Báo cáo khoa học: KCTD5, a putative substrate adaptor for cullin3 ubiquitin ligases docx

As it has been recently reported that deletion of the C-terminus of KCTD5 [16] changes its location to the nucleus, cells were transfected with different deletion mutants of KCTD5.. B Ex

ubiquitin ligases

Yolanda Bayo´n1, Antonio G Trinidad1, Marı´a L de la Puerta1, Marı´a del Carmen Rodrı´guez1, Jori Bogetz2, Ana Rojas3, Jose´ M De Pereda4, Souad Rahmouni5, Scott Williams2,

Shu-ichi Matsuzawa6, John C Reed6, Mariano Sa´nchez Crespo1, Tomas Mustelin2and

Andre´s Alonso1

1 Instituto de Biologı´a y Gene´tica Molecular, CSIC-Universidad de Valladolid, Spain

2 Program of Inflammation, Inflammatory and Infectious Disease Center, and Program of Signal Transduction, Burnham Institute for Medical Research, La Jolla, CA, USA

3 Structural Bioinformatics Group, Centro Nacional de Investigaciones Oncolo´gicas, Madrid, Spain

4 Centro de Investigacio´n del Ca´ncer, CSIC-Universidad de Salamanca, Spain

5 Department of Pathology B-35, University of Lie`ge, CHU of Lie`ge, Belgium

6 Program of Apoptosis and Cell Death, Burnham Institute for Medical Research, La Jolla, CA, USA

The BTB (bric-a-brac, tramtrak and broad

com-plex)⁄ POZ (poxvirus zinc finger) domain is a protein–

protein interaction domain first described in several

proteins of Drosophila melanogaster and poxvirus [1,2]

BTB⁄ POZ domain-containing proteins constitute a

diverse group of proteins involved in transcriptional

repression, cytoskeletal regulation, and ion channel

function [3] More recently, some BTB proteins have

been characterized as substrate-specific adaptors for

cullin(CUL)3-based E3 ligases [4–7] The BTB domain

of these substrate-specific adaptors binds to CUL3, whereas additional domains in these polypeptides, such

as zinc fingers, meprin and traf homology (MATH) domain, and Kelch repeats, work as substrate recogni-tion domains The first protein shown to be regulated

by a CUL3 ligase was MEI-1 in Caenorhaditis elegans This protein is part of the katanin-like microtubule severing complex [5,6] and is recruited to CUL3 by the

Keywords

BTB; cullin; E3 ligases; KCTD; ubiquitin

Correspondence

A Alonso, Instituto de Biologı´a y Gene´tica

Molecular, CSIC-Universidad de Valladolid,

c ⁄ Sanz y Fore´s s ⁄ n, 47003 Valladolid, Spain

Fax: +34 983 184800

Tel: +34 983 184839

E-mail: andres@ibgm.uva.es

(Received 7 April 2008, revised 30 May

2008, accepted 3 June 2008)

doi:10.1111/j.1742-4658.2008.06537.x

Potassium channel tetramerization domain (KCTD) proteins contain a bric-a-brac, tramtrak and broad complex (BTB) domain that is most simi-lar to the tetramerization domain (T1) of voltage-gated potassium chan-nels Some BTB-domain-containing proteins have been shown recently to participate as substrate-specific adaptors in multimeric cullin E3 ligase reac-tions by recruiting proteins for ubiquitination and subsequent degradation

by the proteasome Twenty-two KCTD proteins have been found in the human genome, but their functions are largely unknown In this study, we have characterized KCTD5, a new KCTD protein found in the cytosol of cultured cell lines The expression of KCTD5 was upregulated post-trans-criptionally in peripheral blood lymphocytes stimulated through the T-cell receptor KCTD5 interacted specifically with cullin3, bound ubiquitinated proteins, and formed oligomers through its BTB domain Analysis of the interaction with cullin3 showed that, in addition to the BTB domain, some amino acids in the N-terminus of KCTD5 are required for binding to cullin3 These findings suggest that KCTD5 is a substrate-specific adaptor for cullin3-based E3 ligases

Abbreviations

AU, arbitrary unit; BTB, bric-a-brac, tramtrak and broad complex; CT, cycle threshold; CUL, cullin; GFP, green fluorescent protein; GST, glutathione S-transferase; HA, hemagglutinin; IL-2, interleukin-2; KCTD, potassium channel tetramerization domain; MATH, meprin and traf homology; PBL, peripheral blood lymphocyte; PHA, phytohemagglutinin; PMA, 4b-phorbol 12-myristate 13-acetate; POZ, poxvirus zinc finger;

Ub, ubiquitin.

BTB protein MEL-26 In mammalian cells, a few other

BTB proteins, e.g SPOP, a BTB-MATH protein, and

KEAP1, a BTB-KELCH protein, have been described

as adaptors of CUL3-based E3 ligases [8] CUL3 is

one of the seven cullins found in the human genome

(CUL1, CUL2, CUL3, CUL4A, CUL4B, CUL5 and

CUL7), and most of them bind to adaptors through

their BTB domains, which, in turn, bind to additional

proteins that work as substrate-specific adaptors Thus,

in SKP1–CUL1–F-box, the archetypical cullin E3

ligase, CUL1 binds on the N-terminus to the adaptor

Skp1 that associates with an F-box protein working as

substrate-specific adaptor, and on the C-terminus to

the RING domain-containing protein Roc1⁄ Rbx ⁄ Hrt

[9] Cullin E3 ligases are multimeric RING E3 ligases

that participate in protein ubiquitination, a process

mediated by a three-step enzymatic cascade Ubiquitin

(Ub) is initially activated by the Ub-activating enzyme

(E1) and then transferred to a Ub-conjugating enzyme

(E2), which associates with a third protein, the Ub

ligase (E3), involved in recruiting the substrates for

ubiquitination and, therefore, providing specificity to

this process [10] Ubiquitination is involved in a wide

range of cellular functions, such as cell proliferation,

differentiation, and apoptosis, mainly by targeting

pro-teins for degradation by the 26S proteasome, but it is

also involved in protein transport and signaling

through additional mechanisms [10,11]

Although the human genome might include about

400 BTB proteins [8], only a few have been shown to

work as substrate-binding proteins for CUL3 E3 ligases

In this connection, potassium channel tetramerization

domain (KCTD) proteins form a group of proteins

containing a BTB domain, the function of which is

lar-gely unknown Herein, we report the characterization

of KCTD5, a new POZ⁄ BTB protein that is a putative

new substrate-specific adaptor for CUL3-based E3

ligases

Results and Discussion

KCTD5 was identified in a yeast two-hybrid screening

for the dual-specificity phosphatase VHR while

look-ing for adaptors that help us to understand how this

phosphatase targets its substrates, Erk and Jnk The

clone obtained in this assay contained a cDNA

sequence present in public databases with the Genbank

accession number NM_018992 Next, we tested VHR

interaction with KCTD5 in mammalian cells, and

could not find evidence for this Nevertheless, we

con-tinued the study of this new protein First, we studied

the expression of this gene, finding that its mRNA was

expressed in all the tissues and cell lines tested

(Fig 1A) On the contrary, protein expression was only observed in transformed cells and was absent from primary cells, such as peripheral blood leukocytes (PBLs), mouse brain cells or human brain cells (Fig 1B, lanes 6, 9 and 10), thus suggesting that its expression was upregulated post-transcriptionally Prompted by these results, especially by the differences observed between the expression of mRNA and pro-tein in PBLs, we hypothesized that KCTD5 might be induced by mitogens such as as phytohemagglutinin (PHA) and interleukin-2 (IL-2) in these cells Using these stimuli, we observed a 2.5-fold increase (Fig 1D, lane 5) in mRNA expression and an 84.7-fold increase

in protein expression at 48 h (Fig 1C, lane 7) in PBLs stimulated with PHA To investigate whether other stimuli known to induce T-cell proliferation increase KCTD5 protein, 4b-phorbol 12-myristate 13-acetate (PMA) plus ionomycin and a combination of antibod-ies for the T-cell and CD28 receptors, which mimic antigen stimulation, were used As shown in Fig 1E, these stimuli increased KCTD5 protein to an extent similar to that observed for PHA As RT-PCR assays lack enough sensitivity to detect changes in the amount

of mRNA, quantitative PCR assays were conducted in order to detect subtle changes in KCTD5 mRNA As shown in Fig 1F, there was only a slight decrease of KCTD5 mRNA after PHA stimulation of PBLs These data suggested that these stimuli regulated KCTD5 at

a post-transcriptional level, by increasing either the translation or the stability of KCTD5 protein In the latter case, this would imply that KCTD5 is an unsta-ble protein in the absence of stimuli In this regard, treatment of resting PBLs with MG132, a proteasome inhibitor, has no effect on KCTD5 protein (data not shown), meaning that stimulus-dependent translation is involved in increasing the quantity of KCTD5 protein

in PBLs Altogether, these data suggest that KCTD5 expression is mainly regulated by a post-transcriptional mechanism in PBLs, possibly at the translational level Several databases were searched to find homologs of KCTD5, using as query its BTB domain Although the BTB domain is present in proteins from all eukaryotic groups, when the query included the KCTD5 C-termi-nal region in addition to the BTB domain, homologs were only found among the metazoans However, no protein was found with a BTB domain followed by the C-terminus of KCTD5 in plants and fungi An align-ment of KCTD5 orthologs in several species is shown

in Fig 2A Among BTB proteins, KCTD5 is grouped with potassium channels The similarity with potas-sium channels is restricted to the T1 domain, which is

a BTB domain Whereas cullins are present in all eukaryotes, KCTD5-like proteins appeared later in

evolution in multicellular organisms, most likely to

ful-fil a new function, which is at the present time

unknown Searches for human paralogs, using as query

the BTB domain of KCTD5 to generate a phylogenetic

tree (Fig 2B), gave 22 sequences Some of these

human paralogs are found in highly similar groups

with conserved sequences out of the BTB domain used

for this analysis, e.g the group constituted by KCTD5,

KCTD2, and KCTD17 Elements recently cloned have

been included in two groups: (a) the group formed by

polymerase d and proliferating cell nuclear

antigen-interacting proteins, tumor necrosis factor-a-induced

protein 1 [12], KCTD13 product polymerase

delta-interacting protein 1 [12], and KCTD10 [13]; and (b)

the group formed by the leftover-related proteins

KCTD8, Pfetin (predominantly fetal expressed T1

domain) (KCTD12), and KCTD16, which are involved

in development [14] For the remaining sequences there

are clear paralogy relationships, which indicate close

relationships within the sequences, as in the case of

KCTD3 and Q8TBC3, a human homolog of mouse

seta-binding protein-1 [15], KCTD1 and KCTD15,

and KCTD21 and KCTD6 Most of these sequences

remain uncharacterized This analysis of KCTD

sequences shows that they form a group clearly

differ-entiated from the voltage-gated potassium channels,

not only by the absence of transmembrane domains, but also on the basis of the differences in BTB sequences

To determine the subcellular localization of KCTD5, green fluorescent protein (GFP)–KCTD5 was trans-fected and detected by confocal microscopy (Fig 3A) Whereas GFP alone is found in the nucleus as well as

in the cytosol, fusion of KCTD5 to GFP restricts the expression of the fusion protein, GFP–KCTD5, to the cytosol Furthermore, HEK293 cells were transfected with a plasmid expressing myc–KCTD5, and this protein was detected by immunocytochemistry in the cytosol (Fig 3B) As it has been recently reported that deletion of the C-terminus of KCTD5 [16] changes its location to the nucleus, cells were transfected with different deletion mutants of KCTD5 Immunocyto-chemistry of these cells showed that these constructs were again detected in the cytosol (Fig 3B) Therefore,

in our hands, KCTD5 is detected only in the cytosol

As we had a specific antibody for KCTD5, we tried several times to reveal the endogenous protein with this antibody, but we could not see any specific bind-ing, so we consider that this antibody is not suitable for immunocytochemistry

Although it has been proposed that all the proteins containing a BTB domain are substrate-specific

Fig 1 KCTD5 expression (A) RNA from different tissues and cell lines was analyzed by RT-PCR using specific primers for KCTD5 A plas-mid encoding KCTD5 was used as a positive control (lane 1) for the RT-PCR (B) Expression of KCTD5 in different cell types detected by immunoblot with antibody to KCTD5 (upper panel) b-Actin was detected by immunoblot on the same membrane as an internal control of protein loading (lower panel) (C) Time course of the expression of KCTD5 protein in PBLs stimulated with PHA (D) Time course of the expression of KCTD5 mRNA in PBLs stimulated with PHA, where numbers indicate hours of stimulation (E) Expression of KCTD5 protein in PBLs subjected to various stimuli (F) Levels of KCTD5 mRNA assayed by quantitative PCR in PBLs stimulated with PHA TCR+CD28 indi-cates antibodies specific for T-cell receptor plus CD28.

adaptors for cullin ubiquitin ligases [5,6], in the case of

CUL3, most of the adaptors described so far belong to

the Kelch group Thus, we investigated whether

KCTD5 could interact with CUL3 To test this inter-action, HEK293 cells were transfected with plasmids encoding CUL1, CUL2, CUL3, CUL4A and CUL4B

A

B

Fig 2 Analysis of KCTD5 homologs (A) Multiple protein sequence alignment of various KCTD5 orthologous sequences from different spe-cies (B) Phylogenetic tree built from human paralogs of KCTD5 using the BTB domain of 23 peptides The BTB domain (T1 domain) of the voltage-gated potassium channel KCNC1 protein is included in the analysis to root the tree.

along with KCTD5 Total lysates were prepared from these cells and used for immunoprecipitation assays A specific interaction of KCTD5 with CUL3 was observed (Fig 4A, lane 6), but not with the other cul-lins (Fig 4A, lane 5 for CUL1 and data not shown) This interaction was confirmed in primary cells by car-rying out immunoprecipitation assays in lysates from PBLs stimulated with PHA for 2 days Under these conditions, CUL3 was detected by immunoblot in KCTD5 precipitates (Fig 4B, lane 2), but not when the immunoprecipitation was carried out with an irrel-evant antibody (Fig 4B, lane 1) Then, the ability to form a functional E3 ligase complex with CUL3 and Rbx1 was assayed Expression vectors for these pro-teins were transfected into HEK293 cells, and cell lysates were subjected to immunoprecipitation with antibody to myc As KCTD5 was precipitated when CUL3 was present in the lysate (Fig 4C, lane 5), this result indicates that KCTD5 is part of a canonical cul-lin-based E3 ligase complex A faint band is also seen

in Fig 4C (lane 2) that is probably due to the interac-tion of Rbx1 with endogenous CUL3

We also addressed whether KCTD5 could be ubiqui-tinated, based on the fact that other BTB adaptor pro-teins have been shown to be substrates of E3 ligases

To do this, we transfected cells with expression vectors for myc–Ub and hemagglutinin (HA)–KCTD5, and cell lysates were immunoprecipitated with an antibody specific for HA The precipitates showed the presence

of ubiquitinated proteins (Fig 4D) by immunoblotting

To distinguish between covalent and noncovalent Ub binding to KCTD5, we repeated this experiment, lys-ing the cells with a highly denaturlys-ing buffer containlys-ing

8 m urea Under these conditions, no smear was detected in the KCTD5 immunoprecipitation and nor was a KCTD5 ladder observed in Ub precipitates, which is typical of ubiquitinated proteins (data not shown) In addition, we could not detect KCTD5 ubiquitination in in vitro assays (data not shown) Thus, unlike to what has been described for other BTB proteins that work as substrate-specific adaptors, KCTD5 is not ubiquitinated

The interaction between BTB proteins and CUL3 is considered to be mediated by the BTB domain and the N-terminal region of CUL3 [8], mainly on the basis of assays in which deletion of the BTB domain and the N-terminal region of CUL3 is accompanied

by loss of binding To analyze in detail KCTD5 bind-ing to CUL3, pull-down and immunoprecipitation assays with a series of deletion mutants of KCTD5 and CUL3 were carried out (Fig 5A,C) These exper-iments showed that the C-terminal region of KCTD5 was dispensable for CUL3 interaction, whereas the

B

A

Fig 3 Subcellular localization of KCTD5 (A) Left panels:

fluores-cence images of HEK293 cells transfected with either GFP or GFP–

KCTD5 Right panels: phase contrast images of the same cells (B)

Immunofluorescence staining of HEK293 cells transfected with

plasmids encoding KCTD5 and several deletion mutants with mAb

to myc followed by a secondary antibody labeled with Alexa

Fluor 594.

BTB domain alone (45–145 amino acids), although

essential for this interaction, was not sufficient

(Fig 5B) In fact, it required additional amino acids

(40–45) on the N-terminus, outside of the BTB fold,

as the 40–145 amino acid peptide is the smallest

moi-ety able to interact with CUL3 Studies with other

BTB proteins, e.g SPOP [17]or the BTB protein

At1g21780 from Arabidopsis thaliana [18], have also

shown that other parts of their sequence, in addition

to the BTB domain, are involved in the association with CUL3 On the other hand, the CUL3 region involved in this interaction was the N-terminus, as described for other BTB proteins, because a deletion

of 75 amino acids in the N-terminus of CUL3 com-pletely abrogated the binding of KCTD5 to CUL3 (Fig 5D, lane 8) Therefore, this detailed study on the interaction of KCTD5 with CUL3 shows that the sole BTB domain of KCTD5 does not support this

Fig 4 KCTD5 interacts with CUL3 and ubiquitinated proteins (A) HEK293 cells were transfected with plasmids encoding myc–KCTD5, HA– CUL1, and HA–CUL3, as indicated Cell lysates were subjected to immunoprecipitation (IP) with antibody to myc followed by immunoblotting with antibodies to HA and myc Expression of the tagged proteins is shown in the lower panels as WCL (whole cell lysate) (B) Lysates from PBLs treated with PHA for 2 days (upper panel) were immunoprecipitated with either KCTD5 or an irrelevant IgG antibody and then blotted with antibodies to CUL3 (upper panel) and KCTD5 The panels marked WCL show the expression levels of KCTD5 and CUL3 in the PBL whole cell lysates (C) HEK293 cells were transfected with plasmids encoding myc–Rbx1, HA–CUL3 and GST–KCTD5, lysates from these cells were processed for pull-down with GST beads, and the presence of KCTD5 in the precipitates was checked by western blotting with antibody to GST, followed by anti-HA and anti-myc blots to detect HA–CUL3 and myc–Rbx1 WCLs were immunoblotted with antibodies to GST, HA and myc to assess the expression of the tagged proteins (D) HEK293 cells were transfected with plasmids encoding for HA–KCTD5 and myc–ubiquitin Cell lysates were immunoprecipitated with antibody to HA, and ubiquitinated proteins that interact with KCTD5 were detected with antibody to myc WCLs were immunoblotted with the antibodies to HA and myc to assess the expression of the tagged proteins.

association and requires additional amino acids in the

N-terminus of this domain

As the BTB domain is responsible for

homo-oligo-merization in BTB proteins [3], we addressed whether

KCTD5 might form homo-oligomers For this purpose,

HEK293 cells were transfected with different constructs

of KCTD5 to show this association by either

immuno-precipitation or pull-down assays (Fig 6A,B) We found

the BTB domain to be essential for KCTD5

oligomeri-zation, as peptides expressing the KCTD5 N-terminal

region (N55) or the C-terminal sequence (POZCO,

amino acids 145–234) could not interact with themselves

(Fig 6B) As the POZ⁄ BTB domain of KCTD5 is

dis-tantly related to the T1 domain of voltage-gated

potas-sium channels, this fact was taken as a hint that KCTD5

could also tetramerize To address this issue, gel exclu-sion chromatography was run with recombinant KCTD5 protein and KCTD5 was collected in fractions consistent with the estimated molecular mass of an oct-amer (Fig 6C), which in turn can be explained by the formation of two tetramers

Taken together, our results show that the BTB domain of KCTD5 is not able to bind alone to CUL3, indicating that although it is critical for this associa-tion, other sequences contribute to the binding of sub-strate-specific adaptors to CUL3, namely, five amino acids in the N-terminus of the BTB domain In addi-tion to the BTB fold, KCTD5 presents two other regions: 40 amino acids in the N-terminal sequence, which include a low-complexity region (12–33 amino

Fig 5 Mapping the interaction of KCTD5 with CUL3 (A) Schematic diagram of the several KCTD5 deletion mutants used in this study (B) Plasmids for KCTD5 and different deletion mutants expressed as GST fusion proteins were transfected, along with HA–CUL3, in HEK293 cells Lysates were subjected to pull-down assays with glutathione–Sepharose beads, and the presence of CUL-3 in the precipitates was detected by immunoblot with antibody to HA, followed by antibody to GST The expression of the proteins was checked in the whole cell lysate (WCL) by western blot with antibodies to HA and CUL3 (C) Schematic diagram of the CUL3 deletion mutants used in this study (D) myc–KCTD5 was expressed in HEK293 cells along with different deletion mutants of HA–CUL3 The presence of the different CUL3 pep-tides was checked in the myc immunoprecipitates by western blot with an antibody to HA myc–KCTD5 was detected in the immunoprecipi-tates by immunoblot with antibody to myc The same antibodies were used to show the expression in the WCL (lower panels).

acids), and 88 amino acids in the C-terminus

(PO-ZCO) Taking into account that KCTD5 could be an

adaptor of CUL3 E3 ligases, we favor the hypothesis

that the POZCO region could participate in substrate

recognition, and that this could be a new protein inter-action domain conserved through evolution, as seen in orthologs The fact that KCTD5 can form octamers and the recent description of heterodimerization of CUL3 [19] would indicate that complexes of higher order could be formed among CUL3 and BTB substrate adaptors, implying the recruitment of a great number of substrates by these E3 ligases

Although scarce, the information available about KCTD proteins suggests that these proteins might be involved in development and cellular differentiation For example, in zebra fish, three members of this group – lov (leftover), ron (righton), and dex (dexter) – are expressed asymmetrically in the left and right zebrafish diencephalons [14] Pfetin, a human ortholog of lov and rongenes, encoded by human gene KCTD12, is detected

as mRNA preferentially expressed in fetal organs [20], with the highest expression levels in the cochlea Another KCTD protein, KCTD11⁄ REN, is also regu-lated developmentally in the nervous system [21], and it has been implicated in the regulation of the Hedgehog pathway [22] The information presented in this article would indicate that KCTD proteins might function by recruiting specific substrates involved in development and cellular differentiation for ubiquitination by CUL3

Ub ligases and degradation by the proteasome As regards KCTD5, there is another report that shows its ability to interact with two viral regulatory proteins, Rep68 and Rep78, of the adeno-associated virus type 2, which are essential for viral DNA replication and gene expression [16], although no relationship was established with CUL3

In summary, in this study we present evidence that KCTD5 is a new substrate-specific adaptor for CUL3-based Ub ligases Our data indicate that a relevant mechanism underlying the physiological role of KCTD proteins includes recruitment of proteins to CUL3-based E3 Ub ligases for degradation in the protea-some As identification of substrates recruited to the proteasome would be very valuable for understanding the function of these proteins, we are pursuing the identification of KCTD5-interacting proteins, especially those that are ubiquitinated

Experimental procedures

Antibodies and reagents Tissue culture reagents were from Cambrex (Verviers, Belgium) The 12CA5 mAb against HA was from Roche (Indianapolis, IN, USA), anti-HA clone HA.11 was from Covance (Berkely, CA, USA), anti-glutathione S-transferase (GST) and mAb against myc (9E10) were from Santa Cruz

Fig 6 KCTD5 oligomerization (A) HEK293 cells were transfected

with plasmids encoding for GST–KCTD5 and different deletion

mutants of KCTD5, and cell lysates were subjected to pull-down

with Glutathione–Sepharose beads and immunoblotted with a

spe-cific antibody to FLAG followed by antibody to GST (B) HEK293

cells were transfected with HA–KCTD5 and several plasmids that

expressed different deletion mutants of KCTD5 as GST-fusion

pro-teins Anti-HA immunoprecipitates of the cell lysates were analyzed

by immunoblot with antibody to GST followed by antibody to HA.

(C) Gel filtration chromatography of KCTD5 recombinant protein

produced in bacteria The presence of KCTD5 in the fractions was

analyzed by immunoblot with antibody to KCTD5 Numbers under

the arrows indicate the chromatography fractions in which

mole-cular mass markers are eluted.

Biotechnology Inc (Santa Cruz, CA, USA), anti-cullin 3 was

from Abcam (Cambridge, UK), and mAbs against b-actin,

PHA, FLAG M2 mAb and PMA were from Sigma Chemical

Co (St Louis, MO, USA) Antibodies against CD3

(UCHT1) and CD28 (clone CD28.2) were from BD

Pharm-ingen (Franklin Lakes, NJ, USA) MG-132 was from

Calbio-chem (Darmstadt, Germany) IL-2 was from PreprotechEC

(Rocky Hill, NJ, USA) Goat anti-(mouse IgG) conjugated

with Alexa Fluor 594 was from Molecular Probes (Eugene,

OR, USA) A mouse mAb was raised against recombinant

full-length KCTD5 Human MTC panel II was from

Clontech (Mountain View, CA, USA)

Plasmids and mutagenesis

Standard molecular biology techniques were used to

gener-ate the different constructs used in this study All constructs

were verified by nucleotide sequencing KCTD5 from a

Jurkat cDNA library obtained from Origene (Rockville,

MD, USA) was cloned in the pEF plasmid and served as a

template for the different KCTD5 plasmids used in this

study HA–cullin1 and HA–cullin3 expression plasmids

were a kind gift of C Geisen (Department of Medical

Oncology, Dana-Farber Cancer Institute, Boston, MA,

USA) Cullin4A and cullin4B were generously provided by

K Tanaka (Department of Molecular Oncology, Tokyo

Metropolitan Institute of Medical Science, Japan) [23]

Cell culture and transfections

PBLs were isolated from buffy coats of healthy donors by

centrifugation at 700 g for 30 min on Ficoll–Hypaque (GE

Healthcare) cushions Monocytes⁄ macrophages were

elimi-nated by adherence to plastic for 1 h at 37C Proliferation

was induced by PHA and IL-2, which was added after 48 h

with PHA, antibodies to CD3 plus antibodies to CD28, or

PMA plus ionomycin Jurkat T-leukemia cells were kept at

logarithmic growth in RPMI-1640 medium supplemented

with 10% fetal bovine serum, 2 mm l-glutamine, 1 mm

sodium pyruvate, nonessential amino acids, 100 UÆmL)1

penicillin G, and 100 lgÆmL)1 streptomycin Transfection

of Jurkat T cells was performed as described previously

[24] HEK293 cells were maintained at 37C in DMEM

supplemented with 10% fetal bovine serum, 2 mm

l-gluta-mine, 100 UÆmL)1 penicillin G, and 100 lgÆmL)1

strepto-mycin For transient transfection, HEK293 cells were

transfected using the calcium phosphate precipitation

method [25]

Immunoprecipitation, GST pull-down,

SDS⁄ PAGE, and immunoblotting

These procedures were performed done as reported

previ-ously [24] Briefly, cells were lysed in 20 mm Tris⁄ HCl,

pH 7.5, 150 mm NaCl, 5 mm EDTA containing 1% NP-40,

1 mm Na3VO4, 10 lgÆmL)1 aprotinin and leupeptin, and

1 mm phenylmethanesulfonyl fluoride, and clarified by centrifugation at 16 000 g for 10 min The clarified lysates were preabsorbed on protein G-Sepharose and then incu-bated with antibody for 2 h; this was followed by overnight incubation with protein G-Sepharose beads Immune com-plexes were washed three times in lysis buffer and resus-pended in SDS sample buffer Proteins resolved by SDS⁄ PAGE were transferred to a nitrocellulose membrane, and immunoblotted with optimal dilutions of specific anti-bodies followed by the appropriate anti-IgG–peroxidase conjugate Blots were developed by the enhanced chemilu-minescence technique (ECL kit; GE Healthcare) according

to the manufacturer’s instructions Pull-down of GST fusion proteins was performed with glutathione–Sepharose beads (GE Healthcare) incubated with the clarified lysates for 2 h The complexes were then washed and processed as explained above for the immunoprecipitation Some blots, after being developed by chemiluminescence, were visual-ized with a Bio-Rad VersaDoc chemiluminescence imager

In this case, quantitation was carried out using quantity onesoftware from Bio-Rad

RT-PCR Total cellular RNA was extracted by the TRIzol method (Life Technologies, Grand Island, NY, USA) The condi-tions for cDNA first-strand synthesis and PCR reaccondi-tions were as described previously [26] To address more exactly the expression of KCTD5 mRNA, real-time RT-PCR was carried out in RNA samples treated with DNase (Turbo-DNA freeTM; Ambion, Austin, TX, USA) The resulting cDNA was amplified in a PTC-200 apparatus equipped with a Chromo4 detector (BioRad Laboratories), using SYBR Green I mix containing HotStart polymerase (ABgene, Epsom, UK) b-Actin was used as a housekeeping gene to assess the relative abundance of KCTD5 mRNA, using the comparative cycle threshold (CT) method for relative expression This method allows the relative expression for a given cDNA using the formula: 2)DCT, where DCT¼ DCKCTD5

T  DCbactinT [27] Therefore, one arbi-trary unit (AU) corresponds to the expression of b-actin

Indirect immunofluorescence and confocal microscopy

HEK293 cells were cultured on coverslips and transiently transfected with the indicated plasmids Cells transfected with GFP plasmids were fixed with 3.7% paraformaldehyde and mounted on microscope slides, and GFP was then visu-alized on an MRC-1024 confocal laser scanning microscope (Bio-Rad) Phase contrast images were also taken Immuno-fluorescence staining of transfected KCTD5 was performed

as described previously [24] HEK293 cells were washed in

NaCl⁄ Pi, fixed in 3.7% formaldehyde, permeabilized with

0.1% saponin in NaCl⁄ Pi, and blocked in the same medium

supplemented with 2.5% normal goat serum for 30 min at

room temperature Primary and secondary antibodies were

diluted in the same buffer and incubated with the cells for 1 h

each at room temperature After three washes with NaCl⁄ Pi,

the cells were mounted onto glass slides and viewed under a

confocal laser scanning microscope

Gel filtration chromatography

For gel filtration chromatography, we used recombinant

KCTD5 produced in bacteria as His6-KCTD5 after

removal of the His-tag with thrombin The protein solution

was fractionated through a Superdex 200 fast protein liquid

chromatography column (GE Healthcare), and collected in

fractions of 500 lL Protein was precipitated with 10%

tri-chloroacetic acid and washed with acetone before addition

of SDS sample buffer and analysis by 10% SDS⁄ PAGE

Sequence analysis and alignments

For sequence retrieval, the BTB domain of human KCTD5

was used as query to retrieve the orthologs from the

UniPROT (http://www.ebi.uniprot.org/index.shtml)

data-base using the blast algorithm [28] psi-blast [29] searches

retrieved 22 human paralogs Multiple sequence alignments

of the BTB domain were conducted using muscle [30] and

probcons [31] in both the orthologs and the paralogs To

generate reliable phylogenetic trees, Bayesian inference

using mrbayes v3.1.2 software was applied [32] Multiple

alignments were done in two independent runs, with four

independent Markov chains in each run One thousand five

hundred samples were used to estimate the posterior

proba-bility distribution The amino acid model is a fixed rate

model using a mixture of fixed models To compute a

con-sensus tree, we sampled 2502 from a total of 3002 trees in

two independent files (thus discarding 16% of the initial

samples prior to convergence) To root the tree, the

sequence of the BTB domain (T1) of the voltage potassium

channel KCNC1_HUM is included in the analysis

Acknowledgements

We are grateful to Dr Keiji Tanaka for the CUL4A

and CULB cDNAs, to Dr Cristoff Geisen for the

CUL1 and CUL3 plasmids, and to Dr Joan Conaway

for the myc–Rbx1 plasmid We thank the staff of

Centro de Hemoterapia y Hemodonacio´n de Castilla y

Leo´n for its help with the separation of leukocytes

This work was supported by a grant from Programa

Nacional de Biologı´a Fundamental (Grant

BFU2006-01203⁄ BMC), Red Cardiovascular from Instituto de

Salud Carlos III Y Bayo´n is under contract within the Ramo´n y Cajal Program of the Ministerio de Edu-cacio´n y Ciencia of Spain, co-funded by the European Social Fund through FEDER-FSE

References

1 Bardwell VJ & Treisman R (1994) The POZ domain: a conserved protein–protein interaction motif Genes Dev

8, 1664–1677

2 Zollman S, Godt D, Prive GG, Couderc JL & Laski

FA (1994) The BTB domain, found primarily in zinc finger proteins, defines an evolutionarily conserved family that includes several developmentally regulated genes in Drosophila Proc Natl Acad Sci USA 91, 10717–10721

3 Stogios PJ, Downs GS, Jauhal JJ, Nandra SK & Prive

GG (2005) Sequence and structural analysis of BTB domain proteins Genome Biol 6, R82, doi: 10.1186/ gb-2005-6-10-r82

4 Furukawa M, He YJ, Borchers C & Xiong Y (2003) Targeting of protein ubiquitination by BTB–Cullin 3–Roc1 ubiquitin ligases Nat Cell Biol 5, 1001–1007

5 Pintard L, Willis JH, Willems A, Johnson JL, Srayko

M, Kurz T, Glaser S, Mains PE, Tyers M, Bowerman

B et al (2003) The BTB protein MEL-26 is a substrate-specific adaptor of the CUL-3 ubiquitin-ligase Nature

425, 311–316

6 Xu L, Wei Y, Reboul J, Vaglio P, Shin TH, Vidal M, Elledge SJ & Harper JW (2003) BTB proteins are sub-strate-specific adaptors in an SCF-like modular ubiqu-itin ligase containing CUL-3 Nature 425, 316–321

7 Geyer R, Wee S, Anderson S, Yates J & Wolf DA (2003) BTB⁄ POZ domain proteins are putative sub-strate adaptors for cullin 3 ubiquitin ligases Mol Cell

12, 783–790

8 Petroski MD & Deshaies RJ (2005) Function and regu-lation of cullin-RING ubiquitin ligases Nat Rev Mol Cell Biol 6, 9–20

9 Pickart CM (2001) Mechanisms underlying ubiquitina-tion Annu Rev Biochem 70, 503–533

10 Hochstrasser M (1996) Ubiquitin-dependent protein degradation Annu Rev Genet 30, 405–439

11 Hershko A & Ciechanover A (1998) The ubiquitin system Annu Rev Biochem 67, 425–479

12 Wolf FW, Marks RM, Sarma V, Byers MG, Katz RW, Shows TB & Dixit VM (1992) Characterization of a novel tumor necrosis factor-alpha-induced endothelial primary response gene J Biol Chem 267, 1317–1326

13 He H, Tan CK, Downey KM & So AG (2001) A tumor necrosis factor alpha- and interleukin 6-inducible pro-tein that interacts with the small subunit of DNA poly-merase delta and proliferating cell nuclear antigen Proc Natl Acad Sci USA 98, 11979–11984