ELEGCOCT 19i

Am.. J. Trop. Mled. Hyg.. 3203). 1983, pp. 577-S89Cop)Tight 1983 by The American Society of TropicaJ Medicine and Hygiene

RNA FINGERPRINTING AS A METHOD FOR DISTINGUISHINGDENGUE 1 VIRUS STRAINS*

PATRICIA M. REPIK,t JOEL M. DALRYMPLE,t WALTER E. BRANDT,tJACK M. McCOWN,t AND PHILIP K. RUSSELL*

tU.S. Army Medical Research Institute of Infectious Diseases, Fort Detrick,Frederick, Maryland 21701, and t Walter Reed Army Institute of Research,

Washington, D.C. 20012

Abstract. Virion RNAs of 12 geographically distinct dengue type I (DEN-I) virus isolateswere clearly unique by RNA fingerprinting. Isolates from the same geographic area were verysimilar but differed from those of other areas, allowing us to establish three geographicalgroupings based upon percent shared oligonucleotides. Three Caribbean strains were virtually

~ identical (85-91% homologous oligonucleotides) whereas Pacific/S.E. Asian strains exhibitedconsiderably less homology to one another (44-49%). The Pacific/S. E. Asian strains exhibitedlittle relationship (20-30%) to the Caribbean and African strains. A Sri Lankan isolate dis-played a relatively high degree of homology to Nigerian isolates (60-66% homologous oligo-

S nucleotides), suggesting that the Sri Lanka DEN-I infection originated from Africa. A 1978Nigerian DEN-I isolate and the 1969 Sri Lankan strain each exhibited greater than 50%homology with a 1977 Jamaican strain. The similarities observed between the African/SriLankan and Jamaican strains suggest that the DEN-I virus which caused the 1977 Jamaicanepidemic may have originated from Africa or Sri Lanka. The RNA fingerprint is a uniquecharacteristic of DEN-I strains from a particular geographic region, suggesting this techniqueas a useful tool for dengue epidemiological investigations. J

Dengue (DEN) viruses types 1, 2, 3, and 4 are analysis to determine whether different geograph-distinguishable serologically by neutralization and, ical and temporal isolates of DEN type I (DEN-in some circumstances, by complement-fixation 1) virus display unique and distinguishable fin-tests. 3 More recently, oligonucleotide finger- gerprint patterns. Such capability should provideprint analysis has shown that prototype viruses of a powerful tool for the epidemiological study ofeach of the four DEN serotypes exhibits a unique dengue virus outbreaks.RNA fingerprint. The fingerprints were quite dis-similar and, on average, only 7% of the large re- MATERIALS AND METHODSsolvable oligonucleotides were shared amongthem.4 RNA fingerprint analysis has also been used Virus propagation and purificationto successfully differentiate viruses within a se-rotype, e.g., vesicular stomatitis viruses (Indiana The origins and passage histories of DEN virus

and New Jersey serotypes),8 La Crosse viruses,0 strains used in this study are listed in Table 1. All

influenza A HINI viruses,7 St. Louis encephalitis viruses were originally isolated from human se-

viruses,' Rift Valley fever viruses,' and Pichinde rum. Following the passages indicated in the ta-

viruses.' ble, each isolate was passaged two additional timesThe purpose of these investigations was to char- in Aedes albopictus cell culture prior to RNA

acterize DEN virus RNAs by RNA fingerprint preparation. Ae. albopictus cells, derived fromIgarashi clone C6/36" were grown at 2g*C in thepresence of CO2 in 150-cm2 flasks containing Dul-

Accepted 9 September 1982. becco's modified Eagle medium (high glucose)(Gibco) supplemented with 5% heat-inactivated

* The views of the authors do not purport to reflect fetal bovine serum (FBS) and 0. 1 mM non-essen-the positions of the Department of the Army or the De- tial amino acids (Gibco).partment of Defense. Dengue-I virus strains were propagated by in-

Address reprint requests to: Dr. Patricia Repik, Vi- feeting Ae. albopictus cell monolayers (2 × l07rology Division, U.S. Army Medical Research Instituteof Infectious Diseases, Fort Detrick, Frederick, Mary- cells) at an input multiplicity of 0.08 to 0.3 plaque-land 21701. forming unit (PFU) per cell. Virus was allowed to

577

OM FILE COP" 83 10 19 065

C-'

578 REPIK ET AL.

TABLE I

Dengue I strains used in this study

Odjinal Wisoo

Designation yeau Location passage htstoryO

CV 1636/77 1977 Jamaica 5-sm, 1-LLC-MK2 tc777849 1977 Bahamas I-AP.61 tc purified large plaque,

1-LLC-MK2 tc

CV 58/81 1981 Jamaica I-AP-61 tc, 1-C6136 tclB-H 28326 1968 Nigeria 21I-smIB-H 13689 1978 Nigeria 3-AP-61 tc, 1-C6/36 tcHAWAIIAN 1945 Hawaii 131-sm, 1-LLC-MK2 tc17646 1974 Nauru Island 7-FRhL tc17672 1974 -,Nauru Island 7-FRhL tc, 5-smD-74-061 1974 Bangkok I-Tx. sltendensD-74-063 1974 Bangkok 1-Tx. splendensD-7S-001 1975 Bangkok 3-sm, 1-LLC-MK2 tc, 1-sm691475 1969 Sri Lanka 2-sm, 2-Tx. splendens,

2-LLC-MK2 tc

Abbemations: sin, su.ckling mouse passage; tc, tissue culture passage; LLC.MK., continuous rhesus mokey kidney cell; AP-6i, Aedes pseudo-

sruteltan, cells; C6/36, Aedes aibopictus cloned cells; Taxoekynchilies splesndens, mosquites.

adsorb for 90 min at room temperature, followed sucrose gradient in TSES buffer (0.02 M Tris-by the addition of 30 ml per flask of growth me- HCI, pH 7.4, 0.1 M NaCI, 0.002 M EDTA, 0. 1%dium containing 170 juCi (11P) phosphate (NEN) SDS) for 4 hours at 270,000 x g and 20*C. Theper mi. Following incubation at 285C for 7-8 days, 42S viral RN.A was collected, diluted, and follow-the infected cell supernatants were clarified by ing the addition of 100 jg tRNA, was re-extractedcentrifugation at 3,000 x g for 10 min. Virus sus- with the phenol mixture as before, air-dried, re-pensions were precipitated at 4*C for 6 hours by suspended in 20 ILI of TE buffer (0.02 M Tris-

the addition of polyethylene glycol 6,000 and NaCI, HCI, pH 7.4, 0.002 M EDTA) and stored at7% and 2.3% final concentrations, respectively. -20TC. Final RNA samples contained 0.1-6 xThe precipitated virus was recovered by centrif- 10" cerenkov cpm and were tested for size andugation at 6,000 X g for 30 min, resuspended in purity by electrophoresis of an aliquot (2 to 5 x4 mi TSE (0.01 Tris-HCI, pH 7.8, 0. 12 M NaCl, 104 cpm) on non-denaturing 2.4% polyacrylamide0.001 M EDTA) and purified by density gradient gels'- 11 prior to digestion and fingerprint analysis.centrifugation at 208,000 x g for 18 hours in a Samples of 0.1-1.0 x 106 cpmn of purified 42S30% (w/v) glycerol to M0 (wlv) potassium tar- DEN-I viral RNAs were digested with RNase TItrate gradient in TSE. Virus was recovered from and the resulting oligonucleotides. were resolvedvisible bands, diluted, and repurifled by rate zonal by two-dimensional polyacrylamide gelcentrifugation at 2 70,000 X g for 90 min on a 2 0- electrophoresis"4 as modified by Clewley et aL.'s

70% (w/v) sucrose gradient in 1 M NaCl, 0.01 M Following electrophoresis, the gels were autora-Tris-HCI buffer, pH 8.5. Virus bands were re- diographed at - 70*C using Kodak X-Omnat XAR-covered, diluted fourfold with HSB (0.4 M NaCI, 5 X-ray film.0.01 M Tris-HCI, pH 7.4) and the RNA was ex-tracted. Plaque assays and virus neutralization assays

Preparation and fingerprinting of viral RNA Plaq~ue assays were performed by infecting con-fluent BHK2 I-clone IS" cell monolayers (on 60-

For RNA extraction, sodium dlodecyl sulfate mm plastic petri dishes) with 0.2 mi of a virus(SDS) was added to 1% final concentration and dilution (in Eagle's MEM containing 5% FBS).the suspension was extracted with a phenol: crc- Following adsorption for I hour at 36*C, the cells

sol:8-hydroxyquinoline: chloroform mixture."1 were overlaid with Eage's MEM containing 5%Following ethanol precipitation, the viral RNA FBS and 0.4 agarose (Seakem). Assays were in-was purified by centrifugation in a 15-30% (wlv) cubated at 36*C in an atmosphere of 5% C02 for

RNA FINGERPRINTING OF DENGUE VIRUS 579

6-7 days until plaques were visible. Agarose over-lays were then removed and the monolayers stainedwith crystal violet using the procedure of Hollandand McLaren. 17 Using this method, plaques couldbe counted immediately. Virus neutralization as-says were performed using a plaque reductionneutralization test. 18

RESULTS

The 12 DEN-I isolates used in these studies(Table 1) represent three locations where DEN-I 0is endemic or has been epidemic. These were the •Caribbean (Jamaica and Bahamas), African (Ni- 0geria), and Pacific/S.E. Asian [Hawaii, Nauru Is- Aland (Nauru), Bangkok and Sri Lanka] regions.In addition, certain of the isolates represented atemporal separation of 1 to 10 years (Bangkok-1 year, Jamaica-4 years, and Nigeria--10 years).

Caribbean strains

Two different human isolates from the 1977-1978 Caribbean epidemic were chosen for study- 0a Jamaican strain and a Bahamian strain, the lat-ter being a plaque-purified virus isolated from Ae.pseudoscutellaris tissue culture harvests.' RNAfingerprints of the two 1977 strains and a recent1981 Jamaican isolate are shown in Figure 1. These *••

strains appear very similar; for clarification, sche- I0 0matic representations of coincident fingerprints are Bshown in Figure 2. Fingerprints of the two 1977isolates, Jamaican and Bahamian are depicted inFigure 2A. The unique oligonucleotides numbered1, 3, and 5 originated from the Jamaican strain,while Nos. 2, 4, and 6 were contributed by thelarge-plaque Bahamian isolate. It is very probablethat one pair of oligonucleotides, Nos. 5 and 6,are homologous with the exception of a single basechange (e.g., U -- C) which would be capable of

causing the observed lateral change. Aside fromthese minor differences, the two 197 7 isolates were _similar, sharing 63 of a total of 69 well-resolved 0oligonucleotides, or 91% homology. (Percent ho-

-5 C '

FIGURs 1. Oligonucleotide fingerprint analysis ofthree Caribbean DEN-I strains. The ribonuclease TI- mension from bottom to top. The positions of the two 'resistant oligonucleotides were derived from 35P-labeled dye markers, bromophenol blue (upper) and xylene cy-42S virion RNA and separated by two-dimensional anol FF (lower) are indicated (X). A, plaque-purifiedpolyacrylamide gel electrophoresis. Migration in the first large-plaque isolate from the Bahamas, 1977; 3, Ja-dimension is from left to right, and in the second di- maica isolate, 1977; C, Jamaica isolate, 1981.

C_

580 REPIK ET AL.

mology in this and in all subsequent references isused to define the percentage of well-resolved oli- . / "'gonucleotides which migrated coincidentally uponco-electrophoresis.) . . .

In Figure 2B, comparison is made between two /DEN-i viruses isolated from the same location "- ' " "but at different times--1977 and 1981 Jamaican, . .:strains. In this case, a greater number of unique moligonucleotides was detected. Six oligonucleo-tides appeared unique to the 1977 isolate, whilefive were contributed by the 1981 isolate. Again,one pair of oligonucleotides may in fact be ho- 0mologous with the exception of a single basechange, Nos. 3 and 4. Although less similar than •the two Caribbean strains isolated in 1977 (Ja- 0maican and Bahamian), the two Jamaican strainsisolated 4 years apart share 70 well-separated oli- JAMAICA 77

gonucleotides of a total of 81, or 86% homology. AIIAs • 77

African strains

RNA fingerprints of two DEN-I strains from -

Africa (Ibadan, Nigeria) isolated over a 10-year , , , -interval are presented in Figure 3. To determinethe extent of homology between the two viruses, "RNA from each was simultaneously digested and "Y." °' "" ,'... °•

co-electrophoresed. Origins of the large oligonu- , .' ,r .cleotides in the co-electrophoresed samples were 00 --- "", .,determined by comparing them with their respec- • '-.* " otive single viral RNA digests. Co-electrophoresis • 4'T .of the 1968 and 1978 Nigerian isolates is shown Isin Figure 3C, together with a schematic represen- 0 9 :0o 6tation (Fig. 3D). The 1978 strain retained 60 co- 0 - A

migrating oligonucleotides with the 1968 strain

from a total of 76 well-resolved oligonucleotides, X0 0exhibiting oligtnucleotide homology of 79%. o ',,A

Pacific/S.E. Asian strains BFIGURE 2. Schematic representation of oligonucleo-

Purified 42S virion RNAs of DEN-I Pacific/S. E. tide fingerprint analyses of mixtures of the three DEN-Asian strains (from Nauru, Hawaii, Bangkok, and I Caribbean strains. Oligonucleotides derived from theSri Lanka) were similarly compared by oligonu- Jamaica 1977 isolate are indicated by open circles (0),cleotide fingerprint analysis (Fig. 4). Each finger- while those from the large-plaque Bahamas 1977 or Ja-

maica 1981 isolates are indicated by hatched circles (T).print pattern is clearly unique and distinguish- Filled circles (Ol in this and in all subsequent schematicsable. Fingerprints of the three Bangkok isolates indicate oligonucleotide positions from the two RNAstudiled (two from 1974 and one from 1975) ap- species which migrated coincidentally.peared identical (data not shown); therefore, onlyone Bangkok isolate fingerprint is shown (Fig. 4D).Interestingly, DEN virus isolated from two dif- fingerprints of the two Nauru Island isolates isferent patients during the same epidemic in the shown in Figure 5. The unique oligonucleotideswestern Pacific (Nauru Island, 1974) did exhibit number 1, 3, 5 originated from Isolate I, whileminor differences in their fingerprint patterns (Fig. Nos. 2 and 4 were contributed by Isolate 2. Again,4, A, B). A schematic representation of coincident two pairs of oligonucleotides (No. 1-2, and 3-4)

. a,.--

RNA FINGERPRINTING OF DENGUE VIRUS 581

* ".

soo

0

00 3 0

Ap 0 0,

0* :05

" ,. .. G (PAe"G,- -.

FIGVRE . Oligonucleotide fingerprint analysis of two African (Nigerian) DEN- I strains separated by a 10-yearinterval in isolation. A, 1968 isolate; B, 1978 isolate; C, mixture of the 1968 and 1978 isolates; D, schematic of(C).

may be homologous, with the exception of a single Superimposition of the two S.E. Asian isolatesnucleotide change. Oligonucleotidle No. 5, present yielded a surprising result: only 22 oligonucleo-in Isolate 1, is absent in Isolate 2. Aside from these tides from the Sri Lankan and Bangkok isolatesminor changes, the two Nauru Island isolates are migrated coincidentally from a total of 107 oli-similar and share 53 of a total of 58 well-resolved gonucleotides (2 1% homology) (Fig. 6iB). The Srioligonucleotides, or 91% homology. Lankan isolate contained 46 unique oligonucleo-

To determine the extent of homology among tides, whereas the Bangkok isolate contained 39.other Paciic/SE. Asian isolates, RNA from each Upon closer inspection, the Sri Lankan isolate ap-of two isolates was simultaneously digested and peared to be much more similar to the Nigerianco-electrophoresed. Co-electrophoresis of two Pa- isolates. The Sri Lankan/Bangkok fingerprintcific isolates, Nauru Island No. 2 and prototype comparison represents the only instance whereHawaii (Fig. 6iA) resulted in 49 co-migrating ohi- isolates from a geographic proximity exhibited sogonucleotides from a total of 104, or 47% homol- little similarity.ogy. The Nauru Island isolate contained 27 unique To determine if either of the S. E. Asian strainsoligonucleotides while the Hawaiian contained 28. was more closely related to the Pacific strains than

q ..,

• • o•

582 REPIK ET AL.

9.0

C D

0

.4

-C

RNA FINGERPRINTING OF DENGUE VIRUS 583

Homologies between strains from different," "geographical areas-" c ; ,-;? ) / tI.

(' _. .'," ).-.t ,! Co-electrophoreses (schematic representations)A '- ,, , " of several geographically separated strains of DEN-

.l* . , 4.-- 0-- 1 are presented in Figure 7. For the most part,~ W,,~ -'~~~ .- , ._0 the Pacific/S.E. Asian strains exhibited little re-

-_ .I ) semblance (20-30% oligonucleotide homology) toS O 0 the African and Caribbean strains (Fig. 7, A-C).O00 so 0 The Sri Lankan strain shared a high degree of

01 2lj * O C oligonucleotide homology with the African iso-

XO lates. Co-electrophoresis of the Sri Lankan '69 and[1 :..i o.i Nigerian '78 strains resulted in the co-migration

SA uu2 of 59 of 89 oligonucleotides, or 66% homology (Fig.7D).

FIGURE S. Schematic representation of RNA finger- Fingerprint comparisons between the 1977 Ja-print analyses of a mixture of the two Nauru Island maican strain and all other intergeographic strainsisolates studied, No. I and No. 2. revealed relatively high homologies between only

two other isolates, the Sri Lankan '69 and Nige-rian '78 viruses (Fig. 7E, F). Co-electrophoresisof Jamaican '77 and Nigerian '78 isolates resulted

to each other, co-electrophoreses were performed in 45/85 co-migrating oligonucleotides (53% ho-between the Bangkok and Pacific strains as well mology), while co-electrophoresis of Jamaican '77as between the Sri Lankan and Pacific strains, and Sri Lankan '69 isolates resulted in an evenFigure 6C depicts the result obtained from co- greater number of co-migrating oligonucleotides,electrophoresis of the Bangkok and Hawaiian 55/93 (59% homology). These fingerprint compar-strains. Forty-nine percent of the well-separated isons clearly demonstrated that while these virusesoligonucleotides were common to both, while 17 were similar to one another they were not iden-oligonucleotides were unique to the Bangkok strain tical; therefore the 1977 Caribbean epidemic wasand 19, to the Hawaiian isolate. Co-electropho- probably not caused by either of these particularresis of the Bangkok and Nauru Island No. 2 iso- virus strains from Africa or Sri Lanka. However,lates resulted in the co-migration of 40/91 oligo- the high degree of oligonucleotide homology dis-nucleotides or 44% homology (data not shown). played between the Jamaican and African/SriQuite different results were obtained upon finger- Lankan isolates is striking.print comparison of the Sri Lankan with the Pa- A summary of the percent oligonucleotide ho-cific strains. Only 27% oligonucleotide homology mologies displayed between the various DEN-iwas detected between the Sri Lankan and Nauru strains is presented in Table 2.Island No. 2 strains (Fig. 6D), but somewhat sur-prisingly oligonucleotide homology of 38% was DISCUSSION

exhibited between the Sri Lankan and prototypeHawaiian strain (data not shown). It therefore ap- Virion RNAs of the four DEN prototype sero-pears that the Pacific/S.E. Asian strains (Hawai- types were previously shown to be distinct fromian, Nauru, and Bangkok) are all related to a sim- one another by oligonucleotide fingerprint analy-ilar degree (44 to 49% oligonucleotide homology), sis.4 On an average, only 7% of the large resolv-whereas the Sri Lankan strain appears not to be- able oligonucleotides were shared among the fourlong to this group genetically. Even though a prototype viruses. We similarly analyzed the vi-greater relationship appears to exist between the rion RNAs of temporally and/or geographicallySri Lankan and Hawaiian isolates, 38% homol- separated DEN-I strains and found them also toogy, than between the isolates from Sri Lanka and be unique, rendering the method of oligonucleo-Bangkok or Nauru, 21 to 27% oligonucleotide ho- tide mapping an important tool for distinguishingmology, the relationship is less than that observed virus strains that are serologically indistinguish-among the other Pacific/S. E. Asia virus isolates. able.

aim mil II illall~il-

584 REPIK T AL.

S 0, I. -'I , 0

0

o * s e 0 0* 0 . 0 0 0 0 0• 0 0

0 0

00 0*

0 00B GOK

- . .-. 0, ".- . -

* , -'., I,

0o PO.. f .p . 0

0 - . : ...

0 00

ol,. "o,."o 0 s *o o ":- NA° RU° *

o • 0 0

FIGURE 6. Schematic representations of oligonucleotide fingerprint analyses of mixtures of DEN- I Pacific/S.E.Asian strains.

The DEN-i virus genome does not appear to dropped further still to 79% over a 10-year inter-be as stable as either the VSV-Indiana or influenza val in isolation (Nigeria, '68 and Nigeria, '78virus genomes. 1, The Bangkok DEN- I strains strains). Since some of the virus isolates used asisolated I year apart were indistinguishable from examples of temporal genetic evolution were alsoone another by fingerprint analysis (100% homol- separated geographically, the possibility that theogy). However, oligonucleotide homology dropped rate of genetic change had varied in different en-to 86% when isolations were made 4 years apart demic areas could not be excluded; however, the(Jamaica, '77 and Jamaica, '81 strains), and DEN- I genome appeared to be undergoing a slow

° o 0

RNA FINGERPRINTING OF DENGUE VIUwS 585

-0

0~00, 0b j)

0o 0 0 0 oO 0

0 0 00

0 0

0* 000 0

0 0

c0000 0 o 0 -0

(b 0 0 0000 4

q 00

JAMAIC J1

10 of0

00 0000 0 1

0 40 0

-X 0 0Q 00 0

*1. F 0 0.

FIGURE 7, Schematic representations of RNA fingerprint analyses (co-electrophoreses) of several intergeograph-ical strains of DEN-I.

586 REPIK ET AL.

TABLE 2

Summary of oligonucleotide homologies between Dengue I virus strains

Georphic group Strains (Year) % homology

Caribbean Jamaica ('77) + *Bahamas ('77) 91Jamaica ('77) + Jamaica ('81) 86

African Nigeria ('68) + Nigeria ('78) 79Pacific/S.E. Asian Bangkok ('74) + Bangkok ('75) 100

Nauru No. 1 ('74) + Nauru No. 2 ('74) 91Nauru No. 2 ('74) + Hawaii ('45) 47Hawaii ('45) + Bangkok ('74) 49Bnkok .. ('74) + Nauru No. 2 ('74) 44Sri Lanka ('69) + Hawaii ('45) 38Sri Lanka ('69) + Nauru No. 2 ('74) 27Sri Lanka ('69) + Bangkok ('74) 21

Intrageographic Hawaii ('45) + Jamaica ('77) 20comparisons Hawaii ('45) + Nigeria ('78) 20

Bangkok ('74) + Jamaica ('77) 20Bangkok ('74) + Nigeria ('78) 24Nauru No. 2 ('74) + Jamaica ('77) 30Nauru No. 2 ('74) + Nigeria ('79) 30Sri Lanka ('69) + Nigeria ('g) 66Sri Lanka ('69) + Nigeria ('68) 60Sri Lanka ('69) + Jamaica ('77) 59Sri Lanka ('69) + Jamaica ('81) 55Jamaica ('77) + Nigeria ('78) 53

M Jamaica ('77) + Nigeria ('68) 49 J

but continuous evolution. Although it would have in turn showed 49% and 44% oligonucleotide ho-been more desirable to determine genetic stability mologies, respectively, with the Bangkok isolates.over time using only isolates from a single region, We believe that the reduced oligonucleotide ho-such isolates were not available. mologies displayed among viruses within the Pa-

In the course of our analyses, it was observed cific/S.E. Asian group are due in part to the largethat a given DEN-I fingerprint pattern appeared geographical area encompassed by this groupingcharacteristic of a particular geographical area, as compared to the more geographically limitedallowing us to establish three geographical group- Caribbean and Nigerian groups.ings based upon percent shared oligonucleotides. The Sri Lankan isolate displayed a relativelyWith one exception, virus isolates from each of high degree of homology (60 to 66%) with Nige-these three groups (i.e., Africa, the Caribbean, rian isolates rather than with those isolates fromand Pacific/S.E. Asia) retained 45-100% oligo- its geographically proximal neighbor, Bangkoknucleotide homology with other isolates within that (21% homology). Because normally only 20-30%region. oligonucleotide homology could be detected be-

Unlike viruses in the African and Caribbean tween the three geographical virus groups, the Srigroups, the seven viruses studied from the Pacific Lankan isolate, although recovered from the S. E.and Southeast Asian regions displayed a substan- Asia region, appeared more like a member of thetially lesser degree of intragroup oligonucleotide African group. Because Africa is the ancestralhomology. Only two or more viruses isolated from home of Ae. aegypti and it is thought that endemicspecific areas within these regions exhibited a very dengue in Africa may have been the origin of var-high degree of homology, i.e., the three 1974-1975 ious Asian and American strains,'" it seemsBangkok isolates were identical (100% homology) probable that the Sri Lanka DEN-l virus couldwhile the two 1974 Nauru Island isolates were have been imported from Africa either directly orvery similar (91% homology). However, the Pa- indirectly.cific isolates from Hawaii and Nauru Island dis- Because the fingerprint technique was able toplayed only 47% homology to ac other, and the w- distinguish between genomes of viruses isolated

Cq

RNA FINGERPRINTING OF DENGUE VIRUS 587

from geographically different areas, it was applied been performed by computer-simulation, " 27 andto answer epidemiological questions concerning the has provided a means of estimating statistical con-origin of the 1977 Caribbean epidemic. Epide- fidence limits to be used when quantitatively com-miological investigations of the sudden appear- paring fingerprints of different RNA molecules.ance of DEN-i in Jamaica suggested an African Large oligonucleotides (greater than II nucleo-or Asian origin. 2 Analysis of pertinent RNA fin- tides in length) apparently represent approximate-gerprints revealed that the 1969 Sri Lankan and ly 30% of the entire genome, and on an average,the 1978 Nigerian isolates exhibited 59% and 53% only 85, 50, or 25% of large oligonucleotides re-oligonucleotide homologies, respectively, with the main in common when two overall RNA base se-1977 Jamaican strain. Although none of these vi- quences (full-length genomes) differ by 1, 5 or 10%,ruses approached identity, the striking homologies respectively. RNA fingerprinting is sensitive todisplayed between the Sri Lankan or African strain small genome changes and is particularly usefuland the Jamaican strain suggest that the 1977 Ja- only when comparing closely related RNAs or re-maican epidemic may have resulted from a virus gions of RNAs that have overall base sequenceintroduced from a region near either Sri Lanka or homologies greater than 90%. Through extrapo-Africa. Although it might appear that virus from lation of data, computer-simulation can also bethe Sri Lanka region was the more suspect source applied to help estimate the overall sequence ho-of the Jamaican epidemic, the Sri Lankan and mology of two RNAs based upon percent oligo-Jamaican isolations were also separated by a span nucleotide homologies. Thus, viruses belonging toof 8 years. During those years the 1969 Sri Lankan the Caribbean and African groups were estimatedstrain may have evolved, but without a later Sri to have greater than 99% and 98.5% overall RNALankan isolate this cannot be determined. genome homologies, respectively, while those in

In similar studies, separate geographical iso- the Pacific/S.E. Asian group (Nauru Island, Ha-

iates of DEN-2 virus strains also exhibited unique waii, Bangkok) were estimated to have 93-95% IRNA fingerprints (P. Repik, unpublished data; D. genome homologies. Estimated total genome ho-Trent, personal communication), as have separate mologies between viruses of the three differentisolates of DEN-3 strains (D. Gubler, personal geographic groupings appeared to be in the 85-communication). These results all suggest that the 89% range.technique of RNA fingerprinting can be a useful An additional observation from these studies wastool for flavivirus epidemiologists, allowing more that the majority of oligonucleotides which appearprecise identification of dengue isolates belonging to be conserved are those enriched in cytosine andto a particular subtype as well as monitoring ge- adenine (left side of fingerprint), while those oli-nome changes that can occur over time and geo- gonucleotides which appear more subject to changegraphic distribution. Such studies would require contain a higher proportion of uracil (right side ofnumerous additional dengue virus strains from all fingerprint). No portion of the dengue genome hasendemic as well as epidemic areas, been sequenced aside from the 5' terminus and

The technique of oligonucleotide mapping or penultimate base of DEN-3 virion RNA.4 There-RNA fingerprinting is within the capabilities of fore, the spatial organization of the oligonucleo-many arbovirus laboratories and does measure tides within the viral genome and subsequent cod-certain changes in nucleotide sequence; however, ing assignments are unknown. The importance ofthere are rather serious limitations and it is by no oligonucleotide composition in resistance or sus-means as precise as nucleotide sequence analysis. ceptibility to change awaits detailed genetic char-Oligonucleotide fingerprints are highly reproduc- acterization of the viral particle and a better un-ible in duplicate analyses, and oligonucleotides derstanding of replication strategies under bothfound to be common by their migration on poly- natural and laboratory pressures.acrylamide gels actually do have identical nucleo-tide sequences. 4 In general, only the larger oi- ACKNOWLEDGMENTS

gonucleotides (10 to =40 nucleotides in length) canbe compared since they presumably represent The authors wish to thank Dr. Donald Burke,unique RNA sequences. Such oligonucleotides had Armed Forces Research Institute of Medical Sci-been thought to comprise approximately 10-15% ences, Bangkok, Thailand, for providing theof the viral genome. 5 2 More recently, theoretical Bangkok and Sri Lankan DEN-I strains, Esmieanalysis of the RNA fingerprinting technique has Rose and Dorothy King of the University of the

588 REPIK ET AL.

West Indies, Kingston, Jamaica, for providing the group viruses. Contr. Epidem. Biostal., 3: I-Bahamian and Jamaican DEN-I strains and Dr. 20.

Robert Shope, Vale Arbovirus Research Unit, New 10. Bishop, D. H. L., Beaty, B. J.. and Shope, R. E..1980. Recombination and gene coding assign-

Haven, Connecticut, for supplying material from ments of bunvaviruses and arenaviruses. Ann.which the Nigerian 1978 DEN-I strain was iso- N.Y. Acad. Sci., 354: 84-106.

lated. Dr. Tissa Vitarana of the Medical Research 11. Igarashi, A., 1978. Isolation of a Singh's AedesInstitute, Columbo, Sri Lanka, kindly provided albopictus cell clone sensitive to dengue and Chi-additional informato one Sri Lanka sain. d kungunya viruses. J. Gen Viral., 40: 531-544.additional information on the Sri Lankan strain. 12. Repik, P., Flamand, A., and Bishop, D. H. L.,P.M.R. was a recipient of a postdoctoral fellow- 1976. Synthesis of RNA by mutants of vesicular

ship award from the National Research Council. stomatitis virus (Indiana serotypel and the abilityof wild-type VSV New Jersey to complement theVSV Indiana ts G 1-114 transcription defect. J.

REFERENCES Viral., 20: 157-169.13. Bishop, D. H. L., Claybrook, J. R., and Spiegel-

I. De Madrid, A. T., and Porterfield, J. S., 1974. The man, S., 1967. Electrophoretic separation of vi-flaviviruses (Group B arboviruses: a cross-neu- ral nucleic acids on polyacrvlamide gels. J. Mot.tralization study. J. Gen. Virol., 23: 91- 96. Biol., 26: 373-387.

2. King, S. D., Rose, E., Bancroft, W. H., McCown, 14. De Watcher, R., and Fiers, W., 1972. PreparativeJ. M., Woodall, J., and Sather, G., 1979. The two-dimensional polyacrylamide gel electrophore-laboratory diagnosis of dengue in Jamaica, 1977. sis of

32P-labeled RNA. Anal. Biochem., 40:

Pages 153-158 in Dengue in the Caribbean, 1977. 184-197.(Sci. Publ. No. 375.1 Pan American Health Or- 15. Clewley, J., Gentsch, J., and Bishop, D. H. L.,ganization, Washington, D.C. 1977. Three unique viral RNA species of snow-

3. Bancroft, W. H., McCown, J. M., Mas Lago, P., shoe hare and La Crosse bunyaviruses. J. Virol..Brandt. W. E., and Russell, P. K., 1979. Iden- 22: 459-468.tification of dengue viruses from the Caribbean 16. Eylar, 0. R., and Wisseman. C. L., Jr. 1975. /by plaque-reduction neutralization test. Pages 17.1- Thermal inactivation of type I dengue virus178 in Dengue in the Caribbean, 1977. Pan Amer- strains. Acta Virol., 19: 167-168.

ican Health Organization, Washington, D.C. 17. Holland, J. J., and McLaren, L. C., 1959. Im-4. Vezza, A. C., Rosen L., Repik, P., Dalrymple, J., proved me~hod for staining cell monolayers for

and Bishop, D. H. L., 1980. Characterization virus plaque counts. J. Bacterial., 78: 596-597.of the viral RNA species of prototype dengue vi- 18. Russell, P. K., Nisalak. A., Sukhavachana, P., andruses. Am. J. Trap. Med. Hyg., 29: 643-652. Vivona, S., 1967. A plaque reduction test for

5. Clewley, J. P., Bishop, D. H. L., Kang, C.-V., dengue virus neutralizing antibodies. J. Immu-Coffin, J., Schnitzlein, W. M., Reichmann, M. no,., 99: 285-290.E., and Shope, R. E., 1977. Oligonucleotide fin- 19. Ehrenkranz, N. J., Ventura, A. K., Cuadrado, R.gerprints of RNA species obtained from rhabdo- R., Pond, W. L., and Porter, J. E., 1971. Pan-viruses belonging to the vesicular stomatitis virus demic dengue in Caribbean countries and thesubgroup. J. Virol., 23: 152-166. southern United States--past, present, and po-

6. El Said, L. H., Vorndam, V., Gentsch,J. R., Clew- tential problems. N. Engl. J. Med., 285: 1460-ley, J. P., Calisher, C. H., Klimas, R. A., 1469,Thompson, W. H., Grayson, M., Trent, D. W., 20. Carey, D. E., Myers, R. M., and Reuben, R., 1964.and Bishop, D. H. L., 1979. A comparison of Dengue types I and 4 viruses in wild-caught mos-La Crosse virus isolates obtained from different quitoes in South India. Science, 143: 131-132.ecological niches and an analysis of the structural 21. Carey, D. E., Causey, 0. R., Reddy, S.. and Cooke,components of California encephalitis serogroup A. R., 1971. Dengue viruses from febrile pa-viruses and other bunyaviruses. Am. 1. Trop. tients in Nigeria, 1964-1968. Lancet, 1: 105-106.Med. Hyg., 28: 364-386. 22. Halstead, S. B., 1980. Dengue haemorrhagic fe-

7. Young, J. F., Desselberger, U., and Palese, P., 1979. ver-a public health problem and a field for re-Evolution of human influenza A viruses in na- search. Bull. W.H.O., 58: 1-21.ture: sequential mutations in the genomes of new 23. Dengue in the Caribbean. 1977, 1979. ProceedingsHINI isolates. Cell, 18: 73-83. of a workshop held ins Montego Bay, Jamaica (8-

8. Trent. D. W., Monath, T. P., Bowen, G. S., Vorn- II May 1978). Pan American Health Organiza-dam. A. V., Cropp, C. B., and Kemp, G. E., tion, Washington, D.C.1980. Variation among strains of St. Louis en- 24. Hagen, F. S., and Huang, A. S., 1981. Compari-cephalitis virus: basis for a genetic, pathogenetic, son of ribonucleotide sequences from the genomeand epidemiologic classification. Ann. N. Y. .4 cad. of vesicular stomatitis virus and two of its defec-Si., .354: 219-237. tive interfering particles. J. Viral., 37: 363-371.

9. Cash, P., Robeson, .- 'Flick, B. J., and Bishop, 25. Pedersen, F. S., and Haseltine, W. A., 1980. Anal-D. H. L., 1981. 'Biochemical characterization of ysis of the genome of an endogenous, ecotropicRift Valley fever and other phlebotomus fever retrovirus of the AKR strain of mice: micrometh-

-- --. ,.

RNA FINGERPRINTING OF DENGUE VIRUS 589

od for detailed characterization of high-molecu- tion of Negative Strand Viruses. Elsevier/North-lar-weight RNA. J. Virol., 33: 349-365. Holland, New York.

26. Young, ]. F., Taussig, R., Aaronson, R. P., and 27. Aaronson, R. P., Young, J. F., and Palese, P., 1982.Palese, P., 1981. Advantages and limitations of Oligonucleotide mapping: evaluation of its sen-the oligonucleotide mapping technique for the sitivity by computer-simulation. Nucl. Acidsanalysis of viral RNAs. Pages 209-215 in D. H. Pos., 10: 237-246.L. Bishop, and R. W. Compans, eds., Replica-

I

Accession For

DTCNTIS GR;AU1ELECTE DTIC TAB

Unannounced []

OCT 1 9 19M Justification

0 By

B Distribution/

Availability Codes

Avail and/orDist Special

C_