Hu-K4 is a ubiquitously expressed type 2 transmembrane
protein associated with the endoplasmic reticulum
Antonia Munck, Christopher Bo
¨
hm, Nicole M. Seibel, Zara Hashemol Hosseini and
Wolfgang Hampe
Center of Experimental Medicine, Institute of Biochemistry and Molecular Biology II: University Hospital Eppendorf, Hamburg, Germany
The Hu-K4 protein was first identified as a human
homologue of the K4L protein of vaccinia virus [1].
K4L is a nonessential protein in the life cycle of the virus
and has unknown function. Both Hu-K4 and K4L con-
tain two HXKXXXXD ⁄ E (HKD) motifs which make
them members of the superfamily of HKD proteins
together with phospholipase D proteins and phospholi-
pid synthases [2]. The closest homologues of Hu-K4 are
found in other mammals, murine SAM9 [3] has 93%
identical amino acid residues (Fig. 1). More distantly
related proteins are found in Xenopus (54%) and Dro-
sophila (48%) and in vaccinia virus. In addition to the
viral K4L protein (48%) this virus also encodes the clo-
sest relative of Hu-K4 with known function, the most
abundant viral protein p37 (21%). Other members of
the HKD superfamily are the phospholipase D iso-
forms. Like the other proteins shown in Fig. 1 they
harbour two HKD motifs which are involved in the
catalytic process [4]. For this reason Hu-K4 was named
phospholipase D3 in the GenBank entry NP_036400
although outside the HKD motifs no similarity exists.
Phospholipase D enzymes catalyse the hydrolysis of
membrane phospholipids, e.g. of phosphatidyl choline
to choline and phosphatidic acid which was ascribed

doi:10.1111/j.1742-4658.2005.04601.x
Hu-K4 is a human protein homologous to the K4L protein of vaccinia
virus. Due to the presence of two HKD motifs, Hu-K4 was assigned to the
family of Phospholipase D proteins although so far no catalytic activity
has been shown. The Hu-K4 mRNA is found in many human organs with
highest expression levels in the central nervous system. We extended the
ORF of Hu-K4 to the 5¢ direction. As a consequence the protein is 53
amino acids larger than originally predicted, now harbouring a putative
transmembrane domain. The exon ⁄ intron structure of the Hu-K4 gene
reveals extensive alternative splicing in the 5¢ untranslated region. Due to
the absence of G ⁄ C-rich regions and upstream ATG codons, the mRNA
isoform in brain may be translated with higher efficacy leading to a high
Hu-K4 protein concentration in this tissue. Using a specific antiserum pro-
duced against Hu-K4 we found that Hu-K4 is a membrane-bound protein
colocalizing with protein disulfide isomerase, a marker of the endoplasmic
reticulum. Glycosylation of Hu-K4 as shown by treatment with peptide
N-glycosidase F or tunicamycin indicates that Hu-K4 has a type 2 trans-
membrane topology.
Abbreviations
EST, expressed sequence tag; GST, glutathione S-transferase; Hu-K4, human K4L homologue; PNGaseF, peptide N-glycosidase F.
1718 FEBS Journal 272 (2005) 1718–1726 ª 2005 FEBS
wrapping of infectious intracellular mature virions by
cisternae derived from virus-modified trans-Golgi or
endosomal membranes to form intracellular enveloped
virions [7]. The integrity of the p37 HKD motifs is
required for the formation of the intracellular en-
veloped virion membrane [8]. During this process p37
shuttles between plasma membrane and intracellular
organelles [9] and is involved in the trafficking of
integral membrane proteins from the Golgi apparatus.

Hu-K4 mRNA is most highly expressed in brain, but
at a lower level in almost all tissues. These data are in
agreement with those of Pedersen et al. [3], who found
the mRNA of the murine homologue SAM9 mainly in
brain, but also in other tissues. In situ mRNA hybrid-
ization showed a neuronal expression in the adult and
developing murine brain [3]. The human multiple-
tissue expression array shows a weak signal for Hu-K4
in the corpus callosum, which contains mainly glial
but no neuronal cell bodies, indicating that also in the
human brain mainly neurons express Hu-K4.
1 60
Hum MKP KLMYQELKVPAEEPANELPMNEIEAWKAAEKKARWVLLVLILAVVGFGAL.MTQL
Mur MKP KLMYQELKVPVEEPAGELPLNEIEAWKAAEKKARWVLLVLILAVVGFGAL.MTQL
Xen MSS KVEYKPIQ.PHEEAENHFLQHELHKVKA.RKYYRCALVVAIIITLVFCIL.ASQL
K4L MNPDNTIA
dro MPEYKKLEDQESDVENANRTTVQNTATVQDAGEGQRQAAGQQAGQMVTVSLFMLLFLGSS
p37 M
61 120
Hum FLWEYGDLHLFGP N QRPAPCYDPCEAVLVESIPEGLDFPNASTGNPSTSQAWLG
Mur FLWEYGDLHLFGP N QRPAPCYDPCEAVLVESIPEGLEFPNATTSNPSTSQAWLG
Xen LLFPFLSITSQTT ETVLNKDIRCDDQCRFVLVESIPEGLVYDANSTINPSIFQSWMN
K4L VITETIPIGMQFDKV YLSTFNMWRE
Dro YFQPRPRLHQYKGGRGHGLLEK FD.CNIQLVESIPIGLTYPDGSPRFLSTYEAWLE
p37 WPFASVPA GAKC RLVETLPENMDFRSD HLTTFECFNE
121 180
Hum LLAGAHSSLDIASFYWTLTNNDTHT.QEPSAQQGEEVLRQLQTLAPKG VNVRIAV
Mur LLAGAHSSLDIASFYWTLTNNDTHT.QEPSAQQGEEVLQQLQALAPRG VKVRIAV
Xen IITNAKSSIDIASFYWSLTNEDTQT.KEPSAHQGELILQELLNLKQRG VSLRVAV
K4L ILSNTTKTLDISSFYWSLSD EVGTNFGTIILNEIVQLPKRG VRVRVAV

421 480
Hum VPADEAQARIPYARVNHNKYMVTERA.TYIGTSNWSGNYFTETAGTSLLVTQNGRGG
Mur VPTDESQARIPYARVNHNKYMVTERA.SYIGTSNWSGSYFTETAGTSLLVTQNGHGG
Xen VPATEAQKKIPYARVNHNKYMVTDRV.AYIGTSNWSGDYFINTAGSALVVNQTQSAGTSD
K4L VP DADPPIPYSRVNHAKYMVTDKT.AYIGTSNWTGNYFTDTCGASINITPDDGLG
Dro VPTDSSQEKIPFGRVNHNKYMVTDRV.AYIGTSNWSGDYFTDTAGIGLVLSETFETETTN
p37 IQ NNTKLLIVDDEYVHITSANFDGTHYQNHGFVSF NSIDK
481 540
Hum .LRSQLEAIFLRDWDSPYSHDLDTSADSVGNACRLL
Mur .LRSQLEAVFLRDWESPYSHDLDTSANSVGNACRLL
Xen TIQMQLQTVFERDWNSNYSLTFNTLSSWKEK.C.IF
K4L .LRQQLEDIFMRDWNSKYSYEL YDTSPTKRCKLLKNMKQCTNDIYCDEIQPEKEIPEY
Dro TLRSDLRNVFERDWNSKYATPL V
p37 QLVSEAKKIFERDWVSSHSKSLKI
541
K4L SLE
Hum
B
A
Mur
Xen
K4L
Dro
p37
Fig. 1. Homology of Hu-K4 with other members of the HKD super-
family. (A) Protein alignment performed with the
CLUSTAL method [23].
Highly conserved residues found in at least five of the six proteins are
boxed. The two HKD motifs are overlined. (B) Phylogenetic tree of
the alignment in (A). Hum, human Hu-K4 (AAH36327); Mur, murine

ally truncated Hu-K4 isoforms. The additional 53
N-terminal amino acids, which are not present in the
original database entry, are highly homologous to
Hu-K4 from mouse (GenBank BC076586) and rat
(XM_341811) indicating a high evolutionary pressure
on this sequence and supporting the hypothesis that
this part of the mRNA is translated.
Eleven exons encode the ORF of human Hu-K4
(Fig. 4A, exon 5–15). The analysis of more than 100
GenBank EST clones did not reveal any alternative
splicing in the ORF or in the 3¢ untranslated region
(UTR). To explain the splice variants observed in the
Northern blot we analysed the 5¢-UTR of the available
several dozen EST clones which turned out to be
highly variable (Fig. 4B). Two out of these cDNA
clones, both derived from the same adult female breast
cDNA library, start with exon 2, all other clones start
with exon 1 but skip exon 2 indicating that there might
be two different promotors. The clones containing
exon 1 are very diverse in their exon composition
before exon 5. They might or might not bear the exons
3 or 4, part or the entire region between exons 3 and 4
(Fig. 4B, 3¢ and 4¢), or extended exons 1 or 5 (Fig. 4B,
1¢ and 5¢) and therefore differ in size. Most often
clones with 44, 258 or 422 bp extensions between
exons 1 and 5 are found, nicely explaining the mRNA
isoforms seen in the Northern blot (Fig. 2A). The
brain
heart
skeletal muscle

B
C
Fig. 2. Hu-K4 mRNA distribution. (A) Northern blot analysis using a
32
P-labelled Hu-K4 cDNA fragment. (B) Human multiple tissue
expression array.
Hu-K4 A. Munck et al.
1720 FEBS Journal 272 (2005) 1718–1726 ª 2005 FEBS
short isoform, which is abundantly expressed in brain,
probably includes exons 1 ⁄ 1¢⁄5–10 as 10 out of the 15
respective clones (67%) originate from the fetal or
adult central nervous system. In contrast, only 15%
and 20% of the clones with the 258 or 422 bp exten-
sions are derived from brain, respectively.
The different 5¢-UTRs might function in transla-
tional control. In most cases, 5¢-UTRs that enable effi-
cient translation are short, have a low GC content and
do not contain upstream ATG codons [13]. The lon-
gest isoform of the Hu-K4 mRNA is 457 bp longer
than the shortest version. Exon 4 is the largest exon,
alone accounting for 214 nucleotides. Especially exons
1¢,4¢ and 4 have a high G ⁄ C content (Table 2).
Upstream ATG codons are found in exons 1, 2, 3, 4¢
and 4, but only those in exons 3, 4¢ and 4 are located
in an adequate context for translational start
(Table 2). Taken together, these data suggest that the
smaller mRNA variant from brain which lacks exons
2, 3, 3¢,4,4¢ and 5¢ might be more efficiently trans-
lated than the larger isoforms which predominate in
other tissues.

the stop codon. Two putative polyadenylation signals
(AATAAG and AATAAC) are present about 20
nucleotides in front of the poly(A) tail (Fig. 3A). Both
are not identical to the most often used signal AAT
AAA which is found in 65% of human mRNAs, but
especially variants with a single pyrimidine substitution
like AATAAC also seem to be functional [14].
Expression and topology of the Hu-K4 protein
In order to raise an antiserum against the Hu-K4 pro-
tein two rabbits were injected with a mixture of the
two peptides underlined in Fig. 3. One of the rabbits
produced an antiserum staining a 65-kDa band in
a western blot of Hu-K4-transfected COS-7 cells,
whereas the respective preimmune serum was negative
(Fig. 5A). Preincubation of the antiserum with the
C-terminal peptide LDTSADSVGNACRLL alone or
with both peptides used for immunization prevented
staining, indicating that the antiserum recognizes the
carboxy tail of Hu-K4. The apparent molecular mass
of 65 kDa is higher than the calculated molecular mass
of 55 kDa for Hu-K4 hinting at a post-translational
modification, e.g. glycosylation, of Hu-K4 in cultured
cells. The absence of additional bands in the western
blot indicates that at least in COS-7 cells only the first
start codon (Table 1) is used.
As the Hu-K4 mRNA is most abundant in brain
and since the small mRNA isoform found in brain is
probably translated with highest efficiency, we chose
membranes from rat brains to check whether the
Hu-K4 antiserum recognizes endogenous Hu-K4.

at a geranylgeranyl anchor. Alternatively, Hu-K4 could
harbour a transmembrane domain formed by a stretch
of 17 hydrophobic amino acids (Figs 3A and B; psort
ii predicts a transmembrane domain but not a cleaved
signal peptide). Since several basic amino acids are
present N terminal to the hydrophobic stretch but
none on the C-terminal side, the first 38 amino
acid residues are expected to be cytoplasmic, whereas
the large C-terminal domain including the two
HXKXXXXD ⁄ E-motifs would be luminal or extracel-
lular [17]. This domain inherits seven putative glycosy-
lation sites which could only be glycosylated if it enters
the endoplasmic reticulum. The N-terminal 38 amino
acid residues lack consensus sites for N-glycosylation
(Fig. 3A). To differentiate between the two topologies,
we deglycosylated Hu-K4 heterologously produced in
COS-7 cells (Fig. 5D). Indeed, we found a reduction in
the apparent molecular mass of Hu-K4 after treatment
with peptide N-glycosidase F (PNGaseF) showing that
Hu-K4 is a type 2 transmembrane protein. These data
are confirmed by a reduced molecular mass of Hu-K4
in cells that have been grown in the presence of the
glycosylation inhibitor tunicamycin (Fig. 5D). Differ-
ential glycosylation also explains the different apparent
molecular masses found for Hu-K4 in cultured cells
and brain.
Table 1. Putative start codons of Hu-K4. Comparison of the optimal
translational start site given by Kozak [12] and the putative start co-
dons in the Hu-K4 mRNA. The most important nucleotides of the
Kozak consensus sequence are indicated in bold. A weak context

Hybridization
Commercially available human multiple tissue Northern
blot and multiple tissue expression array (Clontech) were
A
B
Fig. 4. Exon–intron structure of the Hu-K4 gene. (A) Position of the exons on the BAC clone CTC-492K19 (GenBank AC010271.8). The first
and the last nucleotide of each exon are given. The positions of the polyadenylation site, the stop codon and the putative start codons ATG
1
and ATG
4
are indicated. The size of exons 1–5 is drawn in scale. (B) Alternative splicing of the 5¢-UTR. The splice variants, the number of
nucleotides between exon 1 and exon 5 (size) and the number of expressed sequence tags relating to each variant (#EST) are given.
A. Munck et al. Hu-K4
FEBS Journal 272 (2005) 1718–1726 ª 2005 FEBS 1723
hybridized using a [
32
P]-labelled human Hu-K4 probe com-
prising the first 916 nucleotides shown in Fig. 3A as des-
cribed [18].
Bioinformatics
DNA and protein analysis were performed using the pro-
gram dnastar. Hu-K4 encoding ESTs were identified in
GenBank using the basic local alignment tool (blast)of
the National Centre for Biotechnology Information (http://
www.ncbi.nlm.nih.gov/BLAST). Analysis of transmembrane
domains and protein motifs was carried out using psort ii
().
Antibody production, western blotting and
immunocytochemistry
Two rabbits were immunized with a mixture of the follow-

60933
60939
GTA ATG C
TGC ATG T
TCC ATG G
G
Adequate
Weak
Adequate
13 aa
33 aa
31 aa
3¢ 76 49% –
4¢ 56 61% 61041 GGA ATG T Adequate 1 aa
4¢ 61066 GCC ATG T Adequate 24 aa
4 214 73% –
5¢ 35 54% –
ADB
C
105
kDa
15
50
35
75
105
kDa
15
50
35

cDNA were analysed by western blotting using the Hu-K4 antiserum (1 : 2000) or the respective preimmune serum (PIS, 1 : 2000). Preincu-
bation of the antiserum with the C-terminal peptide used for immunization inhibited labelling of Hu-K4. (B) Detection of endogenous Hu-K4.
Membranes from rat brain were analysed by western blotting using the Hu-K4 antiserum in the absence or presence of the C-terminal pep-
tide used for immunization. (C) Hu-K4 is membrane bound. COS-7 cells transfected with the Hu-K4 cDNA or with vector alone (mock) were
disrupted by sonification, separated into a membrane (Mem) and a soluble (Sol) fraction by ultracentrifugation and analysed by western blot-
ting using the Hu-K4 antiserum. In the first lane a blot of conditioned medium of Hu-K4-transfected cells is shown. (D) Deglycosylation.
Membranes from COS-7 cells transfected with the Hu-K4 cDNA were incubated in the absence (–) or presence (+) of PNGaseF. As a control
(Con) nontreated membranes are shown. The two lanes on the right show the western blot analysis of Hu-K4 from membranes of trans-
fected COS-7 cells growing for 24 h in the absence (–) or presence (+) of the N-glycosylation inhibitor tunicamycin (1 lgÆmL
)1
).
Hu-K4 A. Munck et al.
1724 FEBS Journal 272 (2005) 1718–1726 ª 2005 FEBS
cysteine residue). After several injections one rabbit pro-
duced an antiserum appropriate for western blotting and
immunocytochemistry.
Western blotting and immunocytochemistry were per-
formed as described [19] using rabbit anti-(Hu-K4) Ig
(1 : 2000 for western blot, 1 : 1000 for ICC) or mouse anti-
(protein disulfide isomerase) Ig (1 : 100, StressGen). To
prove specificity, the diluted Hu-K4 antiserum was incuba-
ted with the indicated peptides ( 10 lgÆmL
)1
) for 1 h at
37 °C prior to incubation.
Heterologous expression, sample preparation
and deglycosylation
The image-cDNA clone 159455 (GenBank H15746 and
H15747) encoding full-length Hu-K4 was supplied by the
RZPD Deutsches Ressourcenzentrum fu

4
pH 7.4] and heated again to 95 °C for
5 min. After addition of 30 U PNGaseF (Roche) or an
equivalent volume buffer A, deglycosylation was allowed to
proceed for 10–14 h gently agitated at 37 °C. After a sec-
ond addition of PNGaseF or buffer A, the incubation was
repeated. Proteins were then precipitated using methanol ⁄
chloroform [22], separated by SDS ⁄ PAGE and Hu-K4
detected by western blotting.
Acknowledgements
We thank Prof Ulrike Beisiegel and Prof Chica Schal-
ler for discussion and providing the laboratory equip-
ment and Susanne Hoppe for technical assistance. This
work was supported by the Deutsche Forschungsge-
meinschaft (SFB 444 B10).
References
1 Cao JX, Koop BF & Upton C (1997) A human homolog
of the vaccinia virus HindIII K4L gene is a member of
the phospholipase D superfamily. Virus Res 48, 11–18.
2 Sung TC, Roper RL, Zhang Y, Rudge SA, Temel R,
Hammond SM, Morris AJ, Moss B, Engebrecht J &
Frohman MA (1997) Mutagenesis of phospholipase D
defines a superfamily including a trans-Golgi viral pro-
tein required for poxvirus pathogenicity. EMBO J 16,
4519–4530.
3 Pedersen KM, Finsen B, Celis JE & Jensen NA (1998)
Expression of a novel murine phospholipase D homolog
coincides with late neuronal development in the fore-
brain. J Biol Chem 273, 31494–31504.
4 Cummings R, Parinandi N, Wang L, Usatyuk P &

11 Baek SH, Kwak JY, Lee SH, Lee T, Ryu SH, Uhlinger
DJ & Lambeth JD (1997) Lipase activities of p37, the
major envelope protein of vaccinia virus. J Biol Chem
272, 32042–32049.
12 Kozak M (1995) Adherence to the first-AUG rule when
a second AUG codon follows closely upon the first.
Proc Natl Acad Sci USA 92, 2662–2666.
13 Meijer HA & Thomas AA (2002) Control of eukaryotic
protein synthesis by upstream open reading frames in
the 5¢-untranslated region of an mRNA. Biochem J 367,
1–11.
14 Graber JH, Cantor CR, Mohr SC & Smith TF (1999)
In silico detection of control signals: mRNA-3¢-end-pro-
cessing sequences in diverse species. Proc Natl Acad Sci
USA 96, 14055–14060.
15 Sugars JM, Cellek S, Manifava M, Coadwell J & Ktis-
takis NT (2002) Hierarchy of membrane-targeting sig-
nals of phospholipase D1 involving lipid modification of
a pleckstrin homology domain. J Biol Chem 277,
29152–29161.
16 Husain M, Weisberg A & Moss B (2003) Topology of
epitope-tagged F13L protein, a major membrane com-
ponent of extracellular vaccinia virions. Virology 308,
233–242.
17 Hartmann E, Rapoport TA & Lodish HF (1989) Pre-
dicting the orientation of eukaryotic membrane-span-
ning proteins. Proc Natl Acad Sci USA 86, 5786–5790.
18 Hampe W, Rezgaoui M, Hermans-Borgmeyer I &
Schaller HC (2001) The genes for the human VPS10
domain-containing receptors are large and contain many