The fate of intracellular peptides and MHC class I antigen presentation
Neijssen, J.J.
Citation
Neijssen, J. J. (2008, February 6). The fate of intracellular peptides and MHC class I antigen presentation. Retrieved from https://hdl.handle.net/1887/12591
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/12591
Note: To cite this publication please use the final published version (if applicable).
A major role for TPPII in trimming proteasomal degradation products for MHC class I antigen presentation.
Immunity
2004 Apr;20(4):495-50.
A preview was written by J. Yewdell and M. P Princiotta.
Immunity
2004 April;20(4):495-506.
A Major Role for TPPII in Trimming Proteasomal Degradation Products for MHC Class I Antigen Presentation
subsequently degraded (Koopmann et al., 2000; Roelse et al., 1994). MHC class I molecules usually bind pep- tides of nine amino acids with appropriate anchor resi- dues (Rammensee et al., 1993). The TAP transporter also prefers to translocate peptides of around nine Eric Reits,1Joost Neijssen,1Carla Herberts,1
Willemien Benckhuijsen,2Lennert Janssen,1 Jan Wouter Drijfhout,2and Jacques Neefjes1,*
1Division of Tumor Biology The Netherlands Cancer Institute
amino acids (Momburg et al., 1994a), although substrate Plesmanlaan 121
peptides up to 40 amino acids can be translocated into 1066 CX Amsterdam
the ER lumen as well (Koopmann et al., 1996). TAP shows The Netherlands
minor specificity for peptide sequences (Momburg et al.,
2Department of Immunohematology
1994b), which is not unexpected since it has to provide and Blood Transfusion
different class I molecules with peptide epitopes.
Leiden University Medical Center
In vitro experiments suggest that most peptides gen- 2300 RC Leiden
erated by the proteasome will probably have an incor- The Netherlands
rect size for direct class I loading (Kisselev et al., 1999;
Toes et al., 2001). N-terminal extended epitopes can be trimmed to the correct size for MHC class I binding by Summary
cytoplasmic and ER luminal aminopeptidases, but most peptides will be completely recycled into free amino Intracellular proteins are degraded by the proteasome,
acids (Reits et al., 2003). Several cytoplasmic amino- and resulting peptides surviving cytoplasmic peptidase
and endopeptidases have been identified, including leu- activity can be presented by MHC class I molecules.
cine aminopeptidase (LAP) (Beninga et al., 1998), tripep- Here, we show that intracellular aminopeptidases de-
tidyl peptidase II (TPPII) (Geier et al., 1999; Tomkinson, grade peptides within seconds, almost irrespectively
1999), thimet oligopeptidase (TOP) (Mo et al., 1999), of amino acid sequence. N- but not C-terminal exten-
bleomycin hydrolase (BH) (Bromme et al., 1996; Stoltze sion increases the half-life of peptides until they are
et al., 2000), and puromycin-sensitive aminopeptidase 15 amino acids long. Beyond 15 amino acids, peptides
(PSA) (Johnson and Hersh, 1990; Stoltze et al., 2000).
are exclusively trimmed by the peptidase TPPII, which
Overexpression of LAP (Reits et al., 2003) or TOP (York displays both exo- and endopeptidase activity. Sur-
et al., 2003) reduces class I expression, suggesting a prisingly, most proteasomal degradation products are
role for these peptidases in trimming peptides for MHC handled by TPPII before presentation by MHC class I
class I molecules. In addition, the ER-associated amino- molecules. We define three distinct proteolytic activi-
peptidase is also involved in peptide generation for MHC ties during antigen processing in vivo. Proteasome-
class I molecules but also their destruction (Serwold et generated peptides relevant for antigen presentation
al., 2002; York et al., 2002). Since cells lack cytoplasmic are mostly 15 amino acids or longer. These require
carboxypeptidase activities (Reits et al., 2003), the pro- TPPII activity for further trimming before becoming
teasome should generate the correct C terminus of pep- substrates for other peptidases and MHC class I. The
tides for the MHC class I peptide binding groove (Cascio heterogeneous pool of aminopeptidases will process
et al., 2001), unless endopeptidases exist. TPPII appears TPPII products into MHC class I peptides and beyond.
to be essential for the generation of a particular HIV epitope (Seifert et al., 2003) and exhibits endopeptidase Introduction
activity in vitro (Geier et al., 1999; Seifert et al., 2003).
These data suggest that various peptidases modify the The proteasome is the dominant cellular protease de-
pool of class I peptides. Although there may be special- grading intracellular proteins (Kloetzel, 2001; Rock et
ization in location, substrate specificity, and amounts, al., 1994), including many newly synthesized proteins
their collective activity may determine the outcome of (Reits et al., 2000; Schubert et al., 2000). Only a small
a MHC class I response, since more than 99% of the fraction of these peptides is translocated by the trans- peptides are destroyed by peptidases (Princiotta et al., porter associated with antigen processing (TAP) into the 2003; Reits et al., 2003). The few peptides escaping lumen of the endoplasmic reticulum (ER), where they peptidase activity may represent a small pool of particu- can bind newly synthesized MHC class I molecules (re- lar peptides that resist proteases due to amino acid viewed by Yewdell et al., 2003). The vast majority of sequence and peptide length. Alternatively, a nonselec- generated peptides are however degraded by cyto- tive process that degrades peptides at random may plasmic peptidases before being able to interact with be operational with some inefficiency in the form of TAP (Fruci et al., 2003; Reits et al., 2003). In addition, escaping peptides. This is, however, hard to determine peptides can be trimmed by the ER-associated amino- on the basis of the few known in vitro-determined pepti- peptidase ERAAP or ERAP1 before binding to MHC dase specificities.
class I molecules (Saric et al., 2002; Serwold et al., 2002; Here, we have studied the specificity of the collective York et al., 2002), while other peptides are translocated cytoplasmic peptidase activities using internally quenched back into the cytoplasm by the Sec61 complex and peptides introduced into living cells. Peptides are rapidly degraded by aminopeptidases without dramatic se- quence specificity. N-terminal but not C-terminal exten-
*Correspondence: j.neefjes@nki.nl
62
Peptidase Activity in Living Cells
Figure 1. Detecting Peptidase Activity in Liv- ing Cells
(A) Principle of detecting peptidase activity.
Internally quenched peptides are injected in cells. The emission of the fluorescein side chain (F) is absorbed by the quencher Dabcyl moiety (Q). Emission of fluorescein can only be detected when the two groups are spa- tially separated after degradation.
(B) Representative degradation curves of two peptides in vivo. Model peptide T[K-Dabcyl]- NKTER[C-Fluorescein]Y (black line) was mi- croinjected (indicated by the arrow) in a single melanoma cell (Mel JuSo), and the appear- ance of fluorescence was monitored at 520 nm. This experiment was repeated with the same peptide with an N-terminal D amino acid (D-T)[K-Dabcyl]NKTER[C-Fluorescein]Y (gray line). The figure shows a merge of the two experiments. Fluorescence appears im- mediately after introduction in the cell and goes to completion.
sion increases the half-life of the reporter peptide, which the 9-mer peptide was no substrate for proteasomes (data not shown). This agrees with the observation that is in line with observations that carboxy peptidase activ-
ity is absent in the cytoplasm. Beyond 15 amino acids, N-terminally protected peptides were stable in vivo. Ap- parently, small peptides are digested in the cytoplasm peptides become the exclusive substrate for TPPII that
removes N-terminal sequences, including nine potential exclusively by aminopeptidases and not by carboxy or endopeptidases (Reits et al., 2003).
amino acid epitopes. However, epitopes located at the C terminus of these larger peptides are presented less
efficiently suggesting that TPPII can generate but also Sequence Specificity of the Intracellular Peptidase Pool
destroy epitopes. We show that TPPII plays an important
role in antigen processing, as most proteasomal prod- Some aminopeptidases like LAP have defined substrate specificities (Turzynski and Mentlein, 1990). Variation in ucts require further processing by TPPII for MHC class
I presentation. As a consequence, peptide generation the amino acid sequence of the reporter peptide may therefore affect the degradation rate in vivo. This is im- for MHC class I is severely hampered when TPPII activity
is inhibited. portant since an increased peptide half-life should theo-
retically correlate with a better chance to reach the ER lumen, resulting in an improved class I presentation. To Results
examine whether the cytoplasmic peptidases collec- tively show sequence selectivity, we systematically var- Aminopeptidases Degrade Peptides In Vivo
An internally quenched 9-mer peptide was used to de- ied amino acids positioned at the first, second, third, or last position within an internally quenched reporter tect peptidase activity in vivo. The quenched reporter
peptide (T[K-Dabcyl]NKTER[C-Fluorescein]Y) contains peptide sequence (Figures 2A–2D). The variable amino acids (indicated as “X”) were representatives of every a quenching Dabcyl group and a fluorescein group cou-
pled to amino acid two and eight, respectively. The chemical group of amino acids. With the exception of the third amino acid position, the variable amino acids quencher efficiently absorbs emission of the nearby flu-
orescein group, and fluorescence will only be detected were positioned outside the quencher-fluorophore box (indicated by “K” and “C”). Peptide degradation always when amino acids two and eight are separated in space
as the result of peptidase activity (Figure 1A). As shown started immediately after its introduction in living cells by microinjection and went to completion. Surprisingly, before (Reits et al., 2003), fluorescence appeared almost
immediately after microinjection of the quenched re- the different amino acids hardly affected the peptide half-life (Figures 2A–2D, the t1/2of full peptide degrada- porter peptide into living Mel JuSo cells, and degrada-
tion went to completion with a half-life of a few seconds tion is plotted), indicating that the collective pool of peptidases is able to efficiently degrade every peptide (t1/2 4.1 0.9 s; Figure 1B). The peptide was degraded
exclusively by aminopeptidases, as blocking the free N sequence almost irrespective of the type of amino acid at these positions. The differences in peptide half-life terminus of the peptide by a protective group inhibited
degradation (Reits et al., 2003). Treatment of cells with of the tested combinations are within a factor 2. The rate of degradation decreased only when the N-terminal lactacystin to inhibit proteasomal degradation had no
effect on the rapid rate of degradation, indicating that amino acid was replaced by an unnatural D amino acid
amino acid is not cleavable. In summary, the heteroge- neous pool of cytoplasmic peptidases in cells efficiently digests peptides with no major sequence selectivity un- less stereoisomeric D amino acids are introduced.
Cytoplasmic Peptidase Activity, Substrate Length, and Antigen Presentation by MHC Class I
Since the second amino acid of the original reporter peptide contained the quencher group, we repositioned the quencher molecule to the third position and ex- tended the N terminus by one additional amino acid to examine the effect of amino acid variations at the second position. While again no major effect of the sequence variation was observed, the half-life of the reporter pep- tide increased (Figure 2D). Apparently the presence of two additional N-terminal amino acids reduced proteo- lytic access to the quencher-fluorophore reporter. To examine the relationship between peptide length and in vivo half-life, the reporter peptide was extended either at the C or N terminus with a repeat of amino acids again representing the various chemical groups (Figure 3). When the C terminus was extended with three, six, or twelve amino acids, no increase in the half-life of the reporter sequence was observed (Figure 3A). Even a further extension by 18 amino acids did not increase the short half-life of the N-terminal reporter sequence in vivo. Apparently, trimming aminopeptidases are de- grading the first N-terminal amino acids of the reporter sequence at similar rates, irrespective of the length of the C terminus.
If correct, then addition of amino acids at the N termi- nus should increase the reporter’s half-life, as trimming aminopeptidases have to remove more amino acids be- fore encountering the internally quenched reporter. In- deed, extending the N terminus up to three additional amino acids increased the half-life of the reporter se- quence (Figure 3B). Surprisingly, further extension by 6, 11, or 18 additional amino acids did not lead to a further increase in half-life. To exclude that the presence of one or more proline residues in the N-terminal extension would influence aminopeptidase activity (Yellen-Shaw et al., 1997), we replaced these residues for threonine residues. The half-life of these model peptides was not affected by the proline to threonine exchange (data not shown). Despite more N-terminal amino acids, the half- life of the reporter sequence remains similar when pre- Figure 2. Sequence Specificity of the Collective Peptidase Activities
in Living Cells ceded by more than five amino acids, which might sug-
(A–D) 9-mer model peptides with systematic amino acid variations gest that another cytoplasmic peptidase is involved in at position 1 (A), 3 (B), and 9 (C) were tested for in vivo degradation peptide degradation for substrates over 15 amino by Mel JuSo cells. The sequences are shown in every panel with K
acids in length.
representing Lys-Dabcyl, C the Cys-fluorescein, and the enlarged
Although reporter peptides that are N-terminally ex- and underlined amino acid X the variable amino acid. A 10-mer
tended have an increased in vivo half-life, this does not peptide was used to test the effect of variable amino acids at posi-
necessarily mean that N extended antigens are better tion 2 (Figure 2D). Kat position 2 was preceeded by F. The variations
(introduced at position X) represent the different chemical groups presented by MHC class I. To test whether extensions of amino acids with, in addition, P as imino acid and G as the affected antigen presentation by MHC class I molecules, smallest amino acid. L represents the hydrophobic; D, acidic; K,
a HLA-A2-restricted influenza M epitope was expressed basic; F, aromatic; and T, neutral amino acids. D-[T] is the D amino
with N- or C-terminal extensions (Table 1) corresponding acid T. Shown is the half-life in seconds for every peptide (SEM).
to the sequences used in the model peptides tested in Figures 3A and 3B. The various minigenes were cloned in a vector upstream of an IRES sequence followed isomer (Figures 1B and 2A). This suggests that a large
portion of the collective peptidase activities cleaves be- by GFP to ensure cotranscriptional expression. These constructs were stably expressed in HLA-A2-expressing tween the first and second amino acid, as the peptide
bond between the N-terminal D and the following L Mel JuSo cells and sorted for equal GFP expression
64
Peptidase Activity in Living Cells
Figure 3. Substrate Length Specificity of the Collective Peptidase Activities
(A) C-terminal extensions. The 9-mer reporter peptide sequence T[K-Dabcyl]NKTER[C-Flu- orescein]Y containing the quencher-fluoro- phore combination was C-terminally ex- tended with 3, 6, 11, or 18 amino acids. The extension is a repeat of the different amino acids as tested in Figure 2. The peptides were introduced in living cells, and the half-life was determined. Half-life (SEM) is depicted.
(B) N-terminal extensions. The reporter pep- tide Y[C-Fluorescein]RETKN[K-Dabcyl]T has the reverse sequence as the reporter used for the C-terminal extensions. Also, the se- quence of the N-terminal extensions is mir- rored. The reporter peptide was extended with 1, 3, 6, 11, or 18 amino acids, and the half-life was determined in cells (SEM).
(C) The effect of N- or C-terminal extensions on HLA-A2-restricted antigen presentation of the influenza NP-epitope. Cells expressing both HLA-A2 and the influenza M-epitope, the epitope with one or three C-terminal amino acid extensions (C 1 or C 3), or the epitope with N-terminal extensions varying from 1 to 15 additional amino acids (N 1 to N 15) (see Table 1). Equal expression of the minigen was controlled through GFP ex- pression from the same transcript. Activation of the influenza M epitope-specific, HLA-A2- restricted T cell clone was measured by IFN secretion after o/n culture at different ef- fector/target cell ratios. The experiments were performed in triplicate, and means (SEM) are indicated.
before determining the T cell response. The influenza Moreover, extending the N terminus with 15 amino acids epitope with a C-terminal extension of one or three resulted in a reduced presentation of the influenza epi- amino acids were not presented. Mel JuSo cells are tope (Figure 3C). Since the sequence of the shorter ex- apparently unable to generate the properly sized 9-mer tensions is repeated in the longest sequence, these data influenza epitope in the absence of carboxypeptidase indicate that N-terminal extensions do not necessarily activity (Reits et al., 2003). Contrary to what may be improve antigen presentation. The longest extension expected, extending the N terminus of the influenza generates a minigene product of 28 amino acids and epitope with one or five amino acids did not increase may be a substrate for the previously observed pepti- presentation of influenza M epitope to specific T cells. dase activity that is active on larger fragments and may generate, though less efficiently, this specific epitope.
To investigate the role of this peptidase in more detail, Table 1. Peptide Epitopes Expressed by Minigenes for CTL Assay peptides of 27 amino acids with the reporter sequence
Epitope (M) GLIGFVFTL at various positions were microinjected (Figure 4A). The
fluorophore-quencher box at the C terminus of a 27- Epitope (C 1) (M) GLIGFVFTL P
mer peptide was degraded relatively slowly as expected Epitope (C 3) (M) GLIGFVFTL PGL
Epitope (N 1) (M) D GLIGFVFTL for N-terminal-extended peptides. When the reporter
Epitope (N 3) (M) FKD GLIGFVFTL was placed at the N terminus, fluorescence appeared
Epitope (N 5) (M) GPFKD GLIGFVFTL almost immediately due to rapid separation of the
Epitope (N 15) (M) FKDLGPFKDLGPFKD GLIGFVFTL
quencher and the fluorescein group. However, when
Figure 4. Characterizing Peptidase Activity for Longer Peptides
(A) Positional variations. The half-life of the 27-mer peptide shown in Figures 3A and 3B was compared with a 27-mer containing the reporter sequence in the middle. The half- life of the peptides was determined in living cells (SEM).
(B) The effect of N-terminal modifications on degradation of the 27-mer peptide. Amino acids one or one and two were exchanged for D amino acids (shown in bold and under- lined) in the 27-mer model peptide with the reporter in the middle (as in Figure 3D). Alter- natively, the free N terminus of this peptide was blocked with naftylenesulfone (indicated by an asterisk). The peptides were introduced in living Mel JuSo cells and degradation mea- sured. The half-life (SEM) is depicted. No degradation of the N-terminally blocked pep- tide was observed within a period of 200 s.
the same fluorescent-quencher reporter sequence was tabindide is a water-soluble and highly specific revers- ible and competitive inhibitor for TPPII (Breslin et al., placed in the center of the 27-mer peptide, degradation
was as fast as when the reporter sequence was placed 2002; Ganellin et al., 2000; Renn et al., 1998; Rose et al., 1996) and is a small indole-based structure with at the N terminus (Figure 4A). This is unlikely due to
gradual N-terminal trimming, since peptides with only some inherent instability in solution (Breslin et al., 2002).
We tested the effect of butabindide on the degradation three to six amino acids N-terminally of the reporter
sequence are degraded considerably slower than this of a 20-mer internally quenched peptide microinjected into cells (Figure 5A). The addition of 104M butabindide peptide (Figure 3B). Surprisingly, a corresponding 18-
mer peptide with the reporter sequence at the C termi- to the medium was sufficient to completely inhibit the degradation of the 20-mer peptide in living cells. Since nus was degraded almost as fast as the 27-mer with
the centered reporter sequence, suggesting that they cells were always microinjected under serum-free con- ditions, the effect of serum on the stability of butabindide are not substrates for trimming amino peptidases but
that the 18- and 27-mer are possibly targeted by an was tested. Even small amounts of fetal calf serum pre- vented the inhibitory effect of butabindide (Figure 5A).
endopeptidase that cleaves these peptides into large
pieces, thereby separating the quencher and fluoro- To our knowledge, butabindide has not been success- fully used in in vivo experiments before to inhibit peptide phore. To examine the requirements for an unmodified
N terminus, the 27-mer peptide with the reporter se- degradation by TPPII, which may be explained by the apparent inhibitory effect of serum on butabindide treat- quence in the middle was modified by either placing one
or two D amino acids at the N terminus or by blocking the ment. To test the stability of butabindide in our experi- mental system, the 20-mer internally quenched peptide N terminus (Figure 4B). Replacing the first two amino
acids by D amino acids did not influence the half-life of was microinjected in cells (in serum-free medium) at various times after addition of the inhibitor at 104M.
the reporter sequence. However, blocking the free N
terminus prevented degradation of the 27-mer peptide Complete inhibition of peptide degradation was ob- served until1.5 hr after butabindide addition, after in vivo. These data suggest that long peptides are han-
dled by distinct peptidases that can remove larger which peptide degradation was recovering (data not shown).
pieces but require a free N terminus.
We subsequently measured the degradation rates of peptides, varying in size between 9 and 27 amino acids, Long Peptides Are Selectively Degraded by TPPII
Since proteasomal activity may be involved in degrading by microinjecting these in cells cultured in the presence or absence of butabindide. The rate of degradation for large peptide fragments, we measured peptide degrada-
tion in the presence of proteasome inhibitor, but no peptides up to 15 amino acids was not significantly affected by inhibition of TPPII, although the N-terminally change in peptide half-life was observed (data not
shown). Another candidate might be the large peptidase extended 15-mer peptide was degraded less efficiently under these conditions. Presentation of the 14-mer influ- complex TPPII, which has reported exo- and endopepti-
dase activities (Geier et al., 1999). To test whether TPPII enza epitope expressed from minigenes was not af- fected when cells were first stripped to dissociate most was the peptidase specialized in the degradation of the
long peptides, the inhibitor butabindide was used. Bu- surface class I molecules by acid wash and cultured for
66
Peptidase Activity in Living Cells
Figure 5. TPPII and the Generation of Peptides by TPPII in Living Cells
(A) The effect of serum on the stability of the TPPII inhibitor butabindide. Cells were cultured in the presence or absence of FCS while incubated with 100 M butabindide. Subsequently, the 20-mer peptide T[K-Dabcyl]NKTERR[C-Fluorescein]YPGLDKFPGLDK was microinjected and degradation was measured.
(B) The effect of butabindide on the presentation of an expressed 14-mer peptide. HLA-A2-expressing Mel JuSo cells were transfected with a 14-mer peptide (N 5; see Figure 3C). Cells were acid stripped followed by culture for 16 hr in the presence or absence of butabindide (refreshed every 60 min). Activation of the influenza M epitope-specific, HLA-A2-restricted T cell clone was measured by IFN secretion after o/n culture at E/T ratio 1:1. The experiments were performed in 6-fold, and means (SEM, corrected for background) are indicated.
(C) The effect of inhibition of TPPII on peptide degradation in living Mel JuSo cells. Peptides of different length with the reporter sequence at different positions (as indicated, for sequence see Figure 3) were microinjected in cells cultured under serum-free conditions and in the absence or presence of 100 M butabindide. Degradation of the peptides was determined by fluorescence emission at 540 nm and the half- life of the peptides (SEM) is depicted. D.N.O., degradation not observed within the time of analysis (200 s).
16 hr in the absence or presence of butabindide followed dependent on TPPII activity (Figure 5C) irrespective of the position of the reporter sequence, since no degrada- by the CTL assay (Figure 5B). This indicates that buta-
bindide treatment did not inhibit synthesis, assembly, tion was observed (D.N.O.) in the presence of butabin- dide. Apparently, peptides longer than 15 amino acids and transport of MHC class I molecules, nor presenta-
tion of small peptides. However, the degradation of pep- can only be efficiently degraded by TPPII in vivo.
To test whether butabindide affected proteasome tides larger than 15 amino acids appeared to be critically
Figure 6. Specificity of TPPII Inhibition (A) The effect of butabindide on proteasome activity. Cells were incubated for 2 hr with either proteasome inhibitor or butabindide, and accumulation of cyclin D1 was deter- mined by Western blotting. DR chain was probed as loading control.
(B) TPPII-specific siRNA inhibit degradation of peptides longer than 15 amino acids. Mel JuSo cells were transfected for 72 hr with TPPII-specific siRNA from pSUPER coex- pressing GFP. Control and GFP-expressing cells were microinjected with various reporter peptides in the presence of competing pep- tides. The upper panel shows representative degradation curves in control cells; the lower panel shows representative degradation curves in cells expressing TPPII siRNA. The 16-mer peptide has a C-terminal reporter, resulting in a slow degradation when compared to all other peptides with N-terminal reporters.
degradation under these conditions, we cultured Mel This was determined after titration of “cold” 27-mer pep- tides (containing a quencher but no fluorophore) with JuSo cells in the absence or presence of the proteasome
inhibitor MG132 or butabindide for 2 hr in serum-free constant amounts of quenched peptides to saturate TPPII activity. At three times molar excess of cold 27- medium. Accumulation of the proteasome substrate
cyclin D1 was detected only after inhibiting the protea- mer peptide, an 8-fold reduction in rate of degradation of the quenched 27-mer peptide was observed. This some (Agami and Bernards, 2000) and not when cells
were cultured in the presence of butabindide (Figure concentration of competing cold peptides was coin- jected with the internally quenched peptides in control 6A). To confirm that the effects of butabindide were due
to inhibition of TPPII, the peptide degradation experi- and siRNA expressing cells. The 9-mer reporter peptide was rapidly degraded independent of TPPII saturation ments were repeated after transient downregulation of
TPPII using a vector containing GFP and a siRNA con- or additional inhibition by siRNA (Figure 6B). Due to competition with the cold 27-mer, the 16-mer, 20-mer, struct directed against TPPII (Seifert et al., 2003). Two
days after transfection, cells showed a downregulation and 25-mer reporter peptides were slower degraded when compared to Figure 5C, but hardly any degrada- of TPPII expression up to sixty percent (data not shown).
Since TPPII is not fully downregulated, we measured tion was observed when microinjected cells expressed siRNA. Although slower, the 15-mer reporter peptide peptide degradation under saturated conditions in vivo.
68
Peptidase Activity in Living Cells
could apparently still be degraded by other peptidases peptides generated by the proteasome are lost before encountering the peptide transporter TAP (Fruci et al., in siRNA-expressing cells (Figure 6B). Similar to the re-
2003; Princiotta et al., 2003; Reits et al., 2003). These sults obtained with butabindide treatment, specific
peptides are degraded within seconds by cytoplasmic downregulation of TPPII results in the inhibition of deg-
peptidases and most fail to associate with TAP within radation of peptides longer than 15 amino acids.
this time window (Reits et al., 2003). A number of cyto- plasmic peptidases have been characterized that may TPPII and the Generation of MHC Class I Peptides
contribute to peptide processing for MHC class I presen- TPPII may be specialized in degrading large peptide
tation. For example, LAP is a peptidase that is upregu- fragments but at the same time it may also destroy and
lated by IFN and able to degrade peptides destined create antigenic peptides. To examine the importance
for antigen presentation (Beninga et al., 1998; Reits et of TPPII in the proteolytic pathway resulting in class I
al., 2003). TOP is another peptidase known to affect the peptides, we first assayed the cleavage pattern of TPPII.
class I peptide pool (Saric et al., 2001) and MHC class Therefore, a series of 16-mer peptides was generated
I molecules are upregulated when TOP is inactivated by with the reporter sequence at the C terminus and with
siRNA (York et al., 2003). TPPII is reported to be involved an increasing number of D amino acids at the N terminus
in the generation of a HIV epitope (Seifert et al., 2003) (Figure 7A). Degradation of these peptides was inhibited
and may generate C termini of epitopes that are not by butabindide, implying that the peptides were handled
made by the proteasome. It has been suggested that by TPPII (data not shown). Proteases can only cleave
TPPII can compensate for some activities of the protea- the peptide bond between natural L amino acids and
some (Glas et al., 1998) but how, if at all (Princiotta et not between D and L or between D amino acids. Replac-
al., 2001), is unclear. The specificity of the cytoplasmic ing the first one or two N-terminal amino acids reduced
peptidases may be an important determinant in the out- the rate of degradation of the reporter segment, proba-
come of MHC class I antigen presentation if particular bly by affecting the exopeptidase activity of TPPII that
peptide sequences would be more resistant. Although can remove one to three amino acid parts from the N
the substrate specificity of some peptidases is known, terminus (Tomkinson, 1999). However, no further de-
it is unknown whether the combined in vivo activities and crease in the rate of degradation was observed when
enzyme concentration destroys different cytoplasmic more L amino acids were replaced by D amino acids.
peptides equally.
Apparently, TPPII has two activities: one activity that
Here, we show that variation of amino acids at the removes small fragments from the N terminus and a
first, second, third, or last position does not result in second but slower activity that generates larger frag-
major differences in the half-life of introduced peptides.
ments. Since a fragment containing eight D amino acids
Although the quencher and/or fluorophore might affect has to be cleaved after the following L amino acid, TPPII
the ability of some peptidases to hydrolyse the reporter is apparently able to generate 9-mer fragments (or
peptide, our data suggests that the heterogeneous pool longer).
of cytoplasmic peptidases destroys small peptides rap- To determine the relevance of TPPII in peptide genera-
idly but fairly irrespective of amino acid sequence. Since tion for MHC class I molecules, both Mel JuSo cells and
peptides are degraded exclusively by aminopeptidases freshly isolated peripheral blood lymphocytes (PBLs)
(Reits et al., 2003), additional N-terminal amino acids were stripped to dissociate most surface class I mole-
should increase the half-life of an antigen, as has been cules by acid wash. The cells were subsequently cul-
suggested previously (Cascio et al., 2001). The half-life tured in the presence or absence of the proteasome
of the 9-mer reporter sequence increases from 4 to 30 s inhibitor lactacystin, butabindide, or a combination of
when three additional amino acids are present at the N the two inhibitors, all under serum-free conditions. Bu-
terminus. Apparently, every additional amino acid in- tabindide was readded every hour. After 4 hr, MHC class creases the reporter’s half-life with around 6 s until the I molecules at the cell surface were labeled with the peptide becomes longer than 15 amino acids. Still, when antibody W6/32, followed by FACS analysis (Figure 7B). minigene constructs with N-terminal extensions are ex- Surprisingly, butabindide treatment resulted in a similar pressed, no increased level of influenza M-epitope pre- reduction in cell surface MHC class I as treatment with sentation is observed. In fact, a long N-terminal exten- proteasome inhibitor, while the combination of the two sion (15 additional amino acids) results in decreased inhibitors only had a marginal additive effect. Identical levels of antigen presentation. This may be surprising, effects were obtained with the less specific TPPII inhibi- since the short N-terminal sequences were included in tor AAF-CMK (data not shown). Given the substrate the long N-terminal sequence, but it suggests that the specificity of TPPII, our data suggests that a major por- antigenic sequence within the long peptide is partially tion of proteasomal products consists of long peptides degraded by an endopeptidase. Extending the total (over 15 amino acids) that are subsequently targeted by length of the reporter peptide beyond 15 amino acids TPPII. Since these peptides will be too long for MHC does not reduce the rate of degradation any further, class I loading, TPPII will act as a crucial intermediate which suggests that these peptides were handled differ- between proteasomes, cytoplasmic or ER peptidases, ently from shorter peptides by the cytoplasmic amino- and MHC class I molecules. peptidases. Extending the C terminus of the NP-epitope prevents presentation by MHC class I, which confirms
Discussion the observed absence of cytoplasmic carboxypeptidase
activities (Reits et al., 2003).
The efficiency of antigen presentation by MHC class I The enzyme involved in long peptide degradation should display aminopeptidase activity, since blocking molecules is amazingly low, since more than 99% of the
Figure 7. Major Role for TPPII in Antigen Presentation
(A) TPPII and the generation of 9-mer peptides. A 16-mer peptide was synthesized with the reporter sequence Y[C-Fluorescein]RETKN [K-Dabcyl]T at the C terminus. Either one, two, five, or eight N-terminal amino acids were exchanged for their corresponding D stereoisomers that cannot be handled by natural proteases. The peptides were introduced in living cells, and degradation of the reporter sequence was measured.
(B) TPPII and the generation of peptides for MHC class I molecules. Mel JuSo cells (left) and human PBLs (right) were acid stripped followed by culture in the presence or absence of lactacystin, butabindide, or a combination of the two inhibitors. Butabindide was refreshed every 60 min. After 4 hr, the surface expression of MHC class I molecules was determined by FACS analysis. The inhibitors used are indicated in the figure.
(C) Model of the first proteolytic steps in MHC class I antigen presentation. The proteasome releases mainly products of 16 amino acids or longer. These long peptides are handled almost exclusively by TPPII. TPPII can destroy but also generate antigenic peptides for TAP by either removing small packages of amino acids or making large fragments, thereby creating the C terminus. Smaller peptides can be produced directly by the proteasome and after further trimming by TPPII. Other peptidases (represented as X, Y, and Z) that include TOP and LAP target these substrates. Antigenic peptides translocated by TAP can be generated exclusively by the proteasome, by the proteasome and TPPII, and by the proteasome, often TPPII, and other peptidases. TPPII appears to be an important intermediate between the proteasome and the rest of the peptidase pool involved in trimming fragments to antigenic peptides and mainly free amino acids.
70
Peptidase Activity in Living Cells
the N terminus also protects the long peptides. Thimet MHC class I molecules after an acidic wash in the pres- oligopeptidase has been shown in vitro to select peptide ence of proteasome inhibitors or butabindide. Surface substrates between 8 and 16 amino acids in length (Saric expression of MHC class I was equally affected by both et al., 2001), but it is unclear whether it can degrade inhibitors, and hardly any additive effect of the combina- longer peptides as well. TPPII is a protease that is even tion of the two inhibitors was observed. Peptide supply larger than the proteasome (5–9 MDa) and is able to for MHC class I molecules but not peptide presentation generate antigenic fragments by endopeptidase activity is affected by butabindide treatment, suggesting that (Geier et al., 1999). Although TPPII can degrade small TPPII is critical for MHC class I peptide generation. In trimer substrates in vitro (Geier et al., 1999), it is unclear vitro data suggest that the proteasome generate an whether this reflects the activities of the complex in vivo. array of peptides between three and twenty-two amino Regulating subunits may not be copurified with the TPPII acids with two-thirds too short to function in antigen core, similar to the situation with the proteasome where processing (Kisselev et al., 1999). However, it is likely usually the 20S core is purified without the 19S subunits. that peptides reenter the proteasome multiple times dur- TPPII is upregulated when various activities of the pro- ing in vitro digestion, possibly resulting in shorter prod- teasome are inhibited (Geier et al., 1999; Princiotta et ucts than generated in vivo. Peptides are no substrate al., 2001), which can be envisaged when under these for the proteasome in living cells (Reits et al., 2003), and conditions the proteasome is generating larger break- reentry into the proteasome will not be a relevant route in down intermediates requiring further fragmentation by vivo. Our data imply that a major portion of proteasomal TPPII. Alternatively, TPPII could replace the proteasome products is larger than 15 amino acids and require fur- but should then be able to unfold and deubiquitinate ther degradation by TPPII before becoming relevant for substrates. Whether this is possible is unknown. TPPII MHC class I. It has been suggested that the proteasome is at least unable to degrade the proteasomal substrate usually generates the proper C terminus of class I bind- cyclin D1. When we repeated the peptide degradation ing peptides (Cascio et al., 2001) with aminopeptidases experiments in the presence of the specific TPPII inhibi- only trimming from the N-terminal side. Our data pro- tor butabindide or after downregulation of TPPII by pose a role for TPPII in this process since TPPII can siRNA, we observed that TPPII was selectively involved generate at least 9-mer (antigenic) fragments containing in the degradation of peptides of over 15 amino acids, a new C terminus. TPPII may thus generate but also whereas the degradation of smaller peptides was not destroy antigenic peptides, which may explain why a affected. Substrates with the reporter located nine long minigene-expressed peptide is less efficiently pre- amino acids from the N terminus were very efficiently sented than minigene-expressed peptides shorter than degraded, suggesting that TPPII cleaves somewhere 15 amino acids.
between the ninth and eleventh residue. Experiments By visualizing peptidase activity in living cells, we have with D amino acid-containing substrates suggest that defined TPPII as a critical player in antigen processing TPPII can remove blocks of one to three amino acids and presentation by MHC class I molecules. Our data for the N terminus but also fragments of nine amino suggests the following sequence of events to occur for a acids or more (albeit less efficient). This corresponds successful class I response (Figure 7C). The proteasome to the reported in vitro activities of TPPII, where TPPII digests proteins mainly in fragments larger than 15 requires a free N terminus and preferably removes frag- amino acids. TPPII trims these peptides into smaller ments of two to three amino acids (Tomkinson, 1999). fragments that may include proper class I binding pep- It is unclear whether there is a maximal size for the tides. A variety of peptidases (including TOP, LAP, and peptide substrates for TPPII, but it may cleave at various ERAAP) then continues the digestion into even smaller positions within a substrate, since a reporter sequence products. These three proteolytic steps act in a process in the middle of a 27-mer peptide is degraded as rapidly degrading proteins down to small fragments and amino as an N-terminally positioned reporter. The observed
acids, occasionally generating a class I binding peptides differences in half-life of C-terminally positioned report-
instead of destroying it. In every step, peptides may ers in peptides longer than 16 amino acids may be the
escape further trimming by cytoplasmic peptidase activ- result of cleavage by TPPII, resulting in differently sized
ity by being transported into the ER by TAP for consider- fragments containing the reporter sequence that need
ation by ERAAP and MHC class I. Most peptides will subsequent degradation by aminopeptidases. How
however be degraded into single amino acids.
TPPII decides to remove small or long fragments is un- clear. One possibility is that additional complexes regu-
Experimental Procedures late the activities, but whether TPPII has regulatory part-
ners determining recognition and degradation and Synthetic Peptides
selecting long peptides is unclear. Still, TPPII may play The various fluorescent peptides were synthesized by solid-phase an important role in the degradation process of proteins strategies using an automated multiple peptide synthesizer (Syro II, MultiSyntech, Witten, Germany) using Fmoc chemistry. Fluorescein back to amino acids. The proteasome is specialized
was covalently coupled to the cysteine residue using fluorescein- in the digestion of ubiquitinated proteins that require
5-iodoacetamide (Molecular Probes, Leiden, The Netherlands).
unfolding (Glickman and Ciechanover, 2002), whereas
Fmoc-L-Lys(Dabcyl)-OH was obtained from Neosystems (France).
TPPII targets large peptide substrates and possibly un-
All peptides were HPLC purified (95% pure) and validated by folded proteins. TOP and other peptidases may then mass spectrometry.
degrade peptides that are too small for TPPII.
We tested the role of TPPII both in peptide generation Peptide Degradation Analysis
in vivo but also in peptide generation for MHC class I Cells on coverslips were placed on an inverted Zeiss Axiovert 135 microscope equipped with a dry Achroplan 63 (NA 0.75) objective.
molecules by following cell surface reappearance of
Peptides were quantified by their red appearance (Dabcyl) using Cancer Society (NKB 2001-2416, NKB 2003-2861) and the Nether- lands Organization for Scientific Research.
photospectrometry and mixed with Fura Red (Molecular Probes, Leiden, The Netherlands) as a microinjection marker. Excitation of
both fluorescein and Fura Red was at 475 nm. Emission of fluores- Received: September 22, 2003 cein and Fura Red was measured after splitting the emitted light Revised: March 11, 2004 using a 580 nm dichroic mirror and simultaneously detected with Accepted: March 14, 2004 PTI model 612 analog photomultipliers. For data acquisition, FELIX Published online: March 25, 2004 software (PTI Inc., USA) was used. Fura Red was used as a control
for microinjection and cell leakage. The relative fluorescence of References fluorescein was expressed as the ratio of fluorescein to Fura Red
signal, as described (Reits et al., 2003). To inhibit proteasomal activ- Agami, R., and Bernards, R. (2000). Distinct initiation and mainte- ity, cells were incubated in the presence of 10 M lactacystin for nance mechanisms cooperate to induce G1 cell cycle arrest in re- 30 min at 37C. To inhibit TPPII activity, cells were cultured at 37C sponse to DNA damage. Cell 102, 55–66.
in serum-free HEPES-buffered Iscoves medium in the presence of Beninga, J., Rock, K.L., and Goldberg, A.L. (1998). Interferon-gamma 100 M butabindide-oxalate (Tocris, Avonmouth, UK) dissolved in can stimulate post-proteasomal trimming of the N terminus of an 10 mM Tris-HCl (pH 8.0) prior to immediate use. Microinjection of antigenic peptide by inducing leucine aminopeptidase. J. Biol.
the internally quenched peptides was performed within 1.5 hr after Chem. 273, 18734–18742.
addition of butabindide. Peptide half-life was determined in at least
Breslin, H.J., Miskowski, T.A., Kukla, M.J., Leister, W.H., De Winter, three independent experiments (generally at least five successful
H.L., Gauthier, D.A., Somers, M.V., Peeters, D.C., and Roevens, P.W.
microinjections per experiment) and depicted as mean value SEM.
(2002). Design, synthesis, and tripeptidyl peptidase II inhibitory ac- The siRNA-targeting TPPII (5-AAGTGGCGATGTGAATACTGCT-3)
tivity of a novel series of (S)-2,3-dihydro-2-(4-alkyl-1H-imidazol-2- (Seifert et al., 2003) was cloned into the vector pSUPER (Brummel-
yl)-1H-indoles. J. Med. Chem. 45, 5303–5310.
kamp et al., 2002) containing GFP (kindly provided by R. Schotte
Bromme, D., Rossi, A.B., Smeekens, S.P., Anderson, D.C., and and H. Spits). Cells were transiently transfected with the siRNA
Payan, D.G. (1996). Human bleomycin hydrolase: molecular cloning, construct, and 3 days upon transfection cells expressing GFP were
sequencing, functional expression, and enzymatic characterization.
microinjected with a mixture of quenched peptides of different
Biochemistry 35, 6706–6714.
lengths in combination with three times the amount of nonfluores-
cent 27-mer peptides (to measure TPPII activity under saturated cir- Brummelkamp, T.R., Bernards, R., and Agami, R. (2002). A system
cumstances). for stable expression of short interfering RNAs in mammalian cells.
Science 296, 550–553.
FACS Analysis Cascio, P., Hilton, C., Kisselev, A.F., Rock, K.L., and Goldberg, A.L.
Peptides were stripped of surface MHC class I molecules on cells (2001). 26S proteasomes and immunoproteasomes produce mainly by mild acid treatment as described (Sugawara et al., 1987). In brief, N-extended versions of an antigenic peptide. EMBO J. 20, 2357–
cells were incubated on ice for 10 min, washed twice with cold 2366.
medium (Iscove’s medium without phenol red), incubated for 2 min Fruci, D., Lauvau, G., Saveanu, L., Amicosante, M., Butler, R.H., with ice-cold citric acid-Na2PO4buffer, followed by two washes with Polack, A., Ginhoux, F., Lemonnier, F., Firat, H., and van Endert, ice-cold Iscoves medium. Subsequently, cells were cultured at 37C P.M. (2003). Quantifying recruitment of cytosolic peptides for HLA in medium alone or in the presence of 100 M butabindide, 10 M class I presentation: impact of TAP transport. J. Immunol. 170, 2977–
lactacystin, or a combination of both inhibitors. Butabindide was 2984.
replaced every hour, and after 4 hr cells were harvested, put on ice,
Ganellin, C.R., Bishop, P.B., Bambal, R.B., Chan, S.M., Law, J.K., and stained with FITC-coupled W6/32. Fluorescence was measured
Marabout, B., Luthra, P.M., Moore, A.N., Peschard, O., Bourgeat, by flow cytometry (Becton Dickinson).
P., et al. (2000). Inhibitors of tripeptidyl peptidase II. 2. Generation of the first novel lead inhibitor of cholecystokinin-8-inactivating pep-
Western Blotting tidase: a strategy for the design of peptidase inhibitors. J. Med.
Mel JuSo cells were cultured in the absence or presence of 10 mM
Chem. 43, 664–674.
MG132 or 100 M butabindide-oxalate for 2 hr before analysis of
Geier, E., Pfeifer, G., Wilm, M., Lucchiari-Hartz, M., Baumeister, W., cell lysates by 10% SDS-PAGE and Western blotting. MHC class II
Eichmann, K., and Niedermann, G. (1999). A giant protease with molecules were detected by the mAb 1B5 and cyclin D1 by the Ab
potential to substitute for some functions of the proteasome. Sci- M-20.
ence 283, 978–981.
Glas, R., Bogyo, M., McMaster, J.S., Gaczynska, M., and Ploegh, CTL Assay
H.L. (1998). A proteolytic system that compensates for loss of pro- Mel JuSo cells were stably transfected with HLA-A*0201 and
teasome function. Nature 392, 618–622.
pcDNA3 containing a minigene-IRES-GFP sequence. Cells were
Glickman, M.H., and Ciechanover, A. (2002). The ubiquitin-protea- FACS sorted for equal HLA-A2 and GFP expression. For the experi-
some proteolytic pathway: destruction for the sake of construction.
ment with butabindide, target cells were acidic stripped and cultured
Physiol. Rev. 82, 373–428.
for 16 hr in the presence or absence of butabindide. The human
T cell clone (Inf A 13 TGA) recognizing the Flu-M57–65peptide Johnson, G.D., and Hersh, L.B. (1990). Studies on the subsite speci- (GILGFVFTL) in the context of HLA-A*0201 was cultured o/n with ficity of the rat brain puromycin-sensitive aminopeptidase. Arch.
target cells at various E:T ratios in the presence of IL-2 (20 IU/ml). Biochem. Biophys. 276, 305–309.
The amount of IFN secreted by the T cells was detected in the Kisselev, A.F., Akopian, T.N., Woo, K.M., and Goldberg, A.L. (1999).
supernatant by Elisa according to the manufacturer’s protocol (San-
The sizes of peptides generated from protein by mammalian 26 and quin, Pelipair reagent set human IFN M9333). Controls for sponta- 20 S proteasomes. Implications for understanding the degradative neous IFN release and recognition of the specific peptide were in- mechanism and antigen presentation. J. Biol. Chem. 274, 3363–
cluded.
3371.
Kloetzel, P.M. (2001). Antigen processing by the proteasome. Nat.
Acknowledgments
Rev. Mol. Cell Biol. 2, 179–187.
Koopmann, J.O., Post, M., Neefjes, J.J., Hammerling, G.J., and We thank Marieke Bax, Diana Nijholt, and Rachid el Betar for experi-
Momburg, F. (1996). Translocation of long peptides by transporters mental support; K. Jalink for support with biophysical techniques;
associated with antigen processing (TAP). Eur. J. Immunol. 26, 1720–
C. Slicer (Tocris) for information about butabindide-oxalate; M. Ma- 1728.
succi (Karolinska, Sweden) for samples of butabindide; and R.
Schotte and H. Spits (AMC, The Netherlands) for the vector pSUPER- Koopmann, J.O., Albring, J., Huter, E., Bulbuc, N., Spee, P., Neefjes, J., Hammerling, G.J., and Momburg, F. (2000). Export of antigenic GFP. This research was supported by grants from The Netherlands
72
Peptidase Activity in Living Cells
peptides from the endoplasmic reticulum intersects with retrograde Sugawara, S., Abo, T., and Kumagai, K. (1987). A simple method to eliminate the antigenicity of surface class I MHC molecules from protein translocation through the Sec61p channel. Immunity 13,
117–127. the membrane of viable cells by acid treatment at pH 3. J. Immunol.
Methods 100, 83–90.
Mo, X.Y., Cascio, P., Lemerise, K., Goldberg, A.L., and Rock, K.
Toes, R.E., Nussbaum, A.K., Degermann, S., Schirle, M., Emmerich, (1999). Distinct proteolytic processes generate the C and N termini
N.P., Kraft, M., Laplace, C., Zwinderman, A., Dick, T.P., Muller, J., of MHC class I-binding peptides. J. Immunol. 163, 5851–5859.
et al. (2001). Discrete cleavage motifs of constitutive and immuno- Momburg, F., Roelse, J., Hammerling, G.J., and Neefjes, J.J. (1994a).
proteasomes revealed by quantitative analysis of cleavage prod- Peptide size selection by the major histocompatibility complex-
ucts. J. Exp. Med. 194, 1–12.
encoded peptide transporter. J. Exp. Med. 179, 1613–1623.
Tomkinson, B. (1999). Tripeptidyl peptidases: enzymes that count.
Momburg, F., Roelse, J., Howard, J.C., Butcher, G.W., Hammerling,
Trends Biochem. Sci. 24, 355–359.
G.J., and Neefjes, J.J. (1994b). Selectivity of MHC-encoded peptide
Turzynski, A., and Mentlein, R. (1990). Prolyl aminopeptidase from transporters from human, mouse and rat. Nature 367, 648–651.
rat brain and kidney. Action on peptides and identification as leucyl Princiotta, M.F., Schubert, U., Chen, W., Bennink, J.R., Myung, J.,
aminopeptidase. Eur. J. Biochem. 190, 509–515.
Crews, C.M., and Yewdell, J.W. (2001). Cells adapted to the protea-
Yellen-Shaw, A.J., Wherry, E.J., Dubois, G.C., and Eisenlohr, L.C.
some inhibitor 4-hydroxy-5-iodo-3-nitrophenylacetyl-Leu-Leu-leu-
(1997). Point mutation flanking a CTL epitope ablates in vitro and cinal-vinyl sulfone require enzymatically active proteasomes for
in vivo recognition of a full-length viral protein. J. Immunol. 158, continued survival. Proc. Natl. Acad. Sci. USA 98, 513–518.
3227–3234.
Princiotta, M.F., Finzi, D., Qian, S.B., Gibbs, J., Schuchmann, S.,
Yewdell, J.W., Reits, E., and Neefjes, J. (2003). Making sense of Buttgereit, F., Bennink, J.R., and Yewdell, J.W. (2003). Quantitating
mass destruction: quantitating MHC class I antigen presentation.
protein synthesis, degradation, and endogenous antigen pro-
Nat. Rev. Immunol. 3, 952–961.
cessing. Immunity 18, 343–354.
York, I.A., Chang, S.C., Saric, T., Keys, J.A., Favreau, J.M., Goldberg, Rammensee, H.G., Falk, K., and Rotzschke, O. (1993). Peptides
A.L., and Rock, K.L. (2002). The ER aminopeptidase ERAP1 en- naturally presented by MHC class I molecules. Annu. Rev. Immunol.
hances or limits antigen presentation by trimming epitopes to 8–9 11, 213–244.
residues. Nat. Immunol. 3, 1177–1184.
Reits, E.A., Vos, J.C., Gromme, M., and Neefjes, J. (2000). The major
York, I.A., Mo, A.X., Lemerise, K., Zeng, W., Shen, Y., Abraham, substrates for TAP in vivo are derived from newly synthesized pro-
C.R., Saric, T., Goldberg, A.L., and Rock, K.L. (2003). The cytosolic teins. Nature 404, 774–778.
endopeptidase, thimet oligopeptidase, destroys antigenic peptides Reits, E., Griekspoor, A., Neijssen, J., Groothuis, T., Jalink, K., van
and limits the extent of MHC class I antigen presentation. Immunity Veelen, P., Janssen, H., Calafat, J., Drijfhout, J.W., and Neefjes, J.
18, 429–440.
(2003). Peptide diffusion, protection, and degradation in nuclear and cytoplasmic compartments before antigen presentation by MHC class I. Immunity 18, 97–108.
Renn, S.C., Tomkinson, B., and Taghert, P.H. (1998). Characteriza- tion and cloning of tripeptidyl peptidase II from the fruit fly, Drosoph- ila melanogaster. J. Biol. Chem. 273, 19173–19182.
Rock, K.L., Gramm, C., Rothstein, L., Clark, K., Stein, R., Dick, L., Hwang, D., and Goldberg, A.L. (1994). Inhibitors of the proteasome block the degradation of most cell proteins and the generation of peptides presented on MHC class I molecules. Cell 78, 761–771.
Roelse, J., Gromme, M., Momburg, F., Hammerling, G., and Neefjes, J. (1994). Trimming of TAP-translocated peptides in the endoplasmic reticulum and in the cytosol during recycling. J. Exp. Med. 180, 1591–
1597.
Rose, C., Vargas, F., Facchinetti, P., Bourgeat, P., Bambal, R.B., Bishop, P.B., Chan, S.M., Moore, A.N., Ganellin, C.R., and Schwartz, J.C. (1996). Characterization and inhibition of a cholecystokinin- inactivating serine peptidase. Nature 380, 403–409.
Saric, T., Beninga, J., Graef, C.I., Akopian, T.N., Rock, K.L., and Goldberg, A.L. (2001). Major histocompatibility complex class I-pre- sented antigenic peptides are degraded in cytosolic extracts primar- ily by thimet oligopeptidase. J. Biol. Chem. 276, 36474–36481.
Saric, T., Chang, S.C., Hattori, A., York, I.A., Markant, S., Rock, K.L., Tsujimoto, M., and Goldberg, A.L. (2002). An IFN-gamma-induced aminopeptidase in the ER, ERAP1, trims precursors to MHC class I-presented peptides. Nat. Immunol. 3, 1169–1176.
Schubert, U., Anton, L.C., Gibbs, J., Norbury, C.C., Yewdell, J.W., and Bennink, J.R. (2000). Rapid degradation of a large fraction of newly synthesized proteins by proteasomes. Nature 404, 770–774.
Seifert, U., Maranon, C., Shmueli, A., Desoutter, J.F., Wesoloski, L., Janek, K., Henklein, P., Diescher, S., Andrieu, M., de la Salle, H., et al. (2003). An essential role for tripeptidyl peptidase in the generation of an MHC class I epitope. Nat. Immunol. 4, 375–379.
Serwold, T., Gonzalez, F., Kim, J., Jacob, R., and Shastri, N. (2002).
ERAAP customizes peptides for MHC class I molecules in the endo- plasmic reticulum. Nature 419, 480–483.
Stoltze, L., Schirle, M., Schwarz, G., Schroter, C., Thompson, M.W., Hersh, L.B., Kalbacher, H., Stevanovic, S., Rammensee, H.G., and Schild, H. (2000). Two new proteases in the MHC class I processing pathway. Nat. Immunol. 1, 413–418.
