Báo cáo khoa học: The heat shock protein 70 molecular chaperone network in the pancreatic endoplasmic reticulum ) a quantitative approach potx

Post-translational protein transport into the yeast ER involves the Sec complex in the membrane, comprising the Sec61p subcomplex [1,2], the putative signal peptide receptor subcomplex [

in the pancreatic endoplasmic reticulum ) a quantitative approach

Andreas Weitzmann, Christiane Baldes, Johanna Dudek and Richard Zimmermann

Medizinische Biochemie und Molekularbiologie, Universita¨t des Saarlandes, Homburg, Germany

The initial step in the biogenesis of approximately

30% of eukaryotic proteins is their integration into

the membrane or their transport into the lumen of the

endoplasmic reticulum (ER) Protein integration or

transport into the ER can occur cotranslationally or

post-translationally, and typically requires signal

pep-tides at the N-terminus of the precursor proteins and

the transport machinery Post-translational protein

transport into the yeast ER involves the Sec complex

in the membrane, comprising the Sec61p subcomplex

[1,2], the putative signal peptide receptor subcomplex

[3,4], the heat shock protein (Hsp) 40, termed Sec63p

[5], and the luminal Hsp70 proteins Kar2p [6] and

Lhs1p [7] Sec63p and Kar2p are essential proteins in yeast and, together with Lhs1p, are also involved in cotranslational protein transport into the ER [8,9] Cotranslational protein transport into dog pancreas microsomes involves a similar Sec61 complex [10–12] Furthermore, protein transport into the mammalian

ER involves Hsp70-type molecular chaperones and their Hsp40-type cochaperones: Hsp70-type molecular chaperones of the ER lumen, IgG heavy chain-binding protein [BiP (also termed glucose-regulated protein (Grp) 78, HspA5) and Grp170 (Orp150) are involved

in cotranslational and post-translational insertion of precursor polypeptides into the Sec61 complex [13]

Keywords

BiP; endoplasmic reticulum; J-domains;

nucleotide exchange factors; molecular

chaperones

Correspondence

R Zimmermann, Medizinische Biochemie

und Molekularbiologie, Universita¨t des

Saarlandes, D-66421 Homburg, Germany

Fax: +49 6841 1626288

Tel: +49 6841 1626510

E-mail: bcrzim@uks.eu

(Received 15 June 2007, revised 9 August

2007, accepted 10 August 2007)

doi:10.1111/j.1742-4658.2007.06039.x

Traditionally, the canine pancreatic endoplasmic reticulum (ER) has been the workhorse for cell-free studies on protein transport into the mamma-lian ER These studies have revealed multiple roles for the major ER-lumi-nal heat shock protein (Hsp) 70, IgG heavy chain-binding protein (BiP), at least one of which also involves the second ER-luminal Hsp70, glucose-reg-ulated protein (Grp) 170 In addition, at least one of these BiP activities depends on Hsp40 Up to now, five Hsp40s and two nucleotide exchange factors, Sil1 and Grp170, have been identified in the ER of different mam-malian cell types Here we quantified the various proteins of this chaperone network in canine pancreatic rough microsomes We also characterized the various purified proteins with respect to their affinities for BiP and their effect on the ATPase activity of BiP The results identify Grp170 as the major nucleotide exchange factor for BiP, and the resident ER-membrane proteins ER-resident J-domain protein 1 plus ER-resident J-domain pro-tein 2⁄ Sec63 as prime candidates for cochaperones of BiP in protein trans-port in the pancreatic ER Thus, these data represent a comprehensive analysis of the BiP chaperone network that was recently linked to two human inherited diseases, polycystic liver disease and Marinesco–Sjo¨gren syndrome

Abbreviations

BiP, IgG heavy chain-binding protein; ER, endoplasmic reticulum; ERj, endoplasmic reticulum-resident J-domain protein; Grp, glucose-regulated protein; GSH, glutathione; GST, glutathione S-transferase; Hsp, heat shock protein; RM, rough microsome; SPR, surface plasmon resonance.

BiP was also identified as a luminal protein that is

involved in the completion of protein translocation

[14,15] Furthermore, BiP was shown to seal the

lumi-nal end of the mammalian Sec61 complex in the

absence of protein translocation and at several stages

during cotranslational translocation of preproteins

[16–18] BiP was shown to involve an unidentified

resi-dent ER Hsp40 and could not be substituted by its

lian ortholog of yeast protein Sec63p was shown to be

an abundant protein in canine pancreatic microsomes and was found in association with the Sec61 com-plex [19–21] The Sec63-related protein ER-resident J-domain protein (ERj) 1 was observed to be asso-ciated with translating ribosomes on the ER surface [22–24] and to be able to complement a yeast mutant that is deficient in Sec63p [25]

In the ER of Saccharomyces cerevisiae, four Hsp40 proteins with a luminal J-domain have been identified: the two membrane proteins Sec63p [5] and Scj2p [26], and the two luminal proteins Scj1p [27] and Jem1p [28] In the ER of various mammalian cells, five Hsp40 proteins have been identified: the three membrane pro-teins Sec63 (alternative names: ERj2, DnaJC2) [19–21], ERj1 (Mtj1p, DnaJC1) [22,29,30], and ERj4 (MDG1, DnaJB9) [31,32], and the two luminal proteins ERj3 (HEDJ, Dj9, DnaJB11) [33–35] and ERj5 (JPDI, DnaJC10) [36,37] (Fig 1, Table 1) Furthermore, nucleotide exchange factors) Sil1p in yeast [38,39] and Sil1 (also termed BAP) in mammalian cells [40]) have been identified, and Lhs1p and Grp170 were shown to be alternative nucleotide exchange factors for Kar2p and BiP, respectively [41,42] Both LHS1 and SIL1 are nonessential genes in yeast How-ever, simultaneous deletion of both genes results in synthetic lethality

Here we quantitatively characterized the Hsp70 chaperone network in the rough ER of a single

J domain

N

C

N

C

cytosol

ER lumen

J-domain J-domain

N

C

N

TRX Cys

GF

C

Grp170 BiP

Sil1

N

N N

C

C C

ER membrane

ATPase

domain

ATPase

domain

PB

domain

PB

domain

Fig 1 The established network of Hsp70-type molecular

chaper-ones in the lumen of the mammalian ER The cartoon summarizes

data from different cell types The putative domain organization of

the various proteins is indicated (PB, peptide-binding domain; GF,

glycine ⁄ phenylalanine-rich region; Cys, cysteine-rich region; TRX,

thioredoxin-like domains) The quantitative aspects are summarized

in Table 1.

Table 1 Hsp70 chaperones and their cochaperones of the ER in the canine pancreas The concentrations refer to a suspension of RMs, with a concentration of heterotrimeric Sec61 complexes of 2.12 l M [54] The affinities of Hsp40s for BiP are based on the SPR experiments shown in Fig 4 or were determined previously [21,22] The ATPase experiments shown in Fig 5 are the basis for the stimulatory effects of Hsp40, Sil1, and Grp170 The experimental details are given in Experimental procedures ND, not determined.

Protein

(alternative name)

Concentration

in suspension

of RMs (l M )

Recombinant protein (amino acid residues)

Affinity for BiP

in the presence

of ATP (K D in l M )

Stimulation of ATPase activity

of BiP (fold)

Further stimula-tion of BiP ATPase by Grp170 (fold)

Sil1 (fold)

ERj1 (Mtj1p, DnaJC1) 0.36

ERj2 (Sec63p, DnaJC2) 1.98

mammalian tissue, canine pancreas, which

predomi-nantly comprises exocrine cells Except for ERj4, all

mammalian Hsp40s of the ER were detected As Sil1

was found at very low concentrations, Grp170

(ortho-log of yeast Lhs1p) appears to act as the predominant

nucleotide exchange factor for BiP–ADP in the

pancre-atic ER The interactions of the various J-domains

with BiP were characterized by pull-down experiments,

surface plasmon resonance (SPR) spectroscopy, and

ATPase experiments These data provide the first

comprehensive and quantitative analysis of a

chaper-one network that was recently linked to two human

hereditary diseases, autosomal dominant polycystic

liver disease (OMIM 174050) and the

neurodegener-ative Marinesco–Sjo¨gren syndrome (OMIM 248800)

[43–46]

Results

Two Hsp70s, four Hsp40s and Sil1 form a

chaperone network in the pancreatic ER

Previously, we had determined the concentrations of

BiP, Grp170, ERj1, ERj2 and ERj3 in suspensions of

dog pancreas microsomes [21,22,33,34] In order to

determine how abundant ERj4, ERj5 and Sil1 are in

these pancreatic microsomes, purified glutathione

S-transferase (GST) hybrid proteins and specific

anti-bodies were employed in western blotting, according

to the established procedure (Fig 2; Table 1) In all

cases, two different antibodies were used that

recog-nized the respective recombinant protein In the case

of ERj5 and Sil1, these antibodies recognized an

iden-tical band in the pancreatic microsomes We

deter-mined concentrations of 2 lm for ERj5 and 5 nm for

Sil1 in the microsomal suspensions In the case of

ERj4, the antibodies failed to identify a common

antigen in the pancreatic microsomes Therefore, we

conclude that ERj4 is not an abundant protein in

dog pancreas microsomes under physiological

condi-tions This view is also supported by the fact that we

failed to characterize ERj4 in proteomic analysis of

these microsomes It follows from the concentrations

of Hsp70 and Hsp40 proteins in the ER lumen

(Table 1) that all Hsp40s can be associated with BiP

at any given time

The J-domains of all mammalian ER-resident

Hsp40s productively interact with BiP but differ

in their affinities for BiP

To determine whether the mammalian ER-resident

Hsp40s contain a functional J-domain, hybrid proteins

0 500 1000 1500

0 100 200 300

400

A

B

microsomes (µL)

protein (µg)

Fig 2 Quantitation of proteins in dog pancreas microsomes Serial dilutions of BSA (filled squares) were run on SDS polyacrylamide gels in parallel with two samples of purified recombinant protein (arrowheads, A) The proteins were stained with Coomassie Brilliant Blue, and the staining intensity was quantified by densitometry (Per-sonal Densitometer; Applied Biosystems, Krefeld, Germany) The same purified protein (arrowhead) was run on SDS polyacrylamide gels in parallel with serial dilutions of dog pancreas microsomes (filled circles, B) Subsequently, the proteins were transferred to poly(vinylidene difluoride) membranes and incubated with rabbit antibodies that were directed against the protein of interest and with a peroxidase conjugate of goat anti-(rabbit IgG) serum The bound antibodies were made visible by incubation with enhanced chemiluminescence (ECL) and exposure to X-ray film The intensity

of silver precipitation was quantified by densitometry The calcula-tion of the molar concentracalcula-tion of the respective protein in micro-somal suspensions was based on the predicted molecular mass of the protein, as calculated by the protean option of the LASERGENE DNASTAR sequence analysis software (GATC, Konstanz, Germany).

A B C

D

G

comprising GST and the respective protein or

J-domain were constructed, purified, and subjected to

three activity assays In the case of the membrane

pro-teins ERj1, ERj 2, and ERj4, the transmembrane

domains were absent form the GST hybrids (Table 1)

Thus, only the ER-luminal domains were analyzed in

these cases In the case of the two luminal Hsp40s

ERj3 and ERj5, GST hybrids were analyzed that

con-tained either the J-domains or the full-length proteins

When compared to each other, the two types of GST

hybrids behaved quite similarly in the functional assays

that were employed here (Table 1)

In the first series of experiments, ‘pull-down assays’

were carried out with detergent extracts of dog

pan-creas microsomes as described previously [21] GST

served as a negative control GST or GST hybrids

were immobilized on glutathione (GSH)–Sepharose

and incubated with detergent extracts of dog pancreas

microsomes in the absence or presence of ATP The

bound proteins were eluted and subjected to

SDS⁄ PAGE and subsequent staining with Coomassie

Brilliant Blue (Fig 3) All GST hybrids selectively

pulled down BiP from the detergent-solubilized

micro-somal proteins in the presence of ATP and less

efficiently in its absence (Fig 3A–F, lane 6 versus

lane 4) From our results, we conclude that BiP

inter-acts with all ERjs in a productive manner, as: (a) GST

did not pull down BiP (Fig 3G, lanes 4 and 6); and

(b) the other major molecular chaperones, present in

the detergent extract of dog pancreas microsomes

(such as Grp94 and calreticulin), did not bind to the

GST hybrids (Fig 3A–F, lane 6 versus lane 5) We

note, however, that the different GST hybrids were

dif-ferent in their BiP pull-down efficiencies (see below)

We next characterized the interaction of BiP with

the GST hybrids by SPR spectroscopy as described

previously [21] (Fig 4) We determined the apparent

affinities in the presence of ATP (KD), which are given

in Table 1 In summary, BiP has an approximately

10-fold higher affinity for ERj1 and ERj5 as compared to

ERj2, ERj3 and ERj4 However, we note that these

apparent affinities have to be treated with caution, as

the kinetics could not be fitted perfectly to a 1 : 1

binding model The generally accepted explanation for

this fact is that after J-domain-mediated ATP

hydroly-sis and in the absence of a real substrate, BiP binds

Hsp40 as a substrate [47] Accordingly, this interaction

is not seen when only the ATPase domain of BiP is employed instead of full-length BiP (data not shown)

We note that it was also observed for yeast Kar2p plus Sec63p that a stable interaction between this Hsp70– Hsp40 pair is possible in the absence of any substrate polypeptide [47], and stable interactions were also seen previously between BiP and mammalian ERj2⁄ Sec63 [21] and ERj1 [22], in both pull-down and SPR experi-ments

Next, we investigated whether the GST hybrids stim-ulate BiP’s ATPase activity under steady-state condi-tions, i.e in the presence of 500 lm ATP BiP was incubated with [32P]ATP[cP] in the absence or presence

of GST hybrid After various times of incubation, the samples were analyzed by TLC and phosphorimaging (Fig 5A–E; Table 1) According to the time-dependent hydrolysis of ATP under the different conditions, all J-domains stimulated the ATPase activity of BiP GST had no such stimulatory effect, even at much higher concentrations [21] Therefore, it seems to be unlikely that the observed stimulation of BiP’s ATPase activity

by the GST–J-domain hybrid was due to a BiP–sub-strate rather than a BiP–cochaperone interaction In the case of ERj1, ERj3, and ERj4, the stimulatory effects of the ERjs correlated with their affinities for BiP However, for unknown reasons, this was not the case for ERj2⁄ Sec63 and ERj5

Grp170 serves as an efficient and general nucleotide exchange factor for BiP Our previous attempts to purify Grp170 with ATP-affinity chromatography resulted in a mixture of Grp170 and BiP [13] Here, we employed gel filtration chromatography in the absence or presence of ATP as

a subsequent and final purification step (Fig 6A,B) In the absence of ATP, a proportion of BiP

cofractionat-ed with Grp170, with an elution maximum of both proteins at a position that corresponded to a molecular mass of 240 kDa (Fig 6A) In addition, the vast majority of BiP was observed at a position that corre-sponded to its monomeric molecular mass (70 kDa)

In the presence of ATP, however, there was hardly any overlap of the two proteins, and they more or less eluted according to their theoretical molecular masses (70 and 140 kDa; Fig 6B) Apparently, the two Hsp70-type molecular chaperones can form a heterodimeric

Fig 3 Functional characterization of Hsp40-type cochaperones of the ER: selective binding of BiP GST or GST hybrid was immobilized and incubated with detergent extract of microsomes in the absence or in the presence of ATP as described in Experimental procedures The unbound and bound proteins were collected and subjected to SDS ⁄ PAGE and subsequent staining with Coomassie Brilliant Blue We note that: (a) the band that is labeled BiP was identified as such by western blotting; and (b) we failed to detect Grp170 in the bound fractions under these conditions.

Fig 4 Functional characterization of Hsp40-type cochaperones of the ER: affinity for BiP SPR analysis was carried out with immobilized ERj3 (A), ERj3J (B), ERj5 (C), ERj5J (D), and ERj4 (E) and recombinant mouse BiP as described in Experimental procedures We note that for all Hsp40s there was no interaction observed in the absence of ATP or when ATPcS was used instead of ATP (not shown) The calculated affinities are given in Table 1 We note that prior to application of BiP, dissociation of previously applied BiP was allowed to reach completion (not shown).

complex in the absence of free ATP) i.e when at

least BiP can be expected to be in the ADP form –

and this complex is dissociated in the presence of an

excess of free ATP

In order to confirm this interpretation, an

immobi-lized antibody that recognizes native Grp170 was

employed A fraction from the gel filtration

chroma-tography in the absence of ATP that contained

approximately stoichiometric amounts of both

chaper-ones was incubated with immobilized antibodies to

Grp170 either in the absence or in the presence of

ATP Subsequently, the antibody-bound and unbound

proteins were analyzed by SDS⁄ PAGE and protein

staining (Fig 6C, lanes 2 and 3 versus lanes 4 and 5)

The antibody to Grp170 coimmunoprecipitated BiP

more efficiently in the absence than in the presence of

ATP; that is, a significant amount of BiP remained in

the unbound fraction in the presence of ATP (Fig 6C,

lane 3) Thus, the two chaperones are indeed able to

form a stable complex in the absence of ATP

In the next experiment, the ability of Grp170 to interact with Hsp40 was examined BiP served as an internal control for this experiment In order to keep the two Hsp70 chaperones from forming a complex, ATP was present A mixture of both Hsp70-type chap-erones was incubated with an immobilized J-domain (ERj1J) Subsequently, the J-domain-bound and unbound proteins were analyzed by SDS⁄ PAGE and protein staining (Fig 6D, lane 4 versus lane 2) As expected, BiP was efficiently bound by the immobilized J-domain In contrast, Grp170 was not bound by the immobilized J-domain, i.e remained in the unbound fraction (Fig 6D, lane 2) Thus, in contrast to BiP, Grp170 appears to be unable to form a stable complex with Hsp40-type proteins

Grp170 has a low basal ATPase activity that was hardly stimulated by ERj1J and serves as a nucleotide exchange factor for BiP in the presence of ERj1 [42] In order to analyze the nuleotide exchange activity of Grp170 in the presence of the other ERjs, steady-state

A

E

I

B

F

J

C

G

K

D

H

L

Fig 5 Functional characterization of Hsp40-type cochaperones and nucleotide exchange factors of the ER: effect on the ATPase activity of BiP ATP hydrolysis assays were carried out under steady-state conditions as described in Experimental procedures The concentrations were: ATP, 500 l M ; Sil1, 2 l M ; BiP and Kar2p, 2 l M ; Hsp40, 2 l M ; and Grp170, 0.25 l M

ATPase assays were carried out that involved BiP and

Grp170) in a physiological ratio ) plus Hsp40

(Fig 5A–E; Table 1) The established nucleotide

exchange factor Sil1) at about two-fold molar

excess) served as a positive control in these

experi-ments (Fig 5F–J) Under conditions of stimulation of

BiP’s ATPase activity by any ER-resident Hsp40,

Grp170 led to further acceleration of ATP hydrolysis

Thus Grp170 can serve as a nucleotide exchange factor

for BiP after stimulation of BiP by any ER-resident

Hsp40 Grp170 was more efficient than Sil1 in this

respect We note that Sil1 had previously been shown

to stimulate the ATPase activity of BiP in the presence

of ERj4 under slightly different conditions [40]

There-fore, the apparent lack of nucleotide exchange factor activity of Sil1 in the presence of ERj3, Erj4 and ERj5

in our experiments should not be taken as an indication

of a specialized function of Sil1 (Fig 5H–J; Table 1) However, the data point to the fact that the two nucleo-tide exchange factors have different efficiencies

Both ERj1 and ERj2⁄ Sec63 as well as Grp170 functionally interact with the yeast BiP ortholog Kar2p

The yeast ortholog of BiP that is termed Kar2p was observed to be unable to substitute for BiP in facili-tating Sec61 channel gating in canine pancreatic

- ATP

Grp170

Grp170 + BiP

fraction number

BiP BiP

+ ATP

100

50

0

fraction number

300

150

0

Fig 6 Characterization of Grp170: functional interactions Superose 6 gel filtration was carried out as described in Experimental procedures

in the absence (A) and presence (B) of ATP Fractions were collected and subjected to SDS ⁄ PAGE and subsequent staining with Coomassie Brilliant Blue Staining intensity was quantified by densitometry Grp170 (open squares) and BiP (filled circles) were identified as such by western blotting An aliquot of fraction 5 of the gel filtration in the absence of ATP (termed input and shown in lane 1) was incubated with immobilized antibodies to Grp170 in the absence or presence of ATP as indicated (C) The unbound (lanes 2 and 3) and bound (lanes 4 and 5) proteins were collected and subjected to SDS ⁄ PAGE and subsequent staining with Coomassie Brilliant Blue An aliquot of an ATP eluate

of ATP–agarose chromatography was incubated with GSH–Sepharose (– ERj1J) or immobilized ERj1J (+ ERj1J) in the presence of ATP (D) Subsequently, unbound (lanes 1 and 2) and bound (lanes 3 and 4) material were separated by centrifugation, and analyzed by SDS ⁄ PAGE and protein staining with Coomassie Brilliant Blue The protein ladder was run on the same gel (lane 5).

microsomes [18] In fact, it even had a dominant

nega-tive effect on BiP in these experiments Furthermore,

this gating activity of BiP was shown to involve an

unidentified resident ER Hsp40 [18] In analogy to the

situation in yeast, the respective Hsp40 is expected to

be a membrane protein in pancreatic microsomes As

lack of interchangeability of various Hsp70s has been

observed previously [48,49], it seemed reasonable that

the failure of Kar2p to support a complete ATPase

cycle in concert with one of the two most likely

candi-date Hsp40s in channel gating, ERj1 and ERj2, or the

major pancreatic nucleotide exchange factor, Grp170,

was responsible for the effects of Kar2p in channel

gat-ing in mammalian microsomes At first, we addressed

the question of whether Kar2p functionally interacts

with the two relevant J-domains Kar2p was stimulated

in its ATPase activity by the two J-domains to an

extent that is comparable to their stimulation of BiP

(Fig 5K,L) The stimulation by ERj1J was 3-fold and

that by ERj2J was 2.5-fold, as compared to 5.2-fold

and 3.5-fold (Table 1) Next, we deterrmined whether

Grp170 functionally interacts with Kar2p after

stimula-tion of its ATPase activity by ERj1J or ERj2J Grp170

stimulated the ATPase activity of Kar2p to an extent

that is comparable to the stimulation of BiP

(Fig 5K,L) The stimulation by Grp170 in the presence

of ERj1J was 2.4-fold and that of ERj2J was 2.2-fold,

as compared to 5.6-fold and 5.7-fold (Table 1) The

observed differences appear to be too small to provide

an explanation for the inability of Kar2p to substitute

for BiP in Sec61 gating in mammalian microsomes

Thus, lack of interchangeability between BiP and

Kar2p at the level of the Hsp40s ERj1 and ERj2 and at

the level of the nucleotide exchange factor Grp170 does

not appear to be responsible for the observed effect of Kar2p in channel gating experiments [18]

Discussion

The pancreatic network of ER-luminal chaperones under steady-state conditions

From the concentrations of the various chaperones and cochaperones in the lumen of canine pancreatic rough microsomes (RMs), one can extrapolate to the situation in the corresponding rough ER (Table 1, Fig 7) It has to be taken into account that during preparation of microsomes, about 50% of the luminal content is lost into the postribosomal supernatant and that the luminal volume of the microsomes is only a minor fraction of the total volume of the microsomal suspension (we estimate 1 : 500) Taken together, we estimate that in the rough ER, the concentrations of the most abundant chaperones such as BiP and ERj5

or their J-domains, such as in the case of ERj2⁄ Sec63, are in the low millimolar range Furthermore, the total amounts of Hsp70 and Hsp40 proteins in the ER lumen (Table 1) and the observed affinities (Table 1) allow one to conclude that, in principle, all J-domains can be associated with BiP at any given time In real-ity, however, a large proportion of BiP will be engaged with polypeptide substrates Therefore, one can assume that under these steady-state conditions, the deter-mined affinities become relevant On the basis of the analogies with yeast, and the facts that human ERj1 can complement deletion of the SEC63 gene in yeast, and that ERj4 appears to be absent from pancreatic rough ER under nonstress conditions, the two Hsp40s

ERj1 0.18m M ERj2

0.29m M * ERj4

2m M *

UPR BiP

5mM*

sensors:

PERK IRE1 ATF6

Grp170

0.005m M

Fig 7 The BiP network in the pancreatic ER The estimated ER-luminal concentration and the measured affinities are indicated The observed concentrations for RMs (Table 1) were corrected for the facts that about 50% of the luminal proteins leak out of the organelle dur-ing tissue homogenization and that the inner volume of the RM is small as compared to the total volume of the RM suspension (approxi-mately 1 : 500), in order to estimate the concentration of the luminal proteins in the ER of pancreatic cells UPR, unfolded protein response;

*UPR-inducible.

dates for alternatively cooperating with BiP in protein

transport This may explain why disruption of

SEC63⁄ ERJ2 in autosomal dominant polycystic liver

disease is not lethal, in contrast to the situation in

yeast [44] Grp170 was characterized as an alternative

nucleotide exchange factor for BiP [42] This is in

per-fect agreement with the fact that disruption of SIL1⁄

BAP in humans and mice does not cause lethality

[45,46,50] Thus, as in yeast, the presence of the

addi-tional nucleotide exchange factor Grp170 may

com-pensate for the loss of Sil1

The network of ER-luminal chaperones under

stress conditions

When misfolded proteins accumulate in the ER,

vari-ous signal transduction pathways are activated that

increase the biosynthetic capacity and decrease the

bio-synthetic burden of the ER This phenomenon is

termed the unfolded protein response Most of the

members of the chaperone network discussed here are

under control of the unfolded protein response

(indi-cated by an asterisk in Fig 7) Therefore, one would

expect ERj4 to be present in pancreatic microsomes

under stress conditions Furthermore, the GRP170 gene

would be expected to be overexpressed after disruption

of SIL1⁄ BAP in humans and mice Thus,

overexpres-sion of GRP170 may compensate for the loss of Sil1 in

most tissues However, for unknown reasons, this does

not seem to work for certain areas of the cerebellum in

patients suffering from Marinesco–Sjo¨gren syndrome

[45,46] and in the so-called woozy mice [50]

The ATPase cycle of BiP

In principle, BiP’s ATPase cycle follows the

well-estab-lished, paradigmatic functional cycle of DnaK [51]

Briefly, BiP–ATP has a low affinity for polypeptide

substrates Proteins with BiP-reactive J-domains have a

high affinity for BiP–ATP and can therefore bind to

the underside of the ATP-binding cleft Owing to their

additional domains, the Hsp40s ERj3, ERj4 and ERj5

may be able to bind polypeptide substrates and deliver

them to peptide-binding domains of BiP in the course

of their interaction with BiP In contrast, in the case of

ERj1 and ERj2, BiP appears to be recruited to the

substrate polypeptides by spatial proximity to the

Sec61 complex and ribosomes, respectively In any

case, interaction of the J-domain with the ATP-binding

cleft triggers ATP hydrolysis and a subsequent

confor-mational change in the peptide-binding domain

Apparently, in the in vitro analysis in the absence of

of the J-domain or neighboring parts of Hsp40 in the peptide-binding domain In the presence of BiP sub-strates, Hsp40s dissociate from BiP, and the polypep-tide substrates are trapped by BiP Next, BiP-bound ADP is exchanged for ATP, and the above-mentioned conformational change in the peptide-binding domain

is reversed Substrate is released, and BiP is ready for the next round of the cycle Typically, ADP–ATP exchange is catalyzed by a nucleotide exchange factor, such as Grp170 in the pancreatic ER, that has a high affinity for BiP–ADP We note that our observation of

a stable complex of Grp170 and BiP is perfectly in line with previous observations of chaperone complexes in the ER lumen [52,53]

Experimental procedures

Materials The protein ladder (10–200 kDa) was obtained from Life Technologies (Grand Island, NY, USA) ATP-C8-agarose, thrombin and peroxidase conjugate of goat anti-(rabbit IgG) serum were obtained from Sigma Chemical Company, Tauf-kirchen, Germany) [32P]ATP, ATP, GSH–Sepharose 4 Fast Flow, protein A–Sepharose and protein G–Sepharose 4 Fast Flow, Superose 6B, X-ray films and the enhanced chemilu-minescence (ECL) were obtained from GE Healthcare (Frei-burg, Germany) Poly(vinylidene difluoride) membranes and Centricon devices were obtained from Millipore (Schwal-bach, Germany) Chaps was obtained from Calbiochem (Schwalbach, Germany) Hepes and Coomassie Brilliant Blue were purchased from Serva (Heidelberg, Germany)

Purification of proteins from dog pancreas Dog pancreas microsomes were prepared as previously described [13] The microsomes were stripped with respect

to ribosomes according to published procedures [11] After reisolation of microsomes by centrifugation, the pellets were resuspended in extraction buffer (20 mm Hepes⁄ KOH,

pH 7.5, 400 mm KCl, 1 mm EDTA, 1.5 mm MgCl2, 2 mm dithiothreitol, 15% w⁄ v glycerol, 0.65% w ⁄ v Chaps), result-ing in a crude extract Typically, the ribosomes were pelleted by centrifugation for 30 min at 2C and 240 000 g

in a Beckman TLA 100.3 rotor Purification of ATP-bind-ing proteins was carried out on ATP-C8-agarose as described previously [13] Where indicated, the ATP eluate was concentrated in Centricon devices and subjected

to A¨kta chromatography in a Superose 6B column (16· 500 mm) (GE Healthcare) The running buffer was identical to the extraction buffer, except that glycerol was omitted, KCl was reduced to 200 mm, and 4 mm Mg-ATP was added where indicated