Virology 293, 75–85 (2002) doi:10.1006/viro.2001.1235, available online at http://www.idealibrary.com on
Structural Studies of Human Enteric Adenovirus Type 41 Anne-Laure Favier,* Guy Schoehn,† Michel Jaquinod,* Charlotte Harsi,‡ and Jadwiga Chroboczek* ,1 *Institut de Biologie Structurale, 41 rue Jules Horowitz, 38027 Grenoble, France, †EMBL, Grenoble Outstation, 6 rue Jules Horowitz, BP180, 38041 Grenoble, France; and ‡Departamento de Microbiologia, Instituto de Ciencias Biomedicas, University of Sao˜ Paulo, Sao˜ Paulo, SP, 05508-900, Brazil Received July 24, 2001; returned to author for revision September 5, 2001; accepted October 10, 2001 Enteric adenoviruses of serotypes 40 and 41 possess some specific structural features, one of which is the presence on the virion of two fibers of different lengths and primary sequences. These viruses are notoriously difficult to grow under laboratory conditions. In this paper the successful growth and purification of Ad41 are presented in detail. Structural Ad41 proteins were analyzed by biochemical methods, mass spectrometry, and electron microscopy (EM), in order to identify and localize them on polyacrylamide denaturing gels and to assess the proportion of short and long fibers in the virion. Surprisingly, the three proteins composing virus short and long pentons did not totally enter the denaturing polyacrylamide gels, which is probably due in part to their high pI. The pentons were separately purified and their dimensions were estimated from EM data. The EM images suggest that there are the same amounts of short and long fibers in each virion. © 2002 Elsevier Science (USA)
one of the reasons for the narrow and specific tropism of enteric serotypes is the presence in the gastrointestinal tract of a receptor interacting with the short fiber. Adenovirus internalization is mediated by another viral protein, the penton base, interacting with integrins ␣v. It was demonstrated for several Ad serotypes that an —ArgGly-Asp— (RGD) motif, probably localized in the flexible loop region of the penton base, is involved in virus cell entry through interaction with ␣v␤3 and ␣v␤5 integrins (Bai et al., 1994; Mathias et al., 1994; Schoehn et al., 1996; Stewart et al., 1997; Wickham et al., 1993). However, both Ad40 and Ad41 lack the RGD motif in their bases; instead Ad40 carries RGAD and Ad41 carries IGDD sequences (Albinsson and Kidd, 1999). Furthermore, the uptake of Ad41 in cells which are efficiently producing commonly used serotypes was shown to be delayed, suggesting that Ad41 entry is independent of ␣v integrins (Albinsson and Kidd, 1999). We are interested in the structural proteins of Ad41. To obtain these proteins we need reliable virus preparations. Following the published protocols and introducing some improvements we were able to culture the virus in cells commonly used in virology laboratories, purify it with comparatively high yield, and study some interesting properties of its structural proteins. We identified and localized structural viral proteins on polyacrylamide denaturing gels. In addition, we estimated the ratio of the short to long fibers in Ad41 virions.
The human enteric adenoviruses are studied for two main reasons. First, they are an important human pathogen, accounting for nearly 20% of infantile diarrheas. Second, as they do not normally cause disease outside of the gastrointestinal tract (Brown et al., 1992), they are proposed as gene transfer vectors for intestinal epithelial cells (Croyle et al., 2000). The human enteric adenovirus serotypes 40 and 41 are known to grow poorly under laboratory conditions. Several defects in their life cycle have been observed that might explain poor growth, albeit probably only in part (Mautner et al., 1995, 1999; Tiemessen and Kidd, 1995). In addition, these serotypes have some specific structural features which suggest that the virus interaction with the host cell at the beginning of infection might be different from that of other human adenoviruses. Adenovirus (Ad) 2 fiber protein is responsible for the initial attachment of the virus to primary receptors on cell surfaces. However, enteric Ad particles contain two fibers of different primary sequences and different lengths. It was shown that the longer fiber, but not the shorter fiber, interacts with CAR, a protein that is a cell receptor for several human serotypes (Roelvink et al., 1998). It is possible that
1 To whom correspondence and reprint requests should be addressed. Fax: (33) 4 38 78 54 94. E-mail: [email protected]
2 Abbreviations used: Ad, adenovirus; FBS, fetal bovine serum; FFU, focus fluorescent unit; Freon, trifluorotrichloroethane; IGDD, Ile-GlyAsp-Asp; MALDI, matrix-assisted laser desorption/ionization; MS, mass spectrometry; OPU, optical particle unit; RGD, Arg-Gly-Asp; RGAD, Arg-Gly-Ala-Asp; Versene, PBS containing 0.53 mM EDTA; QTof, quadrupole time-of-flight.
RESULTS Ad41 growth and purification Enteric Ad40 and Ad41 are known as fastidious adenoviruses since they do not grow well in conventional 75
0042-6822/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.
FAVIER ET AL. TABLE 1 Comparison of Viral Preparations of Ad41 and Ad2 OPU/ml
FFU/cell 1.5 10
1.9 ⫻ 10 11 5.5 ⫻ 10 11
9 ⫻ 10 8 16 ⫻ 10 8
0.31 ⫻ 10 3 3.43 ⫻ 10 3
1.9 ⫻ 10 11 6.0 ⫻ 10 10 1.1 ⫻ 10 11 2.9 ⫻ 10 11
9 ⫻ 10 8 8.17 ⫻ 10 7 2.25 ⫻ 10 8 1.0 ⫻ 10 9
0.47% 0.13% 0.20% 0.34%
0.31 ⫻ 10 3 0.10 ⫻ 10 3 0.40 ⫻ 10 3 0.60 ⫻ 10 3
cell lines. The described methods use 293 (human embryonic kidney), Hep2 (human larynx carcinoma), CaCo2 (human colon adenocarcinoma), Chang (conjunctival), and PLC/PRF/5 (human liver hepatoma) cell lines. The method for Ad41 growth described here makes use of known protocols for culturing fastidious enteric adenoviruses under laboratory conditions, using routine cell cultures, with some careful adjustments. We were unable to grow Ad41 in HeLa cells and so 293 cells were used instead. Preparation of Ad41 was compared with the preparation of Ad2, known to give a good yield in cell cultures. Ad2 was purified from Ad2-infected cells, using the same purification protocol (Mittereder et al., 1996). Both viruses were titered on 293 cells. The lower band of Ad41 in a cesium chloride gradient was obtained at a density of 1.36 g/cm 3 and the upper band, containing defective particles, had a density of 1.31 g/cm 3. Using for quantification of adenovirus the formula of Mittereder et al. (1996), we obtained 5.5 ⫻ 10 11 OPU/ml (16 ⫻ 10 8 FFU/ml) for Ad2 from eight 175-cm 2 flasks and 1.9 ⫻ 10 11 OPU/ml (9 ⫻ 10 8 FFU/ml) for Ad41 from thirty 175-cm 2 flasks (Table 1). In the Ad41 preparation the ratio of infectious to physical particles seems to be comparable with that of Ad2 but virus yield per cell is about 10 times lower for Ad41 than for Ad2 (Table 1). However, the ratio of infectious to physical particles is about 10 times higher than in other methods describing Ad41 preparation (Brown et al., 1992). Titers of other Ad41 preparations used during this study are also given in Table 1. To achieve the high yield of infectious Ad41 the following parameters seem to be of importance. Use of trypsin for 293 cell splitting should be avoided, cells used for virus production should be well attached, infected cells should be recovered when cytopathic effect (cell clustering) starts to be observed, and virus purification should be carried out rapidly on large amounts of infected cells to avoid virus losses by temperature effect and surface attachment. Enteric adenoviruses Ad40 and Ad41 cause a very specific cytopatic effect in cell cultures, resulting in cell clustering followed by, or concomitant with, cell detachment (Brown et al., 1992). While with Ad2 the saturation of infection was obtained with 0.25 l (4 ⫻ 10 5 FFU) of purified virus, it was not possible to achieve saturation of infection with Ad41 since already above
1.5 0.14 0.3 2.3
0.75 l (6.75 ⫻ 10 5 FFU) of Ad41, cell detachment was observed (Fig. 1). The level of saturation was measured with anti-Ad2 and anti-Ad41 polyclonal antibodies. It should be borne in mind that this reflects the production of structural virus proteins, which is not identical to virus production since not all structural proteins synthesized will be used for virus assembly. Nevertheless, this approach can be useful in the case of viruses which have physical or infectious particles that are difficult to measure (Tiemessen and Kidd, 1994). Identification of viral proteins Our initial goal was the determination of the stoichiometry of long and short fibers in Ad41 virions. For this, we envisaged using densitometry of stained polyacrylamide denaturing gels. However, using published data we were unable to unequivocally localize two fibers among other virus structural proteins. Therefore we decided to identify virion proteins visible on such gels after dye staining.
FIG. 1. Saturation curve of Ad41 versus Ad2. The virion production was analyzed 24 h postinfection by immunofluorescence. At low dilutions the values for Ad41 decreased since near virus saturation the cells detach very easily and are lost at washing.
STRUCTURAL STUDIES OF ADENOVIRUS TYPE 41
In our hands N-terminal amino acid analysis could be applied to two proteins only since the other turned out to be blocked at the amino terminus. Indeed, it has been shown that, for example, the hexon of Ad2 is modified by acetylation (Jornvall et al., 1974). The proteins VI and VII, synthesized as precursors and cleaved by Ad protease at the (L,M)XGG2 sequence (Freimuth and Anderson, 1993), were identified by their processed N-terminal amino acid sequences, respectively, AFS- and AKRRSS-. In the immature precursor proteins, the sequences immediately N-terminal to the sequences identified are LNGG and MYGG, respectively (Grydsuk et al., 1996; Mautner et al., 1995). For mass spectrometry analysis the viral proteins were analyzed on 10% SDS–polyacrylamide gels stained with Coomassie brilliant blue. The appropriate bands were excised and digested by trypsin, and then the crude mixture was analyzed by MALDI-MS. The confirmation was obtained with MS/MS experiments performed with a QTof mass spectrometer. The resulting sequences deduced from extensive series of b and y ions were used to screen the databank by Blast and FASTA programs. The proteins were identified using homologies found with the sequence of Ad41 and Ad40 (Table 2). It should be noted that the only differences in the sequence of Ad41 proteins when compared with those of Ad40 are of conserved type. The comparison of the mobility in SDS–PAGE for Ad2 and Ad41 proteins is shown in Fig. 2. Surprisingly, with mass spectrometry, neither short nor long fibers nor the penton base protein could be localized on such a gel. A certain amount of penton base was found in the stacking gel at the entry into the resolving gel and submitted to mass spectrometry analysis (Table 2). The band visible in Fig. 2a above protein IIIa (with mobility suggesting a long fiber) contains the product of hexon proteolysis. The fragment was unambiguously identified using peptide sequences deduced from data obtained from ES-MS/MS. Four peptides were obtained with sequences FIGTNINK, NTELSYQLML, ANGQTWTADDNY, and DVNMILQSSLGNDLR located at positions 150–157, 342–351, 408–419, and 562–576, respectively, of the Ad41 hexon protein. The bands running below protein VII were not analyzed. The silver staining of a similar gel did not reveal additional bands in the area expected for pentons protein mobility (Fig. 2b). It shows in addition that the hexon and protein VII are poorly stained with the silver procedure. We were able to localize some of these three missing proteins only by applying increased amounts of Ad41 to the gel and by using the sensitivity of the Western blot technique (Fig. 3a). The long fiber (LF) and base run together just above protein IIIa. The molecular masses of these proteins are not very different: approximately 57 kDa for the base protein, 60.5 kDa for the long fiber, and about 63 kDa for protein IIIa. The short fiber (41.4 kDa) is seen between protein V
77 TABLE 2
Identification of Ad41 Structural Proteins by Mass Spectrometry Hexon
Identified by MALDI-MS mapping of serotype 41 peptides NTELSYQLMLDALGDR 353 388 NTELSYQLMLDALGDR-
Identified by MALDI-MS mapping of serotype 41 peptides Identified by MALDI-MS mapping of serotype 41 peptides FSAIEAVV 62 70 FSAIEAVV
104 VLDMASTYFDIR FFDTFQEITSVSR
291 DEQTFQEITSVSR IPSQFTLSTEEER
60 EEAGLLDDGVEFVR QAAQQLGEFAYGK
126 QAAQQVGEFAYGK EVGGLLLDQSNPTP
Identified by MALDI-MS mapping of serotype 41 peptides (Grydsuk et al., 1996) TTVDDVIDSVVADA-R 55 70 TTVDDVIDSVVADAQR 149
QAAAAIASMAQ 160 QAAAAIASMAQ
Note. The amino acid sequences were deduced from the tandem MS/MS spectra using an ES-Q-TOF instrument equipped with a nanospray source. To identify the proteins and to consequently locate the position of these peptides in the polypeptide chain, the experimental sequences (in boldface type) were compared to the viral sequences present in the NCBI data bank using the Blast program (http://www.ncbi.nlm.nih.gov:/BLAST/). The locations of the peptides are given in comparison to the Ad40 and Ad41 amino acid sequences deduced from DNA (GenBank Accession Nos. L19443 and U14517). The amino acid sequences are very similar in both subtypes. Hexon, penton base, and IIIa proteins were additionally identified based on their tryptic peptide mass mapping by the Profound program (http://prowl.rockefeller.edu/ cgi-bin/ProFound) using the masses determined with a MALDI-MS instrument operating in reflectron mode. The specific parameters for peptide identification are given under Materials and Methods.
and protein VI. Surprisingly, Ad41 protein VI could not be revealed with the polyclonal antibody against protein VI of Ad2 despite the fact that there is significant homology in the primary sequences of these proteins (Grydsuk et al., 1996).
FAVIER ET AL.
Results concerning three proteins forming long and short pentons suggested that they enter only partially the classical SDS–polyacrylamide gels. It should be noted that theoretical pI values for Ad41 long and short fibers are 7.51 and 9.13, compared with 5.61 and 5.85 for fibers of Ad3 and Ad2, respectively (Table 3). For the visualization of Ad41 penton proteins we ran a two-way gel without stacking, with the wells made in the middle of the gel. Such a gel allows the visualization of proteins running toward the cathode and toward the anode at the same time. Four protein bands running toward the cathode were observed (Fig. 3b). One of them was identified by mass spectrometry as the penton base protein (pI 5.5), and three others were of an amount not permitting MS analysis. Then we used a two-way gel for immune analysis of Ad41 fibers. Monoclonal antibody 4D2 recognizing the Nterminus of both fibers showed that some fibers run toward the cathode at 30 min, whereas a more sizable portion of fibers is resolved toward the anode at 200
FIG. 2. Absence of Ad41 short fiber, the long fiber, and the penton base proteins on dye-stained denaturing polyacrylamide gels. (a) Analysis and identification of Ad41 proteins on 12.5% gel stained with Coomassie brilliant blue. Samples of Ad2 (7 g) and Ad41 (6 g) were analyzed. The molecular mass of markers is given in kilodaltons. Ad2 proteins were designated according to Maizel et al. (1968). Ad41 proteins were identified by the N-terminal amino acids and by mass spectrometry analyses. (b) Comparison of Coomassie blue and silver stain. In both cases a similar sample of Ad41 (1.3 g) was analyzed.
FIG. 3. (a) Identification of Ad41 virus proteins by Western blot. Samples of Ad41 (0.3 g each) were run on 10% SDS–polyacrylamide gel and the virus proteins were transferred onto PVDF membrane. The Western blot was performed with antibodies obtained against: Ad41, 1; Ad40, 2; Ad3, 3; Ad3 penton, 4; Ad2 penton base, 5; Ad2 IIIa protein, 6; Ad2 fiber N-terminus (4D2), 7. LF—long fiber, SF—short fiber. * denotes base protein. (b) Visualization of Ad41 proteins by two-way electrophoresis and Coomassie brilliant blue staining. Ad41 (4 g) was applied to the well of 10% a SDS– polyacrylamide gel (prepared as described under Materials and Methods) and run for 90 min. Bands of proteins running toward the anode were excised and analyzed by MALDI-MS. (c) Visualization of Ad41 fibers by two-way electrophoresis and Western blot. Samples of Ad41 (1 g each) were run on 15% SDS–polyacrylamide gel (prepared as described under Materials and Methods). Samples were applied at different times, to obtain indicated run times for the individual sample. On the left is presented the PVDF membrane after the electrotransfer, stained with Coomassie blue. The right part of the figure presents the Western blot analysis performed with 4D2 monoclonal antibody specifically recognizing the conserved Ad fiber N-terminus. Marker positions are shown for a 200-min run.
min (Fig. 3c, right). Similarly as in the case of classical gels used for Western and dye staining (Figs. 2 and 3a), this analysis strongly suggests that under the
STRUCTURAL STUDIES OF ADENOVIRUS TYPE 41 TABLE 3 Theoretical pI of Penton Proteins Fiber Ad41 Ad2 Ad3
Long 7.51 Short 9.13 5.85 5.61
Penton base 5.50 5.05 5.26
Note. The theoretical pI of the fiber and the penton base proteins was obtained with the software http://www.up.univ-mrs.fr/⬃wabim/ d_abim/compo-p.html.
gallery, trimeric (asterisk and enlarged contouring in Figs. 4B and 4C), which is in agreement with the fiber head images of other serotypes (Albiges-Rizo et al., 1991; van Raiij et al., 1999). When the base sits on its side, at this resolution (approximately 15 Å), the shape of the penton is very similar to the pentons of other serotypes (reviewed by Chroboczek et al., 1995). The lengths of the fibers were 340 ⫾ 13 Å for the long fiber and 200 ⫾ 13 Å for the short fiber. These dimensions are consistent with those of fibers and penton bases of some other human adenoviruses (Table 4). Ratio of fibers in Ad41 virions
conditions used two fibers’ pools exist, behaving differently. Purification of major structural proteins of Ad41 The CsCl gradient fraction recovered above the virus band contains free viral structural proteins and the immature/defective virions which can be disrupted by dialysis against water. This fraction was used for the purification of Ad41 pentons — complexes of base and fiber. The long and the short pentons could not be purified on the same column, possibly because of their difference in charge as described above. And so, the short penton was purified on S-Sepharose at pH 6.1, and the long penton was purified on Q-Sepharose at pH 7.5. The hexon could not be purified directly on the Q-Sepharose since, surprisingly, we have seen it all along the salt gradient. However, after the passage of the protein mixture on the S-Sepharose, on which it was recovered in the flow-through fraction, the hexon was easily purified on Q-Sepharose. When purified pentons, which on electron microscopy (EM) images have shown the correct structure (see Fig. 4), were run on classical PAGE, a small amount of penton base could be identified by mass spectroscopy at the appropriate position below the 62-kDa MW marker. However, no fiber was detected (results not shown). Electron microscopy By using the negative staining technique we were able to visualize the entire virus (Fig. 4A), a group of nine hexons (Crowther and Franklin, 1972) (data not shown), and the purified long and short pentons (Figs. 4B and 4C). The base of the penton can attach differently to the carbon giving top or side views (the lower and upper panels of Figs. 4B and 4C, respectively) and the fibers, which are quite flexible, cannot necessarily point along the fivefold axis. When the base sits on the carbon in a top view position its pentagonal shape is clearly visible but the fiber is bent just at the junction between the base and the fiber (lower panels of Figs. 4B and 4C), like it is in the image of Ad41 particles (Fig. 4A). The fiber heads are clearly globular and, in some of the images from the
The densitometric scanning of the denaturing polyacrylamide gels to estimate the ratio of fibers in Ad41 virions could not be used in view of the fact that the pentons only partly enter such gels. The ratio was therefore estimated by electron microscopy. By counting the number of short and long fibers in 20 negatively stained virions (like that in Fig. 4A, zoom out) we estimate the proportion of long to short fiber to be 1:1. For this estimation only nonambiguously attributed fibers (long or short) were taken into account. Of 20 virions, the same number of short and long fibers was observed in 5 virions. This does not mean that the proportions in other virions were different, only that it is not possible to use these images to unambiguously estimate the ratio. DISCUSSION The goal of this work was to study the structural properties of enteric Ad41. For this, rather large amounts of purified virus were required. However, Ad41 is known to grow poorly under laboratory conditions and an improved protocol for its growth and purification had to be established. One of the problems encountered, described also by others (Brown et al., 1992), was the profound cytopathic effect exerted by Ad41 infection on host cells. It results in the detachment of Ad41-infected cells and their subsequent loss during washing, a phenomenon not observed during growth of Ad2 or Ad3 under laboratory conditions. Some adjustments allowed us to improve the yield of Ad41. First, trypsin was not employed for cell splitting since we found that its use impaired subsequent infection. Second, we have taken care to have the 293 cells well attached before infection and, once infected, to recover them just before the step of spontaneous cell detachment. Finally, virus was extracted from large amounts of infected cells and for its purification we employed the method of Kanegae et al. (1994), which permitted the virus to be purified in 1 day. This decreased the virus loss in terms of physical as well as infectious particles, thanks to a smaller temperature effect and diminished surface attachment of virus. Thus, by carefully adjusting the growth and purification conditions we were able to routinely obtain the Ad41 prepa-
FAVIER ET AL.
FIG. 4. (A) Electron microscopy of Ad41 negatively stained with ammonium molybdate. Two different fibers are clearly seen, one short and one longer. The zoom shows the long and short fibers attached to different bases in the Ad41 virion. (B) Gallery of long pentons negatively stained with sodium silicotungstate. (Top) The base in side view; (bottom) the base in top view. In some images the trimeric head of the fiber is clearly seen (white asterisk). (C) Gallery of short pentons negatively stained with sodium silicotungstate. (Top) Pentons with the base in side view; (bottom) pentons with the base in top view. In some images the trimeric head of the fiber is clearly seen (white asterisk). The unambiguously identified long and short fibers are indicated by small white squares or circles, respectively, in the enlarged view.
STRUCTURAL STUDIES OF ADENOVIRUS TYPE 41
TABLE 4 Dimensions of Fibers and Penton Bases of Some Human Adenoviruses, as Measured by EM Using Negative Stain Technique (Adapted from Chroboczek et al., 1995) Penton length (Å) Serotype Ad2 Ad5 Ad3 Ad40 Ad40 Ad41 Ad41
Long Short Long Short
Fiber length 373 ⫾ 7 372 ⫾ 12 160 ⫾ 14 — — 340 ⫾ 13 a 200 ⫾ 13 a
End-on-view 400 ⫾ 9 — 181 ⫾ 8 390 ⫾ 13 263 ⫾ 18 375 ⫾ 13 235 ⫾ 13
Side view 431 ⫾ 6 — 227 ⫾ 11 432 ⫾ 9 319 ⫾ 11 405 ⫾ 13 264 ⫾ 13
Number of repeats
Shaft length (Å) per repeat
21.5 21.5 5.5 20.5 11.5 21.5 11.5
13.5 13.4 14. 13.9 14.3 13.5 13.1
Note. The fiber length was measured on free fiber except for Ad41. a The length of the fiber was measured in the virus so this measure does not take into account the last 20 or so N-terminal amino acids. To calculate the length per shaft repeat, the length of the head domain, 49 Å, was subtracted (Chroboczek et al., 1995). End-on-view penton lengths were measured from the far side of the base, lying end-on to the tip of the fiber head (lower panels of Figs. 4B and 4C). The side view penton length was measured from the bottom of the base to the tip of the fiber (top panels of Figs. 4B and 4C).
rations comparable with those of Ad2. The virus yield per host cell when compared with that of Ad2 was 10 times lower but the final titer was significantly higher than for Ad41 prepared by others (Brown et al., 1992). The general structure of human adenoviruses is similar. They have icosahedric capsids made of 240 hexons, with 12 vertices each containing the complex of two proteins, called the penton. The penton is made of a base and an outwardly extending fiber protein. However, the enteric adenoviruses of serotypes 40 and 41 (subgroup F) contain two different fibers, each encoded by a separate gene (Kidd et al., 1993; Yeh et al., 1994). It was of interest to know the ratio of long to short fibers in Ad41 virions. Initially, we planned to assess it by densitometry of protein bands visible after virion electrophoresis on denaturing polyacrylamide gels. Trying to identify precisely the position of these proteins on the gel using biochemical techniques of protein analysis and mass spectrometry, we have not been able to find the fibers or the base protein using Coomassie brilliant blue. They were also not visible with silver stain. Eventually, we were able to localize these proteins using rather concentrated samples of virus and highly reactive specific antibodies. It has been observed subsequently that, depending on the preparation, variable amounts of these proteins, usually below the amounts judged to be stoichiometrically reasonable, enter the gel. Even when the preparations used for EM, where long pentons were visible, were run on classical PAGE, we were able to identify by MALDI-MS only the penton base and not the fiber protein. This erratic behavior on classical denaturing polyacrylamide gels was in part understood when a basic pI of Ad41 fiber proteins was taken into account. To visualize penton proteins we ran two-way polyacrylamide denaturing gels. With dye stain we identified by MALDI-MS a penton base protein running, surprisingly,
toward the cathode. Subsequently, on similar gels used for Western blot with strongly fiber-reactive antibody, two pools of fibers could be observed, one running toward the anode and another toward the cathode (Figs. 3a and 3c). It is possible that the conditions of classical PAGE are not well adjusted for charged and quite hydrophobic pentons, resulting, for example, in their partial precipitation. Moreover, it should be borne in mind that the molar amount of fiber protein in Ad virions is not very high. Ad virions contain 240 copies of trimeric hexons, 74 copies of monomeric protein IIIa, and only 12 copies each of pentameric base protein and trimeric fiber (Van Oostrum et al., 1987). Finally, in the case of Ad41 and Ad40, the molar amounts of each fiber are twice lower than that for fibers of other human Ads. Assuming that a certain amount of fibers is lost due to the precipitation, it could explain our difficulties in the identification of these proteins on polyacrylamide gels. In addition, taking into account that also base protein was seen in two different locations depending on the sensitivity of the visualization technique (Figs. 3a and 3b) it is likely that dissociation of the penton components under denaturing conditions is less effective for Ad41 than for other human Ads. This aberrant behavior of Ad41 pentons excluded the use of densitometry for estimation of the ratio of the fibers. The fiber ratio in Ad41 virions was therefore assessed by electron microscopy. When nonambiguously identified fibers (short or long) were taken into account, the mean ratio was found to be 1:1. In the virion, the homotrimeric fiber is attached to the homopentameric penton base through the fiber N-terminal region (Devaux et al., 1987). This mismatch concerns all human adenoviruses and the structural features permitting it are not known. The first 40 amino acids of both Ad41 fibers are quite similar in both primary sequence and charge, and both possess the conserved motif 10FNPVYPYD 17, which
FAVIER ET AL.
was proposed to be involved in the specific recognition between the fiber and the penton base through complementary ionic and hydrophobic interactions (Caillet-Boudin, 1989). It is likely that under conditions of a similar level of protein synthesis for both fibers and assuming the similar affinity of both fibers to the penton base protein, the 1:1 ratio of fibers in Ad41 virion will be achieved spontaneously. Understanding the mechanism of fiber sorting must await some additional data and in particular the estimation of the level of the cellular synthesis of fiber molecules which are competent for interaction with the bases in the virion. Fibers mediate adenovirus attachment to the host cell. The majority of human adenoviruses use coxsackie and adenovirus receptor (CAR) protein as a primary receptor. However, Ad3 fiber (subgroup B), Ad40 and Ad41 short fibers (subgroup F), and Ad37 fiber (subgroup D) do not attach to CAR (Devaux et al., 1987; Roelvink et al., 1998). The major determinant of virus tropism is the primary receptor. Indeed, the cell tropism of these serotypes is rather different—CAR binding serotypes are predominantly respiratory viruses, whereas Ad3 is respiratory and ocular, Ad40 and Ad41 are enteric, and Ad37 is ocular. The short fibers of Ad40 and Ad41 are nearly as unrelated to their long fibers as they are to fibers of other subgroups (Chroboczek et al., 1995), which substantiates the data on the lack of interaction of these short fibers with CAR. Since the enteric tropism of Ad40 and Ad41 cannot be explained by the interaction of the long fiber of these viruses with ubiquitously expressed CAR, it prompts us to think that this is the interaction of short fibers with a digestive tract-specific component that is responsible for the observed tropism. The structural and functional studies on these viruses and their structural components will help us to understand the mechanism of their cell entry and specifically to find the rationale behind their particular cell tropism. MATERIALS AND METHODS Virus culture and purification Ad41 strain Tak was obtained from Dr. R. Glass (CDC, Atlanta, GA). 293 cells, a transformed human embryonic kidney cell line, were grown in EMEM supplemented with 2 mM glutamine and 10% FBS. They were divided twice a week at a split ratio of 1:3, using versene. The first virus preparation was obtained by infecting a 25-cm 2 dish with a small amount of untitered Ad41, collecting the cells after 3 days, and preparing the infected cell extract by three cycles of freezing and thawing. Dishes of 175 cm 2 were seeded with 7 ⫻ 10 6 cells and at 90% confluency each dish was infected with infected-cell extract obtained from the 25-cm 2 dish. The resulting infectious
virus (henceforth called the inoculum) was estimated by plaque focus fluorescence assay carried out on 293 cells. We routinely obtained the nonpurified virus suspensions with a titer of 5 ⫻ 10 7 FFU/ml. For subsequent infections, 293 cells in 175-cm 2 flasks were rinsed twice with 10 ml PBS containing 0.5 mM Ca 2⫹ and 0.5 mM Mg 2⫹ and infected at a m.o.i. of 1.5–2, which amounted to about 250 l of viral inoculum diluted in 3 ml of EMEM containing 2 mM glutamine, 0.2% FBS, 50 IU/ml (50 g/ml) streptomycin, 50 IU/ml (50 g/ml) penicillin. Infection was carried out in the CO 2 incubator for 1 h with mixing every 15 min. Without removal of viral inoculum, each of the dishes was filled with 22 ml of EMEM/0.2% FBS and left in the incubator under 5% CO 2 for approximately 3 days. The infected cells were routinely recovered before the majority of cells detached. Infected cells were detached by tapping the dishes and collected by a 5-min centrifugation at 1000 rpm. Cells from 30 flasks were suspended in 8–10 ml of EMEM/0.2% FBS and stored at ⫺20°C. They were disrupted by three cycles of freezing and thawing and then glass beads (approximately 2 ml, 425–600 m, Sigma) were added and the mixture was vortexed five times for 1 min in the cold. The mixture was extracted with an equal volume of freon and clarified by centrifugation at 3000 rpm for 10 min. The resulting supernatant was centrifuged for 10 min at 10,000 rpm. The virus was purified from the supernatant according to Kanegae et al. (1994) with some modifications. Briefly, the clarified cell extract was applied on top of the noncontinuous cesium chloride gradient. For this a 12-ml tube with 2.5 ml of 4 M CsCl in 20 mM Tris buffer, pH 7.4, containing 1 mM EDTA, 50 mM NaCl (TNE), and 5 ml of 2.2 M CsCl in TNE was filled with clarified extract and centrifuged in a Beckman SW41 rotor at 40,000 rpm for 1 h. The lowest opalescent band of virus (approximately 450 l, density of 1.36 g/cm) was collected, mixed with an equal volume of saturated CsCl, and placed at the bottom of a 2.2-ml tube of a TLS55 rotor of a Beckman Optima TL ultracentrifuge. It was overlaid with 0.62 ml of 4 M buffered CsCl followed by 0.7 ml of 2.2 M buffered CsCl and centrifuged for 3 h at 50,000 rpm. The opalescent virus band was collected, dialyzed against three changes of 200 mM potassium phosphate buffer, pH 7.8, and finally mixed with an equal volume of 2 M sucrose. Virus (9 ⫻ 10 8 FFU/ml) was stored in glass tubes at ⫺20°C (Croyle et al., 2000). Saturation curve 293 cells were seeded at a density of 3 ⫻ 10 4 cells per well in 96-well culture dishes, in such a way that infection could be done after 2 days at 90% confluence. We found that such a prolonged contact of cells with plastic increased 293 cell adherence, which was impor-
STRUCTURAL STUDIES OF ADENOVIRUS TYPE 41
tant in view of the easy detachment of these cells at high m.o.i. Cells were rinsed two times with PBS–Dulbecco’s containing 0.5 mM MgCl 2 and 0.5 mM NaCl 2. A cascade of 1:2 dilutions of Ad41 in EMEM/0.2% FBS was prepared and 50 l was added to each well (done in triplicate). All solutions were warmed to 37°C before use. One hour after infection, 100 l of EMEM/0.2% FBS was added without supernatant removal. After 24 h, the virion production was analyzed by immunofluorescence as follows. After the supernatant was removed, the cells were fixed for 10 min with cold methanol containing 15% PBS, saturated for 30 min at 37°C with PBS/1% BSA, incubated for 1 h at RT with anti-Ad41 antibody (diluted 1:500 in PBS/1% BSA), washed two times with PBS, incubated with the secondary antibody FITC (Fluorescein-Conjugated AffiniPure Goat Anti-Rabbit IgG, Jackson ImmunoResearch, diluted 1:100) for 1 h in the dark at RT, and washed two times with PBS. Finally, cells were overlaid with 50 l of PBS before the fluorescence was measured for 1 s at 485 nm:580 nm. Antibodies For Western blot analysis the polyclonal rabbit antibodies were used at the following dilutions: anti-Ad40 (from A. Kidd), anti-Ad41 (made by authors), anti-Ad2 protein VI (made by authors), and anti-Ad2 protein IIIa (from P. Boulanger), at 1:10,000; anti-Ad3 fiber ⫹ base (made by the authors), at 1:50,000; anti-Ad3 (made by the authors), at 1:2000, anti-Ad2 base protein (from P. Boulanger), at 1:20,000. Monoclonal antibody 4D2 recognizing the FNPVYPY fragment present in the N-terminal part of all fibers (from J. Engler) was used at 1:50,000. The ECL detection system (Amersham Pharmacia) was used throughout this work. Purification of Ad41 structural proteins Part of the CsCl gradient localized above the virus band after the first centrifugation was recovered and dialyzed against water to disrupt virions (Laver et al., 1969). In order to obtain short pentons (complex of the penton base and short fiber) a portion of this preparation was dialyzed against 40 mM MES buffer, pH 6, and chromatographed on a S2-Sepharose column equilibrated with the same buffer. Elution was performed using a 30 ml linear NaCl gradient, 0 to 500 mM in the same buffer. The viral proteins in the column fractions were visualized by Western blot with the anti-Ad41 antibody and the fractions containing short pentons (at 150–180 mM NaCl) were pooled. Under these conditions the hexon did not attach to the resin and was recovered in the flow-through fraction. It was further purified on the Q-Sepharose column, from which it was eluted at about 190 mM NaCl. The long pentons were obtained from the initial water-treated preparation dialyzed against 20 mM
Tris buffer, pH 7.5, and fractionated on the Q2-Sepharose column equilibrated with the same buffer, using conditions as described above. Fractions containing long pentons (at 60–80 mM NaCl) were pooled. Electrophoretic analysis of Ad41 proteins Virus or viral proteins were analyzed on the SDS– polyacrylamide gels and stained with Coomassie brilliant blue or with silver stain. For the visualization of short and long pentons, two-way gels without stacking were used. For this SDS–polyacrylamide gels were cast in the DNA electrophoresis apparatus (20 cm long) with sample wells formed at approximately half of the gel length. The gel was run at 40 mA for the indicated times and stained with Coomassie brilliant blue or analyzed by Western blot. N-terminal amino acid analysis N-terminal amino acid analysis was carried out by Edman degradation performed on Problott membranes (Applied Biosystems) containing viral proteins after PAGE and electroblotting in 3-[cyclohexylamino]-1-propanesulfonic acid buffer. Mass spectrometry analysis The general protocol of Shevchenko et al. (1996) was followed with some modifications. Acrylamide denaturing gels containing viral proteins were stained with Coomassie brilliant blue. Excised protein spots were washed twice with 50 mM acetate ammonium buffer, pH 7.5, dehydrated by treatment with acetonitrile, and dried in a vacuum centrifuge. Gel pieces were swollen in 20–25 l of 50 mM NH 4HCO 3 containing 10% acetonitrile and 0.1–0.5 g of trypsin (Sigma). Proteolytic digestion was carried out for 1–4 h at 37°C. Mass spectra of the tryptic digests were acquired on a Voyager Elite X1 (Perspective Biosystems) MALDI-TOF mass spectrometer equipped with delayed extraction and operated in the reflectron mode. One microliter of each digest was deposited directly onto the sample probe and mixed with 1 l of saturated solution of 2,5-dihydroxybenzoic acid in acetonitrile:water:TFA (60:39:1). A mass list of peptides was obtained for each protein digest. For protein identification the peptide mass fingerprint was analyzed by computer, using two different programs: MS-FIT, http:// prospector.ucsf.edu/ucsfhtml3.2/msfit.htm, or PROFOUND, http://126.96.36.199/prowl-cgi/ProFound.exe. Masses of peptides from the proteolytic digestion were compared to the masses of a peptide database calculated from the NCBI nr protein database with mass tolerances of ⫹/⫺ 0.2 Da and taking into account the fact that the methionine residues are partially oxidized [(mass/charge (m/z) values listed in the PROFOUND and MS-FIT programs)]. Other
FAVIER ET AL.
parameters were as follows: taxonomy, all viruses; protein mass range of 10–120 kDa; pI of 5–12; unmodified (SH) cysteines; a maximum of three missed tryptic cleavage sites. For proteins that could not be identified by MALDI-MS the digest mixture was pooled, dried in a vacuum centrifuge, desalted with ZipTip C18 (Millipore), and submitted to the nanospray MS/MS analysis. The MS/MS spectra were acquired with ES-Q-Tof instrument (Micromass, Manchester, UK) used with a Z-Spray ion-source working in the nanospray mode. About 3 to 5 l of the desalted sample was introduced via a medium sample needle (PROTANA Inc., Odense, Denmark), with the capillary voltage set to an average of 1000 V and that of the sample cone set to 50 V. Glu-fibrinopeptide was used to calibrate the instrument in the MS/MS mode. The MS/MS spectra were transformed using MaxEnt3 (MassLynx, Micromass Ltd.) and the amino acid sequences were analyzed using PepSeq (BioLynx, Micromass, Ltd.). Amino acid sequences, sequence tags, or peptide ion fragments that could be determined were used to screen the protein data bases for homologous sequences (http://www.ncbi.nlm.nih.gov/blast/blast). Electron microscopy Samples at approximately 0.1 mg protein/ml were applied to the clean side of carbon on mica (carbon/mica interface) and negatively stained with 2% ammonium molybdate, pH 7.4 (native virus), or 1% sodium silicotungstate, pH 7.0 (long and short pentons). Micrographs were taken under low-dose conditions with a Jeol 1200 EX II microscope at 100 kV and a nominal magnification of 40,000 times. ACKNOWLEDGMENTS Anne-Laure Favier is a recipient of a Ph.D. fellowship from ARC (French Cancer Society). We are indebted to Jean-Pierre Andrieu and Mathilde Louwagie for N-terminal amino acid analysis. We are grateful to Alistair Kidd, Pierre Boulanger, and Jeff Engler for sera and to Roger Glass for the Ad41 initial sample. We thank Jerome Garin for the use of mass spectrometry facilities in his laboratory, Richard Wade for discussions, and Rob Ruigrok for encouragement.
REFERENCES Albiges-Rizo, C., Barge, A., Ruigrok, R. W., Timmins, P. A., and Chroboczek, J. (1991). Human adenovirus serotype 3 fiber protein. Comparison of native and recombinant proteins. J. Biol. Chem. 266(6), 3961–3967. Albinsson, B., and Kidd, A. H. (1999). Adenovirus type 41 lacks an RGD alpha(v) integrin binding motif on the penton base and undergoes delayed uptake in A549 cells. Virus Res. 64, 125–136. Arnberg, N., Kidd, A. H., Edlund, K., Olfat, F., and Wadell, G. (2000). Initial interactions of subgenus D adenoviruses with A549 cellular receptors: Sialic acid versus alpha(v) integrins. J. Virol. 74(16), 7691–7693. Bai, M., Campisi, L., and Freimuth, P. (1994). Vitronectin receptor antibodies inhibit infection of HeLa and A549 cells by adenovirus type 12 but not by adenovirus type 2. J. Virol. 68, 5925–5932.
Brown, M., Wilson-Friesen, H. L., and Doane, F. (1992). A block in release of progeny virus and a high particle-to-infectious unit ratio contribute to poor growth of enteric adenovirus types 40 and 41 in cell culture. J. Virol. 66, 3198–3205. Caillet-Boudin, M. L. (1989). Complementary peptide sequences in partner proteins of the adenovirus capsid. J. Mol. Biol. 1208, 195–198. Chroboczek, J., Ruigrok, R. W., and Cusack, S. (1995). Adenovirus fiber. Curr. Top. Microbiol. Immunol. 199, 163–200. Crowther, R. A., and Franklin, R. M. (1972). The structure of the groups of nine hexons from adenovirus. J. Mol. Biol. 68, 181–184. Croyle, M. A., Stone, M., Amidon, G. L., and Roessler, B. J. (1998). In vitro and in vivo assessment of adenovirus 41 as a vector for gene delivery to the intestine. Gene Ther. 5, 645–654. Croyle, M. A., Yu, Q. C., and Wilson, J. M. (2000). Development of a rapid method for the PEGylation of adenoviruses with enhanced transduction and improved stability under harsh storage conditions. Hum. Gene Ther. 11(12), 1713–1722. Devaux, C., Caillet-Boudin, M. L., Jacrot, B., and Boulanger, P. (1987). Crystallization, enzymatic cleavage, and the polarity of the adenovirus type 2 fiber. Virology 161, 121–128. Freimuth, P., and Anderson, C. W. (1993). Human adenovirus serotype 12 virion precursors pMu and pVI are cleaved at amino-terminal and carboxy-terminal sites that conform to the adenovirus 2 endoproteinase cleavage consensus sequence. Virology 193, 348–355. Grydsuk, J. D., Fortsas, E., Petric, M., and Brown, M. (1996). Common epitope on protein VI of enteric adenoviruses from subgenera A and F. J. Gen. Virol. 77, 1811–1819. Jornvall, H., Ohlsson, H., and Philipson, L. (1974). An acetylated Nterminus of adenovirus type 2 hexon protein. Biochem. Biophys. Res. Commun. 56, 304–310. Kanegae, Y., Makimura, M., and Saito, I. (1994). A simple and efficient method for purification of infectious recombinant adenovirus. Jpn. J. Med. Sci. Biol. 47, 157–166. Kidd, A. H., Chroboczek, J., Cusack, S., and Ruigrok, R. W. (1993). Adenovirus type 40 virions contain two distinct fibers. Virology 192, 73–84. Laver, W. G., Wrigley, N. G., and Pereira, H. G. (1969). Removal of pentons from particles of adenovirus type 2. Virology 39, 599–604. Maizel, J. V., White, D. O., and Scharff, M. D. (1968). The polypeptides of adenovirus II. Soluble proteins, cores, top components and the structure of the virion. Virology 36, 126–136. Mathias, P., Wickham, T., Moore, M., and Nemerow, G. (1994). Multiple adenovirus serotypes use alpha v integrins for infection. J. Virol. 168, 6811–6814. Mautner, V., Bailey, A., Steinthorsdottir, V., Ullah, R., and Rinaldi, A. (1999). Properties of the adenovirus type 40 E1B promoter that contribute to its low transcriptional activity. Virology 265, 10–19. Mautner, V., Steinthorsdottir, V., and Bailey, A. (1995). Enteric adenoviruses. Curr. Top. Microbiol. Immunol. 199, 229–282. Mittereder, N., March, K. L., and Trapnell, B. C. (1996). Evaluation of the concentration and bioactivity of adenovirus vectors for gene therapy. J. Virol. 70, 7498–7509. Roelvink, P. W., Lizonova, A., Lee, J. G., Li, Y., Bergelson, J. M., Finberg, R. W., Brough, D. E., Kovesdi, I., and Wickham, T. J. (1998). The coxsackievirus–adenovirus receptor protein can function as a cellular attachment protein for adenovirus serotypes from subgroups A, C, D, E, and F. J. Virol. 72, 7909–7915. Schoehn, G., Fender, P., Chroboczek, J., and Hewat, E. A. (1996). Adenovirus 3 penton dodecahedron exhibits structural changes of the base on fiber binding. EMBO J. 15, 6841–6846. Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. (1996). Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 68, 850–858. Stewart, P. L., Chiu, C. Y., Huang, S., Muir, T., Zhao, Y., Chait, B., Mathias, P., and Nemerow, G. R. (1997). Cryo-EM visualization of an exposed RGD epitope on adenovirus that escapes antibody neutralization. EMBO J. 16, 1189–1198.
STRUCTURAL STUDIES OF ADENOVIRUS TYPE 41 Tiemessen, C. T., and Kidd, A. H. (1994). Adenoviruses type 40 and 41 growth in vitro: Host range diversity reflected by differences in patterns of DNA replication. J. Virol. 68(2), 1239–1244. Tiemessen, C. T., and Kidd, A. H. (1995). The subgroup F adenoviruses. J. Gen. Virol. 76, 481–497. van Oostrum, J., Smith, P. R., Mohraz, M., and Burnett, R. M. (1987). The structure of the adenovirus capsid. III. Hexon Packing determined from electron micrographs of capsid fragments. J. Mol. Biol. 198, 73–89.
van Raaij, M. J., Louis, N., Chroboczek, J., and Cusack, S. (1999). Structure of the human adenovirus serotype 2 fiber head domain at 1.5 A resolution. Virology 262, 333–343. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. (1993). Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73, 309–319. Yeh, H. Y., Pieniazek, N., Pieniazek, D., Gelderblom, H., and Luftig, R. B. (1994). Human adenovirus type 41 contains two fibers. Virus Res. 33, 179–198.