Non-orthologous gene displacement

consequence of a lack or excess of DMPK protein. Resolution of the alternative mechanisms underlying the devtlopment of DM requites the production~of transgenic mouse models that include an exoanded CfG ;Meaa;hmust be the next goal of

235-238

6 Hoffmann-Radvayi, H. et al. (1993) Hum. Mol. Gene;. 2.

_ 1263-1266 Krahe. R. et al. (1335) Genomics 28. ’ l-14 D Wang.J.Z. etal. (199j) Hmn. Mol. O Genet. 4,X%606 9. Jansen, G. et al. (1996~ Nat. Genet. 13,316-323 10 Reddy,S. etal. (1996) Nat. Genet.

Admowledgements 13.3i4-335 We thank Gill Bates and Sita Reddy for helpful discussions. Work in our II Justice, R.W. et al. (1995) Genes Dev. 3. i34-546 Iabontory is funded by the Muscular 12 Wang, Y.H. et al. (1994) Science Dystrophy Association. ReferenceS f Harper. P. (1989)Mrotonic L@stmply. Majorproblems in 2

:&r~rnlo~r21. Saunders Brook, I.D. etal.(1992) CeN68.

799-noi3

3 Carango, P., Noble,J.E.. Marks. H.G. and Funanage.V.L. (1993)

Gerronrics18.340-348 etal. t 19933 Nat. Gene:. 4. 233-238 5 Fu. Y.H.et al. (1993)Science 260,. 4 Sabouri, LA.

18

Jansen, G. et nl. (1995)Hum.

Mol.

Genet. 4,843-852

19 Shaw. DJ. ct al. (1993)Genomics 18.67%679 20

StJohnston. D. (1995) Ce//81. 161-170.

21 Rastinejad. F., Conboy, M.J.. Rando.

T.A. and Blau. H.M. (1993) Ce1175, 1107-1117 22 Rastineiad. F. and Blau. H.M.(1993) Cell72,‘903-917

Tanela. K.L.et al. C1WjJ.l. Cell Ho!. 128.995-1002 24 Willems. P. (1994)Nut. Genet. 8. 213-215 25 Zeitlin. S. et nl. (1995)NN/. Genet 265.669-671 11.155-163 13 Pearson, C.E. and Sinden. R.R. 26 Duyao. M.P. et al. (1995)Scieuce (1996)Biochemistty35, 5041-5053 269.407-410 14 Timchenko, L.T., Timchenko. N.A.. Caskey.CT. and Roberts, R. (1996) 27 Nasir, J. et nl. (1995) Celi81.81l-823 28 BurriRht.E.N. et al. (1995)Cc//82. Hum. MO/. Genet. 5. 115-121 937-948 IS Boucher. C.A. er al. (1995)Hum. 29 Ike& H. etal. (1996)NW. Genet. Mol. Gem?. 4. 1919-1925 13.196-202 16 Otten. A.D. and Tapscott. SJ. (1995) Pmt. Nat/. Acad. Sci. U. S. A. 12. 30 Bingham. P.M. etal.(1’9%) Nat. Genet. 9. 191-196 54654469 31 The Dutch-Be&n FragileX 17 Shaw, DJ. et al. (1993)J. Med. Consortium (1994)Cell78 23-33 Genet.30. 189-192 23

LETTER

Non-orthologous genedisplacement ,- _+N Unrelated proteins with the same biochemical activity can evolve independentlyt. Without complete genome information, however. it is not possible to ascertain whether a particular function is encoded by homologous or non-homologous genes in two different species. The chance always remains that the counterpart to a given gene in one species has not been sequenced in the other. In an attempt to prove and to assess quantitatively the postulate that dis:inct proteins in different species can be responsible for rhc same hmcticn, we compared the first two genomes to be sequenced completely, those of the parasitic bacteria Huemop~iltcs inJluenzae2 and M>roplasma genituhrm:‘. For this problem to be addressed, the function of the gene in question has to be essential in both organisms. Despite this ]iIIIitatiOn, we found a considerdhle number of cases where apparently indispensable cell functions are encoded by non-onhologous genes. 8y definition, orthologs are genes that are related by vertical

descent from a common ancestor and encode proteins with the same function in different species. By contrast, paralogs are homologous genes that have evolved by duplication and code for proteins with similar, but not identical, functions’. Comparison of the 1703 putative proteins encoded by the H. inJlfrmtzaegenome2.s with the 469 M. genitalitrm protein sequences, revealed that 233 showed sufficiently high similarity for them to be considered s;C~ologs (the criteria for distin_+tishing orthologs from ~an!ogs a.p= _ discussed in Ref. 5, in which the genome of H. hlfltrenrue was compared with the available 75% of the ficberfchia co/i genome). In addition, by comparing the remaining M. genitulitrm and H. ittflttrertzae proteins with those in the non-redundant sequence database, we identified 12 clear-cut cases where essential cellular functions in the two bacteria were encoded by non-orthologous (i.e. unrelated or paralogous) genes (Table 1). We call this phenomenon non-orthologous gene TIC SEPTEMBER1996 VOL. 12 No. 9

displacement, although its origins can be quite diverse. Nucleoside diphosphate kinase (NDK) appears to be a case where an essential function has been taken over by an unrelated protein. This example shows the impact of the concept of non-orthologous gene displacement on the analysis of complete genome sequences. Knowing that a function, in this case the formation of nucleoside triphosphates from diphosphates catalyzed by NDK, is indispensable but that the given genome does not encode homologs of the respective proteins from other organisms, one shot&J sccecn functionally unassigned gene products for likely candidates. In M. genifulitrm, we found two plausible candidates for a novel NDK, both of which contain sequence motifs typical of nucleoside and nucleotide kinases. The case of the repair DNA polymerase shows the complexity of the scenarios leading to non-orthologous gene displacement. It appears that a duplication of the DNA polymense 111(Pal 111)gene in the common ancestor of Gram-positive and Gram-negative bacteria twitich also encoded the

LETTER TABLE 1.Non-orthobgous genes coding forthe samefunction in Mycopbzma genftulfum Ifaemopbilus in~ae

and

H. fnJluarrne

M. g4?mwiunl

- -._ No fnqwmce similarity between M. genitahn Phosphoglycerate

MG43O

PMGI_BACSU PMGI_ECOLI PMGI_MAIZE

and H fn~errzae~ieirr~ HI0757 PMGl_ECOLI PMGMJIUMAN Q@mA) not in G(+)

MG46iI

LDH_BACSU ~HM_HU~

tfctDor ffd~~

LLDDJXOLI G(+f The HI enzyme is distantly related to eukatyotic thee BZ

LPLAJXOLI seu~Lo46W_l

HI0027 tl@EJ

LIPB_ECOLI 551458 (yeast)

E. coli and yeast encode both

None

N10876 ( rrdk)

NDK_ECOLI NDKB_HUMAN

The two predicted ldnases in MG are candidates for this indispensable activity

DPOl_ECOLI DPO1J~CIX.Z

MG encodes two homoloe of DNA polymense III. MG261 is the likely repair polymemse as it belongs ta a putative repair opero&

mm&e

L-lactate dehydrogenase

H11739B

Escheticbia coli encode5 both types of enzymes

Lipoate-protein ligase

MG270

Nucleoside diphosphate kinase

MG?&F’

DNA polymerase. repair

MG261 (dna.8

DP3A_~IN DP3A_ECOLI

HI0856

RNase H

MG262?d

DPOI_BACCA DPOI_HAEIN

HI0138 ( rf?bA): HI1059 ( nrhBl

RNH_ECOLI RNHl_YFAST RNH2_ECOLI MC326_1 @I. caprtcoi..) SC23CDS_13 (yeast)

MG262 is homologous to the 5’-3’ exonuclease domain of DNA polymerase I. It is predicted m replace the two un~l~ted RNases H of HI in primer removal during DNA replicadon

Glycyl-tRNA syntbewse

MG251

SYG_HUMAN

NIO927 0 HIQ924 Q$vs1

SYGA_ECOLI SYGB_ECOLI CFUZOj47_I (Chl3mydia) G(-1

The MG enzyme conrains one subunit. the HI counterpart two

Yeast encodes both types of enzymes

MG26@

fpolA)

types of enzymes

l?uw@s hi AI genitatkm and H. infiwnzae Prolyl-tRNA synthetase

MG283

Cytidine deaminase

MG052

Pyruvate

MG273

dehydrogenase El component LInnd p subunits

MG274

YHIO_YEAST

CDD_BACSU CDD_HU~

HI0729

SYP_ECOLI

@rvs,

YER7_YEAST

HIl3;O

CDD_ECOLI

The MG cytidine deaminase is more closely related to euka~otlc enzymes thzn to those from Gt+) bacteria

ODPl_ECOLI G(-1

Two subunits in C(*) barrerla (including MG) and eukaryotes, once

(cd&

ODPA_BACSU HI1233 ODPA_HUMAN (aceJ.?l ODPB_BACSiJ ODPB_HUMAN

subunit in CC-) bacteria

3 The M. genil~iirf~ and H. fjt~t~~zae proteins are designated as in Refs 3 and 2. respectively. The onhol~ous E. cofl @ne is indicated in parentheses where available. bThe SWISS-PROT code is given wherever available. GW and Gf.+) indicate the species range of the orthologs CGramnegative and Gram-Positive, respectively). C Abbreviations: MG. M. genitalittm, HI, H. inJ7trenzac. d? indicates that there are no detectable ortholop, but the functions are evidently essential for the organisms: thus. the detected similarity to other proteins with similar functions is suggestive but nerds IO be vpritied experimentally.

repair enzyme Pol If. was foIlowed by 3 subsequent elimination of one of the Pal 111pamlogs in the Gram-negative lineage and of Pol I in the Mycoptasmalineage’). As a result, one of the Poi 111pamlogs, which is encoded in the same

operon as two pu~tive repair enzymes, has probably ovenaken the repair polymemse functinn in the mycoplasmas (Ref. 6 and Table 1). Theoreticaliy. one may imagine that in a minimal organism. a single polymerase would he responsible

TIG SEPTEMBER 1996 VOL. 12 No. 9

335

for repli~tion and repair DNA synthesis-. The anatysis of the actual

dt. geuitdimn genes. hcnvcqxz makes a compelling cxse for non-onholo~gous displacement. The situation with IWase H is even mcxe inrtic;rte. fl. it1JllIPilzClL~

and E. colieach encode two RNases H. These enzymes are involved in the removal of RNA primers during DNA replication. which is evidently essential for the ~mpletion of replication. In M. genitnlirtm. there is no homolog of any know RNase H. and the activity is likely to be provided by the ~G262 prorein that is homologous to the N-terminal. Y-3 exonuclease domain of Pof I (Ref. 6 and Table I). This example shows the limitations of the classical defmition of 0~010~ as even though the MG262 protein is more closeiy related to PO! I than to any other protein in H. j?I~~~~zueor other bzicteiia. it is obvious that the functions of these proteins only partially overlap. FiialIy. as two ws of aminoacyl-tRNA synthetases of J*Lgenifalitrm have only eukaryotic orthvIogs, horizontal gene transfer with subsequent displacement of the bacterial enzymes seems to be a plausible mechanjsm (TabIe Il. In addition to the I1 instances of non-orthologous gene displacement unrelated

summarized in Table 1. we found more than IO cases in which it remains unclear whether the ,&I.genitalitrm and H. ittJ7ttmzae genes responsible for a particular function are orthologs or paralogs (data not shown). Even though the respective proteins are highly similar in M. ~enif~lft~~~ and H. inJi’trenzae, they show approximately the same level OF similarity to their eukaryotic or archaeal counterparts, suggesting the possibility of non-orthologous displacement. However, alternative hypotheses. such as an unusual ‘burst-like’ mode of evolution of these genes, cat-m& be ruled out. It is remarkable that as many as 11 non-onhologous genes with the same functions have been found compared with the 233 apparent orthologs in H. itlflrcerrzae and M.genilalirrm. In larger genomes. there are likely to he hundreds of such cases, and they will require special considemtion in the reconstruction of metabolic

pathways and in comparative genome analysis. EUGENE V.KOONIN*, ARUDYIR MUSHEGIAN’ ANDF’EER BORK~ ~rk~embl-heidelberg.de

References 1 Dcolittie. R.F.(19941Tnwh Biochem. Sci. 19, Ii-18 2

Fleischmann, R.D. etal.(199%

3

Science 269. ;iclc,--?12 Fmser, C.M.et al. (1995) Science 270. 397-403

4

Fitch. W.M.f 1970)$@. Zeal. 19, 99-106

6 Koonin. E.V.and Eork. P. (1996) TrendsBiochent. Sci. 21. 128-129 7 MusheRian.A.R.and Koonin. E.V. Pfvc. A&f!.Acad Sri. C! S. A.
r

The 1997 Bfst IEI~onic Chmf&xmce in Cd l3io~ogy dfororgarnizers

7he September 1996issue of TisSis a thematic issue on the RNApoiymen.se II uanscription machinery.Soreral review articles have been specially commissioned on key areas OFthis rapidly expanding topic, including: RNApoIynaerase II triu~~Iptian @ R. Korniqy

’ This September issueof

trends in CELL BIOLOGY i cat&s an interview with Barry Hardy on virtual c0nfetences. In this in~view, Dr Ha&~ issues an invi~& to scienti~s working in cell biology and i .nzki*&?d&Ul*#ines t;J ‘&come involved in the 1997 Fii International Ceil Biology Electmnic Conference. This is a new venture, and requires a cell biological advisory ~~~ to assist r&h organizing the meeting sessions and inviting peters. Anyone interested in participating, at any level, should contact either Barry Hardy ([email protected]) or the W&S fn CELLBIOLOGY editorial team (at tcb~e~~ier.co.uk or the editorial address listed at the front of thjs journal>.

TAFs mediate tramzriptional activation and ~~~~ 4 C.P. Z%i@zerti.%dR. 7Jfati The

RNA polymcrww Bl gene?& elon@ion Iactors by D. Reines, J. W. Conuway and R.C. ~o~~u~~v Thehumangencralc0&sors b K. Miser and M. Mefrferentst

The

multiple rotas of tramicdptiad~ factor TFnH 19J.Q. S~~tn~, P. Vi& and&M. i@y Mediator of trans&flonnl rqulation t.y S. BjcWdundand Y-S,Ktm ~e~le~~~~~~f~~ mnscription by RNApol~merase II @ R.G. Roeder

TIG SEMIZMIWI 1996 Vol.. 12 No. 9

336

an&o1