<<

. 13
( 61 .)



>>

the polylmker at either side of the insert, and distilled water to 20 pL Incubate
for 2 h at the appropriate temperature.
10 Electrophorese the digests on a 2% agarose gel, load as markers DNA fragments
m the range of 0.1-2 kbp (see Note 52) If not two sizes of inserts (representing
de Wmd, van Z&i/, and Berns
78
fragments flanking both sidesof the oligonucleotide) are found, onemay charac-
terize more clones (steps7-10, see Note 53)
3 3.2. Determination of the Orientation
of the Flanking Probes with Respect to the Inserted Oligonucleot/de
If the orientation of the subclones with respect to the oligonucleotide is to be
determined, this can be done by hybrtdization with a blot of a gel contammg
the correspondmg mutant cosmrd that is triply digested with the diagnosttc
enzyme for the vn-us plus the ohgonucleottde cleaving enzyme plus the enzyme
that cuts the cloning linker (the blot may be generated m Section 3.2.4., see
Note 45). Cut out each lane from the blot (containing a digested mutant cosmid)
and hybridize the mmiblots with the one of the two correspondmg, flanking
clone contammg, radiolabeled plasmtds. After exposure of each autoradiogram,
two bands will be visible: a band representing the cosmid vector, cross-
hybridizing with the radiolabeled plasmid vector, and a second band that repre-
sents the side of the ohgonucleottde that is represented in the flanking probe.
3 3.3. Crosshybndizatlon Studies with DNA
from a Prototype Herpesvirus
The genomes of one or more members of all three subfamilies of her-
pesviruses have now entirely been sequenced (reviewed in ref. I), yielding a
wealth of data on gene location and (putative) gene function for each subfam-
ily. Therefore, a feasible way to gam mstght mto the global layout of the
genome of a less-well-known herpesvirus is by investigating crosshybrtdization
with the genome of a prototype herpesvirus from the same subfamily A defined
set of well-characterized small probes, derived from large genomic regions of
the vu-us under study 1san ideal tool to perform thesecrosshybridization studies
We successfully investigated crosshybridization between small flanking
probes derived from a mutagemzed 41 -kbp subgenomic region from PRV
(derived from cosmid clone c-448, Fig. 5) and cloned fragments form HSV- I
Briefly, a series of cloned KpnI fragments of HSV- 1were digested with BamHI
plus SspI, and with BarnHI plus EcoRI plus SspI (see Note 54). These specific
KpnI fragments of HSV-1 were chosen on the basis of previous cross-
hybridization experiments by others, using large PRV fragments as probes,
estabhshmg a rough synteny map between both viruses The digested HSV-1
clones were loaded manifold on agarose gels, electrophoresed, and blotted onto
nylon membranes. The blots were subsequently cut m small blots, each con-
taming a series of digested HSV-1 clones. Flanking small PRV probes were
cut out of their vector, isolated by preparative gel electrophoresis and labeled
by random priming (see Note 55). After hybridization at low stringency (6X
SSC, 25% formamide, 42”(Z), blots were washed by subsequent multiple
Characterization of a Herpesvirus Genome 79

washes at increasing stringency: 6X SSC, 60°C; 6X SSC, 65°C; 3X SSC, 65°C;
0.1X SSC, 65OC.Between each series of washings, blots were exposed to auto-
radiography film. Bands that crosshybridized after washing at low stringency,
but disappeared at the highest stringency washing were marked cross-
hybridizing. In this way we were able to define four crosshybridizing regions
and to establish synteny between a large subgenomic region of both viruses (6).
3.3.4. The Use of Small Probes to Gain Information
on Location, Sizes, and Expression of Virus Genes
The availability of a set of well-characterized probes from the virus genome
also allows the gaining of mformation on the transcriptional units of the vuus
genome. Although the herpesvirus genome contains many transcriptional units,
containing several genes with coterminal 3™ ends (21), a thorough analysis of
the mRNAs encoded m the subgenomic region(s) of the virus that are
mutagenized by ohgonucleotide msertion yields information on gene location,
expression, and on the genes that are inactivated by insertion of the oligo-
nucleotide. This analysis IS facilitated by the absence of mtrons m most genes
(21). A chapter of this volume is devoted to the analysis of HSV gene expres-
sion (Chapter 23), so we again provide only an outline of the procedure we
have followed.
RNA was prepared from cells that were mock infected, or infected for 2 and
for 6 h (to be able to discriminate between early, early-late, and late genes,
respectively) with a high multiplicity of wild-type virus. These RNAs were
loaded in manifold on RNA gels, electrophoresed, and blotted. Small frag-
ments of each blot, containing lanes with the three RNA preps were each
hybridized with a flanking probe, under high stringency. Size and class of all
transcripts that were detected unambiguously were listed. If a transcript was
detected with probes derived from both flanks of a specific ohgonucleotide,
the oligonucleotide was assumed to be Inserted mto that transcript. Smce most
herpesvirus transcripts have coterminal3™, and not S, ends, we could in many
casesobtain indications if the oligonucleotide was inserted in the 5™ part of the
transcript (and thus likely in the open-reading frame) or m the 3™ part of the
transcript (and thus possibly in a downstream located open-readmg frame).
3.3.5. The Use of Flanking Probes as Sequence Tags
The ultimate and definite way to identify virus genes of interest is by
sequence analysis. The cloned small virus DNA fragments, flanking the
inserted oligonucleotide in the set of mutant cosmid derivatives, provide an
excellent way to rapidly identify genes, and also to determme m what genes
and at what site the oligonucleotide is inserted in the various mutants. The
inserts of the plasmids may directly be sequenced using double-stranded DNA
80 de Wind, van Zijl, and Berns
If the G C content of the virus is high (as m PRV), smgle-stranded
sequencing.
templates may be preferred for sequencmg. Single-strand templates can etther
be achieved by producmg single-strand DNA of the plasmtd by entire super
infection with a helper phage, rf it contams a single-strand phage replication
ortgm (like, e.g., the pBluescript series from Stratagene), or otherwise by
subcloning in one of the M 13mp series of bacteriophages.
4. Notes
1 It should be noted that a small deletion has occurred here at the Junction of both
fragments during hgatton m VIVO
2 In the origmal protocol, insert fragments were, after digestion, purlfled free from
vector sequences before transfection to avotd the risk of mcorporatton of vector
sequences into the reconstituted vnus by nonhomologous recombmatton or ltga-
non (2). This last purtficatton step has proven to be superfluous since we never
found mcorporation of vector sequences (when the purification step was omtt-
ted) after overlap recombmatton If, however, this purification step turns out to
be requtred when performmg overlap recombmatton with another herpesvn-us,
the protocol for size selection on glycerol gradients (see Section 3 1 3 ) 1s smt-
able for thus purpose.
3 The efficiency of vnus regeneration when the set of five fragments was used for
overlap recombmatton was at least as high as when four fragments are used (2)
This may be related to the larger overlaps between the live subgenomic frag-
ments than between the four fragments It 1s belteved that the frequency of
homologous recombmatron strongly depends on the length of the homology (see
also Note 13)
4 Obviously, this type of mutagenesis is only efficient if most of the genome is
protein encoding For HSV-1, this is 89% of the virus genome (20)
5. The oligonucleotide should be palmdromic to allow double-strandedness. This
also prevents polarity, 1 e , msertton rn either orientation within a gene will be
functional In addttton, tt IS preferred that the enzyme gives sticky (as opposed to
blunt) overhangs after cleavage. The EcoRI site (GAATTC) in this specrfic oli-
gonucleottde 1s absent from the entire PRV genome. Consequently, this single
ohgonucleottde would be suitable for mutagenesis of all four cloned subgenomtc
regions of the vnus. If the restriction site that is present m the mutagemc oligo-
nucleottde would also be present (rarely) m the virus genome, a different oltgo-
nucleotide should be devised for mutagemzation of each region of the V˜-LIS
genome (see also Note 10) Although the amber (TAG) stopcodon 1s the only
stopcodon used in thrs specific oligonucleottde, any of the three stopcodons
(TAA, TAG, and TGA) may be used provided that the aforementioned reqmre-
ments for the oligonucleottde are fulfilled
6. The cleavage site of an enzyme recognizing a 4-base sequence will statistically
be present once m every 256 bp (presumed that the vtrus genome contains 50%
G.C bp) The PRV genome, consistmg of 73% G C bp, contains considerably
more recogmtion sites for these enzymes, which recognize G C rich quadruplets)
Characterization of a Herpesvirus Genome
Although linearization of a plasmrd by such an enzyme will thus not truly be
random, the use of several four-cutter enzymes with different recognition sites to
linearize the plasmid will ensure lmearization at almost (quasi) random sites
7. It should be noted here that the oltgonucleottde msertion mutagenesis as described
here has worked well for cosmids up to 46 kbp total length (I e., including the
vector). Using a cosmid of 5 1 kbp, we found deletions m the plasmid after ohgo-
nucleotide insertion mutagenesis Thus, approx 46 kbp may be the upper size
limit of the cosmid to be mutagenized using this technology.
8. Viable virus mutants are only obtained if the gene carrying the inserted ohgo-
nucleotide IS not essenttal for virus replication in cultured cells. It IS beheved that
at least half of the vnus genes are dispensable for growth of the vnus m tissue
culture (22) If the gene is essential, however, the mutant vn-us may be rescued by
performing the overlap recombination in cells that provide the essential gene
product in tram (see, e.g., refs. 9 and 10, and Chapter 6 of this volume) If the
oligonucleotide IS Inserted at the end of the cloned fragment, m the region
homologous between two adjacent clones, the site containing the oligonucleotide
may be lost m some of the progeny virus generated durmg each overlap recombi-
nation, dependmg on the site of the crossover. To circumvent this, it may be
advantageous to select two complete sets of cosmid clones, each havmg different
overlaps (see also Note 13)
9 Although fragments between approx 33 and 45 kbp may be cloned in cosmid
vectors (depending on the specific vector), the upper hmit m this case IS set by
experimental limitations (see Note 7) to approx 4 1 kbp
10. Since the insert has to be cleaved out of the cosmid vector for transfection, the
restriction site in the lmker to use should preferentially be absent from the virus
genome. If no such site can be found, different regions of the virus genome may
be cloned using different linkers from which the restriction site is absent m that
specific cloned fragment (see also Note 5) However, a number of 8-bp restric-
tion sites, (e.g , AscI, NotI, PacI, PmeI, and @I [all available from New England
Biolabs], Srfl [Stratagene], Sse8387I [Takaraa], or SwaI [Boehrmger]) are now
available. Smce every 8-bp sequence statistically is present only once in every 66
kbp, one w˜fl probably be able to find a restrrctron sate that IS absent from the
entire virus genome. In this case, a single linker suffices for cloning of cosmtds
contaimng the entire virus genome It is important to note here that the restriction
site in the linker used for cloning the virus genome in cosmids should be another
site than the site in the mutagenic oligonucleotide that is later inserted at random
sites in the cloned vu-us genomic fragment. The lmker may be purchased com-
mercially However, if no linker containing the desired restriction site is avail-
able, it should be newly designed. This linker should be a palindrome, contammg
the desired restriction site flanked by two bases (preferentially C or G residues)
on both sides.
11. The procedure might be shortened by blunting of the vn-us DNA fragments and
addition of the cloning linker (as described in Section 3.1 S., steps l-6) directly
after shearmg (Sections 3 1 1 and 3.1.2 ) and before the glycerol gradient (Sec-
82 de Wncf, van ZJ/, and Berm
tton 3.1 3.) This also obviates the purtficatton of sheared molecules on a pre-
parative gel after linker addition (Section 3 1.5 , steps 7-9)
12. Although the vnus DNA fragments may be also be cloned in low copynumber
plasmtd vectors like pBR322, cloning into a cosmtd vector has the advantage that
tt ts much more efficient, thus facilitating the generation of a set of vu-us clones
that meets the requirements with respect to integrity, insert size, and size of
homologous overlaps between the inserts. Any cosmid vector may be used
for cloning
13 For efficient recombmatton, a homologous region of at least 0 5 kbp is required.
However, this region should be as short as possible to mnumtze the risk of loss of
an oltgonucleottde inserted at the homology at the flank of a vtrus clone (see also
Notes 3 and 8).
14. It 1s sufficient to clone only one complete copy of every inverted repeat present
m the virus genome, its counterpart is regenerated during transfectton (2)
15. The pellet should not dry out completely, as this impairs dtssolutton.
16. Uncleaved phage lambda DNA gives a single band of 48.5 kbp; a HJndIII digest
gives fragments of 23 1, 9 4, 6.6, 4 4, 2 3, 2.0, and 0.6 kbp Heat the lambda
DNA for 5 min at 65°C prior to loading to melt out annealing of the terminal
overhangs of the phage DNA.
17. Gentle vortexmg of DNA fragments smaller than 50 kbp does not shear the frag-
ments stgmficantly further.
18. T4 DNA polymerase IS capable of bluntmg 5™ protrnstons (using tts 5™ + 3™ poly-
merase acttvtty) and also 3™ protrustons (by its 3™ + 5™ exonuclease activity)
Hence, it 1s the enzyme of choice to generate blunt ends for linker addition and
subsequent cloning of the subgenomic fragments of the vnus The Klenow frag-
ment of E colz DNA polymerase I might also be used here, however, the latter
enzyme has a considerably lower 3™ -+ 5™ exonuclease activity than the T4 enzyme.
19. An excess of linker m the ligation will most efficiently drove the reaction toward
addition of linkers to all DNA ends.
20. Any restrtction enzyme dtgestton buffer may substttute for the T4 polymerase buffer
2 1. Theorettcally, steps 5 and 6 are not essential since cosmtd vectors contammg the
novel restriction sue may directly (1 e., after dephosphorylatton) be used to ligate
to the sheared vtrus DNA to whtch the same site 1sadded. However, prior cloning
of the restrictton sate contammg cosmtd vector (steps 5 and 6) 1srecommended as
this improves the efficiency of cosmtd cloning.
22. At thts stage, cosmid vectors that have acctdently retained the ortgmal restriction
site may be eliminated by digestion with the correspondmg enzyme prtor to trans-
formation
23 As a control for transformation, it is advised to also transform 0 1 ng control
plasmid (like pBR322) The efficiency of transformatton should be equal to, or
higher than, 5 x 1O6colomes/mg plasmid.
24 Cells may be concentrated by spmnmg for 2 mm at 4OOOg.
25. If the restriction enzyme with which the cosmid vector was cleaved yields blunt
or 3™ protrudmg termmi, perform the second incubation at 55™C
Characterization of a Herpesvws Genome
26 The amount of cultures to be inoculated depends on practical factors hke, e.g.,
the capacity of the microfuges and gel electrophoresls equipment.
27 Frozen storage of fresh transformant E co12 cells is necessary smce large plas-
mlds are often not well mamtained by E. ˜011.As a consequence, deletions will
occur If transformant colonies are stored too long at 4°C.
28 A digestion buffer and Incubation temperature should be chosen m which both
enzymes have sufficient activity. Most manufacturers provide lists with toler-
ated condltlons for their enzymes If one of both enzymes requires a low salt
buffer and the other enzyme a high salt buffer, one may also cleave first with
the first enzyme, and add salt and the second enzyme after the first dIgestIon has
completed.
29 Do not expose the DNA to strong light if the DNA solution contams ethldmm
bromide
30 Pure DNA has an OD,,,/OD,,, of 1.8 An OD,,O of 1 corresponds to a DNA
concentration of 50 pg/mL.
31 Depending on the amount of plaques obtained, the transfection may be scaled
down to 2 mg DNA In the latter case, 35-mm dishes with cells are used.
32. If separate plaques are to be obtained, different amounts (e.g , 0 1 and 0.9 mL) of
the CaPO, precipitate may be added to different dishes with cells. In addition,
medium containing 1% methyl cellulose, as a solidifying agent, should then be
used Plaques can be picked usmg a mlcroplpet equipped wrth a sterile yellow tip
33 At this time a CaP04 precipitate will be vrslble at a x 100 magmficatlon Although
a precipitate that 1sformed at the right pH should be very fine, we often observe
some lumps that may be related to the size of the DNA molecules m the preclpl-
tate This does not greatly interfere with the efficiency of transfectlon
34. The length of the glycerol shock may vary for optimal results between 0.5 and 3
mm, depending on the cell type
35. The time of incubation is dependent on the length of the viral cycle If this 1slong
and a cytopathtc effect 1snot yet vlslble when the cells are grown to confluency,
the medium may be transferred to a novel dish containing cells at lower density.
36 In PRV we sometimes observe submolar bands that are derived from the repeat
regions. These varlatlons are likely caused by expansion or contraction in simple
sequence repeats wlthm the PRV IR and TR regions. The major form can usually
be plaque purified In naturally occurring isolates of the virus, similar size het-
erogeneity of the repeats 1soften observed
37. Enzymes other than FnuDII, HaeIII, and RsaI may also be used, provided that
the enzymes have 4-bp specificities and yield blunt ends. To this purpose, the
optimal ethldmm bromide concentration and the amount of enzyme to be used
should be determmed experimentally.
38. It 1s tempting to perform this ligation at a higher DNA concentration, to subse-
quently package the resultmg cosmld concatemers. However, this would be
disastrous as each monomer m the concatemers has the cos site at a very different
site (owing to the very nature of the random insertion of the mutagenic ohgo-
nucleotlde). The distance between the cos sites would therefore often not be 33-50
de Wind, van Ztjl, and Berns
84
kbp, the required distance for packaging. Moreover, the rare cosmids that will be
packaged do not carry a contiguous fragment of the vnus genome anymore

<<

. 13
( 61 .)



>>