Immunology and Cell Biology

(2004)

82

, 112–118 doi:10.1046/j.0818-9641.2004.01230.x

© 2004 Australasian Society for Immunology Inc.

Research Article

Type I interferon genes from the egg-laying mammal,

Tachyglossus aculeatus

(short-beaked echidna)

G A V A N A H A R R I S O N ,

1

K E L L Y A M C N I C O L

1

a n d E L I Z A B E T H M D E A N E

2

1

University of Western Sydney, School of Science, Food and Horticulture, BCRI Building, Penrith South DC, NSW 1797 and

2

Division of Environmental and Life Sciences, Macquarie University, North Ryde, NSW 2109, Australia

Summary

The type I IFN are an important group of multifunctional cytokines that have, for whatever reason,evolved to a high level of complexity in eutherian mammals such as humans and mice. However, until recently, littlewas known about the type I IFN systems of the other two groups of extant mammals, the marsupials and the egg-laying monotremes. Preliminary partial type I IFN sequences from the short-beaked echidna were previously foundto cluster only with the IFN-

β

subtype in phylogenetic analyses, but a lack of sequence information madeinterpretation of these results tenuous. Here, we report cloning of the full-length genes of representatives from thetwo previously defined groups of echidna type I IFN by genomic walking PCR. Along with analysis of conservedcysteine placement and promoter elements, phylogenetic analysis incorporating these sequences strongly suggestthat the two groups of echidna type I IFN genes are in fact homologueous to IFN-

α

and IFN-

β

, confirming that theduplication leading to these two major classes of type I IFN occurred prior to the divergence of eutherians andmonotremes some 180 million years ago. Thus, even though there are major differences in gene copy number andheterogeneity, separate IFN-

α

and IFN-

β

gene families are a feature of the cytokine networks of all three groups ofliving mammals.

Key words

:

echidna, evolution, IFN, monotreme, type I interferon.

Introduction

Unlike members of the Eutheria and Marsupialia, the threeextant species of the Monotremata (platypus and two speciesof echidna) are peculiar amongst mammals in retaining thereptilian-like habit of laying eggs.

1

Whilst this and otherunusual features of monotreme anatomy and physiologycould be dismissed as being ‘abnormal’, it remains to be seenhow these characteristics influence, and are influenced by,attributes that have traditionally been viewed as pivotal to thesuccess of mammals, such as developments in the immunesystem. Moreover, along with the marsupials, the mono-tremes represent closely related branches to the extensivelystudied eutherian mammals in the tree of life. This propitiousrelationship has been largely untapped in its ability to providenew information on the molecular evolution and selectiveforces leading to the highly evolved and well documentedimmune system of present-day eutherian mammals. More-over, due to this relatively recent common ancestry witheutherians, monotremes and marsupials also provide testcases for current notions of molecular evolution and genefunction in mammals with differing modes of reproduction.

The IFN are a group of cytokines that have a variety offunctions, most notably conferring a viral-resistant state toreceptive cells.

2

These molecules are divided into two osten-sibly unrelated types, designated type I and type II, based on

physiochemical properties. However, in contrast to the soletype II interferon, IFN-

γ

, and other cytokines, the type I IFNare encoded by a complex gene family in eutherian mam-mals.

3

Members of the type I IFN family are primarilydivided into two broad categories, those related to IFN-

α

andIFN-

β

. The IFN-

α

-related IFN are further divided into foursubtypes: IFN-

α

, -

δ

, -

τ

and -

ω

, which show different degreesof relatedness to each other and in some cases have verydifferent functions.

4

Birds on the other hand, possess a morelimited array of type I IFN that are also more homogeneous innature

5

and probably arose through independent gene duplica-tions to the type I IFN subtypes of eutherian mammals.

6

Atpresent, the reason for the expansion of type I IFN genecomplexity in eutherian mammals is a matter of conjecture.

In a preliminary study based on partial DNA sequences,we showed that while a representative marsupial (tammarwallaby,

Macropus eugenii

) possesses a complement of typeI IFN genes similar in complexity to that of eutherians, theshort-beaked echidna (

Tachyglossus aculeatus

), a monotreme,appears to possess a much more limited inventory of thesegenes.

7

We were also able to demonstrate that the wallabytype I IFN genes fell into two groups, probably analogous tothe eutherian IFN-

α

and IFN-

β

. Whilst the echidna also hadtwo distinct groups of these genes, their relationship to thoseof eutherians remained unclear due to a lack of sequence data.Subsequently, we have isolated the full length genes fromthese two classes of wallaby type I IFN, confirming theiridentities as IFN-

α

and IFN-

β

homologues.

8

Here, we furtherinvestigated the identity of the previously isolated short-beaked echidna partial type I IFN sequences by isolating theequivalent full-length genes by genomic walking PCR.

Correspondence: Elizabeth M Deane, Division of Environmentaland Life Sciences, Macquarie University, North Ryde, NSW 2109,Australia. Email: edeane@els.mq.edu.au

Received 9 September 2003; accepted 13 November 2003.

Echidna type I IFN genes

113

Materials and Methods

DNA preparation and construction of genomic walking templates

Genomic DNA was extracted from the same kidney sample of ashort-beaked echidna (

T. aculeatus

) euthanased at Taronga Zoo(Sydney, NSW, Australia) as described previously.

7

From this DNA,five sets of genomic walking PCR templates based on the restrictionenzymes

Dra

I,

EcoR

V,

Pvu

II,

Sca

I and

Stu

I, were constructed byusing a Universal GenomeWalker kit (Clontech, Palo Alto, CA,USA) as described by us previously.

8

Genomic walking PCR

Previously, we had identified two distinct groups of type I IFN genesin the echidna based on partial DNA sequences.

7

Here, we designedsets of primers targeted to a representative of each of these groups(group I: TaG2; group II: TaG11) in order to amplify the cognatefull-length genes. Genomic walking PCR was carried out in two stepsusing primers internal to the first set in the secondary reactions(nested PCR) as described previously.

8

Briefly, in the initial genomicwalking PCR, 1

µ

L (containing approximately 0.1 ng DNA) of eachof the five restriction enzyme libraries was used in a reaction with10 pmole of the adaptor-specific AP1 primer supplied in the Uni-versal GenomeWalker kit (Clontech) and either the primary senseor antisense primer for each of the targeted sequences. Componentsused in the PCR were according to the instructions supplied with theUniversal GenomeWalker kit (Clontech). PCR temperature cyclingconditions were as follows: 7 cycles of 94

°

C for 30 s, followed by72

°

C for 4 min, 32 cycles of 94

°

C for 25 s, followed by 67

°

C for4 min, and finally, 67

°

C for 7 min. From the initial PCR product foreach enzyme library, a 1:50 dilution was performed using steriledeionized water and 1

µ

L of this used as template in a nestedsecondary PCR. In these reactions, the AP2 nested adaptor primerfrom the kit was used (instead of the AP1 primer) in combinationwith the nested gene-specific primer directed in the appropriateorientation. In the nested PCR reaction, temperature cycling was thesame as described above except that 5 and 20 cycles were performedin the first and second rounds, respectively (rather than 7 and 32).Primers used in the genomic walking PCR were as follows: TaG2:primary forward primer – 5

′

-CATCTTCAGCCAAAATCTGTC-CCAGACC-3

′

; primary reverse primer – 5

′

-GTAACTCTTCAG-GCGCAGTCGAAAAATGCCA-3

′

; nested forward primer – 5

′

-CCATGGACTTCATCAGCAGCTGGAAGAG-3

′

; nested reverseprimer – 5

′

-GTTTCGGATACGTGTCTTCCTCCAAGCAC-3

′

andTaG11: primary forward primer – 5

′

-CCATGGAGTCCATCAGGA-GATGGTGTGG-3

′

; primary reverse primer – 5

′

-TTCTGCCCT-TCAGGTAATCCATCATCCTC-3

′

; nested forward primer – 5

′

-GCTGTTTCTGGAAGAGGAAATGGGGTG-3

′

; nested reverse primer– 5

′

-TAGGCTGATGTCCCTTCTCAGGGTGCTG-3

′

.

Amplification of full-length genes

In order to amplify the targeted full-length genes as a singlecontinuous PCR product, and avoid using 5

′

and 3

′

genomic walkingPCR fragments from different closely related loci (see

8

) to constructthe final sequence, a further round of PCR was necessary withprimers complementary to the extremities of the 5

′

and 3

′

genomicwalking PCR products. For locus TaG11, nested PCR was carried outon 5

µ

L of undigested echidna genomic DNA under the sameconditions described above with the following primers: primaryforward primer – 5

′

-GAAGAGGGGAAGAGAGCTAGTTTGGCAG-3

′

;primary reverse primer – 5

′

-CATATCTCCTCCAAAATGTCTTC-

CCTGAC-3

′

; nested forward primer – 5

′

-ATGTGAGGAGGGAG-GGCATTCCAGGCCAG-3

′

; nested reverse primer – 5

′

-TCCTTACTCTATTCATCTTCCCTTCCGAG-3

′

. For the TaG2locus, this approach was unsuccessful (presumably due to amplifica-tion of fragments from different loci in the genomic walking PCRstep) and so a ‘back tracking’ approach

8

was used instead with thefollowing primers: primary forward primer – 5

′

-GATTCCCTAAGA-GATGCTCCCTGGATAG-3

′

and nested forward primer – 5

′

-GGA-TAGACCCGTAACAGTTTAGTGTCGC-3

′

.

Cloning and sequence analysis

All PCR products were extracted from agarose gels using a Bresa-Clean DNA purification kit (GeneWorks, Adelaide, SA, Australia).Purified PCR fragments were then cloned into the pGem T-easyplasmid vector (Promega, Madison, WI, USA) and transformed into

Escherichia coli

JM109 High Efficiency Competent cells (Promega)according to the manufacturer’s instructions. Individual bacterialcolonies were picked and incubated overnight in LB broth (SigmaChemical, St Louis, MO, USA) at 37

°

C with constant agitation.Plasmid DNA was then extracted by the method of Zhou andcolleagues.

9

Plasmids were sequenced completely in both orientationsusing a BigDye Terminator Cycle Sequencing Ready Reaction Kit(PE Applied Biosystems, Foster City, CA, USA) and the sequencingproducts separated and analysed on a model 377 automated sequencer(PE Applied Biosystems) through the Australian Genome ResearchFacility (University of Queensland, QLD, Australia). All sequenceanalysis was performed by using the Genetics Computer Group(GCG; Madison, WI, USA) software package through the AustralianNational Genomic Information Service website (http://www.angis.org.au).

Phylogenetic analysis

The maximum parsimony method, as employed by the PHYLIPsoftware package (J Felsenstein, University of Washington, Seattle,WA, USA), was used to generate phylogenetic trees based on100 bootstrap resampling data sets as described previously.

7

Treeswere generated from an alignment of nucleotide coding sequencesconstructed by using the GCG PILEUP program with various mam-malian type I IFN sequences (see

8

for GenBank accession numbers).In this analysis, a chicken type I IFN sequence (GenBank accessionX92476) was used as an outgroup. Trees were then displayed usingthe TREEVIEW program.

10

Results

Characterization of a full-length type I IFN group I gene from the echidna

The partial genomic DNA clone TaG2 that we designated asa group I IFN gene sequence from the echidna

7

was targetedin this study for full-length cloning by genomic walking PCR.From the available 220 bp of echidna sequence from theoriginal partial clone, we designed primers to amplify the 5

′

and 3

′

regions of the cognate gene by using this technique.The intention was to then amplify the complete coding regionand flanking sequences by using primers designed to thedistal regions of these fragments. Genomic walking PCR inthe 5

′

direction yielded a number of products (not shown), thelargest of which was an approximately 1 kb

EcoR

V fragmentthat was selected for further analysis. In the 3

′

direction, thelargest fragment obtained was a 1.2 kb

Dra

I fragment (notshown) and this was also sequenced. Comparison of the

114

GA Harrison

et al

.

proximal regions of these fragments to the original TaG2partial sequence revealed several nucleotide mismatches ineach case. Attempts to amplify an internal product based ondistal primers for these fragments failed, indicating that thefragments probably originated from different loci. This wasnot unexpected as previous Southern blotting using the TaG2clone as a probe had indicated approximately three closelyrelated genes.

7

In order to overcome this problem, a secondround of genomic walking PCR was carried out based on adistal 5

′

primer back towards the gene of interest (‘backtrack-ing’; see

8

). This resulted in a 1.6 kb

Stu

I fragment thatsequence analysis showed to contain a 546 bp open readingframe (ORF) with a high degree of similarity to the TaG2partial sequence (98% nucleotide identity) as well as pub-lished type I IFN sequences. In addition to the ORF, this clonealso contained 889 bp and 219 bp of upstream and downstreamregion, respectively. The deduced 182 aa protein was pre-dicted to consist of a 23 aa signal peptide and a 159 aa matureprotein by the Signalase program (N Mantei, Department ofBiochemistry, Swiss Federal Institute of Technology, Zurich,Switzerland). We have designated this clone TaG2FL4 and ithas been assigned the GenBank accession number AY194919.

Characterization of a full-length type I IFN group II gene from the echidna

The partial clone TaG11 designated by us previously as agroup II echidna type I IFN,

7

was targeted for full-length genecloning in this study. Southern blotting with the TaG11 clonehad previously shown that this group appears to consist of asingle gene only. Consequently, it was not surprising thatgenomic walking PCR yielded fragments that perfectly over-lapped the original TaG11 clone sequence, allowing a morestraight-forward approach to cloning the full-length gene. The5

′

genomic walking PCR fragment of most interest consistedof an approximately 3 kb

Stu

I fragment while in the 3

′

direction, an approximately 0.9 kb

Stu

I fragment was selectedfor analysis. Primers designed from the distal regions of thesePCR products were then used to amplify the interveningregion, including the gene of interest, from genomic DNA.This resulted in a 3.2 kb PCR product with an ORF of 561 bpthat was a perfect match with the TaG11 partial clone,preceded by 2.1 kb of 5

′

flanking region and followed by

527 bp of 3

′

flanking sequence. The deduced 187 aa proteinencoded by the ORF was predicted to consist of a 21 aa signalpeptide and a 166 aa mature protein by using the Signalaseprogram. We have designated this clone TaG11FL4 and it hasbeen assigned the GenBank accession number AY194920.

Sequence comparison of echidna type I IFN genes

Table 1 gives the percentage sequence identities between thefull-length echidna genes compared with other publishedmammalian type I IFN genes. Consistent with results basedon the partial sequences,

7

the two echidna genes share approx-imately 64% and 36% identity at the nucleotide and deducedamino acid levels, respectively, in the coding region. Theechidna gene contained in TaG2FL4 gave slightly higherpercentage identities at both the nucleotide and amino acidlevels with eutherian IFN-

α

/

ω

and the marsupial IFN-

α

genesthan for the equivalent IFN-

β

. However, one exception to thiswas that at the protein level human IFN-

β

was more similar tothis gene than human IFN-

α

1 (36.7% and 35.8%, respec-tively). Conversely, the gene contained in TaG11FL4 wasslightly more similar to eutherian and marsupial IFN-

β

thanIFN-

α

/

ω

genes at both the nucleotide and aa levels.

Phylogenetic analysis

Previously, we had found that the sequences for the echidnagroup I and II partial clones both clustered with the eutherianand marsupial IFN-

β

in phylogenetic reconstruction analysis.

7

However, this result was based on a relatively small amountof sequence data and it was not clear whether this wasartefactual as it had been anticipated that IFN of both typeswould be present in monotreme species. In order to moreaccurately assess the relationship between the echidna type IIFN and those of other species, we therefore constructed aphylogenetic tree based on the full-length sequences isolatedin this study (Fig. 1). In contrast to our previous finding, thisplaced the echidna group I sequence (TaG2FL4) in a mono-phyletic clade with the IFN-

α-related genes (including euthe-rian IFN-α, IFN-δ, IFN-ω and marsupial IFN-α) while thegroup II sequence (TaG11FL4) clustered with the IFN-β. Inboth cases, the echidna genes were placed basal to both theeutherian and marsupial genes of each type. In agreement

Table 1 Nucleotide and amino acid percentage identities for echidna IFN gene coding regions

NUCLEOTIDE

TaG2 FL4 TaG11 FL4 Human α1 Murine α1 Human ω1 Wallaby α Human β Murine β Wallaby β Porcine δTaG2 FL4 * 63.8 58.5 54.7 57.6 58.7 56.3 52.4 55.0 54.0TaG11 FL4 36.2 * 55.9 52.8 55.0 58.5 59.7 55.8 58.8 49.0Human α1 35.8 33.0 * 76.0 70.1 58.5 48.7 46.2 53.6 56.1Murine α1 34.2 27.7 62.4 * 64.3 55.2 48.0 46.2 48.8 53.1Human ω1 32.1 27.7 57.1 51.3 * 60.3 50.8 49.5 50.3 58.4Wallaby α 38.4 33.0 37.4 32.6 40.6 * 54.6 48.0 53.0 53.0Human β 36.7 38.8 29.8 28.7 30.9 36.2 * 67.2 58.9 48.8Murine β 29.3 35.1 25.0 21.8 30.3 27.1 49.2 * 54.0 44.3Wallaby β 31.4 36.4 32.4 27.7 32.4 34.0 39.1 31.9 * 47.3Porcine δ 30.3 22.3 37.6 33.7 40.4 35.4 28.5 24.6 27.4 *

AMINO ACID

Nucleotide (above diagonal line of asterisks) and amino acid (below diagonal line of asterisks) percentage identities for type I IFN from differentmammalian species.

Echidna type I IFN genes 115

with the phylogenetic tree constructed using partial sequencespreviously, based on the full-length sequences, porcine IFN-δformed the most basal branch of a eutherian IFN-α-relatedclade that was sister to the marsupial IFN-α genes (albeit withlow bootstrap support: 53/100). Thus, this result suggests thatthe IFN-δ gene diverged from an IFN-α gene ancestor earlyin the evolution of eutherian mammals but, after this lineage,split from the marsupials some 130 million years ago.11

Cysteine residue placement in echidna type I IFN

Typically, type I IFN of the α-related class form two disul-phide bridges (Cys1-Cys99 and Cys29-Cys139) while thoseof the β class form only one (equivalent to Cys29-Cys139 inIFN-α).12,13 This was found to hold not only for eutherian type

I IFN with some rare exceptions (e.g. murine IFN-β andporcine IFN-δ), but also those of the tammar wallaby.7,8 It wastherefore surprising that, based on our previous partialsequences, while both groups of echidna type I IFN clusteredwith the IFN-β, the group I sequences contained a cysteineresidue at a position (Cys99) normally specific to the α-related class.7 Armed with the full-length coding sequencesfor these genes, we can now show that the echidna group Igene (TaG2FL) codes for cysteine residues in all four posi-tions usually involved in forming the two disulphide bridgesof α-related IFN (Fig. 2). Moreover, the echidna group II IFNgene (TaG11FL4) contains only the two conserved cysteinescommon to both IFN-α and IFN-β, lacking the other twoconserved cysteines typically found in IFN-α and their closerelatives. Thus, the cysteine-coding profile of the echidna

Figure 1 Nucleotide parsimonytree of representative full-lengthtype I IFN sequences. Numbersindicate the number of bootstrapreplicates (out of 100) that gavethis branching configuration at aparticular node and hence indicatethe level of confidence in thatbranching arrangement. Note thatthe echidna G2FL4 sequence clus-ters with the eutherian and marsu-pial IFN-α-related genes whereasthe echidna G11FL4 sequenceclusters with the eutherian andmarsupial IFN-β genes. Genbankaccession numbers for sequencesused in the analysis are the sameas those used in 7.

116 GA Harrison et al.

group I gene is more consistent with an IFN-α while thegroup II gene is more consistent with encoding an IFN-β.

Echidna type I IFN gene promoter regions

Like the 5′ promoter regions of eutherian and marsupial typeI IFN genes, both of the echidna genes isolated in this studyalso contain a number of occurrences of the tetranucleotideGAAA sequence which have been implicated in transcriptionfactor binding within 120 bp of the putative transcription startsite (Fig. 3).14,15 However, in general, there is little sequenceconservation between the putative promoter regions of theechidna genes and those of the equivalent eutherian andmarsupial type I IFN genes. A notable exception to this is thatthe region of overlapping positive regulatory domains I and II(PRDI and PRDII) in eutherian IFN-β genes, is moderatelyconserved with the equivalent region in the gene from theechidna TaG11FL4 clone:Human β GAGAAGTGAAAGTGGGAAATTCC

|| ||||| |||| | || |||TaG11FL4 GAAAAGTG.AAGTTTGGAAATCC

PRDII is known to be the binding site for NF-κB16 and,significantly, this site is only found in IFN-β genes ineutherians.17 It is also worth mentioning that the TaG2FL4gene has a unorthodox putative TATA-box compared to IFN-α genes (CATTTAA rather than TATTTAA) though thefunctional significance of this is not known.

Discussion

In the present study, we have isolated representative full-length genes from each of the two groups of echidna type IIFN that had originally been identified based on partialsequencing of these genes. These are the first full-length IFNgenes to be reported from a monotreme and indeed, to ourknowledge, are the first full-length cytokine genes to bedocumented from this group of mammals. The more compre-hensive analysis facilitated by the complete sequences madeavailable in this study have allowed further clarification of thehomologous relationship of these genes to those of previouslyreported mammalian type I IFN. Previously, sequence analy-sis based on the available partial echidna sequences gaveconflicting views on whether the two groups of echidna typeI IFN were related to IFN-α and IFN-β or whether they wereboth more closely related to IFN-β than IFN-α. In the currentstudy, three lines of evidence have suggested that the echidnagroup I IFN are homologous to the IFN-α-related genes whilethe group II gene is related to IFN-β. Firstly, we found thatthe group I gene (from clone TaG2FL4) was slightly moresimilar to the IFN-α-related genes than the IFN-β and clus-tered with a IFN-α-related clade in phylogenetic analysis,whereas the group II gene (from clone TaG11FL4) was moresimilar to, and clustered with, the IFN-β. Secondly, the groupI gene had a cysteine profile more consistent with an IFN-αwhereas the group II gene had a cysteine profile more

Figure 2 Amino acid alignment of echidna type I IFN against those of other mammals. Note that two cysteines (denoted C α/β-1 andCα/β-2) are conserved in IFN of both α-related and β classes (except murine IFN-β), while the α-related class (except porcine IFN-δ) alsocontain another two conserved cysteines (denoted Cα-1 and Cα-2) that are not present in the IFN-β but are present in the echidna group IIFN (TaG2FL4).

Echidna type I IFN genes 117

consistent with an IFN-β. Thirdly, the 5′ promoter region ofthe gene from TaG11FL4 possessed regions similar to PRDIand PRDII (NF-κB recognition) sites that are only present inIFN-β genes in eutherians,17 whereas the group I gene didnot possess the equivalent regions. This, along with evi-dence obtained from Southern blotting using the partialclones as probes,7 suggests that the echidna possessesapproximately three IFN-α genes (previously designated

group I) and probably a single IFN-β gene (previouslydesignated group II).

The isolation of the full-length type I IFN genes from amonotreme now gives us a more complete picture of theevolution of these genes in the three extant groups ofmammals as we have also recently characterized full-lengthtype I IFN genes from a marsupial8 and eutherian genes forthese molecules have been extensively studied over many

Figure 3 Comparison of puta-tive 5′ promoter regions fromechidna, tammar wallaby, mouseand human type I IFN genes (A)IFN-α genes. Sequences similarto the positive regulatory domainI of the human IFN-β gene pro-moter (PRDI-like) are indicatedfor the human α1/murine α4genes as are the ‘TG sequence’and TATA-box.19 Numberingrefers to the murine IFNA4 genepromoter relative to its transcrip-tion start site. (B) IFN-β genes.Positive regulatory domains I toIV (PRDI-IV) are indicated, asare negative regulatory domain I(NRDI), octameric sequence (Oct.)and the TATA-box.14 PRDII isalso the site of binding of thenuclear transcription factor NF-κB. Numbering refers to thehuman IFN-β gene promoterrelative to its transcription startsite. GAAA sequences commonlyfound interspersed in promoterregions of type I IFN and IFNstimulated genes (ISG) are shownin bold.

118 GA Harrison et al.

years.4 Clearly, the type I IFN systems from all of thesegroups of mammals show various similarities. One of themost obvious of these is that all of the mammalian groupspossess a single or a small number of IFN-β genes and alarger number of IFN-α-related genes. In terms of secondarystructure, the type I IFN of the three mammalian groups alsoappear to be similar with IFN-α and their close relativescontaining four conserved disulphide-forming cysteines whilethe IFN-β, in general, have only two of these conservedcysteines. Moreover, the 5′ promoter regions of these genes inall three groups contain multiple copies of the tetranucleotideGAAA (often proceeded by a spacer and the dinucleotideAA) which, in eutherians at least, are thought to be part of thecore binding sites for the interferon regulatory factor (IRF)family of transcription factors14,15 and play a pivotal role intranscriptional regulation of these genes. Although we obtainedinsufficient sequence data from the TaG2FL4 clone in the 3′flanking region to confirm this in the corresponding gene,we note that the 3′ flanking sequence of clone TaG11FL4(IFN-β) corresponding to the 3′ untranslated region, con-tained the AU-rich sequences that have been implicated inrapid mRNA turnover in other type I IFN genes18 and so thistype of post-transcriptional regulation is probably ubiquitousacross the type I IFN of all mammals.

One striking difference between the three groups ofmammals in respect to their type I IFN gene families is thenumber of IFN-α-related genes. Southern blotting had indi-cated that there were probably three group I IFN in theechidna7 and we now know that these are likely to be homo-logous with the IFN-α. Previously, sequence analysis basedon the partial sequences had also shown that within thisgroup, the level of sequence similarity was quite high (92%average nucleotide similarity).7 This is in contrast to thesituation in eutherians and marsupials where upwards of 10IFN-α or genes closely related to IFN-α, are present and thedegree of sequence similarity amongst these genes is consid-erably lower. Given that type I IFN related to IFN-α havetaken on functions associated with reproduction in someeutherian species, it is tempting to suggest that marsupialsand eutherians have a more extensive repertoire of these genesas they have taken on functions involved in viviparity/placentation in these species, whereas the oviparous mono-tremes do not require this level of complexity (7 and referencestherein). Regardless of whether such extraneous functions forthe type I IFN exist in marsupials and eutherians, thesequence data from the echidna reported in this paper, alongwith those of the tammar wallaby isolated in our laboratorypreviously, should prove invaluable in future comparative andevolutionary studies of the type I IFN systems in mammals.

Acknowledgements

We thank the University of Western Sydney and the Cooper-ative Research Centre for the Conservation and Managementof Marsupials for financial support. We also thank the staff at

Taronga Zoo for assistance with tissue collection and KatrinaHarrison for editorial assistance.

References

1 Renfree MB. Monotreme and marsupial reproduction. Reprod.Fertil. Dev. 1995; 7: 1003–20.

2 De Maeyer E, De Maeyer-Guignard J. Type I interferons. Int.Rev. Immunol. 1998; 17: 53–73.

3 Weissmann C, Weber H. The interferon genes. Prog. NucleicAcid Res. Mol. Biol. 1986; 33: 251–300.

4 Roberts RM, Liu L, Guo Q, Leaman D, Bixby J. The evolutionof the type I interferons. J. Interferon Cytokine Res. 1998; 18:805–16.

5 Staeheli P, Puehler F, Schneider K, Göbel TW, Kaspers B.Cytokines of birds: conserved function – a largely different look.J. Interferon Cytokine Res. 2001; 21: 993–1010.

6 Hughes AL, Roberts RM. Independent origin of IFN-α andIFN-β in birds and mammals. J. Interferon Cytokine Res. 2000;20: 737–9.

7 Harrison GA, Young LJ, Watson CM, Miska KB, Miller RD,Deane EM. A survey of type I interferons from a marsupial andmonotreme. Implications for the evolution of the type I inter-feron gene family in mammals. Cytokine 2003; 21: 105–19.

8 Harrison GA, McNicol KA, Deane EM. Interferon α/β Genesfrom a Marsupial, Macropus eugenii. Dev. Comp. Immunol.2004 (in press).

9 Zhou C, Yang Y, Jong AY. Mini-prep in ten minutes. Biotech-niques 1990; 8: 171–3.

10 Page RDM. Treeview: an application to display phylogenetictrees on personal computers. Comput. Appl. Biosci. 1996; 12:357–8.

11 Janke A, Feldmaier-Fuchs G, Thomas WK, von Haeseler A,Pääbo S. The marsupial mitochondrial genome and the evolutionof placental mammals. Genetics 1994; 137: 243–56.

12 Wetzel R. Assignment of the disulphide bonds of leukocyteinterferon. Nature 1981; 289: 606–7.

13 Viscomi GC. Structure-activity of type I interferons. Biotherapy1997; 10: 59–86.

14 Lopez S, Navarro S. Transcriptional repression of the type I IFNgenes. Biochimie 1998; 80: 689–701.

15 Fujii Y, Shimizu T, Kusumoto M, Kyogoku Y, Taniguchi T,Hakoshima T. Crystal structure of an IRF-DNA complex revealsnovel DNA recognition and cooperative binding to a repeat ofcore sequences. EMBO J. 1999; 18: 5028–41.

16 Maniatis T, Du Whittmore LA, W et al. Positive and negativecontrol of human interferon-β gene expression. In:McKnight SL, Yamamoto KR (eds). Transcriptional Regulation,Vol. II. New York: C Press, 1992; 1193–220.

17 MacDonald NJ, Kuhl D, Maguire D et al. Different pathwaysmediate virus inducibility of the human IFN-α1 and IFN-βgenes. Cell 1990; 60: 767–79.

18 Whittemore L-A, Maniatis T. Postinduction turnoff of beta-interferon gene expression. Mol. Cell. Biol. 1990; 64: 1329–37.

19 Bragança J, Civas A. Type I interferon gene expression. Differ-ential expression of IFN-A genes induced by viruses and double-stranded RNA. Biochimie 1998; 80: 673–87.