Proc. Natl. Acad. Sci. USAVol. 81, pp. 6642-6646, November 1984Biochemistry

Characterization of the genes specifying two metacyclic variableantigen types in Trypanosoma brucei rhodesiense

(antigenic variation/monoclonal antibodies/tsetse flies/DNA sequences/Southern hybridization)

MICHAEL J. LENARDO*t, ALLISON C. RICE-FICHT*, GREGORY KELLY*, KLAUS M. ESSERt,AND JOHN E. DONELSON*Departments of *Biochemistry and of tMedicine, University of Iowa, Iowa City, IA 52242; and tDepartment of Immunology, Division of CommunicableDisease and Immunology, Walter Reed Army Institute of Research, Washington, DC 20012

Communicated by William Trager, July 11, 1984

ABSTRACT Bloodstream trypanosomes evade the im-mune system of their mammalian host by sequentially express-ing a large number of different variable surface glycoproteins(VSGs). In contrast, metacyclic trypaposomes, the final devel-opmental stage in the tsetse fly, express a much more restrict-ed set of VSGs. These metacyclic VSGs are the first to beexposed to the immune system of the mammalian host afterinfection and may offer the potential for the eventual develop-ment of a vaccine. We have identified cDNAs for two VSGs incDNA libraries prepared from amplified metacyclic popula-tions of Trypanosoma brucei rhodesiense and show that theycorrespond to two different metacyclic serotypes. Determi-nation of the cDNA sequences shows that metacyclic VSGmRNAs are similar to VSG mRNAs expressed during thebloodstream stage. Southern blots demonstrate that the meta-cyclic VSG genes are located near chromosomal telomeres. Noevidence of gene rearrangement associated with expression ofthese VSGs was found.

Trypanosoma brucei rhodesiense, like other salivarian try-panosomes, has a surface coat comprised predominately of asingle antigenic molecule, the variable surface glycoprotein(VSG). Continual changing of the VSG, a process called4ntigenic variation, permits trypanosomes to escape destruc-tion by their host's immune system (1). Studies on the genet-ic mechanism of antigenic variation reveal that, in some cas-es, an expression-linked extra copy (ELC) of a silent VSGgene is inserted near a chromosomal telomere where it istranscribed (2, 3). In other cases, the transcription of the pre-viously expressed VSG gene stops and transcription of aVSG gene already near a telomere begins (4). The telomere-linked VSG genes are usually flanked by 5- to 30-kilobase(kb) regions that are devoid of restriction endonuclease sites.These regions may be comprised of simple repetitive se-quences or may contain modified nucleotides, or both (2, 5).VSGs first appear on the surface of the trypanosome dur-

ing the metacyclic stage of its life cycle. Metacyclic trypano-somes occur in the salivary glands of the tsetse fly and areinjected into a mammalian host during the fly's bite. Previ-ous studies have shown that metacyclic organisms consist of10-15 different antigenic types (6, 7). In a given serodeme,this restricted set of metacyclic variable antigen types(MVATs) is repeatedly expressed at the metacyclic stage in-dependent of the bloodstream variable antigen type (VAT)that is ingested by the fly.Each tsetse fly possesses several orders of magnitude few-

er metacyclic organisms than are needed to isolate enoughVSG or mRNA for experimentation. However, metacyclicVSGs of T. brucei subspecies are expressed in the blood-

stream of mice up to 5 days after infection begins (8, 9). At 5days the organisms have multiplied so that a sufficient num-ber expressing metacyclic VSGs can be obtained for bio-chemical study. Between day 5 and day 7 of infection, theparasites switch from the expression of MVATs to earlybloodstream VATs.We have constructed cDNA libraries using mRNAs isolat-

ed from day 5 trypanosomes and have identified cDNA se-quences for two VSGs expressed by the metacyclic popula-tion. These sequences were used to determine the primarystructure of these two VSGs and to investigate the molecularbasis for the expression of a restricted subset of VSG typesin metacyclic trypanosomes.

MATERIALS AND METHODSTrypanosomes. WRATats 1.1, 1.12, and 1.14 are clones of

T. b. rhodesiense, stock LVH/75/USAMRU-K/18 (10).Matacyclic trypanosomes were obtained from WRATat 1.1organisms as described (8). MVATs 1 through 14 correspondto WRATats 1.21 through 1.34, respectively, in standard no-menclature (11).

Preparation of Day 5 Trypaposomes Expressing MetacyclicVSGs. Day 5 trypanosome populations were generated fromeither a mixed MVAT population or populations with someMVATs selectively neutralized by monoclonal antibodies(MAbs). To enrich for one MVAT serotype in a day 5 popu-lation, 0.1 ml of each of the MAbs for the MVATs to beexcluded were pooled, diluted to 2.5 ml in M199 medium(M.A. Bioproducts, Walkersville, MD), added to 3 xmetacyclic trypanosomes in 0.5 ml of M199 medium, and in-cubated for 15 min on ice. Thirty male C57 BL/6J mice (TheJackson Laboratory) were inoculated i.p. with 0.1 ml of thetrypanosome-MAb mixture. A similar protocol was used todeplete a metacyclic population of one MVAT.Trypanosomes were isolated and used for indirect immu-

nofluorescent assay as described (8, 10).Recombinant DNA Techniques. Trypanosome poly(A)+

RNA was prepared by standard procedures (12). DNA wascollected from the gradients on which RNA was isolated anddesalted on a Bio-Gel A-15m column. cDNA libraries wereconstructed by using the technique of Land et al. (13) andwere screened by established procedures (12). GenomicSouthern blots and RNA dot blots were conducted as de-scribed (12, 14). The Maxam and Gilbert technique was usedfor DNA sequence determinations (15). BAL-31 nuclease di-gestions were performed with 0.2 units of enzyme (New En-gland Biolabs) per 1 jig ofDNA at 30'C for variable times inthe recommended buffer (12).

Abbreviations: VSG, variable surface glycoproteins; ELC, expres-sion-linked extra copy; VAT, variable antigen type; MVAT, meta-cyclic VAT; MAb, monoclonal antibodies; kb, kilobase.

6642

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Oct

ober

17,

202

1

Proc. Natl. Acad. Sci. USA 81 (1984) 6643

RESULTS

Assignment of VSG cDNAs to Metacyclic Stereotypes. TwocDNA libraries were prepared from day 5 metacyclic popu-lations and screened by a "plus/minus" strategy. Radioac-tive first-strand cDNAs from day 5 metacyclic (expresser)mRNA and from bloodstream clone WRATat 1.12 (nonex-presser) mRNA were used as probes. Several clones that hy-bridized strongly to the day 5 metacyclic cDNA but not tothe bloodstream WRATat 1.12 cDNA were detected. Theclones bearing larger inserts were used in RNA hybrid-selec-tion experiments. Unfortunately, all efforts to show thatmRNAs that hybridized to these cDNA inserts would directthe in vitro translation of proteins recognized by antibodiesagainst whole metacyclic organisms (either monoclonal orpolyvalent) were unsuccessful for unknown reasons. There-fore, another identification strategy was devised.

Table 1 shows that 14 different MVATs arising in tsetseflies infected by T. b. rhodesiense can be defined by a panelof MAbs raised against metacyclic trypanosomes (unpub-lished results). Greater than 95% of the metacyclic popula-tion in the salivary gland of the tsetse fly contains only theseMVATs. In Table 1, the MVAT 7+ population was obtainedby neutralizing metacyclic trypanosomes of all of the majorMVATs except MVAT 7. Conversely, the MVAT 7- andMVAT 6- populations were generated by neutralizing onlythese serotypes. In the enrichment for MVAT 4 organisms(MVAT 4+), the MAbs against MVATs 1 and 6 only partiallyneutralized these organisms as shown in Table 1.To correlate cDNA clones with a given MVAT, RNA dot

blots (14) were performed with mRNAs from the day 5 popu-lations shown in Table 1. Fig. 1 shows some examples ofthese dot blots. The left-most dot blot is a control demon-strating that a cDNA for the IaTat 1.2 VSG (16) hybridizedonly to mRNA isolated from IaTat 1.2 trypanosomes. In thenext dot blot, a probe containing the trypanosome a- and /3-tubulin genes indicates that the amount of RNA from eachpopulation spotted on the filter was approximately the same.Three cDNAs were shown to contain sequences unique tospecific MVATs. The cDNA clones called 3L41 and 11R45hybridized to MVAT 7-enriched mRNA and the 2L11 cDNA

Table 1. MVAT composition of various day 5 preparations of T.b. rhodesiense organisms

Percentage of fluorescing organisms

MVAT MVAT MVAT MVAT MVAT MVATantibody mix 7+ 7- 6- 4+

1 15.0 - 29.02 7.0 -3 4.0 1.0 0.54 7.0 1.0 1.0 24.05 10.0 - 0.5 0.5 1.06 48.0 1.0 69.0 46.07 15.0 87.0 - 47.08 5.09 1.0 1.0 2.0 1.010 0.4 7.0 3.011 3.0 1.012 0.113 ND 2.0 3.0 2.0 2.014 ND ND 7.0 10.0 ND

Total 115.1* 90t 89.5T 66T 104*

Values are based on indirect immunofluorescent assays of be-tween 100 and 1000 organisms. A dash indicates that no organismsfluoresced. ND, not determined.*Total is >100%o because of small accumulative errors in scoringpositive organisms.tTotal is <100% because of the presence of trypanosome VATsother than MVATs.

clone hybridized to MVAT 4-enriched mRNA. The relativehybridization intensities indicate that these cDNAs corre-spond to RNAs that encode major trypanosome proteins.The hybridzation pattern of these cDNAs to other day 5 pop-ulations also reflected the relative amounts of MVAT 7 andMVAT 4 mRNA in these populations (see Table 1). For ex-ample, the 3L41 and 11R45 cDNAs did not hybridize withthe MVAT 7- and MVAT 4+ populations, which were de-void of MVAT 7 organisms.The analysis of the 2L11 cDNA hybridization pattern was

complicated by the relative abundance of MVAT 1 andMVAT 6 organisms in the MVAT 4+ population. It is unlike-ly that the 2L11 cDNA corresponds to MVAT 1 because IFAanalysis showed that this serotype was not present in theenriched population used to construct the cDNA libraryfrom which the 2L11 cDNA clone was isolated (data notshown). Also the 2L11 cDNA did not encode the MVAT 6VSG because it did not hybridize to RNA from the MVAT7- population, which contains 69% MVAT 6 organisms. Fur-ther confirmation of the assignment of 2L11 cDNA toMVAT 4 is the finding that an Escherichia coli-expressedfusion protein containing a 2L11-encoded amino acid se-quence bound a MAb specific for MVAT 4 (unpublished re-sults).Sequence Analysis of the MVAT cDNAs. To compare the

metacyclic VSG structure and mRNA untranslated regionswith those from bloodstream organisms, the sequences ofthe three metacyclic cDNAs were determined and used todeduce the amino acid sequence as shown in Fig. 2. The2L11 cDNA (henceforth called the MVAT 4 cDNA) containsa long open reading frame that encodes a COOH-terminalhomology region as found in bloodstream VSGs (17). At the5' end of the coding strand is the 35-nucleotide leader se-quence thought to occur at the 5' ends of all VSG mRNAs(19). We tentatively have assigned the initiator methioninecodon based on its location and its following sequence,which is characteristic of a signal peptide (20).The sequences of 3L41 and 11R45 cDNAs (henceforth

called MVAT 7 cDNAs) hybridize identically in RNA dotblots but do not overlap. The 11R45 cDNA sequence en-codes a COOH-terminal homology region and an 82-nucleo-tide 3' untranslated region similar to that found in other VSGmRNAs (2, 17). It contains, however, an imperfect copy ofthe usually conserved 16-base sequence found at the 3' endof bloodstream VSG mRNAs-i.e., 2 of the 16 positions aredifferent (2). To prove that the 3L41 and 11R45 cDNAsarose from the same mRNA species, a 0.95-kb (Cla I/BamHI) genomic fragment expected to span the two cDNAswas cloned and sequenced. As shown in Fig. 2, the codingsequence of the MVAT 7 VSG consists of the 3L41 and11R45 cDNAs plus 171 nucleotides deduced from the geno-mic fragment. Furthermore, sequences in the cloned geno-mic fragment flanking these 171 nucleotides exactly matchedthe cDNA sequences, suggesting that this genomic regionencoded the MVAT 7 VSG (data not shown). Interestingly,the MVAT 7 VSG does not have several conserved cysteineresidues usually found in other VSGs of its homology group(17). There remains an unlikely possibility that if an ELCgene were not detected (see Discussion), the sequence pro-vided from the genomic region could differ from the se-quence of the mRNA.

Analysis of Genomic Regions Encoding the MVAT 4 andMVAT 7 Genes. The MVAT 4 and MVAT 7 cDNAs wereused as probes in genomic Southern blots to determine ifrearrangements of the metacyclic VSG genes were associat-ed with their expression as for bloodstream VSG genes (2-4.16). Fig. 3 shows an experiment in which genomic DNAsfrom WRATat 1.1, WRATat 1.14, and MVAT 4' (see Table1) were digested with restriction enzymes and probed with aPst I restriction fragment of MVAT 4 cDNA that comprises

Biochemistry: Lenardo et aL

Dow

nloa

ded

by g

uest

on

Oct

ober

17,

202

1

6644 Biochemistry: Lenardo et al.

1.2 Tubulin 3L41 1 1 R45 2L11

7,0

7

1.2 4.

FIG. 1. Autoradiogram of an RNA dot-blot experiment. Each filter contains 12.5 ng, 6.2 ng, and 3.1 ng (left to right) of total poly(A)+ RNAfrom the MVAT 7+, MVAT 7-, MVAT 6-, and MVAT 4+ populations (see Table 1) and from T. b. brucei clone IaTat 1.2 (16). The fiveradioactive probes from left to right were: the IaTat 1.2 VSG cDNA (1.2), the trypanosome a- and 3-tubulin genes (tubulin), and three clonedcDNAs from the day 5 metacyclic cDNA libraries called 3L41, 11R45, and 2L11.

most of the cDNA sequence. WRATat 1.1 is the blood-stream trypanosome clone ingested by the flies that gave riseto the metacyclic populations and WRATat 1.14 is an ear-ly bloodstream VAT obtained after fly transmission ofWRATat 1.1 (unpublished results).Each of the four restriction enzymes shown in Fig. 3

cleaved once in the MVAT 4 cDNA. The probe hybridizedto two restriction fragments when the genomes were digest-

ed with Pvu II, HindIll, or Sal I, indicating that there is onlyone gene for the MVAT 4 VSG in the genome. It hybridizedto a single genomic Pst I fragment because a Pst I fragmentof the cDNA was used as the probe. There is no apparentdifference in the sizes or number of hybridizing fragments inthe three genomes. The relative intensities of some bandsvaried slightly from one blot to another (for example, com-pare the two bands in the HindIll digest of WRATat 1.14

TAACGCTATTATTAGAACAGMCTGTACTATATTiGAAGTCAAGAGCCAGCGTGGTCGCCGCGAGTATCCATACAAGGGCGTCTTAGGCAGCAAA ACC AGCA GAGG GCTTCCGCG

TAGTTAGTGGArGCGGTTCTG M P T A G A V L L L A T A A V L S R P A A A A N E K K P L T I S A A G A V C G F S N E L K E V A S F A A T K..SEPTMSTRVQQATSCVLLI IGCTNYAASKQQTENNAVCDTPCKCLGR

VNAYLTETEQLGTLAVDFCAA IFNGNGKPTAGEVYAALLAAKTQTDRQAQQQKLVTGALLASSLAVQKAYLEGALEHARQSIKAFASSSRQHTIAAGATNTKL----SGHHPNQGRRGRRRSSS

LCSQQAGHIIESFIIHVFYQAHRDATTTC IYKSDEHQRKAADVPCVNNGGGLNKI TI TTRATKPQARPNNSKGNSPSKRAGGAVRGETPASGRLKAAKLLKSSNYETATFTPTEMTI QNRRACEQVDE

LNSKYKA I VAGTTPNGAAAEKN CKLVN GEQSQNVFLASQTTN I ALKWGDG LLTVP I GDTA ISASKEYRFDELNFDKEQG I PTPALTLE ISMGCQKNPGDTGGACDGTNADDGVVTKLTFTTAAPVA

3L41e&

SNWEPNNKDS IASGNYKECAAALEQVPPRQSTATAAASKLLKLEAEDDPKLEEELI SKNSFYGETSPLVAGAYKKTSKASLS I SNRTEEGRVANSKAAHEAAKRLMA I KQPSDLSTYTDDSS FALLl_ 11 R45

DFP I SNVKI TATDFSSLHSKLKSYRKKHTADGHQLLKAKLQHLEKQMQMNATACKLGAE I SSGVGQLAMKPPLGAELTPALRDQVNKFITDNYGANEGDFRSKFLAKTEQAAVFYLDGKSKKTKKI

Y V VEP STAKTTTSGRCEGKAKTACPKSDYRQRI1EKDGKEECKPKSGEEQKTQITTGAGEGAADKKEJCSELESKSELVTASGYAFFKG I HTEQAEKV NPRSQGNPETAENKKEGGN AAKPFCSTIQNQTUJ

Vlk G LE PE C T P E K IE G E T K D HI V N F T L MI S AA F M. .. .

,G V G T P PTG K V G I E G K Q D S S F LLS Q F AL V V A A F A A LL F * AATCAAATMCCCCCTCAAATTATTTACTTCT

CTAACA,1TTMGCTAMACATATMTrAACACCTAAGAGTTACCCGAAAAAAAAMAAAAA

FIG. 2. Predicted amino acid sequences and flanking untranslated nucleotide sequences for MVAT 4 (2L11 cDNA) on the upper line andMVAT 7 (3L41 cDNA, 11R45 cDNA, and 171 nucleotides of a genomic clone) on the lower line. Dashes indicate a region where compressionsin the MVAT 7 sequencing gels prevented an unambiguous determination of the sequence. Boxes show the conserved 35-base sequence at the5' end of the 2L11 cDNA and conserved amino acids in the COOH-terminal homology regions. The thick vertical line shows the probablecleavage sites of the COOH-terminal hydrophobic tails; arrowheads indicate usually conserved cysteines (17). Underlined is the conserved 16-mer in the 3' untranslated region of 2L11 cDNA; asterisks indicate base changes from the canonical sequence (2, 18).

Proc. NatL Acad Sci. USA 81 (1984)

Dow

nloa

ded

by g

uest

on

Oct

ober

17,

202

1

P/roc. Nat. Acad. Sci. USA 81 (1984) 6645

Pvu Hd Pst Sal1 4+14 1 4 14 1 4 14 1 4 14

11.81118__* *___ _ -11.8

...~.. -51

3.0-

FIG. 3. Autoradiogram of a Southern blot using a 1508-base-pairPst I fragment of the MVAT 4 cDNA as the radioactive probe. Ge-nomic DNAs (2.5 ,ug per lane) from bloodstream trvpanosomeclones WRATat 1.1 and WRATat 1.14 (lanes 1 and 14. respectively)and from MVAT 4- (lanes 4+) were digested with Pi- II. Hindll(Hd). Pst I, or Sal I. Numbers indicate fragment sizes in kb.

DNA in Figs. 3 and 4A), but these differences were not re-producible and their significance, if any, is not known.A cluster of several restriction sites of =10.5 kb beyond

the 3' end of the gene suggested that the gene was near achromosomal telomere as are expressed bloodstream VSGgenes (2-4, 16). To test this possibility, the WRATat 1.14and MVAT 4' genomic DNAs were digested with BAL-31nuclease for various lengths of time prior to cleavage withHindI11 and Southern analysis (Fig. 4A). The 3.0- and 11.7-kb HindI1 fragments were present before BAL-31 digestion

AW14 DNA M4+ DNA

M 0 2 4 8 12 16 0 4 12 M

I.a40

B Cta I

0 2 4 8 12

14.4- _

7.8- 11-~~*Wmam _ _ __ 3.0

_ _

0

FIG. 4. Autoradiogram of Southern blots with the same MVAT 4cDNA probe as in Fig. 3 (A) or the 3L41 cDNA (MVAT 7) as probe(B). In A, genomic DNA isolated from WRATat 1.14 (W14) andMVAT 4+ (M4+) was digested with BAL-31 for 0. 20. 40. 80, or 120min (lanes 0, 2. 4, 8, or 12, respectively). At 120 min. the ratio ofBAL-31 to WRATat 1.14 DNA was doubled, and the incubation wascontinued until 160 min (lane 16). The DNAs then were cleaved withHindIII and applied to the agarose gel. In B. WRATat 1.14 genomicDNA was digested with BAL-31 for up to 120 min and cleaved withCla I, an enzyme that cleaves the cDNA once. Lanes marked Mcontain DNA size markers; sizes of fragments are given in kb.

(the 0 lanes), but the 11.7-kb fragment was rapidly degradedby BAL-31, while the 3.0-kb fragment was affected only af-ter 160 min of digestion. This provides strong evidence thatthe MVAT 4 VSG gene is situated near a naturally occurringdouble-stranded break in the DNA, presumably a chromo-somal telomere (21). Southern analyses with the MVAT 7cDNAs as probes also showed the presence of a single chro-mosomal gene for the MVAT 7 VSG, which appeared onfragments of identical size in all genomes tested (data notshown). As shown in Fig. 4B, BAL-31 digestion indicatesthat the MVAT 7 VSG gene is also telomere-linked.

Fig. 5 shows abbreviated restriction maps that summarizethe results of a large number of Southern analyses withMVAT 4 and MVAT 7 cDNA probes. Note that the regionsbetween the two genes and the telomeres seem to be "bar-ren" of restriction sites but that the regions upstream of thegenes appear to have a normal complement of restrictionsites. The dissimilarity of restriction sites upstream of eachof the two genes suggests that they are located near differenttelomeres.

DISCUSSIONThe study of genes encoding VSGs from trypanosomes atthe metacyclic stage is of interest from two standpoints.First, it permits study of the coordinated expression of a re-stricted subset of VSG genes at a particular time in the devel-opmental cycle of the organism. Second. the limited subsetof VSG types expressed by metacyclic organisms offers theprospect of a vaccine against trypanosomiasis. This reportaddresses the first of these considerations and describes theisolation of cDNAs for two VSG serotypes expressed onmetacyclic organisms. These cDNAs have all of the charac-teristics of those for VSGs expressed at the bloodstreamstages-i.e., NH,-terminal amino acid sequence heteroge-neity. COOH-terminal amino acid homologies and a hydro-phobic "tail." potential sites of glycosylation. and conservedsequence blocks in the 5' untranslated and 3' untranslatedregions of the mRNA (2-6, 16. 17). It is likely. therefore, thatall of these features are important for the expression andproper functioning of the VSG molecule at both the metacy-clic and bloodstream stages. Moreover, these features ap-pear to be conserved between the VSGs of T. h. rihodesiensedescribed here and those of the closely related subspecies T.b. brucei which have been investigated previously (2-6, 16.17).The genes for these two MVATs are located in genomic

regions that are similar to regions bearing VSG genes ex-pressed at the bloodstream stages-i.e., they are telomere-linked, and the region between the VSG coding sequenceand the telomere is nearly devoid of restriction sites. Down-stream "barren" regions are predominately composed of re-peats of the hexamer sequence C-C-C-T-A-A (22, 23). Thenumber of these repeats can change and lead to differencesin the sizes of telomere-linked restriction fragments contain-ing bloodstream VSG genes from one generation to the next(22. 23). We observed no size differences in the telomericrestriction fragments containing the MVAT 4 and MVAT 7genes. If the sizes of these barren regions are expanding andcontracting during replication, it is possible that the cloneswe tested coincidentally have the same-size barren regions.

Unlike most telomere-linked bloodstream VSG genes,these two MVAT VSG genes are not preceded by an up-stream barren region. Genomic Southern blots using manyrestriction enzymes indicate a continuous distribution of re-striction sites in front of these genes (not shown). Compari-son of the pattern of restriction sites upstream of these twoMVAT VSG genes to other telomere-linked bloodstreamVSG genes does not reveal any obvious similarity (2-4).Therefore, each of the MVAT VSG genes appears to be at aunique telomere-linked position.

Biochemistry: Lenardo et al.

11.7

Dow

nloa

ded

by g

uest

on

Oct

ober

17,

202

1

6646 Biochemistry: Lenardo et al.

kb -8

MVAT 4

MVAT 7

-6 -4 -2 0 2 4 , 10 12 14

S Pt P H S P H Pt X)If I I I...anx

C x P R C R PBII I I I

FIG. 5. Abbreviated restriction endonuclease maps in the regions of the MVAT 4 and MVAT 7 VSG genes as derived from Southern blotsincluding those shown in Figs. 3 and 4. Cleavage sites for Sal I (S), Pst I (Pt), Pvu 11 (P), HindIll (H), Xba I (X), Cla I (C), EcoRI (R), andBamHI (B) are shown. The thick portions of the lines indicate the coding regions, the solid circles on the right end indicate chromosomaltelomeres, and the dotted lines on the left end indicate that the DNA extends beyond the regions mapped.

In considering the expression mechanism of these twoMVAT VSGs, one possibility is that they are expressed viaELC genes that have disappeared from the genome in day 5trypanosomes. Unfortunately this possibility cannot be com-pletely discounted because we cannot obtain enough orga-nisms at earlier times of infection (for example, day 3 or 4) toanalyze their genomes. However, the occurrence of a meta-cyclic ELC gene seems unlikely. We have found the VSGmRNA turns over in less than 24 hr (16). Since some para-sites still express metacyclic VSGs as late as day 7 (unpub-lished results), at least a fraction of the organisms should stillhave the ELC present on day 5. Much longer exposures ofthe Southern blots than those shown in Fig. 3 did not revealan ELC gene for either MVAT VSG investigated.The lack of evidence for an ELC gene suggests that these

two MVAT VSG genes are among those genes that can beexpressed without duplicative transposition (18). Anotherexample of this is the bloodstream AnTat 1.6 VSG, whoseantisera cross-react with 6-8% of the corresponding meta-cyclic population and whose telomere-linked gene can be ex-pressed without duplication (24). This particular gene can belost from the genome, which may be related to the reportedinstability of some MVAT repertoires (25).How the trypanosome expresses only a reduced subset of

VSG genes at the metacyclic stage remains unclear. Thesetwo telomere-linked MVAT VSG genes differ from telo-mere-linked bloodstream VSG genes in that (i) they are notpreceded by barren regions, (ii) they are followed by barrenregions that may not undergo much size variation, (iii) their3' untranslated regions can contain mutations in a conserved16-mer sequence usually found in functional VSG mRNAs(18), and (iv) they are among the rare examples of telomere-linked VSG genes for which there are no other related genesin the genome (26). It remains to be seen if any of these fea-tures are related to the developmental expression of thesegenes. Further characterization of these and other metacy-clic VSG genes will be necessary to elucidate the mecha-nisms that regulate their transcription.

The authors thank Maurice Schoenbechler and Margaret Mead-ows for excellent technical assistance; Dr. Lyman Roberts, Dr. JackGingrich, and Larry Macken for production of infected tsetse flies;Kimberly Brown for performing some of the Southern analyses; LynYarbough for providing the cloned tubulin genes; and Dr. JohnHoak for support and encouragement. This work was supported byU.S. Army Grant DAMD 17-82-C-2228; National Institutes ofHealth Grants Al 18954, Al 06935, and HL07344; and the BurroughsWellcome Foundation.

1. Vickerman, K. (1978) Nature (London) 273, 613-617.2. Borst, P. & Cross, G. A. M. (1982) Cell 29, 291-303.3. Pays, E., Van Assel, S., Laurent, M., Dero, B., Michiels, F.,

Kornenberger, P., Matthyssens, G., Van Meirvenne, N., LeRay, D. & Steinert, M. (1983) Cell 34, 359-369.

4. Williams, R. O., Young, J. R. & Majiwa, P. A. 0. (1981) ColdSpring Harbor Symp. Quant. Biol. 45, 945-949.

5. Raibaud, A., Gaillord, C., Longacre, S., Hibner, U., Buck, G.,Bernardi, G. & Eisen, H. (1983) Proc. Natl. Acad. Sci. USA80, 4306-4310.

6. Le Ray, D., Barry, J. D. & Vickerman, K. (1978) Nature (Lon-don) 273, 300-302.

7. Crowe, J. S., Barry, J. D., Luckins, A. G., Ross, C. A. &Vickerman, K. (1983) Nature (London) 306, 389-391.

8. Esser, K. M., Schoenbechler, M. J. & Gingrich, J. B. (1982) J.Immunol. 129, 1715-1717.

9. Hajduk, S. & Vickerman, K. (1981) Parasitology 83, 609-621.10. Campbell, G. H., Esser, K. M., Wellde, B. T. & Diggs, C. L.

(1979) Am. J. Trop. Med. Hyg. 28, 974-983.11. Lumsden, W. H. R. (1982) Syst. Parsitol. 4, 373-376.12. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Labora-tory, Cold Spring Harbor, NY).

13. Land, H. M., Grez, H., Hansen, H., Lindermaier, W. &Schuetz, G. (1981) Nucleic Acids Res. 9, 2251-2265.

14. Thomas, P. (1983) Methods Enzymol. 100, 253-266.15. Maxam, A. M. & Gilbert, W. (1980) Methods Enzymol. 65,

499-560.16. Murphy, W. J., Brentano, S. T., Rice-Ficht, A. C., Dorfman,

D. & Donelson, J. E. (1984) J. Protozool. 31, 65-73.17. Rice-Ficht, A., Chen, K. & Donelson, J. (1981) Nature (Lon-

don) 294, 53-57.18. Borst, P., Bernards, A., Van der Ploeg, L. H. T., Michels,

P. A. M., Liu, A. Y. C., De Lange, T. & Kooter, J. M. (1983)Eur. J. Biochem. 137, 383-389.

19. Dorfman, D. & Donelson, J. E. (1984) Nucleic Acids Res. 12,4907-4920.

20. Kreil, G. (1981) Annu. Rev. Biochem. 50, 317-348.21. De Lange, T. & Borst, P. (1982) Nature (London) 299, 451-

453.22. Blackburn, E. H. & Challoner, P. B. (1984) Cell 36, 447-457.23. Van der Ploeg, L. H. T., Liu, A. Y. C. & Borst, P. (1984) Cell

36, 459-468.24. Laurent, M., Pays, E., Delinte, K., Magnus, E., Van Meir-

venne, N. & Steinert, M. (1984) Nature (London) 308, 370-373.

25. Barry, J. D., Crowe, J. S. & Vickerman, K. L. (1983) Nature(London) 306, 699-701.

26. Parsons, M., Nelson, R. G., Newport, G., Milhausen, M., Stu-art, K. & Agabian, N. (1983) Mol. Biochem. Parasitol. 9, 255-289.

- Ia a a

Proc. NatL Acad Sci. USA 81 (1984)

Dow

nloa

ded

by g

uest

on

Oct

ober

17,

202

1