Research Interests

The endoplasmic reticulum (ER) is a cellular organelle through which a significant proportion of proteins pass on their way to their functional sites in membranes, exocytic and endocytic compartments, or the cell exterior. Far from being a passive traffic way, the ER is home to an array of molecular chaperones, which help proteins to fold and guide their maturation. Despite this support, protein biogenesis is an error-prone process. A considerable fraction of all newly synthesized polypeptides fail to attain their native conformation due to mutations, transcriptional and translational errors, folding defects, or imbalanced subunit synthesis. Mature proteins can be damaged by environmental stress conditions, such as high-energy radiation, chemical insults, or metabolic by-products. Malfunction or aggregation of defective proteins challenges the homeostasis of the ER and the cell as a whole. As a consequence, evolution has produced a protein quality control (PQC) network that operates on several levels to maintain the integrity of the ER.


The work of this group focuses on how ER homeostasis is maintained (Fig. 1). In other words, how quality control pathways selectively dispose aberrant proteins without jeopardizing correctly folded polypeptides. Signals contained in misfolded proteins of the ER-lumen and membranes are decoded by ubiquitin ligases anchored in the ER membrane. Proteins committed for degradation are exported from the ER lumen or membrane in a process termed protein dislocation. Subsequently, substrate molecules are ubiquitylated and degraded by cytoplasmic 26S proteasomes. This process is referred to as ER associated degradation or ERAD. However, misfolded proteins are not the only substrates of this system. It also plays a regulatory role. For example it eliminates rate-limiting enzymes of sterol synthesis in response of the flux through this pathway.

Since the ERAD pathway appears to be conserved from yeast to mammals, we have used the model organism Saccharomyces cerevisiae to investigate the fundamental mechanisms and to identify the key components of this important pathway. These are the HRD ubiquitin ligase and the Doa10 ubiquitin ligase. The HRD-ligase is crucial for turnover of membrane-bound and ER-luminal substrates. Doa10 targets membrane proteins for degradation that carry lesions in their cytoplasmic domains. Both yeast ubiquitin ligases and their identified co-factors are summarized in Fig. 2. The mammalian counterparts of the yeast components are mentioned as well.

In the last years, the group has identified and characterized components of these ligase complexes using genetics, molecular biology, and protein purification strategies. However, to unravel the basic molecular mechanisms how these protein machines work, additional in vitro approaches and quantitative methods have to be established.


Protein homeostasis in the endoplasmic reticulum

Fig. 1: Proteins translocate into the ER via Sec61 (step 1). Unfolded proteins (U) can engage chaperones to fold into their native conformation (step 2). Export factors select correctly folded proteins (F) (step 3) and transport them to the GOLGI (step 4). Retention factors prevent exit of unfolded proteins (step 5). Misfolded proteins (M) are the result of unproductive folding efforts. Chaperones try to remodel these polypeptides into a folding competent state (step 6). Retention factors keep aberrant proteins in the ER (step 7). Eventually, defective proteins (M), along with a fraction of folding intermediates (U), are relegated into the cytosol and disposed (step 8).

ER-associated protein degradation in yeast and mammalian cells

Fig. 2: ER-associated protein degradation in yeast and mammalian cells. Molecular chaperones and proteins of the glycosylhydrolase-47 family (Mns1 and Htm1) detect misfolded polypeptides and direct them to membrane bound ligases (Doa10, RMA1, HRD). After dislocation to the cytosolic face of the ER membrane, substrates are ubiquitylated by an ubiquitin ligase. All ligase complexes comprise a central, catalytic active RING finger protein (E3), ubiquitin-conjugating enzymes (E2), and additional factors. The AAA ATPase Cdc48 releases ubiquitylated molecules from the ER membrane. The adapter proteins Rad23 and Dsk2 escort the ubiquitylated substrates to the 26S proteasome for degradation. Concurrently Png1 deglycosylates glycoproteins through its association with Rad23. Proteins containing a glycan-interaction motif or ubiquitin-binding domains are depicted in red and blue, respectively. Proteins are labeled with their yeast names in blue and green letters indicate the mammalian counterpart.

Figures from: Hirsch, C., Gauss, R., Horn, S.C., Neuber, O., and Sommer, T. (2009) The ubiquitylation machinery of the endoplasmic reticulum. Nature 26, 453-460


Selected Papers:


Ubx2 links the Cdc48/p97-Complex to Endoplasmic Reticulum Associated Protein Degradation

Neuber, O., Jarosch,E.,  Volkwein, C., Walter, J., and Sommer, T. (2005) Nature Cell Biol. 7, 993-998

We and others have demonstrated that the AAA-ATPase Cdc48/p97 plays a crucial role in protein dislocation. However, the precise role in this transport step was not well characterized. In addition, it remained to be clarified how Cdc48/p97 is recruited to the ER-membrane. Using biochemical approaches we were able to demonstrate that the integral membrane protein Ubx2 mediates interaction of the Cdc48/p97-complex with the HRD and Doa10 ubiquitin-ligases. Ubx2 contains an UBX-domain that interacts with Cdc48/p97 and an additional UBA-domain. Both domains are located on the cytoplasmic surface of the ER and are separated by two transmembrane segments. In cells lacking Ubx2, the interaction of Cdc48/p97 with the ligase complexes is abrogated and in turn breakdown of ER-proteins is affected. Thus, protein complexes comprising the AAA-ATPase, the recruitment factor Ubx2, and one of the known ERAD ubiquitin ligases play central roles in ERAD. Furthermore, degradation of a cytosolic/nuclear protein, which is ubiquitinated by Doa10, is disturbed in absence of Ubx2. This demonstrates that different Cdc48/p97 dependent pathways converge at the ER-surface.


The Hrd1 ligase complex – a linchpin between ER-luminal substrate selection and cytosolic Cdc48 recruitment

Gauss, R., Sommer, T., and Jarosch, E. (2006) EMBO J. 25,1827-1835

We have recently developed methods to study extensively the interactions of yeast ER-membrane proteins by co-immunoprecipitation and co-purification. Using this assay, we were able to describe a complex of Hrd1/Der3 and its partner protein Hrd3 with the ER-membrane protein Der1. Genetic data imply that Hrd3 is the major substrate receptor of this heterogenic ligase complex in the ER-lumen. Although Hrd3 and Der1 bind to soluble substrate proteins independently, both proteins are essential to trigger substrate dislocation. At the cytosolic surface of the ER the HRD-complex associates with the AAA-ATPase Cdc48/p97. Cdc48p binding depends, as expected, on Ubx2, but most importantly also on substrate processing by the Hrd1-complex, suggesting that ubiquitination precedes substrate mobilization by the Cdc48/p97-complex.


A complex of Yos9 and the Hrd1-ligase integrates ER-quality control into the degradation machinery

Gauss, R., Jarosch, E., Sommer, T., and Hirsch, C. (2006) Nature Cell Biol. 8,849-854

How are newly synthesized proteins, which are in the process of folding distinguished from terminally misfolded proteins? One answer to this question involves N-linked glycans. These sugar structures are attached to newly synthesized proteins en bloc. After the covalent linkage of these sugar moieties, they are ‘trimmed’ by glucosidases and mannosidases. This trimming is a slow process and proteins with incompletely trimmed glycans are protected from degradation. Thus, the slow trimming event provides a time window in which newly synthesized proteins can fold. Although key lectins that may recognize the terminal glycan structures have been identified, a connection that links them to the ERAD pathway remained elusive. Recently we identified an association between the ER quality control lectin Yos9 and Hrd3. This interaction ties both pathways together. We identified designated regions in the luminal domain of Hrd3 that interact with Yos9 and the HRD ubiquitin ligase. Binding of misfolded proteins occurs via Hrd3, suggesting that Hrd3 recognizes proteins which deviate from their native conformation while Yos9 ensures that only terminally misfolded polypeptides are degraded.


Usa1 Functions as a Scaffold of the HRD-ligase

Horn, S.C., Hanna, J., Hirsch, C., Volkwein, C., Schütz, A., Heinemann, U., Sommer, T., and Jarosch, E. (2009) Mol. Cell 36, 782-793

By biochemical analyses we were able to show that Usa1, a newly identified subunit of the HRD ligase, plays a dual role within this complex: First, it recruits the ancillary factor Der1 and second it mediates oligomerization of the HRD complex. These separate activities mirror different requirements for the processing of malfolded proteins with distinct topology. On one hand, Usa1 promoted oligomerization of Hrd1 which is predominantly required for the breakdown of membrane proteins while it is dispensable for the turnover of soluble polypeptides. On the other hand, Der1 recruitment, in turn, is a prerequisite only for the degradation of soluble substrates. In this collaborative work with the group of Udo Heinemann, we could also identify the relevant domains of Usa1, which are involved in the relevant protein interactions at the ligase complex. The N-terminus of Usa1 binds the very C-terminal 40 amino acids of Hrd1. Binding of Der1 involves the C-terminal region of Usa1. Thus, Usa1 functions as a scaffold that assembles the different activities of the HRD-ligase for processing of different classes substrates that differ in their topology. In a broader sense, our data strengthen the view that scaffold proteins modulate ubiquitin ligase activities rather then being passive devices.


Htm1 protein generates the N-glycan signal for glycoprotein degradation in the endoplasmic reticulum

Clerc, S., Hirsch, C., Oggier, D.M., Deprez, P., Jakob, C., Sommer, T., and Aebi, M. (2009) J Cell Biol. 12, 159-172

N-linked glycans are essential for the breakdown of glycoproteins. The covalently attached oligosaccharide structure is used as a signal to display the folding status of the protein. Newly synthesized proteins receive a Glc3Man9GlcNAc2 modification. Such a glycan structure protects a newly synthesized protein from degradation. Subsequently it is trimmed by glucosidases and mannosidases until a specific signal is generated, which is recognized by the quality control ubiquitin ligase. Since trimming of glycans is slow, these processing steps provide a time window in which a newly synthesized protein can adopt its cognate conformation.

In a collaborative effort we were able to define the function of Htm1 as an ?1,2-specific exo-mannosidase that generates the Man7GlcNAc2 oligosaccharide with a terminal ?1,6-linked mannosyl residue on degradation substrates. This oligosaccharide signal is decoded by the ER-localized lectin Yos9 that in conjunction with Hrd3 triggers the ubiquitin-proteasome dependent hydrolysis of these glycoproteins. The Htm1 exo-mannosidase activity requires processing of the N-glycan by glucosidase I, II and mannosidase I, resulting in a sequential order of specific N-glycan structures that reflect the folding status of the glycoprotein.

Since Htm1 generates the crucial signal that flags a protein for degradation, its activity must be tightly controlled. Thus, we searched for associated factors that could be involved in controlling Htm1 activity. Surprisingly, we co-purified the protein disulfide isomerase Pdi1 together with Htm1. Binding of Pdi1 occurs at the Htm1 C-terminus whose function is unknown. Our results raise the speculation that the activity of Htm1 is linked to incorrect disulfide bridge formation.



Current Scientific Topics:


Hrd3 acts as a ‘Holdase’ and substrate receptor of the HRD-ligase

Franziska Zimmermann and Sathish Kumar Lakshimpathy

In order to understand how substrates bind to the ER-luminal sub-module of the HRD ubiquitin ligase (Hrd3, Yos9) we started to work with purified components. Since the function of Hrd3 does neither rely on its membrane anchor nor on its cytosolic carboxyl-terminus this protein can be expressed in insect cells in a soluble form that comprises the entire ER luminal domain. We also expressed Yos9 in insect cells and purified it. Expressed and purified Hrd3 binds Yos9 from yeast extracts, indicating that the protein has a native fold. Using denatured firefly luciferase we demonstrated a direct interaction of Hrd3 with a misfolded protein. Next, we tested whether Hrd3-binding deploys holdase activity by preventing the aggregation of a denatured protein. To this end we diluted denatured luciferase in buffer containing Hrd3 and monitored its aggregation by static light scattering. This experiment clearly shows that Hrd3 prevents aggregation. Moreover, its ‘holdase’ activity is similar to that of the Hsp70-type ER-chaperone Kar2/BiP. In this assay, Yos9 and a Yos9 mutant defective in the glycan-binding MRH domain also display some ‘holdase’ activity, indicating that Yos9 averts protein aggregation independently of its lectin function.

Our in vitro data imply that Hrd3 directly binds misfolded proteins. In vivo, Hrd3 was shown to interact with the ER-luminal Hsp70-type chaperone Kar2. We favor the idea that Kar2 segregates bound clients from Hrd3 and mobilizes them after a scanning procedure. The structural rearrangement upon ATP hydrolysis of Kar2 could be required for this process because Hrd3 does not hold an ATPase function.

Future experiments will address the function of Hrd3 and Yos9 in protein aggregation and protein folding and its interaction with the chaperone machinery.


The ubiquitination capacity of the HRD-ligase

Katrin Bagola and Maximillian von Delbrück

One of the major, yet unsolved questions in the field is how proteins are extracted from the ER. We previously could demonstrate that ubiquitination of dislocated substrates is essential for their dislocation, suggesting that ubiquitin conjugation contributes to the directionality of the membrane passage. Furthermore we have shown that over expression of ubiquitinK48R delays ERAD significantly. However, this simple picture of ERAD being executed by attachment of a simple K48-linked polyubiquitin chain to substrates is challenged by recent mass. Spectrometric analyses.

To this end, we have established an in vitro ubiquitin-conjugation assay with purified components. The cytosolic domains of all components of the HRD ligase are expressed in E. coli. To make sure that the proteins are soluble, we expressed several of them without their transmembrane segments. Using this system, we were able to demonstrate that ubiquitin-conjugation at the ER-surface is a complex reaction and involves a variety of different ubiquitin chains and not only K48-ubiquitin chains. We will continue this powerful in vitro approach and will include quantitative parameters.


The function of Cdc48/p97 at cellular membranes

Martin Mehnert and Ernst Jarosch

The AAA-ATPase Cdc48, in mammals termed p97 or VCP, is a key player in many ubiquitin-dependent processes like homeotypic membrane fusion, cell cycle control, proteasome-mediated protein degradation and DNA repair. To fulfill these activities, Cdc48/p97 teams up with a large set of diverse ancillary proteins in a temporally and spatially controlled manner. Yeast Ubx2, for example, recruits Cdc48Npl4/Ufd1 to the ER membrane and establishes its interaction with ubiquitin ligase complexes involved in ERAD. In the past this group were among the first to unravel the important function of the yeast Cdc48Npl4/Ufd1 complex in ERAD. They demonstrated that a loss of Cdc48Npl4/Ufd1 significantly delays ERAD. Notably, mutations in Cdc48Npl4/Ufd1 primarily impair the extraction of the substrates from the ER without affecting their polyubiquitylation.

Dfm1 is a small integral protein of the ER membrane that displays strong structural similarities to a subunit of the HRD ligase, Der1. Dfm1 is also associated with at least one of the ERAD ligases. Preliminary results strongly imply that Dfm1 does not bind to Cdc48Npl4/Ufd1 but rather to another Cdc48-subcomplex, which may fulfill functions aside from the proteolysis.

In the future we will establish a functional network of membrane-bound Cdc48/p97-related activities and unravel novel mechanisms by which protein ubiquitylation affects the fate of secretory proteins.


Htm1p is a mannosidase that generates N-linked Man7GlcNAc2 glycans to accelerate the degradation of misfolded glycoproteins

Anett Köhler and Christian Hirsch

N-linked glycans are essential for the breakdown of glycoproteins. The covalently attached oligosaccharide structure is used as a signal to display the folding status of the protein. Newly synthesized proteins receive a Glc3Man9GlcNAc2 modification. Such a glycan structure protects a newly synthesized protein from degradation. Subsequently it is trimmed by glucosidases and mannosidases until a specific signal is generated, which is recognized by the quality control ubiquitin ligase. Since trimming of glycans is slow, these processing steps provide a time window in which a newly synthesized protein can adopt its cognate conformation.

To determine if Htm1p acts as a mannosidase in this pathway, Anett Köhler and Christian Hirsch developed an in vitro system consisting of Htm1p purified from yeast cells and the commercially available bovine pancreatic ribonuclease B (RNAse B) as model substrate. This protein has a single glycosylation site at Asn34, which is occupied by heterogeneous oligomannose-type glycans containing 5–9 mannose residues. Additionally, four disulfide bonds stabilize the structure of RNAse B, which allow conversion of RNAse B into a misfolded quality control substrate by reductive denaturation. Using this assay they could demonstrate that a complex of Htm1p and the oxidoreductase Pdi1p converts Man8GlcNAc2 oligosaccharides of the glycoprotein RNAse B to the Man7GlcNAc2 form, which enhances the elimination of aberrant glycoproteins.