substrate for the Trx system with an apparent reaction rate of the same order as for insulin, i.e. 2 x lO”M-‘set-’ (Holmgren, 1985; Holmgren and Bjornstedt, 1995). Our spectrophotometric and gel electrophoretic data show that both the E. coli and the human Trx system, in a doseand time-dependent manner, reduce interchain disulfide bridges in purified human monoclonal IgG proteins of different subclasses and in polyclonal IgG. The reduction was dependent on Trx(SH)z because in control experiments neither E. coli nor mammalian TR alone cleaved any disulfides. This is consistent with the inability of TR to reduce protein disulfides (Holmgren, 1984. 1985), although mammalian TR reduces a wide range of low mol. wt substrates (Holmgren and Bjornstedt, 1995) including the T-cell effector polypeptide NK-lysin (Andersson et al., 1996). Interestingly, the IgG subclasses showed differences in their susceptibility to Trx-mediated reduction, as verified for both the human and the E. coli system. Thus. while IgGl, IgG3 and IgG4 showed similar but kinetically slightly different cleavage patterns with a ranking order of susceptibility IgG3 > IgGl > IgG4, in contrast IgG2 was not cleaved at all. These findings were consistent using several myeloma proteins of each subclass and independent of the Trx system used. The subclass differences could be explained neither by the type of light chain, by the G2m23 allotype, nor as a dose-response effect, as IgG2 was not cleaved, even at a high Trx concentration (15 PM) or after prolonged incubation (4 hr). Thus, the incomplete reduction of polyclonal IgG observed by SDS-PAGE could be explained by its 35% content of IgG2. The cleavage pattern for IgGl, IgG3 and IgG4 observed by SDS-PAGE showed that the inter-heavy chain disulfides first readily reacted, generating half-molecules comprised of one heavy and one light chain. This
Human IgG is substrate for the thioredoxin system was followed by a slower cleavage of the inter-heavylight chain disulfides. This interpretation is supported by the kinetic data from the spectrophotometric experiments indicating the reduction of four bridges in IgG1, 13 in IgG3 and almost four in IgG4 following 45min of reduction. Together, these results indicate not only that Trx reduces the inter-heavy chain disulfides more efficiently than the inter heavy-light chain disulfides but also that the intrachain disulfides are not affected. The SDS-PAGE results showed that further reduction of alkylated Trx-reduced chains with 2-ME gives products of greater size which were interpreted as intrachain disulfide reduction since the compactness of the domains is lost due to linearization (Jack et al., 1993). Trx has previously been used to produce half-molecules of c(?macroglobulin (Larsson et al., 1988) and may be regarded as a sensitive probe for protein structure (Holmgren, 1984). The differences observed among the IgG subclasses to reduction by Trx might be explained from a structural point of view. As IgG2 is a more compact molecule than IgGl and IgG3, sterical constraints may impede access of Trx to the disulfides of this subclass (Burton, 1990; Jefferis, 1990; Snow and Amzel, 1988). The slower reduction rate of IgG4 is in line with this because IgG4 has intermediate structural characteristics. Accessibility problems may also explain the inability of Trx-mediated reduction of buried intrachain disulfides. It is interesting to note that the sensitivity of IgG subclasses to enzymatic cleavage with papain into Fc and Fab fragments follows the same pattern. IgGl and IgG3 are more readily digested by papain than IgG4, whereas IgG2 is quite resistant (Jefferis, 1990). Although we were unable to detect any Trx-mediated reduction in the three IgG2 proteins tested by SDS-PAGE we observed a small but consistent consumption of NADPH in the spectrophotometric experiments corresponding to the reduction of about one disulfide bridge. This cannot be attributed to contamination from other IgG subclasses but rather we consider that one of the four inter-heavy chain disulfides is more labile, or that a surface-exposed thiol is engaged in a mixed disulfide. IgG2 has been proposed to contain such sulfhydryl groups (Schauenstein et al., 1986) as have light chains (Buchwald, 1971). The functional and physiological implications of our findings have yet to be elucidated but its worth mentioning that the concentrations of Trx used in this study are within the micromolar and nanomolar physiological ranges encountered in mammalian tissues (Grankvist rt al., 1982; Holmgren and Luthman, 1978; Padilla et al., 1995; Rozell et al., 1985) and human plasma (Nakamura et al., 1996), respectively. Moreover, activated lymphocytes secrete Trx (Ericson et ul., 1992) potentially giving higher local concentrations. Higher rates of reduction are also to be expected under less aerobic conditions and at 37% Thus, one may envision that Trx exerts stimulatory or down-regulating physiological effects on circulating Abs and their membrane counterparts. Effector functions, which differ among IgG subclasses (Brtiggemann et al.. 1987; Burton, 1990; Jefferis,
715
1990) may be modulated since the hinge region, where the disulfides are, contributes to structural properties (Brekke et al., 1994; Michaelsen, 1988) which are of importance for segmental flexibility (Oi et al., 1984), different effector functions (Aase et al., 1993; Brekke ct al., 1994; Isenman et al., 1975; Michaelsen, 1988) e.g. complement activation and Fey-receptor interaction. as well as Ig catabolism (Kim et al., 1995). Interestingly, inter-heavy-light chain disulfides are also considered to be important for some effector functions and Ag-binding properties (Dorai et al., 1992). Trx may also down-regulate effector functions by rendering Igs susceptible to proteolytic attack analogous to the effects of papain digestion in the presence of cysteine (Jefferis, 1990). In this context the structural compactness of IgG2, the main subclass involved in immune response to polysaccharide antigens (Hammarstrom and Smith, 1986) could be advantageous in maintaining its effector functions in bacterial infections. In a conceivable mechanism bacterial Trx would be unable to exert any effect on this subclass. The Trx system may also play a more general role in the assembly, disassembly and proper folding of Igs in a way similar to that exerted by the two Trx-like domains of protein disulfide isomerase (PDI) on the folding of various proteins (Lundstrom and Holmgren, 1990). In conclusion, we have shown that the thioredoxin system efficiently cleaves interchain disulfides in human IgG subclasses except for IgG2. These in vitro findings suggest that the Trx system may be involved in structural and functional aspects related to Ig or Ig-like receptors. AcknoLt,lr~gements-We wish to thank Mrs Maggy Magnusson for skilful technical assistance. The study was supported by the Swedish Medical Research Council (grants no. 16X-00105. 13X-3529 and 13X1 1213), the Swedish Foundation for Health Care Sciences and Allergy Research, and the Inga-Britt and Arne Lundberg’s Foundation.
REFERENCES Aase A., Sandlie I., Norderhaug
L., Breeke 0. H. and Michaelsen T. E. (1993) The extended hinge region of IgG3 is not required for high phagocytic capacity mediated by Fc receptors but the heavy chains must be disulfide bonded. Europeun Journal cf immunology 23, 1546 1551. Andersson M., Holmgren A. and Spyrou G. (1996) NK-lysin, a disulfide containing effector peptide of T-lymphocytes is reduced and inactivated by thioredoxin reductase. Implication for a protective mechanism against NK-lysin cytotoxicity. Journal of’Biological Chrmi.rtr~~ 271, 10 I 16-l 0 120. Biguet C.. Wakasugi N.. Mishal Z., Holmgren A., Chouaib S.. increases the Tursz T. and Wakasugi H. ( 1994) Thioredoxin proliferation of human B-cell lines through a protein kinase C-dependent mechanism. Journal of’ Biological Chemistr!, 269,28865-28870. Brekke 0. H., Michaelsen T‘. E., Aase A. and Sandlie I. (1994) Structure-function relationships of human IgG. fn?munobgist 2, 125-130. Brekke 0. H., Sandin R., Michaelsen T. E. and Sandlie I. (1993) Activation of complement by an IgG molecule without a genetic hinge. Nuture 363, 628-630. Brtiggemann M., Williams G. T., Bindon C. 1.. Clark M. R..
716
C. G. M. MAGNUSSON
Walker M. R., Jefferis R., Waldmann H. and Neuberger M. S. (1987) Comparison of the effector functions of human immunoglobulins using a matched set of chimeric antibodies. Journal of E.xperitnental Medicine 166, 1351-1361. Buchwald B. M. (1971) An immunoglobulin light chain of subgroup III of 1 type with a free sulfhydryl group. Canadian Journal OfBiochemistry 49,900-902 Burton D. R. (1990) The conformation of antibodies. In Fc Receptors and the Action of' Antibodies (Edited by Metzger H.), pp. 31-54. American Society for Microbiology, Washington DC. Cseke C. and Buchanan B. B. (1986) Regulation of the formation and utilization of photosynthate in leaves. Biochimica et Biophysics Acta 853, 43-63. Dorai H., Wesolowski J. S. and Gillies S. D. (1992) Role of inter-heavy and light chain disulfide bonds in the effector functions of human immunoglobulin IgG 1. Molecular Immunology 29, 1487-1491. Ericson M. L., Hiirling J., Wendel-Hansen V., Holmgren A. and Rosen A. (1992) Secretion of thioredoxin after in ritro activation of human B cells. Lymphokine and Cytokine Reseurch 5, 201-207. Grankvist K., Holmgren A., Luthman M. and TBljedal I.-B. (1982) Thioredoxin and thioredoxin reductase in pancreatic islets may participate in diabetogenic free-radical production. Biochemical und Biophysical Research Communications 107, 1412-1418. Hammarstriim L. and Smith C. I. E. (1986) IgG subclasses in bacterial infections. Monographs in Allergy 19, 122-133. Holmgren A. (1984) Enzymatic reduction-oxidation of protein disulfides by thioredoxin. Methods in Enzymology 107, 295300. Holmgren A. (1985) Thioredoxin. Annuul Review of Biochemistry 54, 237-271. Holmgren A. (1989) Thioredoxin and glutaredoxin system. Journal of Biologic& Chemistry 264, 13963-13966. Holmgren A. and Bjiirnstedt M. (1995) Thioredoxin and thioredoxin reductase. Methods in Enzymology 252, 199-208. Holmgren A. and Luthman M. (1978) Tissue distribution and subcellular localization of bovine thioredoxin determined by radioimmunoassay. Biochemistry, 17,40714077. Ikeyama S. (1987) Thermally induced disulfide bond exchanges in human IgE. Molecuhr Immunology 24,231-237. Isenman D. E., Dorrington K. J., Painter R. H. (1975) The structure and function of immunoglobulin domains. II. The importance of interchain disulfide bonds and the possible role of molecular flexibility in the interaction between immunoglobulin G and complement. Journul qf Immunology 114, 172661729. Jack H.-M., Sloan B., Grisham G., Reason D. and Wabl M. (1993) Large cytoplasmic inclusion body K-chain has unusual intrachain disulfide bonding. Journal of Immunology 150, 49284933. Jefferis R. (1990) Molecular structure of human IgG subclasses. In The Human IgG Subclusses: Molecular Anulysis qf Structure, Function and Regulation (Edited by Shakib F.), pp. 1530. Pergamon Press, Oxford. Jefferis R., Reimer C. B., Skvaril F., de Lange G., Ling N. R., Lowe J.. Walker M. R., Phillips D. J., Aloisio C. H., Wells T. W., Vaerman J. P., Magnusson C. G. M., Kubagawa H., Cooper M., Vartdal F., Vandvik J. P., Haaijamn J. J., Make12 O., Sarnesto A., Lando Z., Gergely J., Rajnaviilgyi E., L&z16 G., Radl J. and Molinaro G. A. (1985) Evaluation of monoclonal antibodies having a specificity for human IgG subclasses: results of an IUISjWHO collaborative study. Immunology~Letters 10, 2233252.
et al.
Kim J.-O., Tsen M.-F., Ghetie V. and Ward E. S. ( 1995) Evidence that the hinge region plays a role in maintaining serum levels of the murine IgGl molecule. Molecular Immurzo/og~~ 32,467475. King D. J., Adair J. R., Angal S., Low D. C., Proudfoot K. A., Lloyd J. R., Bodmer M. W. and Yarranton G. T. (1992) Expression, purification and characterization of a mousehuman chimeric antibody and chimeric Fab’ fragment. Biochemical Journal 281, 317-323. Larsson L.-J., Holmgren A., Smedsrod B., Lindblom T. and Bjork I. (1988) Subunits of human rz-macroglobulin produced by specific reduction of interchain disulfide bonds with thioredoxin. Biochemistry 27, 983-99 1, Lilja G., Magnusson C. G. M., Oman H., Johansson S. G. 0. (1990) Serum levels of IgG subclasses in relation to IgE and atopic disease in early infancy. Clinical and Experimental Allergy 20,40774 13. Lundstrom J. and Holmgren A. (1990) Protein disulfide isomerase is a substrate for thioredoxin reductase and has thioredoxin-like activity. JournalqfBiolo~gical Chemistry 265, 9114-9120. Michaelsen T. J. (1988) Alteration of the conformation of human IgG subclasses by reduction of the hinge SSS bonds. Moleculur Immunolqyy 25, 639-646. Morel1 A., Skvaril F. and Barandun S. (1976) Serum concentrations of IgG subclasses. In Clinical Immunobiolog~ (Edited by Back F. and Good R.). pp. 35-76. Academic Press, New York. Nakamura H. S., DeRosa S., Roederer M., Anderson M. T.. Dubs J. G., Yodoi J., Holmgren A., Herzenberg L. A. and Herzenberg L. A. (1996) Elevation of plasma thioredoxin levels in HIV-infected individuals. Znternationulfmmunology~ 8, 101-109. Oi V. T., Vuong T. M.. Hardy R., Reidler J., Dangl J.. Herzenberg L. A. and Stryer L. A. (1984) Correlation between segmental flexibility and effector function of antibodies. Nature 307, 136-140. Padilla C. A., Martinez-Galisteo E., BBrcena J. A., Spyrou G. and Holmgren A. (1995) Purification from placenta, amino acid sequence structure comparisons and cDNA cloning of Molecular and Cellular Endocrinolo,yy human glutaredoxin. 85, l-12. Rosen A., Lundman P., Carlsson M., Bhavani K., Srinivasa B. R., Kjellstriim G., Nilsson K. and Holmgren A. (1995) A CD4+ T cell line-secreted factor growth promoting for normal and leukemic B cells identified as thioredoxin. Internutional Immunology 7, 6255633. Rozell B., Hansson H.-A., Luthman M. and Holmgren A. (1985) lmmunohistochemical localization of thioredoxin and thioredoxin reductase in adult rats. Europeun Journal of Cell Biology 38, 79-86. Schauenstein E., Dachs F., Reiter M., Gombotz H. and List W. (1986) Labile disulfide bonds and free thiol groups in human IgG. I. Assignment to IgGl and IgG2 subclasses. International Arc~hicesof Allergy and Applied Immunoiogy~80, 174-179. Snow M. E. and Amzel L. M. (1988) A molecular-mechanics study of the conformation of the interchain disulfide of human immunoglobulin G4 (IgG4). Molecular Immunology 25, 1019-1024. Tagaya Y., Maeda Y., Mitsui A., Kondo N., Matsui H., Hamuro J., Brown N., Arai K.-I., Yokota T., Wakasugi H. and Yodoi J. (1989) ATL-derived factor (ADF) an IL2 receptor/Tat inducer homologous to thioredoxin; possible involvement of dithiol-reduction in the IL2 receptor induction. EMBO Journal 8, 757-764.
Human
IgG is substrate
Thelander L. and Reichard P. (1979) Reduction of ribonucleotides. Annual Review qf‘Biochemistr,v 48, 133-l 58. Wakazugi N., Tagaya Y., Wakasugi H.. Mitsu A., Maeda M., Yodoi J.. Tursz T. (1990) Adult T-cell leukemia-derived factor!thioredoxin produced by both human T-lymphotropic virus type I- and Epstein-Barr virus-transformed lymphocytes acts as autocrine growth factor and synergizes with
for the thioredoxin
system
1 and interleukin 2. Procredinys Academy ofSciences U.S.A. 87, 8282-8286.
interleukin
717 of’ the ,litrlionul
Yodoi J. and Tursz T. (1991) ADF a growth-promoting factor derived from adult T cell leukemia and homologous to thioredoxin: involvement in lymphocyte immortalization by HTLV-I and EBV. Adt~unc~c~s in Ctmcw Rc.wtrrch 57, 38 I 411.