Mammalian HSP40/DNAJ homologs: cloning of novel cDNAs and a proposal for their classification and nomenclature

Abstract

Abstract We have cloned 10 novel full-length cDNAs of mouse and human HSP40/DNAJ homologs using expressed sequence tag (EST) clones found in the DDBJ/GenBank/EMBL DNA database. In this report, we tentatively designated them mHsp40, mDj3, mDj4, mDj5, mDj6, mDj7, mDj8, hDj9, mDj10, and mDj11. Based on the identity of the deduced amino acid sequences, mHsp40, mDj3, and mDj11 are orthologs of human Hsp40, rat Rdj2, and human Tpr2, respectively. We determined that mDj4 is identical with the recently isolated mouse Mrj (mammalian relative of DnaJ). PSORT analysis (a program that predicts the subcellular localization site of a given protein from its amino acid sequences) revealed that hDj9 has an N-terminal signal peptide; hence, its localization might be extracellular, suggesting that there may be a partner Hsp70 protein that acts together with the hDj9 outside of the cell. The same analysis indicated that mDj7 and mDj10 may have transmembrane domains. In order to simplify the complicated and confusing nomenclature of recently identified mammalian HSP40/DNAJ homologs, we propose here some new rules for their nomenclature. This proposed nomenclature includes the name of species with 2 lowercase letters such as hs (Homo sapiens), mm (Mus musculus) and rn (Rattus norvegicus); Dj standing for DnaJ; the name of types with A, B, and C, which were previously classified as type I, II, and III according to the domain structure of the homologs; and finally Arabic numerals according to the chronological order of registration of the sequence data into the database.

INTRODUCTION

It is now widely accepted that the molecular chaperone activity of the various members of the HSP70 proteins is regulated by members of the HSP40/DNAJ family of proteins through stimulating their adenosine triphosphate (ATPase) activity (Caplan et al 1993; Silver and Way 1993; Cyr et al 1994; Hightower et al 1994; Hartl 1996). Members of the HSP40/DNAJ family are known to have 3 distinct domains: (1) a highly conserved J domain of approximately 70 amino acids in size, often found near the amino terminus, which has been known to mediate interaction with HSP70 and regulate its ATPase activity; (2) a glycine and phenylalanine (G/F)-rich region possibly acting as a flexible linker; and (3) a cysteine-rich region (C domain) containing 4 [CXXCXGXG] motifs resembling a zinc-finger domain (Bork et al 1992). To date, 3 HSP40/DNAJ homologs have been identified in Escherichia coli, nearly 20 in Saccharomyces cerevisiae, and more than 15 in mammals (Zuber et al 1998). However, not all HSP40/DNAJ homologs necessarily contain all of these 3 domains. Recently, Cheetham and Caplan (1998) proposed the classification of HSP40/DNAJ homologs in which they classified them into 3 groups: type I homologs have all 3 domains (J, G/F, and C), type II have the J and G/F but not the C domain, and type III have the J domain alone.

The number of identified members of mammalian HSP40/DNAJ homologs is expanding very rapidly, and their nomenclature is very complicated and confusing. For example, one of the type I human homologs is called Hdj2 (Chellaiah et al 1993), Hsdj (Oh et al 1993), or dj2 (Terada et al 1997), while its mouse ortholog is named Hsj2 (Royaux et al 1998) and its rat ortholog is named Rdj1 (Leng et al 1998). Moreover, the name of Hsj2 has also been used for completely different HSP40/DNAJ homologs (Royaux et al 1998; Pei 1999). It is evident that a more comprehensive system of classification and new rules for the nomenclature of mammalian HSP40/DNAJ homologs is needed. Since we are also interested in how many HSP40/DNAJ homologs may exist in mammals, we searched the mouse EST (expressed sequence tag) database using the J domain sequence, especially the so-called “J-box” (His-Pro-Glu, HPD). Here, we have identified and characterized 10 mouse and human full-length HSP40/DNAJ homolog cDNAs.

To denote the protein family, we use here fully capitalized names (thus HSP40, DNAJ, HSP70, etc), and we use an initial capital letter for specific family members (thus Hsp40, DnaJ, Hdj2, etc) according to the recent guidebook (Gething 1997).

MATERIALS AND METHODS

Tentative grouping of HSP40/DNAJ homologs in mouse EST clones

A key word search using “DnaJ” in the DDBJ/GenBank/EMBL database yielded nearly 1000 entries. Most of them are human and mouse EST clones. The nucleotide sequences corresponding to the latter were downloaded from the database. These clones were then grouped by homology using Private Database software (Software Development Co Ltd, Tokyo, Japan). This analysis identified 39 distinct genes (Table 1). They were tentatively named HPD-1, HPD-2, HPD-3, and so on (these have the J-box [HPD] sequence), or HSJ1-1, HSJ1-2, HSJ1-3, and so on (these are described as “similar to human HSJ1” in the database), or NI-1, NI-2, NI-3, and so on (the NI indicating not identified; these clones are deposited in the database as DnaJ homologs, but no J-box sequence was found in our first screening). Analysis of human EST clones resulted in a similar tentative classification. Next we searched for EST clones that contain 5′ translation initiation codons in each group according to the Kozak's rule (Kozak 1987), purchased them from American Type Culture Collection, and sequenced them. If there was no EST clone encoding a 5′ translation initiation codon in a group, we obtained an EST clone that extended near to the 5′ terminus and performed 5′ RACE analysis to determine the nucleotide sequence around the translation initiation codon. The determined nucleotide sequences of the full length cDNAs were again tentatively designated as mHsp40, mDj3, mDj4–mDj11 (Tables 1 and 2; see Results and Discussion). In the present study, we omitted mouse EST clones homologous to mouse Mtj1 (Brightman et al 1995) because none of these EST clones contain a J domain or J-box sequences.

Table 1.

Tentative classification and designation of HSP40/DNAJ homologs in mouse and human EST clones^a

Table 2.

Summary of classification and nomenclature of mammalian HSP40/DNAJ homologs and their PSORT analysis

Sequencing and 5′ RACE cloning

The nucleotide sequence was determined by an autosequencer (Applied Biosystems, Model 373) using vector and internal primers. Cloning by 5′ RACE was carried out with a Marathon cDNA amplification kit (Clontech). GSP-1 primer and RNA source were HSJ1-1; 5′-GGGAATTCAGGATGCCATTTAAGTGCTAC-3′, mouse testis.

The underline indicates an EcoRI site. PCR reactions were carried out using Advantage-cDNA Polymerase Mix (Clontech) in a Perkin-Elmer Model 2400 thermal cycler. The resulting products were cloned into pBluescript and sequenced.

PSORT analysis

PSORT analysis (available at http://psort.nibb.ac.jp:8000/) is a program developed by Dr K. Nakai (Osaka University, Japan) to predict the subcellular localization sites of query proteins from their amino acid sequences. This program uses a k–nearest neighbor (k-NN) algorithm for assessing the probability of localizing at each candidate site (Horton and Nakai 1997). If the k-NN data points contain, for example, nuclear proteins with 50%, the query protein is predicted to be localized to the nucleus with the probability of 50%. This program can also predict the presence of signal sequences that are found in proteins sorted through the so-called vesicular pathway (bulk flow), as well as transmembrane segments (TM domain) and membrane topology.

RESULTS AND DISCUSSION

The identification and cDNA cloning of mammalian HSP40/DNAJ homologs is expanding rapidly. To date, more than 15 such homologs have been identified in mammals. Although intracellular function of some of these homologs, such as Hsp40/Hdj1 (Hattori et al 1993; Frydman et al 1994; Minami et al 1996; Michels et al 1997; Morimoto 1998), Hdj2/Hsdj (Tang et al 1997; Terada et al 1997; Cummings et al 1998; Davis et al 1998; Meacham et al 1999), Hsj1 (Cheetham et al 1994), p58 (Melville et al 1997), Csp (cysteine string protein; Braun et al 1996) and auxilin (Ungewickell et al 1995; Jiang et al 1997) have been investigated extensively, some homologs exist only as sequences in the database. The yeast Saccharomyces cerevisiae contains nearly 20 HSP40/DNAJ homologs, as revealed by the analysis of the nucleotide sequence of the whole genome (Zuber et al 1998). We are interested in determining the number of homologs in the mammalian genome. Therefore, we pursued the identification and isolation of full-length cDNA clones of unknown HSP40/DNAJ homologs from the mouse and human EST databases. We focused more on the mouse EST database to be able to analyze the function of each HSP40/DNAJ homolog using transgenic mouse and knock-out technique in the future. We found approximately 300 mouse EST clones homologous to the HSP40/DNAJ protein family. They were tentatively classified into 39 genes by homology comparison (Table 1), from which we obtained representative EST clones in each group from the ATCC and determined the nucleotide sequences of the respective full-length cDNAs.

EST cloning of mouse and human HSP40/DNAJ homologs

mHsp40 (type II), a mouse ortholog of human Hsp40/Hdj1

Since we previously identified a mammalian Hsp40 and isolated human Hsp40 cDNA and genomic clones (Ohtsuka et al 1990; Ohtsuka 1993, 1997; Hata et al 1996; Hata and Ohtsuka 1998), we first attempted to isolate a full-length mouse Hsp40 cDNA. Mouse Hsp40 (AB028272) is 90% identical to human Hsp40 at the nucleotide level and 95% identical at the deduced amino acid level. We have also isolated a genomic clone of mHsp40, the analysis of which revealed that its structure is very similar to that of the human Hsp40 gene, containing 3 exons divided by 2 introns (Hata and Ohtsuka, personal communication).

mDj3 (type I), a mouse ortholog of human Dnj3/Cpr3/Hirip4 and rat Rdj2

One of the type I homologs, human Hdj2/Hsdj (the mouse and rat orthologs are Hsj2 and Rdj1, respectively), has a CaaX prenylation motif at the carboxy terminus (Chellaiah et al 1993; Oh et al 1993; Royaux et al 1998; Leng et al 1998). Human Hdj2/Hsdj is reported to be a homolog of yeast Ydj1 and has been shown to be involved in protein import into mitochondria (Terada et al 1997). Recently, a second type I homolog, Rdj2, has been identified in the rat, which also has a CaaX motif at the carboxy terminus (Andres et al 1997). The human ortholog of Rdj2 was found in the database and deposited as Dnj3/Cpr3 (AF011793, Edwards et al, unpublished data) and Hirip4 (AJ001309; Lorain et al, unpublished data). A mouse EST clone tentatively classified as a member of the HPD-2 cluster appears to be an ortholog of human Dnj3/Cpr3/Hirip4 and rat Rdj2. We designate this mouse homolog as mDj3 (AB028853), which encodes a protein of 412 amino acids. Based on the identity of deduced amino acid sequences, there are obviously 2 groups within the type I homologs: one group (named here the DjA1 group), which includes human Hdj2/Hsdj, rat Rdj1 and mouse Hsj2/mDj2, and another group (named DjA2 group), which includes human Dnj3/Cpr3/Hirip4, rat Rdj2, and mouse mDj3. The amino acid identity within each group among the different species is 92–100%, and the amino acid identity between the 2 groups is only 52–55% (Table 3).

Table 3.

Amino acid identity among mammalian type I HSP40/DNAJ homologs^a

mDj4/m Mrj (type II)

The mouse EST clone that belongs to the HSJ1-4 cluster encodes a type II protein 242 amino acids in, which we call mDj4 (AB028854). EST clones of the HSJ1-7 cluster also encode the same protein. mDj4 is identical to the recently isolated mouse Mrj (for mammalian relative of Dnaj; Hunter et al 1999), which has been shown to be an essential gene for murine placental development using knock-out techniques. The human ortholog of mDj4/mMrj was also identified as human Mrj (hMrj or hDj4, AB014888; Seki et al 1999). Amino acid identity is 94% between mDj4/mMrj and human Mrj. Recently, Pei (1999) reported the isolation of a novel human DnaJ homolog that associates with a pituitary tumor-transforming gene protein and named it Hsj2 (AF080569). Hsj2 appears to be identical to hMrj/hDj4, since their nucleotide sequences are 99% identical. The designation of Hsj2 used by Pei (1999) should be changed to hMrj or hDj4 because Hsj2 has already been used to denote one of the mouse type I homologs (AF055664; Royaux et al 1998). Analyses using the yeast two hybrid system have revealed that mDj4/mMrj interacts specifically with Mlf2 (myeloid leukemia factor 2; Yamaguchi et al, unpublished data) and human hDj4/hMrj interacts specifically with keratin 18 (Izawa et al, unpublished data).

mDj5 (type II)

Since mouse EST clones in the HSJ1-1 cluster have J domain sequences but do not have 5′ translation initiation codons, we performed 5′ RACE analysis. Thus, we obtained a full-length cDNA that encodes a type II protein of 220 amino acids and named it mDj5 (AB028855). The deduced amino acid sequence of the J domain of mDj5 is highly homologous to that of mDj4/mMrj (90% identity), but their carboxy terminal portions show no homology. There is no ortholog of mDj5 in other species reported so far.

mDj6 (type II)

A mouse EST clone in the HSJ1-2 cluster encodes a type II protein of 227 amino acids, designated mDj6 (AB028856). There are no reports so far of orthologs of mDj6 in other mammals.

mDj7 (type II)

A mouse EST clone which was tentatively classified in the HPD-3 cluster encodes a relatively small type II protein of 222 amino acids, which we call mDj7 (AB028857). The complementary sequence of the NI-6 cluster is identical to the sequence of the HPD-3 cluster. Homology search of the DDBJ/GenBank/EMBL database revealed that the rat ortholog of mDj7 seems to be Mdg1 (microvascular endothelial differentiation gene 1; Proels et al, unpublished data, X98993), since their amino acid sequences are 99% identical. However, Proels et al described that the translation initiation site of rat Mdg1 is the methionine residue 3 amino acids before the HPD (J-box) sequence, and it encodes a protein of 173 amino acids. If this is correct, the rat Mdg1 is a novel protein that has a truncated J domain. Whether or not this is the case awaits further analysis. PSORT analysis of the deduced amino acid sequence of mDj7 predicts that it has a transmembrane domain (amino acids 7–23) with type 2 membrane topology, and that it may be localized in the endoplasmic reticulum (Table 2). Interestingly, amino acid sequence of the carboxy terminus of mDj7 and Mdg1 is —CSGQ, which resembles the CaaX prenylation motif (Cox and Der 1992).

mDj8 (type II)

Mouse EST clones of the HSJ1-8 cluster encode a type II protein of 259 amino acids, designated here mDj8 (AB028858). The deduced amino acid sequence of mDj8 is highly homologous to human Hsj1 (Cheetham et al 1992) not only in the J and G/F domains but also in the carboxy terminus region (overall identity is more than 90%). However, their protein lengths are much different (259 and 351). Therefore, mDj8 seems to be a truncated version of the mouse ortholog of human Hsj1. PSORT analysis indicates that the localization of mDj8 may be nuclear (Table 2).

mDj9/hDj9 (type II)

Mouse EST clones of the HPD-7 and NI-1 clusters encode a type II protein, but the EST clones that encode the 5′ translation initiation site were not available from ATCC. Therefore, the nucleotide sequences of several EST clones in these clusters (eg, AA497706, W67505, AA596749, and AA204094) were combined to construct a full-length cDNA. This cDNA sequence encodes a protein of 358 amino acids, designated here mDj9. In addition, we sequenced a human EST clone of the HPD-7 cluster. The deduced amino acid sequence of human Dj9 (hDj9, AB028859) is 95% identical to that of mDj9. To date, there are no orthologs of mDj9 or hDj9 described in other mammals, but its deduced amino acids sequence is highly homologous to one of the Caenor habditi elegans putative DnaJ homologs (Z47356, gene: T15H9.1), not only within the J and G/F domains but also at the carboxy terminal portion (the overall amino acid identity is 58%; Fig 1). Interestingly, PSORT analysis indicates that mDj9/hDj9 has an N-terminal signal peptide (amino acids 1–22); thus, its localization may be extracellular (Table 2). If this turns out to be the case, mDj9/hDj9 is a novel HSP40/DNAJ homolog that acts outside of the cell, suggesting the existence of a novel partner for a member of the Hsp70 protein family. Similar results were obtained from the PSORT analysis of the C elegans T15H19.1 protein.

Fig 1.

Homology of the deduced amino acid sequences between hDj9 and Caenorhabditis elegans putative DnaJ homolog protein (gene T15H9.1). Underlines indicate putative N-terminal signal peptides predicted by PSORT analysis. Identical and similar amino acids are indicated by asterisks and dots, respectively.

mDj10 (type II)

Mouse EST clones of the HPD-4 cluster encode a type II protein of 376 amino acids, which we call mDj10 (AB028860). Its deduced amino acid sequence is highly homologous to another C elegans putative DnaJ homolog (Z73102, gene: B0035.14), with an overall amino acid identity of 43%. The J domain of mDj10 is found at amino acids 109–178. PSORT analysis predicts that mDj10 has a transmembrane domain (amino acids 247–263), that its membrane topology is type 1b, and that its localization is nuclear (Table 2). PSORT analysis of C elegans B0035.14 putative protein resulted in a similar prediction.

mDj11/mTpr2 (type III), a mouse ortholog of human Tpr2

A mouse EST clone of the HPD-1 cluster encodes a J domain but does not have 5′ translation initiation codon. We looked for other EST clones homologous to the 5′ region of the AA033375 clone by consecutive homology searches in the DDBJ/GenBank/EMBL database. We found an EST clone that encodes a 5′ translation initiation codon and sequenced it. The clone encodes a typical type III protein containing 494 amino acids, named here mDj11 (AB028861). The J domain of mDj11 is found near the carboxyl terminus (amino acids 379–448). The deduced amino acid sequence is 96% identical to the human Tpr2 protein (Murthy et al 1996). Human Tpr2 is one of the tetratricopeptide repeat (TPR)-containing proteins (class III) and has 7 TPR repeats. It is known that 2 other type III homologs, Zrf1/Mida1 (Hughes et al 1995; Shoji et al 1995) and p58 (Lee et al 1994), are also class III TPR proteins.

Other HSP40/DNAJ homologs in the mouse EST database

Recently, the third mammalian type I family member, hTid1, was identified (Schilling et al 1998). HTid1 is a homolog of the Drosophila tumor suppressor protein Tid56 (Kurzik-Dumke et al 1998) and interacts with papillomavirus Type 16 E7 oncoprotein. Drosophila Tid56 has been reported to be a mitochondrial protein. PSORT analysis indicates that hTid1 is also mitochondrial protein (Table 2). Both Tid56 and hTid1 are homologous to yeast mitochondrial Mdj1 (Rowley et al 1994). Therefore, the hTid1 protein is the first example of a mammalian mitochondrial HSP40/DNAJ homolog. In the present study, we found that mouse EST clones of the NI-3 cluster encode an ortholog of human hTid1. However, no clones encoded a 5′ translation initiation site.

We suspect that there should be a fourth type I member in mammals because a mammalian homolog corresponding to yeast Scj1 (Schlenstedt et al 1995) localized in the ER has not yet been reported. However, we could not find mouse EST clones corresponding to this gene in the present study.

Mouse EST clones of the HPD-8 cluster appear to be orthologs of human Hlj1 (human liver DnaJ 1, U40992; Hoe et al 1998), but there are no mouse EST clones that encode a 5′ translation initiation codon. Mouse EST clones of the HPD-9 cluster may correspond to mMcg18/Hspf2 (AF036875; Silins et al 1998). Also, mouse EST clones of the HSJ1-3 cluster appear to correspond to mouse Msj1 (mouse sperm DnaJ homolog 1, U95607; Berruti et al 1998). Mouse EST clones of the HSJ1-9 cluster may encode an ortholog of human Spf31 protein (AF083190; King et al, unpublished data).

In the present study, we could not find a mouse ortholog of human Hsj1 (Cheethem et al 1992). As mentioned previously, mDj8 (HSJ1-8 cluster) seems to be a truncated version of human Hsj1.

Recently, human and mouse Hsp40-3 cDNAs (AF088982 and AF088983, respectively; Chen et al, unpublished data) have been deposited in the DNA database, but we could not find mouse EST clones corresponding to Hsp40-3 by our present key word search.

Other mouse EST clones in the clusters corresponding to HPD-5 (NI-10), HPD-6, HPD-8, HPD-10, HPD-11, HSJ1-6, NI-4, NI-5 (complementary), NI-7 (complementary), NI-9 (complementary), NI-12 (complementary), and NI-17 have J-box (HPD) sequences. However, there are no EST clones encoding 5′ regions in these groups. Mouse EST clones in groups of HSJ1-5, NI-2, NI-8, NI-11, NI-14, and NI-16 are deposited in the database as “homologous to DnaJ protein,” but these clones have no J domain or J-box sequence. Although we did not analyze mouse EST clones homologous to Mtj1 (Brightman et al 1995), there seem to be several more HSP40/DNAJ homologs in this group.

Taking account of these results, we suggest that there may be approximately 20 more unidentified HSP40/DNAJ homologs in mammals and that the total may be more than 40.

Amino acid sequence comparison among HSP40/DNAJ homologs

All members of mammalian HSP40/DNAJ family are summarized in Table 2. Although most of type I and type II members have the J domain at or near the N-terminus, type III members have it at the middle or the C-terminal region. We performed homology analysis of the J domain and the remaining part of these members separately. Comparison of amino acid sequences of the J domain of 23 mammalian members are shown in Figure 2A, indicating that while type I and type II members are highly homologous to each other, the J domains of type III members such as Mtj 1, Zrf1, mMcg18, auxilin, and hSpf31 are less homologous each other or to that of type I and type II members. The J-box (HPD) sequence are definitely found in all members. In addition, alanine at the 51st residue of mHsp40 is also conserved without exception. It has been reported that the alanine residue corresponds to ¹⁷⁹Ala of yeast Sec63p and mutation of ¹⁷⁹Ala to ¹⁷⁹Thr (sec63-1) resulted in the temperature-sensitive growth and translocation defect of the ER in yeast (Scidmore et al 1993).

Fig 2.

Homology comparison among mammalian HSP40/DNAJ homologs. (A) Comparison of J domain. It should be noted that the J-box (HPD sequence) and alanine (51st residue of mHsp40) are conserved without exception. (B) Homology among C-terminal part of type I members. (C) Comparison among C-terminal region of Hsp40 group; mHsp40, Hlj1, mHsp40-3, and hDj9. (D) Homology among C-terminal part of Hsj1 group; Hsj1, mDj8, Msj1, mDj4, and mDj6. Identical and highly conserved amino acids are indicated by asterisks and dots, respectively.

Although hTid1 is a type I protein, its C-terminal region has relatively low homology to other type I members (approximately 28% identity with Hsj2/mDj2 or mDj3; Fig 2B). Homology comparison of the remaining part other than the J domain of type II and type III proteins revealed that they are obviously classified into several groups. First, Hsp40 group includes mHsp40, Hlj 1 and mHsp40-3 (overall amino acid identity is 62–64%). Hdj9 appears to belong to this group, since its sequence is 32% identical to members of this group (Fig 2C). Second, Hsj1 group contains Hsj1, Msj1, mDj4, mDj6, and mDj8. Their sequence homology is 50–60%, although Hsj1 and mDj8 are highly homologous (Fig 2D). Third, Zrf1, p58 and mDj11 are classified into a TPR protein group. N- and C-terminal regions without the J domain of other type II and type III members show no apparent homology.

Why are there so many HSP40/DNAJ proteins?

The nucleotide sequence analysis of whole genome revealed that Saccharomyces cerevisiae contains 14 HSP70 proteins and nearly 20 HSP40/DNAJ proteins. In mammals, only 12 HSP70 proteins have been identified so far (Tavaria et al 1997). On the other hand, there are, at present, 23 HSP40/DNAJ proteins, including the results of our present study. Why are there more HSP40/DNAJ proteins than HSP70 proteins? It is now believed that HSP40/DNAJ proteins usually act together with HSP70 proteins. Thus, one HSP70 protein might form a chaperone complex with several different HSP40/DNAJ proteins. Indeed, it has been shown in yeast that the member of the HSP70 family that is located in the ER, BiP, is able to interact not only with Sec63 (Corsi and Schekman 1997) but also with Scj1 (Silberstein et al 1998) and Jem1 (Brizzio et al 1999). Moreover, cytosolic Hsc70/Hsp70 in mammals has been reported to interact functionally with several HSP40/DNAJ proteins such as Hsp40 (Minami et al 1996), Hsj 1 (Cheetham et al 1994), Hdj2/Hsdj (Terada et al 1997; Meacham et al 1999), auxilin (Jiang et al 1997), Csp (Braun et al 1996), and virus T antigens (Srinivasan et al 1997). Each chaperone complex is known to have a specific function in various aspects of the life cycle of proteins. Thus, it is suggested that HSP40/DNAJ proteins (the remaining part of the protein other than the J domain) in each intracellular compartment could determine the functional specificity of the function of each chaperone complex (Rassow et al 1995).

We have recently isolated genomic clones of human (Hata et al 1996; Hata and Ohtsuka 1998) and mouse Hsp40 genes (Hata and Ohtsuka, unpublished data) and demonstrated that the J domain (70 amino acids) of both proteins is precisely encoded by the first exon. This evidence strongly suggests that the J domain has been incorporated into various proteins including virus large and small T antigens by exon shuffling during evolution (de Souza et al 1996). This in turn might explain why there are more HSP40/DNAJ proteins than HSP70 proteins.

Proposal for the nomenclature of mammalian HSP40/DNAJ homolog proteins

At present, there are 23 HSP40/DNAJ homologs in mammals, including our present results (Table 2). Since the present designation of each homolog is very complicated and confusing, we would like to propose here a new system for their nomenclature. First, HSP40/DNAJ homolog proteins are classified into 3 groups according to their domain structure, as suggested by Cheetham and Caplan (1998). Here we designate type I, type II, and type III groups as DjA, DjB, and DjC, respectively, since Roman numerals cannot be used for the nomenclature of proteins and genes. Second, Arabic numerals should be written after the group's name according to the chronological order of deposition of the sequence data in the database. Third, orthologs of each species must be presented by 2 lowercase letters, such as hs (Homo sapiens), rn (Rattus norvegicus), and mm (Mus musculus), before the group name. For example, human Hdj2/Hsdj would renamed as hsDjA1, rat Rdj2 as rnDjA2, mouse Hsp40 as mmDjB1, and so on. Original names could be used together with the new name for purpose of cross-comparison, eg, hsDjB6/hMrj/Hsj2.

CONCLUSION

We isolated 10 new mouse and human cDNA clones of HSP40/DNAJ homologs in full length from the EST database and proposed a set of new rules for the nomenclature of mammalian HSP40/DNAJ homologs for the convenience of investigators studying these molecular chaperones.

Fig 2.

Continued

Fig 2.

Continued.

Fig 2.

Continued.

Appendix