Organization of the Chemokine Genes in the Human and Mouse Major Clusters of CC and CXC Chemokines: Diversiﬁcation Between the Two Species

Home , Chemokine

Genes and Immunity (2001) 2, 110–113  2001 Nature Publishing Group All rights reserved 1466-4879/01 $15.00 www.nature.com/gene BRIEF COMMUNICATION Organization of the chemokine genes in the human and mouse major clusters of CC and CXC chemokines: diversiﬁcation between the two species

H Nomiyama1, A Mera2, O Ohneda2, R Miura1, T Suda2 and O Yoshie3 1Department of Biochemistry, Kumamoto University Medical School, Honjo, Kumamoto 860–0811, Japan; 2Department of Cell Differentiation, Institute of Medical Embryology and Genetics, Kumamoto University, Honjo, Kumamoto 860–0811, Japan; 3Department of Bacteriology, Kinki University School of Medicine, Osaka-Sayama, Osaka 589–8511, Japan

Chemokines are a family of small cytokines that play essential roles in the directed migration of various types of leukocytes. Based on the arrangement of the conserved cysteine residues, they are classiﬁed into two major subfamilies,

CXC and CC, and two minor subfamilies, C and CX3C. So far, more than 40 members of this family have been identiﬁed in humans. Strikingly, the majority of CXC chemokine genes and that of CC chemokine genes are closely clustered at chromosomes 4q12–21 and 17q11.2, respectively. Similarly, the mouse major CXC and CC chemokine gene clusters are located on chromosomes 5 and 11, respectively. In order to understand the evolutionary processes that generated large numbers of CXC and CC chemokine genes in the respective chromosomal sites, we have constructed BAC and YAC contigs covering the human and mouse major clusters of CXC and CC chemokine genes. The results reveal that the organizations of CXC and CC chemokine genes in the major clusters are quite diverged between the two species most probably due to very recent gene duplications and rearrangements. Our results provide an important insight into the evolutionary processes that generated the major chemokine gene clusters and also valuable information in assigning the orthologues between human and mouse major cluster chemokines. Genes and Immunity (2001) 2, 110–113.

Keywords: chemokine family; gene rearrangement; gene cluster

Chemokines are a family of small, mostly basic, heparin- respectively.3,4 Similarly, their mouse counterparts are binding chemotactic cytokines that regulate the migration clustered on chromosomes 5 and 11.1 These functional and activation of various types of leukocytes.1,2 Chemo- and genetic features are in fact highly unique for the kines are classified into four subfamilies, CXC, CC, C and classical inflammatory chemokines as described below. CX3C, based on the arrangement of the N-terminal con- Recently, novel chemokines have been identified, served cysteine residues. Of these, CXC and CC chemo- mostly through application of bioinformatics on the kines are the major subfamilies. The majority of CXC expressed sequence tag (EST) databases.2 Unexpectedly, chemokines are potent attractants for neutrophils, these novel CXC and CC chemokines are directed at cells whereas the majority of CC chemokines recruit mono- such as lymphocytes and dendritic cells, and are cytes.1 Accordingly, these chemokines are strongly expressed constitutively in various lymphoid and other induced in a wide range of tissue cells by various proin- tissues with individually specific patterns. Furthermore, flammatory stimuli and are important mediators of these novel chemokines display much more specific inflammatory responses.1 Notably, these ‘inflammatory ligand-receptor relationships than those of the classical chemokines’ display highly redundant and promiscuous inflammatory chemokines, and their genes are also separ- ligand-receptor relationships.1 Thus, multiple ligands ately mapped at loci different from the traditional chemo- bind to a single receptor while a single ligand interacts kine gene clusters on chromosomes 4 and 17.2 It is now with multiple receptors. Furthermore, the human genes known that these novel chemokines play important roles of these inflammatory CXC and CC chemokines are regu- in the development, homeostasis, and function of the larly mapped at chromosomes 4q12–21 and 17q11.2, immune system by finely regulated trafficking of diverse subsets of lymphocytes and dendritic cells, and thus may be classified as ‘homeostatic or immune chemokines’.2 Correspondence: Dr H Nomiyama, Department of Biochemistry, Kumam- From the phylogenetic analyses, it is also obvious that oto University Medical School, Honjo 2–2-1, Kumamoto 860–0811, Japan. these immune chemokines are much older than the most Ȱ E-mail: nomiyama gpo.kumamoto-u.ac.jp inflammatory chemokines.5,6 Thus, the chemokines at the This work was supported by Grants-in-Aid for Scientific Research major clusters may have been generated relatively from the Ministry of Education, Science, Sports, and Culture of Japan recently in the evolutionary terms, and their highly Received 13 October 2000; revised 8 December 2000; accepted 11 unique genetic and functional features are likely to be December 2000 related to their recent evolution. In addition, certain Chemokine gene clusters H Nomiyama et al 111 viruses such as herpesviruses and poxviruses carry chemokine-like genes, which might have been acquired by host-virus gene exchange.7 CXC and CC chemokines at the major clusters are likely to have been generated by successive gene duplication events.1 Probably due to recent duplications even after diversification of the primates and rodents, it is sometimes difficult to determine orthologous relationships between the human and rodent member genes. For example, human CXC chemokines GCP-2/SCYB6 and ENA78/SCYB5 are more closely related to each other than to the probable murine counterpart LIX/Scyb5.8 In order to understand the evolutionary processes that generated the major chemokine gene clusters and the orthologous relationships between the human and mouse chemokine genes, we previously revealed the organization of the human CC chemokine gene cluster on chromosome 17 by constructing YAC and BAC contigs.4,9 The cluster contains 17 CC chemokine genes including one pseudogene (LD78␥/SCYA3L2), and is about several Mb long with a space of about one Mb separating the MIP and MCP subregions. The MIP subregion contains the MIP group genes such as MIP-1␣/SCYA3, MIP-1␤/ SCYA4 and RANTES/SCYA5, while the MCP subregion contains the MCP group genes such as MCP-1/SCYA2, MCP-2/SCYA8 and eotaxin/SCYA11.4,10 It is notable that the MIP group chemokines predominantly utilize the CC chemokine receptors CCR1 and CCR5, while the MCP group chemokines predominantly interact with CCR2 Figure 1 Construction of a BAC contig of the mouse MCP group and CCR3.11 These complex ligand-receptor relationships subregion in the major CC chemokine gene cluster. The human probably indicate the coevolution of ligands and recep- MCP group subregion and the human and mouse MIP group subregions are also shown for comparison. Thick lines and open circles tors through multiple gene duplication events and diver- indicate BAC clones and chemokine genes, respectively. The orders 5 sification. Recently, we also constructed a BAC contig of chemokine genes in the brackets are not established. An arrow of the mouse MIP group region.12 In the present study, denotes a duplication unit. Wavy lines indicate the regions covered therefore, we extended our previous studies and determ- by the finished BAC sequences deposited in GenBank. The nucleo- ined the organization of the mouse MCP group region tide sequences of the primers used for isolation of BAC and YAC and also the organizations of the human and mouse clones by PCR in Figures 1 and 2 can be obtained from the corresponding author upon request. major CXC chemokine gene clusters. To construct the physical maps of the clusters, genomic clones were identified from the BAC and YAC pools identified, is not present in the corresponding site of the (CITB human and mouse BAC libraries and a CEPH mouse MCP subregion. human YAC library, average insert sizes are 130 kb and We also constructed the contigs covering the major 1 Mb, respectively; Research Genetics, Huntsville, AL, clusters of human and mouse CXC chemokine genes. USA) by PCR. For estimation of the evolutionary relation- Figure 2 shows the organizations of the human and ship of the human and mouse chemokines, the mapping 5 mouse CXC chemokine genes. Like the CC chemokine data were compared to the phylogenetic data that ana- clusters, humans have more genes (15 genes including a lyzed chemokine sequences of representative vertebrate GRO pseudogene/SCYB1P) than mice (10 genes) so far. species. No human CXC chemokine gene in the major cluster is To compare the organization of the human and mouse included in the finished BAC sequences (as of 1 Sep- CC chemokine MCP subregions, we constructed a BAC tember 2000). However, several human working draft contig of the mouse MCP subregion (Figure 1). The sequences deposited in the GenBank database support physical maps of the human and mouse MIP subregions, our physical map. The map is also highly compatible which we constructed previously, are also shown in the with the partial maps and the mapping data of the figure for comparison. In mice, only 12 genes are ident- human CXC chemokine genes on chromosome 4.13–15 ified in the major CC chemokine cluster, and both The CXC gene cluster can be divided into two subre- humans and mice contain six chemokine genes in the gions termed herein GRO and IP-10 subregions (Figure MCP subregion. Notably, the human orthologue of the 2). The GRO subregion contains the chemokines such as mouse JE/Scya2, often called mouse MCP-1, does not IL-8/SCYB8 and GRO␣/SCYB1 which recognize CXCR1 exist in the corresponding site between MCP-3/SCYA7 and/or CXCR2, while the IP-10 subregion contains the and MCP-1/SCYA2. Mouse MCP-5/Scya12, which is in chemokines such as IP-10/SCYB10 and BLC/SCYB13 fact more homologous to human MCP-1/SCYA2 than which utilize CXCR3 or CXCR5, respectively.11 From the JE/Scya2, corresponds exactly to human MCP-1/SCYA2 insert sizes of the YAC clones (Figure 2) and the average on the map. Similarly, the mouse counterpart of human insert sizes of the BAC library used, the human CXC clus- MCP-4/SCYA13, whose mouse ESTs have not been ter is approximately one Mb long with a spacer region

Genes and Immunity Chemokine gene clusters H Nomiyama et al 112 of about several hundred kb between the two subregions. On the other hand, the mouse CXC cluster may be much smaller than the human cluster. Notably, the human GRO subregion appears to consist of two large direct repeats. One repeat contains PF4var1/PF4V1, PPBPL1/PPBPL1, GCP-2/SCYB6, GRO␣/GRO1 and GRO pseudogene/SCYB1P, and the other repeat contains PF4/PF4, PPBP/PPBP, ENA78/ SCYB5, GRO␥/GRO3 and GRO␤/GRO2. On the other hand, the mouse GRO subregion contains only one set of the genes. Furthermore, the mouse GRO group genes are not organized in parallel to a single unit of the human GRO repeats. Instead, the mouse GRO region appears to match well to the middle part of the human GRO region, overlapping in part with both direct repeats. Notably, each pair of the genes from the human direct repeats (PF4var1/PF4V1 and PF4/PF4, for example) is generally more similar to each other than the potential mouse counterpart genes. This suggests that the two direct repeats were generated during the recent human evolution. On the other hand, the mouse GRO/Gro1 and MIP-2/Scyb2 genes are also more similar to each other than to the three human GRO genes, and thus might have been duplicated during the recent rodent evolution. Fur- thermore, mouse weche/Scyb15 and human IL-8/IL8 have no counterparts in the human and mouse genomes, respectively. Since these genes are likely to have been generated before the separation of humans and mice (see the dendrograms in references5,6 and our web site http://cytokine.medic kumamoto-u.ac.jp), the human counterpart of mouse weche/Scya15 and the mouse counterpart of human IL-8 might have been deleted during species-specific rearrangements. Indeed, there is no mouse EST for IL-8 even though there are more than 200 human ESTs for IL-8 in the database. Collectively, it is likely that multiple species-specific gene-duplications and other rearrangements have occurred in the GRO subregion, and a further understanding on the evolutionary processes that generated the human and mouse GRO subregions requires the determination of the entire nucleotide sequences of this subregion in both species. Contrary to the GRO subregion, the IP-10 subregion is well conserved between the two species. This is probably because the IP-10 group genes were generated before the generation of most genes in the GRO subregion (see the dendrograms in the references above). Notably, the IP-10 group chemokines are mostly lymphocyte-specific, ie, the immune chemokines.2 This reminds us that the most lym- Figure 2 BAC contigs of the human and mouse major CXC chemokine gene clusters. Thick lines and open circles indicate phocyte-specific CC chemokines, which were also gener- BAC (or YAC) clones and chemokine genes, respectively. The ated before the emergence of the major cluster CC chemo- orders of chemokine genes in the brackets are not established. kines and are sporadically mapped on chromosomal sites An arrow denotes a duplication unit. A filled circle indicates an different from the major CC chemokine cluster, are also STS CM4.1453. It is not known whether the human PPBPL1/ well conserved between the two species.2 The conser- PPBPL1 gene is active or not. Since PCR with the primer pair vation of these immune chemokines between the two used for human PPBP/PPBP detects both PPBP/PPBP and PPBPL1/PPBPL1 genes, the contents of human PPBP/PPBP and species is likely to be related to their essential roles in the PPBPL1/PPBPL1 genes were determined by Southern hybridiz- homeostatic traffic of specific subsets of lymphocytes and ation. The probe detected 2.2, 1.4 and 0.5 kb EcoRI fragments for dendritic cells.2 PPBP/PPBP and 10 and 1.8 kb fragments for PPBPL1/PPBPL1. The present study together with our previous studies4,9,12 The contents of the human GRO␣/GRO1, GRO␥/GRO3 and and the phylogenetic analyses by Hughes and Yeager5 GRO (pseudo)/SCYB1P genes were also confirmed by hybridiz- has revealed several interesting features of the human ation. The probe detected 5.6 and 3.6 kb EcoRI fragments for GRO␣/GRO1, 11 and 1.4 kb EcoRI fragments for GRO␥/GRO3, and and mouse chemokine gene clusters. First, although the 7.0 and 2.0 kb EcoRI fragments for GRO pseudogene/SCYB1P. The ancestral chemokine genes are considered to have been sizes of YAC clones were taken from the YAC database of White- generated at the time or before the first vertebrates head Institute for Biomedical Research/MIT Center for Genome emerged,16 the chemokine genes in the major CXC and Research. CC chemokine clusters are likely to have been increased

Genes and Immunity Chemokine gene clusters H Nomiyama et al 113 by relatively recent multiple gene duplication events, subfamily on the long arm of chromosome 4 using radiation resulting in more than 20 members. In general, gene hybrids. Genomics 1998; 47: 136–139. duplication is the most important mechanism for generat- 4 Naruse K, Ueno M, Satoh T et al. A YAC contig of the human ing new members of a gene family, and some multigene CC chemokine genes clustered on chromosome 17q11.2. Genomics 1996; 34: 236–240. families are known to contain 1000 members or more.17 5 Hughes AL, Yeager M. Coevolution of the mammalian chemo- In particular, the chemokine gene duplications might kines and their receptors. Immunogenetics 1999; 49: 115–124. have been accelerated through the need of the host 6 Zlotnik A, Morales J, Hedrick JA. Recent advances in chemo- defense system against the molecular mimicry by viruses kines and chemokine receptors. Crit Rev Immunol 1999; 19: 1–47. as suggested by Murphy.18 7 Lalani AS, Barrett JW, McFadden G. Modulating chemokines: Secondly, most chemokine genes might have been gen- more lessons from viruses. Immunol Today 2000; 21: 100–106. erated by tandem duplications caused by unequal cross- 8 Smith JB, Rovai LE, Herschman HR. Sequence similarities of a ing over events and the paralogous genes have been subgroup of CXC chemokines related to murine LIX: impli- propagated at the two chromosomal loci independently. cations for the interpretation of evolutionary relationships Thus, chemokines that have similar receptor usage tend among chemokines. J Leukoc Biol 1997; 62: 598–603. 9 Nomiyama H, Fukuda S, Iio M, Tanase S, Miura R, Yoshie O. to form a subcluster within the major clusters. This type Organization of the chemokine gene cluster on human chromo- of duplication is different from another type of dupli- some 17q11.2 containing the genes for CC chemokine MPIF-1, cation where paralogous genes located in a certain HCC-2, HCC-1, LEC, and RANTES. J Interferon Cytokine Res chromosomal segment were duplicated in block and 1999; 19: 227–234. hence genes show more sequence similarity to paralog- 10 Maho A, Carter A, Bensimon A, Vassart G, Parmentier M. Physi- ous genes on other clusters than to genes linked on the cal mapping of the CC-chemokine gene cluster on the human same cluster.19,20 Thirdly, the organizations of the chemo- 17q11.2 region. Genomics 1999; 59: 213–223. kine gene clusters are substantially diverged between 11 Murphy PM, Baggiolini M, Charo IF et al. International union humans and mice. Apparently, there are more chemokine of pharmacology. XXII. Nomenclature for chemokine receptors. 52 genes in the human major clusters than the mouse major Pharmacol Rev 2000; : 145–176. 12 Fukuda S, Hanano Y, Iio M, Miura R, Yoshie O, Nomiyama H. clusters. Such species differences in the gene number Genomic organization of the genes for human and mouse CC probably have been caused by frequent gene duplications chemokine LEC. DNA Cell Biol 1999; 18: 275–283. even after the diversification of humans and mice. 13 Erdel M, Laich A, Utermann G, Werner ER, Werner-Felmayer Fourthly, from the relative positions on the physical G. The human gene encoding SCYB9B, a putative novel CXC maps, evolutionary relationships of some human and chemokine, maps to human chromosome 4q21 like the closely mouse chemokines can now be clarified. In some cases, related genes for MIG (SCYB9) and INP10 (SCYB10). Cytogenet human or mouse orthologues at the corresponding Cell Genet 1998; 81: 271–272. chromosomal regions are missing. In such cases, the 14 O’Donovan N, Galvin M, Morgan JG. Physical mapping of the physical maps presented in this study might be useful to CXC chemokine locus on human chromosome 4. Cytogenet Cell Genet 1999; 84: 39–42. search for missing species counterparts, if such genes are 15 Tunnacliffe A, Majumdar S, Yan B, Poncz M. Genes for ␤-throm- ever present, by the ‘positional cloning’ approach. Lastly, boglobulin and platelet factor 4 are closely linked and form part the chemokine major clusters, which are likely to have of a cluster of related genes on chromosome 4. Blood 1992; 79: been generated relatively recently, are likely to retain 2896–2900. some clues on the evolutionary processes that are 16 Najakshin AM, Mechetina LV, Alabyev BY, Taranin AV. Identi- involved in the generation of a large gene family. In fact, fication of an IL-8 homolog in lamprey (Lampetra fluviatilis): we previously found that the CC chemokine early evolutionary divergence of chemokines. Eur J Immunol PARC/SCYA18 gene was generated by a fusion of two 1999; 29: 375–382. tandem copies of MIP-1␣/SCYA3-like genes in humans.21 17 Li WH. Evolution of duplicate genes and pseudogenes. In: Nei Determination of the nucleotide sequences of the major M, Koehn RK (eds). Evolution of Genes and Proteins. Sinauer: Sun- derland, MA, 1983, pp 14–37. clusters in the human and mouse genomes, which are 18 Murphy PM. Molecular mimicry and the generation of host ongoing in the Human and Mouse Genome Projects, will defense protein diversity. Cell 1993; 72: 823–826. help us understand the generation and evolution of the 19 Hughes AL. Phylogenetic tests of the hypothesis of block dupli- chemokine gene family. cation of homologous genes on human chromosome 6, 9, and 11. Mol Biol Evol 1998; 15: 854–870. 20 Bailey WJ, Kim J, Wagner GP, Ruddle FH. Phylogenetic recon- References struction of vertebrate Hox cluster duplications. Mol Biol Evol 1 Rollins BJ. Chemokines. Blood 1997; 90: 909–928. 1997; 14: 843–853. 2 Zlotnik A, Yoshie O. Chemokines: a new classification system 21 Tasaki Y, Fukuda S, Iio M et al. Chemokine PARC gene and their role in immunity. Immunity 2000; 12: 121–127. (SCYA18) generated by fusion of two MIP-1␣/LD78␣-like 3 Modi WS, Chen ZQ. Localization of the human CXC chemokine genes. Genomics 1999; 55: 353–357.

Genes and Immunity