Research News Briefs written by Alan Doveand Holger Breithaupt

NATURE BIOTECHNOLOGY VOL 17 SEPTEMBER 1999 http://biotech.nature.com 839

IN BRIEF

RESEARCH NEWS

Retargeting tropism

German and French scientists have success-fully redirected the cellular tropism ofadeno-associated virus type 2 (AAV2) by dis-playing short peptides on the viral capsidthat are recognized by specific receptors ontarget cells. By studying the uptake of AAV2strains in which a 14-amino-acid ligand, L14,had been inserted into six different domainsof the capsid, they showed that one strainsuccessfully infects AAV2-resistant cells dis-playing L14-specific integrin on its surface.Further experiments demonstrated thatother ligands up to 34 amino acids in lengthcan also be inserted into the AAV2 capsid.Though the specificity of the AAV2–integrinreceptor interaction in this study was notfully characterized, the researchers think thattheir approach could be widely applicable inthe design of vectors specific for other celltypes. “These studies demonstrate that, inprinciple, it is possible to redirect AAV2,”says Anne Girod, the lead author on the arti-cle (Nat. Med. 9, 1052–1056, 1999). Theteam’s next aim is to find peptides that allowvectors to be targeted to specific cells types invivo. According to Girod, modified AAV2vectors that specifically infect T-lymphocytescould be an early application.

Aptamers on the scaffoldIn independent reports published in Proc.Natl. Acad. Sci. USA 6, 8567–8572, 1999)and Science (285, 591–595, 1999), twoteams of researchers have recentlydescribed nearly identical approaches foridentifying novel components of cellularsignaling pathways, technology that shouldprove useful both in basic research and theidentification of new drug targets. Eachgroup screened yeast cells expressing alibrary of aptamers—short random peptidesequences expressed on a larger proteinscaffold—to identify peptides that inhibitspecific intracellular signals in yeast. Onceinhibitory peptides had been identified,they were used as “bait” in two-hybridscreens to determine which yeast geneproducts they were binding. The approachidentified new components of the yeastmating pheromone signaling pathway andthe mitotic spindle checkpoint. Havingproven the concept, the approach couldconceivably be employed for signalingstudies and drug target identification inmammalian cells, which lack the power ofyeast genetics. While emphasizing that sev-eral hurdles remain before the techniquebecomes standard practice, Roger Brent,senior author on the PNAS paper, says that“if the two-hybrid experience is a guide, Iwould guess this is about a year away fromwidespread use by relatively unsophisticat-ed laboratories.”

In silico matchmaking

A computer algorithm has been describedthat may serve as a prototype for in silicoapproaches that assign function to hithertounknown proteins. The method, described inScience (285, 751–753, 1999), is based on theobservation that pairs of interacting proteinsoften have orthologs in other organisms inwhich the two proteins are fused into a singlechain. Sequence homologies with differentsegments of a protein from another organismcan therefore be used to predict an interactionbetween two different proteins. Applying thealgorithm to the 4290 known protein fromEscherichia coli, David Eisenberg and his col-leagues came up with 6809 putative interact-ing or functional-related protein pairs in thebacteria. A similar search resulted in 45,502similar patterns for the 5,800 yeast proteins.Although the approach yielded a high pro-portion (82%) of false positives, Eisenbergexpects that refinements can improve the reli-ability of the method, particularly by elimi-nating homologies like nucleotide-bindingcassettes. “Learning how to filter thesedomains out is very important,” he says.Together with his colleague Todd Yeates,Eisenberg has founded Protein Pathways (LosAngeles, CA) to commercialize the method.

AAV muscles aheadScientists from ARIAD Pharmaceuticals(Cambridge, MA) and Stanford University(Stanford, CA) have shown that adeno-asso-ciated viral (AAV) vectors significantly out-perform adenoviral vectors when used totransfect muscle cells with an inducible geneexpression system. In a head-to-head com-parison, the scientists used the two vectors totransfect mice muscle cells with a constructcontaining the gene for human growth hor-mone (hGH) under the control of arapamycin-inducible expression system (Proc.Natl. Acad. Sci. USA 96, 8657–8662, 1999).Injection of rapamycin into mouse musclewas shown to rapidly induce hGH synthesis,which gradually declined as rapamycin levelsdiminished. Importantly, expression of hGHremained stable when AAV was used as vector,but progressively declined when genes weretransferred using the adenoviral vector.According to Victor Rivera, correspondingauthor on the study, AAV integrates its genesinto infected host cells and thus avoidsimmune responses induced by adenovirusthat eventually kill cells carrying the foreigngene. “We tried both systems,” says Rivera,“and our results demonstrated we should fullyexplore AAV.” He contends that the ease andefficiency of the rapamycin system shouldmake it broadly applicable.

Influenza from scratchInfluenza A virus has been reconstructed entirely fromcloned cDNA segments. Until now, the virus had beennotoriously difficult to alter in vitro because the genomeis distributed among eight subgenomic segments of neg-ative-stranded RNA. Researchers have relied on the useof helper viruses to study the pathogen, but suchapproaches allow alteration of only a small subset ofinfluenza virus genes. In the new work, published inProc. Natl. Acad. Sci. USA (96, 9345–9350, 1999), scien-tists have transfected human cell lines with eight separateplasmids, each carrying the cDNA of a subgenomic seg-ment in front of an RNA polymerase I promoter. Theapproach yielded highly concentrated virus stocks, andthe cDNA plasmids were amenable to alteration by stan-dard recombinant DNA techniques. Yoshihiro Kawaoka,a senior author on the paper, cautions that genetic engineering of influenza should beapproached carefully: “Before initiating such experiments, scientists should weigh therisks and benefits of the studies and the containment levels at which the work should beperformed.” Historically, new strains of influenza have periodically caused deadly pan-demics. Similar work by the teams of George Brownlee and Peter Palese will soon bepublished in the J. Virology (73, in press, 1999). Influenza induces strong mucosalimmunity and does not integrate into the host genome, making it a potentially usefulvector for both vaccines and gene therapy.

Cou

rtesy of Peter Palese

© 1999 Nature America Inc. • http://biotech.nature.com©

199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://b

iote

ch.n

atu

re.c

om